Datasets:

TaylorAI

/

pubmed_commercial

like 12

PMC009xxxxxx/PMC9099598.txt

==== Front Nutrients Nutrients nutrients Nutrients 2072-6643 MDPI 10.3390/nu14091753 nutrients-14-01753 Article Consumer Understanding of the Australian Dietary Guidelines: Recommendations for Legumes and Whole Grains Reyneke Gynette 1 https://orcid.org/0000-0003-2228-9880 Hughes Jaimee 2* Grafenauer Sara 3 An Ruopeng Academic Editor 1 School of Medicine, University of Wollongong, Northfields Avenue, Wollongong, NSW 2522, Australia; glr603@uowmail.edu.au 2 Grains & Legumes Nutrition Council, 1 Rivett Road, North Ryde, NSW 2113, Australia 3 Faculty of Medicine and Health, School of Health Sciences, University of New South Wales, Randwick, NSW 2031, Australia; s.grafenauer@unsw.edu.au * Correspondence: j.hughes@glnc.org.au; Tel.: +61-428-941-664 22 4 2022 5 2022 14 9 175304 4 2022 21 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Dietary guidelines provide evidence-based guidance for healthy individuals to improve dietary patterns, although they are most often based on individual foods or food groups. Legumes are a class of food included in current Australian Dietary Guidelines (ADG), mentioned in two of the five food groups, as a vegetable and as an alternative to meat. Whole grain consumption is encouraged in ADG via the statement focused on cereal grains due to their health-promoting properties. Despite their prominence in guidelines, average legume and whole grain consumption in Australia remains lower than recommendations outlined in the ADG. This exploratory study aimed to understand consumer perspectives of wording utilised in dietary guidelines specifically focused on legumes and whole grains. Based on the analysis, there was a significant preference for the statement “each day, consume at least one serve of legumes either as a serve of vegetables or as an alternative to meat” (p < 0.05), which provides a specific frequency and quantification for legume consumption. For whole grain, the significantly preferred statement was “choose whole grain products over refined grains/white flour products whenever you can” indicating a less prescriptive option. Effective messaging in guidelines could consider greater specificity regarding frequency, quantity and quality of foods recommended. This exploratory study suggests an improvement in the adoption and consumption of legumes and whole grains in the Australian diet may be better facilitated through consumer-tested messaging. legumes whole grains consumer preferences dietary guidelines nutrition communication ==== Body pmc1. Introduction Food based dietary guidelines provide evidence-based practical and actionable recommendations that aim to influence the dietary behaviours of the nation through consumer education and targeted health policies and programs [1]. The placement and classification of specific foods within the dietary guidelines are based upon the traditional dietary pattern of that country [1,2]. Consistent with a majority of other countries, the Australian Dietary Guidelines (ADG) promote proportional consumption of a variety of foods across five food groups: (i) vegetables; (ii) fruit; (iii) grains (staple starchy cereal foods); (iv) meat/meat alternatives (protein foods) and (v) dairy foods [3]. While the World Health Organization provides a generalised recommendation for legumes, emphasising consumption as part of a healthy diet [4], across the globe, there is substantial variability in the classification and grouping of legumes within dietary guidelines [2]. Legumes are predominately classified as part of the protein rich food group, although some countries feature them as part of the vegetable group due to their high dietary fibre, mineral and vitamin content [2]. Conversely, the Dietary Guidelines for Americans [5] and the ADG incorporate legumes in both the vegetable food group, as a good source of dietary fibre, vitamins, minerals and phytochemicals [6], and the meat alternatives group, with a protein rich nutrient profile similar to poultry, fish, lean meat and eggs [7]. While many countries include key messages for legume consumption in their dietary guidelines, there is wide incongruity between countries. In particular, some avoid the use of specific quantitative recommendations, opting instead for action words that suggest proportions, frequency and variety [1,2]. As is the case in Australia, legume consumption is promoted via the statements “consume plenty of vegetables, including different types and colour, and legumes/beans” and “consume lean meat and poultry, fish, eggs, tofu, nuts and seeds, and legumes/beans” [3]. Where countries do include quantitative recommendations, there is broad disparity in terms of units of measure, including serve sizes, portions or weight in grams [1]. Likewise, within these measures, the quantity recommended varies between countries [1,2]. Despite the inclusion of legumes in the ADG since 1992, only 28% of Australians consume legumes regularly [8], suggesting that in order to be effective, guideline recommendations need to be applied in the context of additional external measures that address previously identified barriers to intake, such as poor familiarity, lack of preparation skills and gastrointestinal discomfort [9,10]. Furthermore, studies suggest that there may be confusion around the categorisation and quantification of legume recommendations contained in the current ADG that may further hinder uptake [9]. Similar issues have been identified for whole grains and whole grain foods [11]. A recent global review of whole grain recommendations in dietary guidelines found that 44% of countries promote the consumption of whole grains, discerning these food types from more refined grains [11]. Some countries suggest half of all grain foods should be whole grain, as in the US, while others use more passive language such as “mostly” or “preferably” to promote intake [11]. In Australia, whole grain consumption is promoted via the statement consume “grain (cereal) foods, mostly wholegrain and/or high cereal fibre varieties….” [12]. Similar to legumes, whole grain consumption, globally, falls below recommendations [13], and it has been suggested that more targeted and actionable recommendations outlined in dietary guidelines may be one strategy towards improving intake [14,15]. In Denmark, whole grains are actively promoted, with consumers directed to simply “choose whole grain”, which may have contributed to increased rates of whole grain consumption over time [16]. A recent Australian consumer study that aimed to explore consumer understanding of whole grains identified a need for more definitive recommendations for whole grain food intake within the dietary guidelines [17], thus requiring a deeper understanding of the consumers’ perspective of messaging, which may encourage intake. The National Health and Medical Research Council’s coordinated review of the ADG from 2021, provides opportune timing to undertake meaningful and novel research that may optimise the wording that frames the ADG promotion of legume and whole grain intake. This exploratory study aimed to understand Australian consumer perspectives of wording utilised in dietary guidelines, specifically focused on legumes and whole grains, in order to evaluate potential suggested wording for dietary guidelines. It was hypothesised that consumers would identify potential for improvements to the current dietary guidelines’ statements pertaining to legume and whole grain foods to enhance specificity. 2. Materials and Methods A cross-sectional self-administered anonymous survey of Australian adults (aged 18 years old) was conducted using an anonymous online computer-based survey delivered via Qualtrics (Provo, UT, USA) (n = 314) [18]. The survey was approved by the University of Wollongong Research Ethics Committee (Ethics number: 2021/154). Individuals with a formal nutrition education or profession were excluded from participation due to the potential for bias related to enhanced levels of nutritional knowledge and higher rates of dietary guideline adherence [19,20]. The convenience and purposive recruitment of respondents was conducted via an e-survey link, promoted on social media platforms and the Grains and Legumes Nutrition Council (GLNC) eNewsletter for a period of 12 weeks from June to August 2021. Respondents were required to have internet and computer or smart phone access to complete the questionnaire. The survey was accessed via an advertised link that redirected respondents to an introductory statement and consent form, with tacit consent obtained through agreement to commence the survey. The survey included an age declaration prior to commencement and individuals with nutritional training or qualifications were excluded from participation as previously mentioned. Participation was incentivised with a random prize draw to win one of three recipe book sets valued at AUD75. During the data collection phase, additional advertisements via the promotional tools embedded in the particular social media platforms were conducted to target specific age groups and genders, aiming to increase participation across demographics. Survey questions were developed and pilot-tested in consultation with key stakeholders, including consumers, to test construct and content validity, comprehensibility of survey questions and to establish an estimated completion time of 15–20 min. Survey questions were designed to gauge consumer understanding of the current ADG recommendations and preferences for the language used to encourage legume and whole grain intake. The final survey (Supplementary Materials S1) consisted of 20 questions and utilized an open and closed questionnaire design including a combination of multiple-choice questions, drag and drop ranking, and free text boxes for open responses. Closed multiple-choice questions were used to collect demographic characteristics such as age, gender and habitual dietary pattern (“unrestricted omnivore”, “flexitarian”, “vegetarian”, “pescatarian”, “vegan” and “other”), as well as current consumption, barriers and facilitators to legume and whole grain intake. Closed multiple choice questions were also used to determine consumer understanding of current dietary guideline messages. Respondents were provided with the current statements outlined in Guideline 2 of the ADG related to legumes and whole grains and were asked to select the statement that best explained the recommendations (Supplementary Material S1). To explore potential new dietary guideline statements, respondents were asked to rank six statements that would encourage legume intake and five statements for whole grain food intake, where one question for each food type referred to the respondents most preferred statement and least preferred statement. The suggested messaging statements were formulated reflecting the diversity and incongruity of legume and whole grain recommendations across the globe [1,21]. The hypothetical guideline statement options included a combination of quantitative (e.g., Consume one serve of legumes daily) and qualitative recommendations (e.g., Consume mostly whole grain foods) to determine consumer preferences for the language used. An option to maintain current wording was included for comparison. Statistical Analysis All survey data was exported from Qualtrics (Provo, UT, USA) to Microsoft Excel™ (Version 2202, Washington, DC, USA) for data collation. Descriptive statistics were used to provide frequency counts and percentages for demographic data, multiple choice, drag and drop ranking and Likert scale related questions. Descriptive statistics were most appropriate for this study due to the explorative nature of the analysis, and we conducted a repeated measures ANOVA with a Bonferroni correction for multiple tests in R (Version 3.6.2; The R Foundation for Statistical Computing, Vienna, Austria) for the preference-based statements relating to legumes and whole grains based only on the complete cases (275 participants). Alpha was set at 0.05. In the case of multiple-choice questions related to general understanding of the ADG, more than one response was permitted and as a result, the values presented are the proportions of respondents selecting each option. Content analysis and summary of free text responses were also undertaken. 3. Results 3.1. Respondent Demographics The survey was attempted by 314 eligible respondents and completed in full by 275, providing a completion rate of 88%. Due to the independent nature of each question, all survey responses, inclusive of those partially completed, were included in the final analysis with varied participation numbers reflected in the results. The majority of respondents were female (84%, n = 265) and aged 45 years and over (70%) (Table 1). More than half of respondents reported to follow an omnivorous dietary pattern (56%), and a quarter a flexitarian diet (23%) (Table 1). 3.2. Reported Consumption of Legumes and Whole Grains Respondents most commonly reported to consume legumes several times a week (50%), daily (22%) or weekly (18%) and fewer reported consuming legumes less than once a week (Table 2). Legumes were primarily consumed as a source of protein, as an alternative to meat (82%; n = 224), as opposed to a serving of vegetables (44%; n = 121). In contrast, forty percent reported consuming legumes as both a source of protein and as a serve of vegetables (n = 109). Overall, just 16% of respondents identified legumes as a feature of their traditional diet (n = 45). Daily whole grain consumption was common among respondents (88%; n = 260), most of whom reported consuming 1–2 servings (41%; n = 121) or 3–4 serves (42%; n = 123) per day, with one serving defined according to the ADG. Most commonly, respondents selected at least two reasons for the consumption of whole grains (71%; n = 202); predominantly as a source of dietary fibre (80%; n = 229) and over half enjoyed the taste of whole grains (57%; n = 163). Additionally, respondents reported to consume whole grains as a source of carbohydrate (45%; n = 129); plant-based protein (38%; n = 110) and as part of their traditional diet (23%; n = 67). 3.3. Consumer Understanding of the Australian Dietary Guidelines Overall, the majority correctly identified that the ADG were developed to promote health and wellbeing (84%; n = 257), to serve as a healthy eating guide (68% n = 208) and to support healthy dietary choices (54%; n = 167), while half of respondents associated the ADG with reducing risk of specific chronic diseases (59%; n = 182) and diet-related health conditions (59%; n = 180). A small proportion considered the ADG to promote industry or were unsure of their purpose (12%; n = 38). Commonly, the dietary guidelines were deemed applicable to healthy Australians or individuals seeking to maintain a healthy lifestyle (91%; n = 279); individuals seeking to lose weight (29%; n = 90), those with chronic disease (32%; n = 99) and healthcare professionals (33%; n = 101). Three quarters of respondents indicated that the ADG should be followed on most days (76%; n = 233) as opposed to opting for strict and consistent observation (11%; n = 34); some or vague consideration (12%; n = 37) or not at all (1%; n = 3). 3.4. Australian Dietary Guidelines and Legume Recommendations 3.4.1. Understanding of Legume Messaging Almost all respondents demonstrated good understanding for the highlighted statement in Guideline 2 of the ADG pertaining to legume recommendations. As shown in Table 3, respondents most frequently selected the interpretation that “legumes can be consumed as a vegetable as well as a protein replacement for meat and eggs” (72%; n = 202), followed by “choose a variety of vegetable and protein foods” (61%; n = 171), and “legumes are an important food as they feature in two of five food groups” (47%; n = 132) (Table 3). In response to an open ended question related to the two different serving sizes for legumes (75 g (½ cup) as a vegetable serving and 150 g (1 cup) as an alternative to meat), the majority reported that they found it easy to interpret (70%; n = 196), citing that the categorisation accompanied with a specific serving size was “clearly stated”, “easy to understand and remember” and provided a “clear size guide”. However, one fifth of respondents found the serving size and categorisation of legumes confusing (21%; n = 58). Some respondents stated a preference for the serving size provided as a cup measure, stating that “Cups is easier to visualise for me”, “Yes—easy to interpret if using cup measurements rather than grams”, “yes, the cup measurement means more than the weight though so put it first maybe?”. Others expressed that the differentiation between food groups and the need to measure serves is impractical and not commonly practiced. Further responses are provided in Supplementary Material S2 Table S1. 3.4.2. Consumer Preferences for Legume Based Recommendations In considering alternative wording for legume promotion in the ADG, respondents were required to rank provided statements in sequential order, whereby the most preferred statement ranked first and the least ranked sixth. Statements that emphasised daily consumption or provided a quantitative measure for intake ranked well overall (Figure 1). Overall, preference for the six statements differed significantly (p < 0.05). The statement “Each day, consume at least one serve of legumes either as a serve of vegetables or as an alternative to meat” was the most preferred statement, as it was selected significantly more times than all other statements, except “Eat at least 100 g (½ cup) of legumes 3 or more times per week”, where there was no significant difference. The statement “Eat 50–100 g peas, beans or lentils 3 times per week” was selected the fewest times as the most preferred statement. When asked what would be helpful in achieving an increase in legume intake, the majority preferred legumes to feature either in their own food group or as part of the protein group (Table 4). Respondents reported a preference for recommendations that included frequency of intake and quantifiable cup measures, rather than grams (in other words, how many cups and how often). Maintaining the current guidelines was displaced by these preferred options (Table 4). 3.5. Australian Dietary Guidelines and Whole Grain Recommendations 3.5.1. Understanding of Whole Grain Messaging Similar to legumes, almost all respondents demonstrated good understanding of the statement in Guideline 2 related to whole grain consumption (97%, n = 291). The most commonly selected interpretation was “eat most (more than half) of your grain foods from whole grain choices” (53%; n = 157); followed by “eat from a variety of grain foods including refined and whole grains” (33%; n = 98) and “limit refined and low fibre grain foods” (28%; n = 84) (Table 5). 3.5.2. Consumer Preferences for Whole Grain Recommendations To evaluate preferences towards potential new guideline statements for the promotion of whole grain foods, respondents were required to rank provided statements in sequential order, where the most preferred statement ranked first and the least ranked fifth. Similarly to legumes, there was a significant difference between all statements for whole grains (p < 0.05). The statement “Choose whole grain products over refined grains/white flour products whenever you can” was most preferred and had a significantly higher mean score than all other statements (p < 0.05) (Figure 2). Comparatively, “choose high fibre breads and cereals containing at least fifty percent whole grain in the food label” was ranked the most preferred the fewest times. Statements that emphasised the variety of whole grains also ranked well (Figure 2). Maintaining the current guideline wording received the most votes as the least preferred statement (Figure 2). 4. Discussion While previous Australian studies have explored consumer perceptions and attitudes towards legumes and whole grains [10,22,23], to our knowledge, this is the first Australian study to investigate consumer understanding of, and preferences for, the language used in the ADG for the consumption of legumes and whole grains. This exploratory study provides prerequisite insight into the consumers’ perspective of the current representation of legumes and whole grains within the ADG, including preferences for categorisation, frequency and quantity of intake. The prevalence for dietary patterns such as vegan, vegetarian or pescatarian diets reported in the present study is similar to the findings of Figueira et al. (2019); however, the notable difference here was the purposeful distinction between an unrestrictive omnivorous diet [9] and a flexitarian diet, synonymous with a more plant-based dietary pattern that allows for the flexibility to include some animal sourced foods [24]. Congruent with Figueira et al. (2019), most respondents in the current study reported to follow no specific diet, suggestive of an unrestrictive omnivorous diet. However, one quarter of respondents reported to follow a flexitarian diet, reflective of the growth seen in this dietary pattern [25,26]. The accelerating shift away from animal-sourced foods, namely dairy and red meat, as consumers endeavour to align their diet with environmental initiatives [26,27], positions legumes and whole grains as a valuable source of plant protein, dietary fibre and other key nutrients [21,28]. Recently, there has been a call on national governments for the revision of current dietary guidelines to reflect environmental sustainability objectives in a bid to integrate both environmental and population health [29,30]. If carried out correctly, a global shift towards plant-based foods is considered a dietary strategy that may prove important and beneficial for the health of humans and the planet [30,31]. Reflective of this, the Canadian dietary guidelines have recently shifted towards a plant-based diet, emphasising whole grains, legumes, nuts, seeds and fortified plant-based milk alternatives [32]. The literature shows that adherence to the ADG corresponds with higher diet quality [20] and is consequently beneficial for mood, associated with reduced risk for depression and cardiometabolic diseases [33]. Positively, our findings indicate that respondents had a good awareness and understanding of the ADG, widely acknowledging that the guidelines are aimed at healthy Australians or those who want to maintain a healthy lifestyle. Respondents expressed that the guidelines should be followed on most days; however, responses diverged regarding how strictly they should be observed. Although the results were encouraging, particularly as awareness and knowledge are considered prerequisites of behaviour change [34,35], as supported by Bandura’s social cognitive theory [36], the relationship between knowledge and actioned behaviour is complex and multifactorial and rarely equates to actual dietary change [37,38]. In this study, Guideline 2 was well interpreted for both legumes and whole grains. The current recommendations for legumes were well received, with a majority of respondents able to identify legumes as a source of vegetables and a plant-protein alternative to meat. Respondents had a good, overall understanding of the two different categorisations and serve sizes for legumes as a meat alternative or vegetable. These findings were encouraging as they explore both the comprehension and interpretation of this guideline, an important distinction given that previous studies have demonstrated that dietary guidelines may be frequently misinterpreted and poorly implemented by adults who have concurrently reported to have understood the messaging [39,40]. In the current study, the most common interpretation of Guideline 2 for grain foods was to consume most (more than half) grain foods from whole grains. Research conducted with certain population groups in the US found that despite an understanding of the whole grain statement “Make half your grains whole grains”, there were barriers to usage, related to taste, cost and identification of whole grain products [39], further demonstrating that an awareness or understanding of guidelines has been shown to rarely translate into usage [41]. These findings are not uncommon [17] and, therefore, understanding these influencers is necessary if promotion is to be meaningful [9,38]. As stated, legumes and whole grains are not widely consumed in Australia [8], yet the respondents of this study had a high reported intake, well above that of the average intake in Australia, indicated by the number of servings consumed per day rather than more precise estimations. In the case for legumes, respondents reported regular consumption, daily or at least several times a week, despite not considering them part of their traditional diet, indicating that respondents likely had a positive attitude towards legumes and possibly a good baseline knowledge of how to incorporate them into their diet [9]. In a similar manner, respondents had higher than average whole grain intakes, with 88% reporting daily consumption. From representative data of adults reporting consumption of whole grains from the National Nutrition and Physical Activity Survey (NNPAS), the median whole grain intake was 38.4 g/d, with the 48 g Daily Target Intake (DTI) reached by only 39.7% [42] of adults, with 29.1% not consuming any whole grains on the day of the survey [42]. In the current study, over half of respondents reported to enjoy the taste of legumes and whole grains, possibly contributing to high reported intakes. Taste is a strong determinant for consumption, secondary to health benefits, and is often a cited barrier to regular legume intake [9]. Consistent with Foster et al. (2020), the current study found that respondents who reported consuming whole grains, favoured these foods based on their micronutrient content, as a source of dietary fibre (80%), rather than as source of macronutrients, such as carbohydrates (45%) or protein (23%) [17]. This may have also been a function of the particular participant sample, who appear well informed regarding nutrition. To our knowledge, there is limited published research investigating the effectiveness of messaging, provided within the ADG or any food-based guidelines, with consistency over time [41]. The findings of this exploratory study suggested a consumer preference for legume recommendations that provided an indication of a specific frequency and quantification of intake, with the statement “Each day, consume at least one serve of legumes either as a serve of vegetables or as an alternative to meat” the most preferred among respondents. The option to maintain status quo for the wording outlined in Guideline 2 was poorly received, yet, interestingly, the highest ranked statement, as previously mentioned, was quite similar to the original guideline. This popular option maintained the flexibility of legumes as a serving of vegetables or meat alternative but replaced the descriptive term “plenty of” with specified intake set at one serving daily, which was based on optimal intake modelling [43]. This preference was further supported by the popularity of other statements where quantification and frequency for intake was included in the statement. These findings are supportive of Geiger (2001), demonstrating consumer preference for specific terms relating to intake quantity and frequency in comparison to permissive terms such as “enjoy”, “balance” or “over a few days” and, likewise, with the current study, the format of the existing dietary guidelines was rated poorly [44]. The current ADG provide recommendations for serving sizes in a weighted gram measure with a corresponding cup value, whereby 75 g equates to a ½ cup [3], noting that serving size is a standardized amount of food, whereas portion size is the amount of food consumed [45]. It has been shown that consumers are better able to relate a serving size to a common cup measure rather than a weighted value, which requires higher literacy levels and scales [46]. Congruently here, respondents strongly emphasised their preference for quantifiable recommendations expressed in cup measures, stating that grams were less relevant and poorly visualised. Given the variation in legume and whole grain recommendations across regions, a standardisation of the amount promoted by health authorities and agriculture may be helpful. In the case for legumes, a recent review rationalised that 100 g cooked legumes should equate to ½ cup measure, equivalent to one serving, and further proposed this as the “standardised minimum threshold” for consumption in a single day [1,21]. Notably, in Australia, Canada and Europe, 100 g of cooked legumes satisfies the Food Standards Australia New Zealand (FSANZ) criteria required to qualify nutrient content claims for key nutrients found in legumes [21,47]. Indeed, the standardisation of this minimum threshold for consumption is appropriate for the prevention and management of chronic disease [48,49]. Determining frequency of intake for legumes over a week presents more of a challenge than serving size, as this needs to align with the cultural acceptability of a food, routine dietary patterns, and the infrastructure of the food systems [1,21]. In consideration of these factors, and the wide variety of foods available in Australia, even occasional consumption of this minimum ½ cup recommendation may enhance nutrient intake and contribute towards a healthy diet [31]. In the current study, there was less consensus regarding the categorisation of legumes. Most popular was the suggestion to feature legumes in their own food group, but this was only the case when the statement included a direct reference to quantity and frequency factors, noting that the suggestion was not popular in the absence of these indicators, and, likewise, with the suggestion to feature legumes solely in the meat alternative group. Once again, without the provision of quantity and frequency factors, the statement was poorly received; in fact, consumers expressed preference for the original wording rather than a simple reassignment of food groups in the absence of these qualifiers. Abdullah et al. (2017) suggest that classifying legumes as an “alternative” to meat may have a negative connotation, impacting consumer acceptance, and that a bolder approach is required to promote legumes as an independent food group, similar to whole grains and cereals [50]. Similarly, specific consumption guidelines for grain foods are not included in dietary guidelines, but the suggestion that most of the six servings of grain foods recommended for the 19–50 years age-group should be whole grain [12,45]. This amount would meet the 48 g DTI recommended by GLNC [51], the target developed in consultation with an expert round table and aligned with the target suggested in American Dietary Guidelines [52,53]. This recommendation is supported by evidence from systematic reviews and dose–response meta-analyses of prospective cohort studies by Aune et al. [54,55], and is supported by more recent work by Reynolds et al. [56], where each 15 g consumed assisted with risk reduction. Despite the lack of specificity, the documentation supporting the guidelines note that 70% of grains consumed in Australia are refined grains, such that a 160% increase in current whole grain consumption and a 30% decrease in refined grain (cereal) food consumption has been recommended [57]. Making this change is well supported by the food supply, with an analysis indicating one-third of breads on supermarket shelves were whole grain/wholemeal, with a median whole grain content of 20.2 g per serving (2 slices), almost half of the 48 g DTI [58,59]. This change is also supported by the outcomes of the current study, whereby respondents most preferred the dietary guideline statement “Choose whole grain products over refined grains/white flour products whenever you can”. A quantified prescription for whole grain is more difficult, due to the variability in whole grain content within foods. Instead, this must be supported by education and food labelling initiatives to simplify purchasing decisions. Limitations The current study includes several limitations. Online surveys rely on convenience sampling and therefore are prone to selection bias, as respondents are required to be literate, have internet access, volunteer for participation and are more likely to have an inherent interest in nutrition [60]. Research shows that women are more likely to respond to online surveys than their male counterparts [61] and women are known to have healthier eating habits than men [21,62]. This is further exacerbated by gender-based stereotyping, whereby vegetables are considered a “female suited food”, in comparison to the masculinity associated with meat consumption [63]. Similarly, data from the NNPAS suggests that after adjusting for energy intake, adult females appeared to consume more whole grains than males, whereas slightly more males, than females, consumed no whole grain on the day of the survey (30.9% of males, 27.5% of females) [42]. The current findings, however, relate to a small female dominant sample of respondents, reporting a baseline legume and whole grain intake well above national average consumption levels and, therefore, are not representative of the Australian population [8,60]. In the instance where respondents were asked to rank the most preferable statement for potential new guidelines, it is possible that using the term “maintain current wording”, may have caused respondents to reject this option due to an association with injunctive norms [64]. Studies suggest that consumers are more likely to reject injunctive norms in preference for options that facilitate autonomy and free choice [64,65]. Furthermore, in the present study, sustainability and environmental factors were not considered part of the messaging. This was possibly a missed opportunity to explore other concepts that may enhance the messaging, given that environmental benefits are perceived important beneficial factors for the consumption of plant foods such as legumes and whole grains [1,66]. Finally, the study may have been strengthened by the contribution of communication experts to effectively interpret evidence into messaging appropriate for a range of literacy levels and thereby optimise nutrition communication strategies [67]. This is also an appropriate recommendation for the revision of dietary guidelines, where new statements may benefit from broader community consultation to gauge understanding. There is a need for quality studies conducted to specifically evaluate the effectiveness of the dietary guidelines and their measurable contribution to population health through the translation of recommendations to dietary behaviour and change [39,40]. Finally, it is important to emphasise that the current study is an insight into components of dietary guidelines that influence consumers and does not address the challenges for aligning Australian dietary patterns with recommendations for regular legume and whole grain consumption. 5. Conclusions In order to improve dietary intake for legumes and whole grains, this exploratory study highlights consumer views of suitable wording and commentary regarding the current dietary guidelines, which may help inform the review of dietary guideline statements. Findings suggest that consumers may favour legume recommendations that include quantities provided in cup measures and frequency related to daily or weekly intake. The revision of dietary recommendations may consider aligning with a standardised minimum threshold for consumption; for example, a minimum of ½ cup (100 g) cooked legumes in a single day, while offering the flexibility to consume legumes as part of the vegetable or meat alternative group. In regard to whole grains, it needs to be acknowledged that the current statement pertaining to grain (cereal) foods is reportedly ineffective and convoluted and does not direct consumers to better choices within the food group. Yet, the significance of whole grain foods within dietary patterns, evaluated by the global burden of disease research and the simplicity of emphasizing the swap to whole grain in preference to refined choices, provides the impetus to simplify and strengthen wording, and provide clarity to inform consumers to choose whole grain and high fibre food choices as a priority. Both legumes and whole grains provide economic and sustainable food choices for inclusion in dietary patterns and the range of offerings within the food supply currently would support and facilitate consumption at adequate levels to improve the health of Australians. Acknowledgments Thanks to Stephanie Duncombe, PhD Candidate from the University of Queensland, QLD, who provided statistical advice. In writing this, the Grains & Legumes Nutrition Council would like to acknowledge the many dietitians who have contributed to the evidence base and writing of dietary guidelines in Australia. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/nu14091753/s1, Supplementary material S1: Survey Questions; Supplementary material S2: Table S1. Exemplar Quotes. Click here for additional data file. Author Contributions Conceptualization, S.G.; methodology, G.R., J.H. and S.G.; formal analysis, G.R.; writing—original draft preparation, G.R.; writing—review and editing, J.H. and S.G. All authors have read and agreed to the published version of the manuscript. Funding This research received no external funding but was supported by the Grains & Legumes Nutrition Council, a not-for-profit charity. Institutional Review Board Statement The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the University of Wollongong Human Research Ethics Committee (2021/154). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement All data for this study is contained within the article and Supplementary Materials. Conflicts of Interest J.H. is employed by the Grains & Legumes Nutrition Council, a not-for-profit charity. S.G. was an employee of the Grains & Legumes Nutrition Council at the time of survey development and data collection. Figure 1 Ranking of preferred new guideline statements for the promotion of legume intake (n = 275). Figure 2 Ranking of preferred new guideline statements for the promotion of whole grain foods (n = 286). nutrients-14-01753-t001_Table 1 Table 1 Demographic characteristics and dietary pattern of respondents (n = 314). Demographic Variable Count (%) Gender Female 265 (84) Male 49 (16) Age in years 18–24 17 (5) 25–34 31 (10) 35–44 46 (15) 45–54 76 (24) 55–64 84 (27) 65+ 60 (19) Habitual dietary pattern Unrestricted omnivore 176 (56) Flexitarian 73 (23) Vegetarian 22 (7) Vegan 21 (7) Pescatarian 20 (6) Other 2 (0.6) nutrients-14-01753-t002_Table 2 Table 2 Reported legume and whole grain intake. Count (%) Reported legume consumption n = 280 Several times a week 139 (50) At least once per day 62 (22) Approximately once a week 51 (18) 2–3 times per month 16 (6) Irregularly (less than twice per month) 10 (4) Never 2 (0.7) Reported whole grain consumption n = 295 3–4 serves per day 123 (42) 1–2 serves per day 121 (41) Less than a serve per day 31 (11) 5 or more serves per day 16 (5) I do not eat whole grain foods 3 (1) I do not know 1 (0.3) nutrients-14-01753-t003_Table 3 Table 3 Respondents’ interpretation of the focus of Guideline 2 for legume intake (n = 280) *. Interpretation of Guideline 2 1 Count (%) Legumes can be consumed as a vegetable as well as a protein replacement for meat and eggs 202 (72) Choose a variety of vegetable and protein foods 171 (61) Legumes are an important food as they feature in two of the five food groups 132 (47) Legume intake is optional 10 (4) Eat legumes twice a day 7 (3) Other 4 (1) I do not know 2 (0.7) 1 Guideline 2 for legume intake: Enjoy a wide variety of nutritious foods from these five food groups every day including plenty of vegetables of different types and colours, and legumes/beans and lean meats and poultry, fish, eggs, tofu, nuts and seeds, and legumes/beans. * Question allowed participants to select more than one answer, consequently values presented are the proportion of respondents selecting each point and exceed 100%. nutrients-14-01753-t004_Table 4 Table 4 Consumer preferences for guideline recommendations (n = 275) *. Count (%) Question: If the aim is to increase legume intake, which of the following would you find most helpful in achieving this? Legumes feature in their own food group with recommendations for how much and how often to consume 123 (45) Legumes feature in the meat/meat alternatives group as a source of protein, with recommendations for how much and how often to consume 61 (22) Maintain current guideline. Recommendations are to consume legumes as part of the vegetable group and/or meat/meat alternative group 55 (20) Legumes feature in their own food group 24 (9) Legumes feature in the meat/meat alternatives group 12 (4) Question: In relation to the dietary guidelines, how would you prefer the recommendations for intake to be presented? As a cup measure for each food/food group (e.g., consume ½ cup cooked brown rice) 136 (50) As a suggested frequency (e.g., consume 2–3 times per week) 69 (25) Maintain current format: As the number of serves per day for each food/ food group (e.g., consume 5 serves per day) 53 (19) As the number of grams for each food/food group (e.g., consume 48 g whole grain) 17 (6) * Question allowed participants to select more than one answer, consequently values presented are the proportion of respondents selecting each option. nutrients-14-01753-t005_Table 5 Table 5 Respondents’ interpretation of the focus of Guideline 2 for whole grain intake (n = 295) *. Interpretation of Guideline 2 1 Count (%) Eat most (more than half) of your grain foods from whole grain choices 157 (53) Eat from a variety of grain foods including refined and whole grains 98 (33) Limit refined and low fibre grain foods 84 (28) Enjoy any kind of grain-based food 41 (14) Eat only whole grain and/or high cereal fibre foods 41 (14) I do not know 4 (1) 1 Guideline 2 for whole grain intake: Enjoy a wide variety of nutritious foods from these five food groups every day including grain (cereal) foods, mostly wholegrain and/or high cereal fibre varieties, such as breads, cereals, rice, pasta, noodles, polenta, couscous, oats, quinoa and barley. * Question allowed participants to select more than one answer, consequently values presented are the proportion of respondents selecting each option and combined exceed 100%. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Herforth A. Arimond M. Álvarez-Sánchez C. Coates J. Christianson K. Muehlhoff E. A Global Review of Food-Based Dietary Guidelines Adv. Nutr. 2019 10 590 605 10.1093/advances/nmy130 31041447 2. Painter J. Rah J.-H. Lee Y.-K. Comparison of International Food Guide Pictorial Representations J. Am. Diet. Assoc. 2002 102 483 489 10.1016/S0002-8223(02)90113-6 11985405 3. National Health and Medical Research Council Australian Dietary Guidelines Canberra, Australia 2013 Available online: https://www.eatforhealth.gov.au/guidelines/australian-dietary-guidelines-1-5 (accessed on 3 March 2022) 4. World Health Organisation Healthy Diet World Health Organization Geneva, Switzerland 2020 Available online: https://www.who.int/news-room/fact-sheets/detail/healthy-diet (accessed on 1 August 2021) 5. U.S. Department of Agriculture and U.S. Department of Health and Human Services Dietary Guidelines for Americans, 2020–2025 9th ed. December 2020 Available online: https://www.dietaryguidelines.gov/sites/default/files/2020-12/Dietary_Guidelines_for_Americans_2020-2025.pdf (accessed on 3 March 2022) 6. National Health and Medical Research Council Vegetables and Legumes/Beans|Eat For Health 2019 Available online: https://www.eatforhealth.gov.au/food-essentials/five-food-groups/vegetables-and-legumes-beans (accessed on 17 September 2021) 7. National Health and Medical Research Council Lean Meat and Poultry, Fish, Eggs, Tofu, Nuts and Seeds and Legumes/Beans Available online: https://www.eatforhealth.gov.au/food-essentials/five-food-groups/lean-meat-and-poultry-fish-eggs-tofu-nuts-and-seeds-and#:~:text=Legumes%20provide%20many%20of%20the (accessed on 16 September 2021) 8. Grains & Legumes Nutrition Council Grains and Legumes Consumption and Attitudinal Report 2017 2017 Unpublished Available online: https://www.glnc.org.au/wp-content/uploads/2018/09/GLNC-Consumption-Study-Summary-5.pdf (accessed on 31 July 2021) 9. Figueira N. Curtain F. Beck E. Grafenauer S. Consumer Understanding and Culinary Use of Legumes in Australia Nutrients 2019 11 1575 10.3390/nu11071575 31336897 10. Śmiglak-Krajewska M. Wojciechowska-Solis J. Consumption Preferences of Pulses in the Diet of Polish People: Motives and Barriers to Replace Animal Protein with Vegetable Protein Nutrients 2021 13 454 10.3390/nu13020454 33573021 11. Grafenauer S. Curtain F. Historical and Global Perspectives on Grains and Whole Grains within Dietary Guidelines Cereal Foods World 2020 65 10.1094/CFW-65-3-0029 12. National Health and Medical Research Council Grain (Cereal) Foods, Mostly Wholegrain and/or High Cereal Fibre Varieties Available online: https://www.eatforhealth.gov.au/food-essentials/five-food-groups/grain-cereal-foods-mostly-wholegrain-and-or-high-cereal-fibre (accessed on 25 January 2022) 13. Miller K.B. Review of whole grain and dietary fiber recommendations and intake levels in different countries Nutr. Rev. 2020 78 29 36 10.1093/nutrit/nuz052 32728749 14. Khandpur N. De Morais Sato P. Neto J.R.G. Scagliusi F. Jaime P.C. Developing and refining behaviour-change messages based on the Brazilian dietary guidelines: Use of a sequential, mixed-methods approach Nutr. J. 2020 19 66 10.1186/s12937-020-00585-1 32631339 15. Webb D. Byrd-Bredbenner C. Overcoming Consumer Inertia to Dietary Guidance Adv. Nutr. 2015 6 391 396 10.3945/an.115.008441 26178023 16. Lourenço S. Hansen G. Frank P. Stærk B. Petersen C. The Whole Grain Partnership—How a Public–Private Partnership Helped Increase Whole Grain Intake in Denmark Cereal Foods World 2019 64 10.1094/CFW-64-3-0027 17. Foster S.I. Beck E.J. Hughes J. Grafenauer S. Whole Grains and Consumer Understanding: Investigating Consumer. Identification, Knowledge and Attitudes to Whole Grains Nutrients 2020 12 2170 10.3390/nu12082170 18. Qualtrics XM Software 2005 Qualtrics Provo, UT, USA 2020 Available online: https://www.qualtrics.com (accessed on 8 January 2021) 19. Kothe E. Mullan B. Perceptions of fruit and vegetable dietary guidelines among Australian young adults Nutr. Diet. 2011 68 262 266 10.1111/j.1747-0080.2011.01550.x 20. Spronk I. Kullen C. Burdon C. O’Connor H. Relationship between nutrition knowledge and dietary intake Br. J. Nutr. 2014 111 1713 1726 10.1017/S0007114514000087 24621991 21. Marinangeli C.P.F. Curran J. Barr S.I. Slavin J. Puri S. Swaminathan S. Tapsell L. Patterson C.A. Enhancing nutrition with pulses: Defining a recommended serving size for adults Nutr. Rev. 2017 75 990 1006 10.1093/nutrit/nux058 29202192 22. Bateman R.M. Sharpe M.D. Jagger J.E. Ellis C.G. Solé-Violán J. López-Rodríguez M. Herrera-Ramos E. Ruíz-Hernández J. Borderías L. Horcajada J. 36th International Symposium on Intensive Care and Emergency Medicine, Brussels, Belgium, 15–18 March 2016 Crit. Care 2016 20 94 10.1186/s13054-016-1208-6 27885969 23. Szczebyło A. Rejman K. Halicka E. Laskowski W. Towards More Sustainable Diets-Attitudes, Opportunities and Barriers to Fostering Pulse Consumption in Polish Cities Nutrients 2020 12 1589 10.3390/nu12061589 24. Derbyshire E.J. Flexitarian Diets and Health: A Review of the Evidence-Based Literature Front. Nutr. 2017 3 55 10.3389/fnut.2016.00055 28111625 25. Deloitte LLP Plant-Based Alternatives. London, United Kingdom 2019 Available online: https://www2.deloitte.com/content/dam/Deloitte/uk/Documents/consumer-business/deloitte-uk-plant-based-alternatives.pdf (accessed on 18 August 2021) 26. Wickramasinghe K. Breda J. Berdzuli N. Rippin H. Farrand C. Halloran A. The shift to plant-based diets: Are we missing the point? Glob. Food Secur. 2021 29 100530 10.1016/j.gfs.2021.100530 27. Semba R.D. Ramsing R. Rahman N. Kraemer K. Bloem M.W. Legumes as a sustainable source of protein in human diets Glob. Food Secur. 2021 28 100520 10.1016/j.gfs.2021.100520 28. Margier M. Georgé S. Hafnaoui N. Remond D. Nowicki M. Du Chaffaut L. Amiot M.-J. Reboul E. Nutritional Composition and Bioactive Content of Legumes: Characterization of Pulses Frequently Consumed in France and Effect of the Cooking Method Nutrients 2018 10 1668 10.3390/nu10111668 29. Williams P. Broadening influence on the food supply and environmental sustainability Nutr. Diet. 2019 76 247 249 10.1111/1747-0080.12554 31264382 30. Foley J. Diets for a Better Future-Scientific Report-EAT Knowledge 2021 Available online: https://eatforum.org/knowledge/diets-for-a-better-future/ (accessed on 15 September 2021) 31. Fanzo J. Healthy and Sustainable Diets and Food Systems: The Key to Achieving Sustainable Development Goal 2? Food Ethics 2019 4 159 174 10.1007/s41055-019-00052-6 32. Canada’s Dietary Guidelines Health Canada Ottawa, ON, Canada 2019 Available online: https://food-guide.canada.ca/en/guidelines/ (accessed on 31 July 2021) 33. Opie R.S. Ball K. Abbott G. Crawford D. Teychenne M. McNaughton S.A. Adherence to the Australian dietary guidelines and development of depressive symptoms at 5 years follow-up amongst women in the READI cohort study Nutr. J. 2020 19 30 10.1186/s12937-020-00540-0 32276594 34. Anderson E.S. Winett R.A. Wojcik J.R. Self-regulation, self-efficacy, outcome expectations, and social support: Social cognitive theory and nutrition behavior Ann. Behav. Med. 2007 34 304 312 10.1007/BF02874555 18020940 35. Doerksen S.E. McAuley E. Social cognitive determinants of dietary behavior change in university employes Front. Public Health 2014 2 23 10.3389/fpubh.2014.00023 24765620 36. Bandura A. The self system in reciprocal determinism Am. Psychol. 1978 33 344 358 10.1037/0003-066X.33.4.344 37. Baldwin T.T. Falciglia G.A. Application of Cognitive Behavioral Theories to Dietary Change In Clients J. Am. Diet. Assoc. 1995 95 1315 1317 10.1016/S0002-8223(95)00344-4 7594129 38. Bandura A. Health Promotion by Social Cognitive Means Health Educ. Behav. 2004 31 143 164 10.1177/1090198104263660 15090118 39. Chea M. Mobley A.R. Interpretation and Understanding of the Dietary Guidelines for Americans Consumer Messages Among Low-Income Adults J. Am. Coll. Nutr. 2020 39 63 71 10.1080/07315724.2019.1610918 31084517 40. Sigman-Grant M. Dickin K. Dollahite J. Strom H. Olson B. Wengreen H. Mobley A. Krogstrand K. Qualitative evidence of the disconnect between intent and interpretation of common child obesity prevention messages FASEB J. 2010 24 747.2 10.1096/fasebj.24.1_supplement.747.2 41. Brown K.A. Timotijevic L. Barnett J. Shepherd R. Lähteenmäki L. Raats M.M. A review of consumer awareness, understanding and use of food-based dietary guidelines Br. J. Nutr. 2011 106 15 26 10.1017/S0007114511000250 21385508 42. Galea L.M. Beck E.J. Probst Y.C. Cashman C.J. Whole grain intake of Australians estimated from a cross-sectional analysis of dietary intake data from the 2011–13 Australian Health Survey Public Health Nutr. 2017 20 2166 2172 10.1017/S1368980017001082 28592344 43. Schwingshackl L. Schwedhelm C. Hoffmann G. Lampousi A.-M. Knüppel S. Iqbal K. Bechthold A. Schlesinger S. Boeing H. Food groups and risk of all-cause mortality: A systematic review and meta-analysis of prospective studies Am. J. Clin. Nutr. 2017 105 1462 1473 10.3945/ajcn.117.153148 28446499 44. Geiger C.J. Communicating Dietary Guidelines for Americans: Room for improvement J. Am. Diet. Assoc. 2001 101 793 797 10.1016/S0002-8223(01)00197-3 11478478 45. National Health and Medical Research Council What Is a Serve? Eat For Health 2012 Available online: https://www.eatforhealth.gov.au/food-essentials/how-much-do-we-need-each-day/what-serve (accessed on 25 January 2022) 46. Faulkner G.P. Pourshahidi L.K. Wallace J.M.W. Kerr M.A. McCrorie T.A. Livingstone M.B.E. Serving size guidance for consumers: Is it effective? Proc. Nutr. Soc. 2012 71 610 621 10.1017/S0029665112000766 22877767 47. FSANZ Australia New Zealand Food Standards Code–Schedule 4–Nutrition, Health and Related Claims. Attorney-General’s Department 2017 Available online: https://www.legislation.gov.au/Details/F2017C00711 (accessed on 23 July 2021) 48. Nouri F. Sarrafzadegan N. Mohammadifard N. Sadeghi M. Mansourian M. Intake of legumes and the risk of cardiovascular disease: Frailty modeling of a prospective cohort study in the Iranian middle-aged and older population Eur. J. Clin. Nutr. 2016 70 217 221 10.1038/ejcn.2015.153 26395434 49. Bazzano L.A. He J. Ogden L.G. Loria C. Vupputuri S. Myers L. Whelton P.K. Legume Consumption and Risk of Coronary Heart Disease in US Men and Women Arch. Intern. Med. 2001 161 2573 10.1001/archinte.161.21.2573 11718588 50. Abdullah M. Marinangeli C. Jones P. Carlberg J. Canadian Potential Healthcare and Societal Cost Savings from Consumption of Pulses: A Cost-Of-Illness Analysis Nutrients 2017 9 793 10.3390/nu9070793 51. Curtain F. Locke A. Grafenauer S. Growing the Business of Whole Grain in the Australian Market: A 6-Year Impact Assessment Nutrients 2020 12 313 10.3390/nu12020313 52. Griffiths T. Nestel P. Developing a target for daily wholegrain intake for Australians Food Aust. 2006 58 431 433 53. Griffiths T. Towards an Australian ‘daily target intake’ for wholegrains Food Aust. 2007 59 600 601 54. Aune D. Norat T. Romundstad P. Vatten L.J. Whole grain and refined grain consumption and the risk of type 2 diabetes: A systematic review and dose–response meta-analysis of cohort studies Eur. J. Epidemiol. 2013 28 845 858 10.1007/s10654-013-9852-5 24158434 55. Aune D. Keum N. Giovannucci E. Fadnes L.T. Boffetta P. Greenwood D.C. Tonstad S. Vatten L.J. Riboli E. Norat T. Whole grain consumption and risk of cardiovascular disease, cancer, and all cause and cause specific mortality: Systematic review and dose-response meta-analysis of prospective studies BMJ 2016 353 i2716 10.1136/bmj.i2716 27301975 56. Reynolds A. Mann J. Cummings J. Winter N. Mete E. Te Morenga L. Carbohydrate quality and human health: A series of systematic reviews and meta-analyses Lancet 2019 393 434 445 10.1016/S0140-6736(18)31809-9 30638909 57. National Health and Medical Research Council Australian Dietary Guidelines. Providing the Scientific Evidence for Healthier Australian diets 2013 Canberra Available online: https://www.eatforhealth.gov.au/sites/default/files/content/n55_australian_dietary_guidelines.pdf (accessed on 25 January 2022) 58. Abdullah M.M.H. Hughes J. Grafenauer S. Healthcare Cost Savings Associated with Increased Whole Grain Consumption among Australian Adults Nutrients 2021 13 1855 10.3390/nu13061855 34072326 59. Grafenauer S. Curtain F. An Audit of Australian Bread with a Focus on Loaf Breads and Whole Grain Nutrients 2018 10 1106 10.3390/nu10081106 60. Andrade C. The Limitations of Online Surveys Indian J. Psychol. Med. 2020 42 575 576 10.1177/0253717620957496 33354086 61. Smith G. Does Gender Influence Online Survey Participation? A Record-Linkage Analysis of University Faculty Online Survey Response Behavior ERIC Doc. Reprod. Serv. 2008 1 1 21 62. Bärebring L. Palmqvist M. Winkvist A. Augustin H. Gender differences in perceived food healthiness and food avoidance in a Swedish population-based survey: A cross sectional study Nutr. J. 2020 19 140 10.1186/s12937-020-00659-0 33375947 63. Bradbury J. Nicolaou M. Men eating female/healthy foods are judged as more feminine than men eating male/unhealthy foods Proc. Nutr. Soc. 2012 71 1 10.1017/S0029665112002054 22067138 64. Hardeman G. Font X. Nawijn J. The power of persuasive communication to influence sustainable holiday choices: Appealing to self-benefits and norms Tour. Manag. 2017 59 484 493 10.1016/j.tourman.2016.09.011 65. Stok F.M. de Ridder D.T. de Vet E. de Wit J.B. Don’t tell me what I should do, but what others do: The influence of descriptive and injunctive peer norms on fruit consumption in adolescents Br. J. Health Psychol. 2014 19 52 64 10.1111/bjhp.12030 23406475 66. Bechthold A. Boeing H. Tetens I. Schwingshackl L. Nöthlings U. Perspective: Food-Based Dietary Guidelines in Europe—Scientific Concepts, Current Status, and Perspectives Adv. Nutr. 2018 9 544 560 10.1093/advances/nmy033 30107475 67. Krishnaswamy K.M.D. Developing and implementing dietary guidelines in India Asia Pac. J. Clin. Nutr. 2008 17 66 69 18296304

PMC009xxxxxx/PMC9099599.txt

==== Front Nanomaterials (Basel) Nanomaterials (Basel) nanomaterials Nanomaterials 2079-4991 MDPI 10.3390/nano12091513 nanomaterials-12-01513 Article Two-Channel Charge-Kondo Physics in Graphene Quantum Dots https://orcid.org/0000-0002-0334-930X Minarelli Emma L. 12*† https://orcid.org/0000-0003-3906-7985 Rigo Jonas B. 12*† https://orcid.org/0000-0002-0652-2710 Mitchell Andrew K. 12* Kogan Eugene Academic Editor 1 School of Physics, University College Dublin, Dublin 4, Ireland 2 Centre for Quantum Engineering, Science, and Technology, University College Dublin, Dublin 4, Ireland * Correspondence: emma.minarelli@ucdconnect.ie (E.L.M.); jonas.rigo@ucdconnect.ie (J.B.R.); andrew.mitchell@ucd.ie (A.K.M.) † These authors contributed equally to this work. 29 4 2022 5 2022 12 9 151304 4 2022 27 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Nanoelectronic quantum dot devices exploiting the charge-Kondo paradigm have been established as versatile and accurate analogue quantum simulators of fundamental quantum impurity models. In particular, hybrid metal–semiconductor dots connected to two metallic leads realize the two-channel Kondo (2CK) model, in which Kondo screening of the dot charge pseudospin is frustrated. In this article, a two-channel charge-Kondo device made instead from graphene components is considered, realizing a pseudogapped version of the 2CK model. The model is solved using Wilson’s Numerical Renormalization Group method, uncovering a rich phase diagram as a function of dot–lead coupling strength, channel asymmetry, and potential scattering. The complex physics of this system is explored through its thermodynamic properties, scattering T-matrix, and experimentally measurable conductance. The strong coupling pseudogap Kondo phase is found to persist in the channel-asymmetric two-channel context, while in the channel-symmetric case, frustration results in a novel quantum phase transition. Remarkably, despite the vanishing density of states in the graphene leads at low energies, a finite linear conductance is found at zero temperature at the frustrated critical point, which is of a non-Fermi liquid type. Our results suggest that the graphene charge-Kondo platform offers a unique possibility to access multichannel pseudogap Kondo physics. Kondo effect graphene electronic transport quantum dots Irish Research Council through the Laureate Award 2017/2018IRCLA/2017/169 (AKM/JBR) Enterprise Partnership SchemeEPSPG/2017/343 (ELM) We acknowledge funding from the Irish Research Council through the Laureate Award 2017/2018 Grant IRCLA/2017/169 (AKM/JBR) and the Enterprise Partnership Scheme Grant EPSPG/2017/343 (ELM). ==== Body pmc1. Introduction The Kondo effect [1] was originally discussed in the context of local magnetic impurities such as Fe, embedded in non-magnetic metallic hosts such as Au. By progressively decreasing the temperature T, experimental measurements revealed an unexpected resistivity minimum, attributed to enhanced electronic scattering from the impurity local moments [2]. The low-energy physics of such systems is explained by the deceptively simple Kondo model, which features a single spin-12 local moment exchange coupled to a featureless bath of metallic, non-interacting conduction electrons. The “Kondo effect” refers to the universal physics of this model, appearing at T∼TK with TK an emergent low-energy scale called the Kondo temperature, in which the impurity spin is dynamically screened by a surrounding many-body entanglement cloud of conduction electrons [3]. Since then, variants of the basic Kondo effect that arise when magnetic impurities are embedded in unconventional host materials have been studied. Examples include ferromagnets [4] and superconductors [5], as well as topological materials such as topological insulators [6] or Dirac/Weyl semimetals [7]. However, local moments in graphene have attracted the most attention [8,9]. In neutral graphene, the Dirac point is at the Fermi level [10], and so, a spin-12 impurity couples to a bath of conduction electrons with a density of states (DoS) featuring a low-energy pseudogap ρ(ω)∼|ω|r with r=1. This has dramatic consequences for the resulting Kondo physics [9] due to the depletion of the low-energy degrees of freedom in graphene, which can participate in screening the impurity spin. Deeper insights into strongly correlated electron physics and Kondo physics have been gained from tunable circuit realizations of fundamental models in nanoelectronics devices, made possible by remarkable recent advances in nanofabrication and characterization techniques [11,12]. This provides a route to probing and manipulating quantum matter at the nanoscale in a way that would be impossible in bulk systems. In particular, semiconductor quantum dots (QDs) have been shown to behave like artificial atoms [13], with the extreme quantum confinement producing a discrete level structure and strong electron–electron interactions. Coupling such quantum dots to metallic electrodes at quantum point contacts (QPCs) gives rise to the Kondo effect at low temperatures [14,15,16], with a single local moment trapped on the dot facilitating spin-flip scattering of lead conduction electrons. In such devices, entanglement spreads across the QD and the leads in a macroscopic “Kondo cloud” [17,18], producing the famous Kondo resonance in electrical conductance [19]. The quantum transport properties of QDs can be tuned in situ by applying gate voltages to control the QPC transmissions and dot potential. A bias voltage can drive the system out of equilibrium. Quantum dot devices also allow more complex quantum impurity models to be realized experimentally by controlling the microscopic interactions. As such, they constitute a versatile platform to study a rich range of physics [12], including quantum phase transitions (QPTs) [20,21], emergent symmetries [22,23], and non-Fermi liquid (NFL) physics [24,25,26,27]. The two-channel Kondo (2CK) model [28] is a famous example, which embodies the frustration of the Kondo screening of a single impurity by two distinct channels of metallic conduction electrons and displays all these features [29]. The standard 2CK model Hamiltonian reads, (1) H^2CK=H^leads+J1S^·s^1+J2S^·s^2, where H^leads=∑ασkϵk†cασk†cασk† describes two leads α=1,2, each with spin σ=↑,↓ electrons with momentum k. In the original formulation of the 2CK model, the dispersion ϵk is taken to be linear at low energies such that the electronic DoS of the leads at the impurity position is flat. The metallic flat band approximation is typically employed for the free conduction electrons, ρ(ω)∼∑kδ(ω−ϵk)≡ρ0Θ(|ω|−D), describing a flat density of states ρ0 inside a band of half-width D. In Equation (1), S^ is a spin-12 operator for the impurity and s^α is a spin-12 operator for the spin density in lead α at the impurity position, such that H^2CK possesses SU(2) spin symmetry. The metallic 2CK model supports a QPT, with an NFL critical point at J1=J2 [29]. Signatures of the critical point in this model have been observed experimentally in semiconductor quantum dot devices [24,25]. More recently, a new nanoelectronics paradigm has emerged, based on charge-Kondo quantum dots [30,31,32,33,34]. In the standard setup, a large QD is coupled at QPCs to source and drain leads. These devices realize anisotropic multichannel Kondo models through Matveev’s mapping [35,36] of the macroscopic charge states of the large QD to an effective pseudospin that is flipped by electronic tunnelling at the QPCs. This approach has led to unprecedented control over the frustrated 2CK state and has uncovered the full renormalization group (RG) flow diagram through transport measurements [30]. Motivated by these developments, in this paper, we consider combining the two-channel charge-Kondo setup of [30] with the pseudogap Kondo physics of graphene in [9], to realize a novel two-channel pseudogap charge-Kondo effect. We envisage a charge-Kondo device made from graphene components (see Figure 1), such that the dot charge pseudospin is coupled to two channels of conduction electrons, each with the characteristic linear pseudogap DoS of graphene. This work is a theoretical exploration of such a system and its phase diagram. We characterize the phases and phase transitions through thermodynamic quantities and focus on experimentally relevant physical observables such as the conductance. However, we do not claim to address the practical complexities that will inevitably arise in the experimental realization of a graphene charge-Kondo device. We note that the generic properties of fully spin- and channel-symmetric two-channel pseudogap Kondo models were discussed in [37], although the r=1 linear pseudogap case relevant to graphene was not analysed in detail and a device realization was not proposed. Furthermore, our charge-Kondo implementation leads to crucial differences in the model and transport measurement setup, which have not previously been considered. These differences and our new results are highlighted in the following. 2. Model, Methods, and Observables We consider a two-channel charge-Kondo device in which both the quantum dot and the leads are made from graphene, as illustrated in Figure 1. We note that graphene quantum dots have been the topic of active experimental study recently [38,39,40]. We envisage a large dot tunnel coupled to leads α=1,2 at QPCs with transmission τα, which can be controlled in situ by gate voltages. A plunger gate voltage Vg controls the dot potential, and hence the dot filling. A decoherer is interjected between the leads via an Ohmic contact on the dot (black bar in Figure 1), which gives rise to a long dwell time and an effective continuum dot level spectrum (this was achieved in the experiments of [30] using a metallic component). This results in two effectively independent electronic reservoirs around each of the two QPCs; these form the two independent channels in the 2CK model. However, tunnelling events onto and off of the dot are correlated by the large dot charging energy, Ec. The whole device is operated in a strong magnetic field so that the electrons are effectively spinless (that is, the Zeeman splitting is the largest energy scale in the problem). The model Hamiltonian for the device illustrated in Figure 1 is given by (2) H^=H^leads+∑α=1,2∑kk′Jα(Q^+cαDk†cαLk′+cαLk′†cαDkQ^−)+Wα(cαLk†cαLk′+cαDk†cαDk′)+Ec(N^D−Ng)2, where H^leads=∑ασkϵk†cασk†cασk† describes the distinct conduction electron reservoirs around each QPC labelled by α=1,2 and with σ=L,D corresponding to lead or dot electrons (rather than physical spin ↑,↓). For graphene components, we used the dispersion ϵk±=±t1+4cos2a2kx+4cosa2kxcos3a2ky for the two bands, assumed here to be independent of α and σ, with nearest neighbour tunnelling matrix element t≈2.8 eV and lattice constant a≈2.46 Å [10]. The resulting DoS ρασ(ω)≡ρ(ω) has a bandwidth D=3t and possesses a linear pseudogap ρ(ω)∼|ω| for |ω|≪t. The terms proportional to Jα describe electronic tunnelling at the QPCs between leads and dot. The tunnelling matrix elements Jα are related to the bare (unrenormalized) QPC transmissions via [41] τα(ω)=4π2ραL(ω)ραD(ω)Jα2/[1+π2ραL(ω)ραD(ω)Jα2]2, which are in general energy-dependent for structured leads. States of the isolated dot with a macroscopic number of electrons ND are denoted |ND〉, with corresponding dot number operator N^D=∑αkcαDk†cαDk†≡∑NDND|ND〉〈ND|. Tunnelling at the QPCs changes the dot charge, which we describe [35] using the charge raising and lowering operators Q^±=∑ND|ND±1〉〈ND|. The dot has a finite charging energy that depends on the filling via the term proportional to Ec. The filling can be adjusted by tuning Ng in Equation (2), which was controlled in the experiment by the gate voltage Vg=Vg0−2EcNg/e. We define δVg=−2Ec(Ng−ND0−12)/e such that the macroscopic dot charge states |ND0〉 and |ND0+1〉 are degenerate at δVg=0. Potential scattering at the QPCs is described by the term proportional to Wα. Provided kBT,eδVg≪Ec, only the lowest two dot charge states |ND0〉 and |ND0+1〉 are accessible and relevant for transport. In this case, the dot charge operators become effective pseudospin-12 operators, Q^+→S^D+=|ND0+1〉〈ND0| and Q^−→S^D−=|ND0〉〈ND0+1|. Thus, S^D+ flips the dot charge pseudospin from ⇓ to ⇑, while S^D− flips it back. We also introduced the pseudospin operator S^Dz=12(|ND0+1〉〈ND0+1|−|ND0〉〈ND0|). Finally, we performed a trivial relabelling σ={L,D}→{↑,↓} such that the electronic operators become cαLk→cα↑k and cαDk→cα↓k. With this, we arrive at the effective pseudogap two-channel charge-Kondo (2CCK) model studied in this paper: (3) H^2CCK=H^leads+∑α=1,2∑kk′Jα(S^D+cα↓k†cα↑k′︸S^D+s^α−+cα↑k′†cα↓kS^D−)︸s^α+S^D−+Wα∑σcασk†cασk′+eδVgS^Dz. This model is a variant of the famous 2CK model, Equation (1)—but with a few important differences. Firstly, the DoS of the conduction electrons described by H^leads is not metallic, but has a low-energy r=1 pseudogap. Secondly, tunnelling at the QPCs gives an effective anisotropic exchange coupling between the dot charge pseudospin and the conduction electrons. The SU(2) symmetry of Equation (1) is broken in Equation (3) since the z-component of the coupling is missing. However, we found that this effective spin anisotropy is RG irrelevant in the two-channel pseudogap Kondo problem (just as for the single-channel [2,42] and two-channel [29] metallic case, as well as the single-channel pseudogap case [43]). Only the spin flip terms are important for the Kondo effect, and these are captured by the effective model. It should also be emphasized that the effective exchange couplings Jα originate from the QPC tunnellings; there is no underlying Anderson model, so the Jα need not be perturbatively small. In fact, since they are related to the QPC transmissions, they can become large simply by opening the QPCs [30]. This is important because Kondo physics is only realized in the pseudogap model at relatively large bare coupling strengths. Thirdly, the gate voltage δVg appears as an effective impurity magnetic field. Finally, we have an additional potential scattering term Wα. This is traditionally omitted in Equation (1) because potential scattering is RG irrelevant in the metallic Kondo problem [2]. However, we must keep it because potential scattering is known to be important in the single-channel pseudogap Kondo model [43]. Indeed, we find that it is strongly RG relevant in our two-channel pseudogap variant, Equation (3). Another important difference in terms of the experimental realization is the nature of the transport measurement. As illustrated in Figure 1, a series current of spinless electrons is measured between the physical source and drain leads through the dot, in response to a bias voltage. However, in the mapped spin model, this is an unconventional measurement: we effectively apply a bias between leads α=1,2, but only to the σ=↑ conduction electrons. Even though there is no charge current possible between leads in the original 2CK model Equation (1), the charge-Kondo setup Equation (3) allows an effective spin current to be measured. The AC linear response electrical conductance through the device is defined as (4) GC(ω,T)=〈I^2↑〉Vbias|Vbias→0 due to an oscillating bias described by H^bias=−eVbiascos(ωt)N^1↑ with AC frequency ω. Here, I^α↑=−eddtN^α↑ is the current operator for lead α (and σ=↑), while N^α↑=∑kcα↑k†cα↑k†. We obtain the AC linear conductance from the Kubo formula [44]: (5) GC(ω,T)=−Im〈〈I^1↑;I^2↑〉〉ω,Tω≡2πG0ωIm〈〈N^1↑;N^2↑〉〉ω,T, where 〈〈·;·〉〉 denotes a retarded real frequency correlation function evaluated at equilibrium and G0=e2/h is the conductance quantum (ℏ=1). The second equality in Equation (5) follows from the equations of motion and is found to greatly improve the accuracy of the numerical calculations [45]. Note that the system is not in proportionate coupling, and so, correlated electron transport coefficients cannot be expressed in terms of a Landauer-type formula [46] involving the dot spectral function. In addition to the conductance, we explored the phase diagram and RG fixed points (FPs) of the model using physical thermodynamical observables. We define the dot contribution to a thermodynamic quantity Ω at temperature T as ΩD(T)=Ω(T)−Ω0(T), where Ω(T) is calculated for the full lead–dot–lead system, while Ω0(T) is calculated only for the free conduction electrons (without the dot pseudospin). For the entropy SD(T), we used S(0)=−∂F(0)/∂T, with F(0)=−kBTlnZ(0) the free energy. Recently, this entropy has been extracted experimentally in similar quantum dot devices by exploiting a Maxwell relation connecting the entropy change for a process to measurable changes in the dot charge [33,47]. For the magnetic susceptibility kBTχD(T), we evaluated kBTχ(0)=〈(S^totz)2〉(0)−(〈S^totz〉(0))2 at zero field (δVg=0), with S^totz the z-projection of the total spin of the system. The role of particle–hole asymmetry will be assessed through the conduction electron “excess charge” Nα=〈N^α〉−〈N^α〉0 with N^α=∑σN^ασ. The dynamics of the system are characterized by the channel-resolved T-matrix, which describes how conduction electrons are scattered from the dot pseudospin. The T-matrix equation reads, (6) Gαβ(ω,T)−Gαβ0(ω)=Gαα0(ω)Tαβ(ω,T)Gββ0(ω), where Gαβ(ω,T) and Gαβ0(ω) are, respectively, the full and free retarded electronic Green’s functions at the dot position. Due to decoherence between the QPCs (resulting in separately conserved charge in each channel in Equation (3)), we have Gαβ,Gαβ0,Tαβ∝δαβ, and the T-matrix equation becomes channel-diagonal. Furthermore, −1πImGαα0(ω)=ρ(ω) is the free graphene DoS. In the following, we consider the spectrum of the T-matrix for channel α, defined as tα(ω,T)=−1πImTαα(ω,T). Numerical Renormalization Group The two-channel pseudogap charge-Kondo model, Equation (3), is solved using Wilson’s Numerical Renormalization Group (NRG) technique [3,48,49], which provides numerically exact access to the physical quantities discussed in the previous section. The first step is the logarithmic discretization of the conduction electron DoS and subsequent mapping of H^leads to Wilson chains [3,48]: (7) H^leads→H^leadsdisc=∑α,σ∑n=0∞tnfασn†fασn+1†+fασn+1†fασn†. The dot then couples to the end of the Wilson chains, at site n=0. The logarithmic discretization is parameterized by Λ, with the continuum description being recovered as Λ→1 (in this work, we used a standard choice of Λ=2.5). The key feature of the Wilson chain is the behaviour of the hopping parameters tn. For the metallic flat band, tn∼Λ−n/2 at large n [3]. This exponential energy-scale separation down the chain justifies a numerical scheme based on iterative diagonalization and truncation: starting from the dot, successive sites of the Wilson chain are coupled into the system, and this intermediate Hamiltonian is diagonalized. Only the lowest MK eigenstates at iteration n are used to construct the Hamiltonian at iteration n+1. High-energy states discarded at a given iteration do not affect the retained low-energy states at later iterations because of the ever-decreasing couplings tn. This constitutes an RG procedure since the physics of the system at successively lower energy scales is revealed as more Wilson orbitals are added. The computational complexity is constant as new Wilson orbitals are added (rather than exponentially growing) because the same number MK of states is kept at each step. Importantly, it was shown in [50] that although the detailed structure of the Wilson chain coefficients is modified in the pseudogap DoS case, the energy scale separation down the chain is maintained, and hence, the NRG can still be used in this case. We used the exact graphene DoS in this work rather than a pure pseudogap and kept MK=6000 states at each iteration. Dynamical quantities were calculated using the full-density-matrix NRG approach [49,51], established on the complete Anders–Schiller basis [52]. 3. Results and Discussion Having introduced the model and methods, we now discuss our NRG results in detail, starting with an overview of the phase diagram, RG flow diagram, and fixed point analysis. In the following, we confine our attention to the charge-degeneracy point δVg=0. We also introduce the channel-asymmetry parameter Δ=J2/J1≡W2/W1 and discuss the physics in the space of J≡J1, W≡W1, and Δ. Note that Δ=0 corresponds to the situation in which channel α=2 is decoupled on the level of the bare model, while for 0<Δ<1, both channels are coupled to the dot, but channel 1 couples more strongly. Δ=1 describes the frustrated two-channel situation. We confine our attention to regime 0≤Δ≤1, but it should be noted that Δ>1 simply corresponds to stronger coupling for channel 2, and the results follow from the duality 1↔2 and Δ↔1/Δ. We also assumed W>0. 3.1. Overview and Phase Diagram The schematic RG flow diagram in the space of (J,W,Δ) shown in the left panel of Figure 2 was deduced from non-perturbative NRG results and gives a good overview of the physics of Equation (3). In the right panel, we show the quantitative phase diagram in the (J,W) plane for different Δ, with the exact phase boundaries obtained with the NRG. We first briefly recapitulate the results for the single-channel (1CK) case obtained here for Δ=0 (see Figure 2: front plane in RG diagram on the left and turquoise line in the phase diagram on the right). The basic physics are well known from previous studies of the r=1 pseudogap Anderson and Kondo models [9,43,50,53,54,55,56], although note that our graphene charge-Kondo setup gives a spin-anisotropic model, and we used the full graphene DoS rather than a pure pseudogap. At W=0, there is no Kondo effect: the symmetric strong coupling (SSC) FP is unstable, and a finite potential scattering is required to screen the dot pseudospin. In this case, the system flows to weak coupling and eventually to the free local moment (LM) FP with asymptotically decoupled conduction electrons. At finite W, an LM phase with an emergent particle–hole symmetry can be realized, in which the dot and leads decouple at low temperatures (that is, W,J→0 under RG). However, at sufficiently strong bare W, the model supports a QPT to an asymmetric strong coupling (ASC) Kondo state, in which the dot pseudospin is screened and a single hole forms in the bath (that is, W,J→∞ under RG). However, the coupling J must overcome a critical threshold value (for Equation (3) JCRmin≃1.81D). For J<JCRmin, no Kondo state is possible at any W. For J≥JCRmin, and the ASC FP is stable for WCR−(J)<W<WCR+(J). In the lower branch, we find WCR−(J)∼1/J at large J≫JCRmin, such that infinitesimal particle–hole symmetry breaking W→0 is required at large bare coupling J→∞. Thus, although SSC is unstable, the system can flow arbitrarily close to it, before ultimately crossing over to either ASC or LM. The transition between ASC and LM is first-order [43,55] and controlled by a particle–hole asymmetric critical FP denoted ACR. The full NRG phase boundary for our model at Δ=0 is shown as the turquoise line in the right panel of Figure 2 and shows an interesting re-entrant behaviour back into the LM phase at large W (we are not aware of a detailed discussion of this in the literature, even though the same behaviour arises in the regular spin-isotropic pseudogap Kondo model). This is physically intuitive since J and W work antagonistically: at very large W, the bath orbital f1σ0 becomes depopulated, and hence, the exchange coupling to that site J gets “switched off”. Perturbative arguments suggest that the residual coupling to the f1σ1 bath orbital is then Jeff∼t02J/W2, which is consistent with WCR+(J)∼J for the upper branch of the phase boundary. This is indeed confirmed by NRG calculations. The main focus of this paper is the situation when the coupling to the second channel is switched on, Δ>0, where we find several differences from the pure 1CK case. We discuss 0<Δ<1 first. Importantly, we found that the same LM and ASC phases are accessible, with Δ flowing to zero under RG flow upon reducing the temperature or energy scale. Therefore, even though both channels are initially coupled to the dot (at high temperatures, we have a free channel degree of freedom α=1,2), any channel asymmetry leads asymptotically to the decoupling of the less strongly coupled channel α=2 (this can be regarded as “channel freezing” at low temperatures). In the ASC phase, the dot flows to strong coupling with the more strongly coupled channel α=1, while in the LM phase, channel α=1 also eventually decouples, leaving a free dot pseudospin and free conduction electrons. This is indicated by the flow arrows towards the front plane in the RG diagram, Figure 2 (left). However, at finite Δ, the topology of the phase diagram changes—see Figure 2 (right). We still have a finite threshold value of the coupling to realize ASC physics, JCRmin(Δ)>0 (which increases slightly from ≃1.81D at Δ=0 up to ≃2.47D as Δ→1). However, the critical phase boundary now also develops a finite threshold value of the potential scattering WCRmin(Δ)>0 (which reaches its maximum value ≃D as Δ→1). Even at strong bare coupling J→∞, a finite W is required to access the ASC phase. In fact, WCRmin occurs at an intermediate value of J; at large J, we find WCR−∼J in the lower branch. For even larger W, we again have re-entrant LM behaviour, with an upper branch of the phase boundary. For Δ>0, we therefore have large-J behaviour WCR−(J,Δ)=a−(Δ)J for the lower branch and WCR+(J,Δ)=a+(Δ)J for the upper branch. Interestingly, a+≈1 independent of Δ, while a− increases with increasing Δ, as shown in the inset to the right panel of Figure 2. However, a−(Δ)/a+<1 for all Δ (the ratio reaches its maximum ≈0.2 as Δ→1), meaning that the upper and lower phase boundaries never cross, and the ASC phase persists out to infinite J and W. The finite WCRmin also implies that there is no crossover from SSC to ASC for Δ>0. Finally, we examined the channel-frustrated case Δ=1 (see the middle plane of Figure 2 (left) and the purple line of Figure 2 (right)). Here, symmetry dictates a channel degeneracy down to T=0, and therefore no channel freezing. We found that the model supports an LM phase in which both channels flow symmetrically to weak coupling and to particle hole symmetry. However, the ASC FP is unstable since the Kondo effect and conduction electron hole in ASC occur in only one of the two channels. Instead, we have a frustrated asymmetric strong coupling (FASC) phase, with a free channel degree of freedom (a doubled version of the ASC FP, with the Kondo effect and conduction electron hole forming in either channel α=1 or 2). The critical point separating LM and FASC in the Δ=1 plane is denoted FACR. The frustrated FPs are delicate because they sit precisely on the separatrix between RG flow to states with dominant channel 1 for Δ<1 and flow to states with dominant channel 2 for Δ>1. Any finite perturbation |1−Δ| relieves the channel frustration and leads ultimately to channel freezing on the lowest energy scales. This QPT is also first-order; FACR is therefore tricritical since it sits at the boundary between LM, FASC, and ASC. In Table 1, we summarize the FPs discussed above in relation to Figure 2, classifying them according to their physical properties. These properties are extracted from the limiting behaviour of the full thermodynamic and dynamic observables presented in the following. 3.2. Thermodynamics and Fixed Points The temperature dependence of the dot contribution to entropy SD(T) and magnetic susceptibility TχD(T) are obtained from the NRG [48] and presented in Figure 3 for different channel asymmetries Δ. We focused on the behaviour near the critical points by fixing W and tuning J across the transition. From this, information on the fixed points is deduced. 3.2.1. Frozen Channel Degree of Freedom: 0≤Δ<1 We first consider the regime 0≤Δ<1 (left and middle columns of Figure 3). Solid lines show the behaviour in the ASC phase for J>JCR, dashed lines for the LM phase with J<JCR, and the orange line at the critical point J=JCR. In the LM phase, SD=ln(2) and TχD=14 at T=0 in all cases, characteristic of the asymptotically free spin-12 dot pseudospin. The excess conduction electron charge (not shown) is zero in both channels, suggesting an emergent particle–hole symmetry. This is confirmed by the analysis of the NRG many-particle level spectrum (finite size spectrum) at the LM FP, which is identical to that of the free leads. The dot remains unscreened in LM because of the depleted conduction electron DoS at low energies in graphene [9]. The FP Hamiltonian in the LM phase is therefore given by (8) H^LM=H^2CCKwithJ1=J2=W1=W2=Vg=0 In the ASC phase at T=0, we see quenched dot entropy SD=0 and TχD=0 in all cases, characteristic of Kondo singlet formation. However, the conduction electron excess charge is N1=−1 and N2=0 (for W>0), implying hole formation in the more strongly coupled channel α=1 (W1→∞ under RG), while the less strongly coupled channel α=2 recovers an effective low-energy particle–hole symmetry (W2→0 under RG). This suggests the screening mechanism in the generic two-channel case: as W1 grows under RG, the f1σ0 Wilson orbital becomes depopulated, thereby generating an effective coupling between the dot pseudospin and the Wilson f1σ1 orbital, Jeff∼t02J1/W12. However, the DoS of the f1σ1 orbital is modified by the hole forming at the n=0 site. With ρ(ω)≡−1πIm〈〈f1σ0;f1σ0†〉〉∼|ω| at low energies, we find ρeff(ω)≡−1πIm〈〈f1σ1;f1σ1†〉〉∼1/|ω| at low energies. Therefore, even though Jeff is perturbatively small, the effective DoS is strongly enhanced. The effective dimensionless RG flow parameter j1=ρeffJeff grows under RG and leads to Kondo screening of the dot. The Kondo scale for this process is strongly enhanced because of the diverging effective DoS [57], and we find in practice that throughout the ASC phase, TK∼D (since JCRmin>D). However, no hole forms in the weakly coupled channel, and so, j2=ρJ2 remains small due to the depleted bare DoS in channel 2 and flows under RG to weak coupling. This argument shows that the Kondo singlet must form in the same channel in which the hole forms. This was confirmed by the analysis of the NRG level spectrum. In general, we therefore have two distinct ASC phases and two distinct ASC fixed points, depending on whether Δ<1 or Δ>1. For Δ<1, the hole–singlet complex forms in channel α=1, and channel α=2 decouples (FP denoted ASC1), while for Δ>1 (ASC2), it is the other way around. The ASCα FP Hamiltonian obtained when channel α is more strongly coupled reads (9) H^ASCα=H^leads+JαS^D+fα↓1†fα↑1†+S^D−fα↑1†fα↓1†+Wα∑σfασ0†fασ0†withJα,Wα→∞. We now consider the situation in the close vicinity of the QPT, by fixing W and tuning J. At the critical point itself (orange line, J=JCR), we found SD=ln(3) and TχD=16 at T=0. This suggests a level-crossing transition in which the critical FP ACR comprises uncoupled sectors corresponding to LM and ASC. This gives an overall dot ground state degeneracy of 2+1=3 states (2 for LM, 1 for ASC) consistent with the ln(3) entropy and a magnetic susceptibility (14+14+0)/3=16 (corresponding to the average of (Sz)2 for these three degenerate states). This is further supported by the conduction electron excess charge N1=−13, since a single hole appears in channel 1 for only one of the three degenerate ground states (and N2=0 for the decoupled free channel 2). The first-order transition is also consistent with the linear crossover scale T*∼|J−JCR|, describing the flow from ACR to either LM or ASC due to a small detuning perturbation. This scale is evident in Figure 3 by the sequence of lines for different (J−JCR). Indeed, one can cross the QPT by fixing J and tuning W through WCR, which also gives a linear scale T*. We also checked this behaviour along the entire critical phase boundary lines (JCR,WCR) in Figure 2 for different Δ. We find, (10) T*=b|J−JCR|+c|W−WCR|, where b≡b(JCR,WCR,Δ) and c≡c(JCR,WCR,Δ). This implies a universal scaling in terms of a single reduced parameter T/T*, independent of the combination of bare perturbations that act. The FP Hamiltonian describing the critical point is, (11) H^ACRα=1+τ^z2H^LM+1−τ^z2H^ASCα. where the α label denotes the more strongly coupled lead with which the dot forms the Kondo effect in ASC and τ^z is a Pauli-z operator. In Equation (11), τz=+1 gives the doubly degenerate LM ground state (H^LM given in Equation (8)), while τz=−1 gives the ASC ground state (H^ASC given in Equation (9)). At the ACR FP, the three many-body ground states are degenerate and uncoupled (τ^ has no dynamics). Since ACR is unstable, we also consider the leading RG-relevant perturbations to the FP Hamiltonian, δH^ACRα∼T*τ^z, which has the effect of biasing towards either the LM or ASC ground states on the scale of T*. The qualitative behaviour of the thermodynamics shown in Figure 3 for Δ=0 and Δ=0.8 is similar, but it should be noted that both channels are involved for Δ≠0 at finite temperatures. However, the less strongly coupled channel decouples asymptotically because finite 0<Δ<1 flows to Δ=0 under RG upon reducing the temperature or energy scale (see Figure 2 (left)). 3.2.2. Frustrated Channel Degree of Freedom: Δ=1 We turn now to the frustrated case Δ=1, with pristine channel symmetry—see Figure 3, right column. Although the T=0 entropy is SD=ln(2) everywhere except on the phase boundary (top right panel of Figure 3), the origin of the ground state degeneracy is different in the two phases separated by it. In the LM phase (realized for J<JCR), we again have a free dot pseudospin decoupled from two symmetric baths of free conduction electrons; the ln(2) entropy here derives from the free dot pseudospin-12 degree of freedom. This is confirmed by the magnetic susceptibility in this phase, which reaches TχD=14 (dashed lines, bottom right panel of Figure 3). The other phase (realized for J>JCR) is described by the FASC FP: due to the channel symmetry, the ASC state can form in either channel α=1,2. The ln(2) entropy in this case derives from the free channel degree of freedom [37], which embodies the choice of forming the hole–singlet complex of ASC with either of the two channels. This is reflected in the T=0 value of TχD=0 in the FASC phase (solid lines, bottom right panel of Figure 3), since the dot pseudospin is Kondo screened in both of the degenerate ground states. Furthermore, we found that the average conduction electron excess charge in FASC is Nα=12 for both channels—that is, a single hole forms, with equal probability to be in either channel 1 or 2. A Kondo strong coupling state involving both channels simultaneously is not stable. To see this, consider two holes forming symmetrically in the f1σ0 and f2σ0 Wilson n=0 orbitals (W1=W2→∞) and effective Kondo couplings J1,eff=J2,eff→∞ between the dot pseudospin and the residual Wilson n=1 orbitals f1σ1 and f2σ1, which have an effective DoS ρeff(ω)∼1/|ω|—a channel symmetric version of the usual hole–singlet mechanism in ASC as described in the previous section. However, the dot entropy is not quenched in this case, since the ground state of the complex is a spin-doublet. This effective doublet state couples to the Wilson n=2 orbitals f1σ2 and f2σ2. However, since the DoS of these sites is again ∼|ω|, the effective local moment cannot be screened, and the system flows to the LM FP. The dot pseudospin can only be screened by an asymmetric ASC state. Channel symmetry is restored by having two such degenerate states, one in each channel. The FASC FP Hamiltonian comprises a combination of H^ASC1 and H^ASC2 from Equation (9), controlled by an emergent channel degree of freedom α^: (12) H^FASC=1+α^z2H^ASC1+1−α^z2H^ASC2. Here, α^z is a Pauli-z operator that selects ASC1 when αz=+1 and ASC2 when αz=−1. Restricting to the symmetric Δ=1 plane, FASC is stable. However, there is an instability with respect to breaking channel symmetry (not shown), since then, either ASC1 or ASC2 will be selected on the lowest energy scales. A finite perturbation |1−Δ| generates a flow from FASC to ASC1 or ASC2; from the NRG, we found that this QPT is also first-order. The low-energy scale determining the crossover is TΔ∼|1−Δ|. This can be captured in the effective model by including the leading RG-relevant perturbation to the FASC FP, δH^FASC∼TΔα^z. Finally, we considered the quantum critical point in the Δ=1 plane between LM and FASC. Here, we found a level-crossing (first-order) transition, with entropy SD=ln(4) and magnetic susceptibility TχD=18 at the FACR FP [37], which derives from the composition of uncoupled LM and FASC sectors. We have two spin-12 states from the LM degenerating with two spin-singlet states with a free channel degree of freedom in FASC. The excess conduction electron charge is therefore Nα=−14 per channel. We describe the FACR FP with the Hamiltonian: (13) H^FACR=1+τ^z2H^LM+1−τ^z2H^FASC, where H^LM is given in Equation (8) and H^FASC in Equation (12), and we introduced the operator τ^z to distinguish the sectors, similar to Equation (11). As with ACR, the FP is destabilized by RG-relevant detuning perturbations that favour either LM of FASC, which collectively generate the scale T* given in Equation (10). This leads to an FP correction δH^FACR*∼T*τ^z. This is shown by the sequence of lines in the right column of Figure 3. However, FACR is also destabilized by relieving the channel frustration through the perturbation |1−Δ|, which generates the scale TΔ, since FACR contains an FASC sector with this instability. Therefore, FACR has a second RG relevant correction δH^FACRΔ∼TΔα^z. FACR is in this sense tricritical since it sits between LM, FASC, and ASC. 3.3. Dynamics and Transport We now discuss the low-temperature behaviour of the scattering T-matrix and linear response AC electrical conductance in the graphene 2CCK device—see Figure 4. We first considered the T=0 spectrum of the T-matrix as a function of energy in the top row of Figure 4, for the channel asymmetric case Δ=0.8 (left) and frustrated case Δ=1 (right). In all cases, we identified an emergent low-energy scale λ (which is ≈10−4D for the parameters chosen), which characterizes the RG flow through a crossover behaviour in the pseudogap dynamics [37,43]. Deep in the LM phase (blue solid and dotted lines), the bare potential scattering W modifies the bare conduction electron pseudogap DoS of graphene ρ(ω)∼|ω|, to give an effective DoS ρeff(ω)∼1/|ω| (π/2 phase shift) up to logarithmic corrections. This produces leading behaviour in the T-matrix tα(ω,0)∼1/|ω|, as seen in Figure 4 for |ω|≫λ. However, under RG, W→0 in the LM phase; this flow is controlled by the scale λ. Therefore, on the scale of λ, the effective DoS returns to ∼|ω| (zero phase shift), and hence, tα(ω,0)∼|ω| for |ω|≪λ. Since the emergent particle–hole symmetry in the LM phase occurs in both channels for any Δ, we see the same behaviour for t1(ω,0) and t2(ω,0) for both Δ=0.8 and 1. In the ASC phase for Δ=0.8, the weakly coupled channel α=2 shows the same behaviour as LM since it decouples from the dot and gains particle–hole symmetry. For the strongly coupled channel α=1, we have the hole–singlet mechanism in which both the effective W,J→∞. Counterintuitively, we again see similar dynamical behaviour as for LM. This is because, for |ω|≫λ, we have a developing conduction electron hole, which gives t1(ω,0)∼1/|ω|, while for |ω|≪λ, the Kondo singlet forming with the n=1 Wilson orbital effectively removes a second site from the bath. The remaining conduction electrons experience a π phase shift from the modified boundary, and the effective DoS is back to ∼|ω|. Therefore, in ASC, we also have t1(ω,0)∼|ω| for |ω|≪λ. That we have identical behaviour for Δ=1 (in both channels) confirms that FASC is indeed a superposition of ASC1 and ASC2, as argued above. On the lowest energy scales, we have tα(ω,0)∼|ω| in both channels, at any Δ, and in either phase. Given the bare DoS ρ(ω)∼|ω|, this confirms that both phases are regular Fermi liquids, with well-defined (long-lived) quasiparticles [2,43,55]. More interesting is the behaviour at the critical point (F)ACR, since here, we have both spin and charge fluctuations associated with the degenerate LM and (F)ASC ground states. A new dynamical scale is generated, λCR∼λ2/D (≈10−8D for the chosen parameters), which characterizes the low-energy RG flow [37,43]. We found from the NRG that in the channel-asymmetric graphene 2CCK model (e.g., at Δ=0.8, as shown), the T-matrix of the more strongly coupled channel α=1 diverges at low energies. Specifically, t1(ω,0)∼1/[|ω/λCR|×ln2(|ω/λCR|)] as |ω|→0 (solid green lines in the top panels of Figure 4) [58], indicating that ACR is an NFL FP. The dynamical crossover, and hence the minimum in t1(ω,0), occurs on the scale of |ω|∝λCR. However, the weakly coupled channel α=2 has FL correlations t2(ω,0)∼|ω| as |ω|→0 (dotted green lines), confirming that it decouples from the critical complex formed from the dot and channel 1. In the frustrated case Δ=1, both channels behave identically—and both exhibit the same NFL critical divergence at low energies. This again suggests that FACR comprises two copies of ACR, one in each channel. The enhanced conduction electron scattering at the critical point has implications for the conductance, as now shown. In the bottom row of Figure 4, we plot the T=0 dynamical AC conductance as a function of AC driving frequency ω, for the same set of systems. The DC conductance was obtained in the ω→0 limit, which we considered first. In the charge-Kondo system, series transport proceeds by the following mechanism: an electron tunnels from the source lead onto the dot (say at QPC α=1), thus flipping the dot charge pseudospin from ⇓ to ⇑. A second electron then tunnels from the dot to the drain lead (at QPC α=2), thus flipping the dot charge pseudospin back to ⇓ and “resetting” the device, ready for the transport of another electron. A bias voltage between the source and drain produces a net current flow. The amplitude for such a process depends on the conduction electron density of states ρ(ω) and the tunnelling rate at the QPCs. For graphene, we have ρ(ω)∼|ω| at low energies, suggesting that the low-temperature DC conductance should vanish, since there are not enough low-energy electrons in the graphene leads to tunnel through the nanostructure. On the other hand, the tunnelling rate gets renormalized by the interactions (the energy-dependent scattering at the QPCs is characterized by the T-matrices discussed above). Indeed, strong renormalization of the bare QPC transmission at low temperatures due to Kondo physics was measured experimentally in the metallic leads version of the present system in [30]. The measured DC conductance of the graphene 2CCK involves a subtle interplay between the conduction electron DoS and interaction-renormalized scattering rates. We expect the T=0 DC conductance to vanish in all channel-asymmetric systems because the less strongly coupled channel always decouples on the lowest energy scales. Both leads must remain coupled to ensure a finite series current. This is indeed seen in the ω→0 limit of each of the curves in the bottom left panel of Figure 4 for Δ=0.8. However, in the frustrated (channel-symmetric) case Δ=1, both channels remain coupled down to T=0. Although the scattering rates and bare DoS both vanish as ∼|ω| in the LM and FASC phases, implying a suppression of DC conductance, at the critical point FACR, the electronic scattering diverges as |ω|→0. We found from the NRG that these effects conspire to give a finite linear DC conductance in this case—see the green line in the bottom right panel of Figure 4. For an AC bias, the conductance is measured as a function of the driving frequency ω. Conductance resonances are expected when the AC frequency matches the QPC tunnelling rate. At high energies, the pseudospin flip rate in the 2CCK model is given by the bare J (or effective Jeff). We therefore expect to see a peak in the AC conductance when |ω|∼J,Jeff; this is observed from NRG results in the LM, ASC, and FASC phases in Figure 4. However, at low energies |ω|≪λ, the pseudospin flip rate is renormalized, and we found GC(ω,0)∼ω2 in these cases, independent of Δ. At the critical point ACR for 0<Δ<1, both charge and spin fluctuations give an enhanced AC conductance around |ω|∼λ. However, channel α=2 decouples for |ω|≪λ, and so, the conductance also decays at low frequencies. We found from the NRG a slow attenuation GC(ω,0)∼−1/ln|ω| in this regime. However, in the channel-symmetric case Δ=1 at the critical point FACR, GC(ω,0)∼const. for |ω|≪λCR. The finite dynamical conductance here persists down to the DC limit. This is the smoking gun signature of the NFL-frustrated critical point in the graphene 2CCK system. 4. Conclusions and Outlook In this paper, we proposed a charge-Kondo quantum dot device made from graphene components. The novel feature of such a system is that it realizes a linear-pseudogap two-channel Kondo model in a tunable nanoelectronics circuit. This exotic system has a complex phase diagram in the space of dot–lead coupling strength, potential scattering, and channel asymmetry, which we analysed in detail using the NRG. In particular, we uncovered a channel-frustrated Kondo phase, with a non-Fermi liquid quantum critical point at the first-order quantum phase transition. Despite the depleted electronic density of the neutral graphene leads at low energies, critical fluctuations give rise to diverging scattering rates at the critical point, which produce a finite conductance even as T→0. The model supports other interesting, but as-yet unexplored regimes. We confined our attention to the dot charge degeneracy point δVg=0; however, finite δVg appears in the effective model like a magnetic field on the dot pseudospin. One could also investigate the effect of doping/gating the graphene so that the Fermi level is not at the Dirac point. This will give rise to a quantum phase transition between metallic 2CK and pseudogap 2CK. Other physical quantities could also be investigated, such as thermoelectric transport upon including a temperature gradient between leads. The pseudogap 2CCK model we studied theoretically here is likely a simplified description of any real graphene charge-Kondo nanoelectronics device. There may be complexities and subtleties in an experimental realization that were not included in our model or analysis. For example, we assumed that the conduction electrons on both the leads and dot have the same DoS. In particular, gate voltage tuning of the dot to achieve charge-degeneracy, and the addition of the decoherer, may affect the dot electronic DoS. However, we do not expect our basic results to be qualitatively modified by this because the Kondo exchange interaction derives from energy-dependent QPC transmission τ(ω), and hence involves the DoS of both the lead and dot. Therefore, even if only the lead DoS is pseudogapped at low energies, an effective pseudogap Kondo model should still result. To make a quantitative connection to experiments, such effects would have to be taken into account, as well as the possible involvement of more than just two dot charge states (that is, relaxing the condition T≪EC). We believe the predicted conductance signature of the frustrated quantum critical point should however still be observable in experiments. Author Contributions Conceptualization, A.K.M.; Formal analysis, E.L.M., J.B.R. and A.K.M.; Investigation, E.L.M. and J.B.R.; Methodology, J.B.R. and A.K.M.; Supervision, A.K.M.; Writing—original draft, E.L.M.; Writing—review & editing, A.K.M. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Data available from the authors upon request. Conflicts of Interest The authors declare no conflict of interest. Abbreviations The following abbreviations are used in this manuscript: QD quantum dot QPC quantum point contact QPT quantum phase transition QCP quantum critical point 2CK two-channel Kondo NRG Numerical Renormalization Group FL Fermi liquid NFL non-Fermi liquid DoS density of states RG renormalization group FP fixed point LM local moment (F)ALM (frustrated) asymmetric local moment (F)ASC (frustrated) asymmetric strong coupling (F)SSC (frustrated) symmetric strong coupling (F)ACR (frustrated) asymmetric critical Figure 1 Schematic of the two-channel graphene charge-Kondo quantum dot system. A net current flows from source to drain graphene leads through the large graphene dot in response to a bias voltage. A gate voltage Vg controls the dot filling. The black bar denotes the decoherer. Figure 2 Left: RG flow diagram for the pseudogap 2CCK model Equation (3), in the space of effective exchange coupling J, potential scattering W, and channel asymmetry Δ. Stable (unstable) FPs denoted as blue (red) squares. Δ=0 (1) is the pure single-channel (frustrated two-channel) model. For an explanation of the FPs, see the text. Right: Full NRG phase diagram for different Δ. The enclosed region in each case is the Kondo-screened ASC phase (frustrated FASC for Δ=1); the exterior region is the unscreened LM phase. The inset shows the asymptotic behaviour of phase boundaries; see the text. Figure 3 Dot contribution to thermodynamic quantities for the graphene 2CCK model obtained by the NRG. Top row entropy SD(T); bottom row: magnetic susceptibility TχD(T). Left, middle, and right columns correspond to Δ=0 (pure 1CCK), Δ=0.8 (asymmetric 2CCK), and Δ=1 (symmetric 2CCK), respectively. Shown for fixed W=2D, varying J across the QPT according to the colour scale, with solid lines for J>JCR in the ASC (FASC) phase and dashed lines for J<JCR in the LM phase. Orange lines show the behaviour at the ACR (FACR) critical point. Figure 4 NRG results for dynamics and transport in the graphene 2CCK model at T=0. Top row: channel-resolved spectrum of the T-matrix tα(ω,0). Bottom row: linear response AC electrical conductance GC(ω,0). Left: channel asymmetry Δ=0.8; right: frustrated case Δ=1. Model parameters: J/D=10 with W/D=1 (LM); 12 (ASC); 12 (FASC); ≃8.605 (ACR); ≃8.572 (FACR). nanomaterials-12-01513-t001_Table 1 Table 1 Classification of FPs according to their physical observables, with * denoting unstable FPs. Asymmetry Fixed Point SD(T=0) TχD(T=0) N1 t1(ω,T→0) GC(ω,T→0) ∀Δ LM line ln2 1/4 0 |ω| ω2 0≤Δ<1 ASC 0 0 −1 |ω| ω2 0≤Δ<1 ACR * ln3 1/6 −1/3 1/ωln2(λCR/ω) 0 Δ=1 FASC * ln2 0 −1/2 |ω| ω2 Δ=1 FACR * ln4 1/8 −1/4 1/ωln2(λCR/ω) const Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Kondo J. Resistance minimum in dilute magnetic alloys Prog. Theor. Phys. 1964 32 37 49 10.1143/PTP.32.37 2. Hewson A.C. The Kondo Problem to Heavy Fermions Cambridge University Press Cambridge, UK 1997 Volume 2 3. Wilson K.G. The renormalization group: Critical phenomena and the Kondo problem Rev. Mod. Phys. 1975 47 773 10.1103/RevModPhys.47.773 4. Martinek J. Utsumi Y. Imamura H. Barnaś J. Maekawa S. König J. Schön G. Kondo effect in quantum dots coupled to ferromagnetic leads Phys. Rev. Lett. 2003 91 127203 10.1103/PhysRevLett.91.127203 14525397 5. Franke K. Schulze G. Pascual J. Competition of superconducting phenomena and Kondo screening at the nanoscale Science 2011 332 940 944 10.1126/science.1202204 21596987 6. Mitchell A.K. Schuricht D. Vojta M. Fritz L. Kondo effect on the surface of three-dimensional topological insulators: Signatures in scanning tunnelling spectroscopy Phys. Rev. B 2013 87 075430 10.1103/PhysRevB.87.075430 7. Mitchell A.K. Fritz L. Kondo effect in three-dimensional Dirac and Weyl systems Phys. Rev. B 2015 92 121109 10.1103/PhysRevB.92.121109 8. Chen J.H. Li L. Cullen W.G. Williams E.D. Fuhrer M.S. Tunable Kondo effect in graphene with defects Nat. Phys. 2011 7 535 538 10.1038/nphys1962 9. Fritz L. Vojta M. The physics of Kondo impurities in graphene Rep. Prog. Phys. 2013 76 032501 10.1088/0034-4885/76/3/032501 23411583 10. Neto A.C. Guinea F. Peres N.M. Novoselov K.S. Geim A.K. The electronic properties of graphene Rev. Mod. Phys. 2009 81 109 10.1103/RevModPhys.81.109 11. Heinrich A.J. Oliver W.D. Vandersypen L.M. Ardavan A. Sessoli R. Loss D. Jayich A.B. Fernandez-Rossier J. Laucht A. Morello A. Quantum-coherent nanoscience Nat. Nanotechnol. 2021 16 1318 1329 10.1038/s41565-021-00994-1 34845333 12. Barthelemy P. Vandersypen L.M. Quantum dot systems: A versatile platform for quantum simulations Ann. Phys. 2013 525 808 826 10.1002/andp.201300124 13. Kastner M.A. Klein O. Lyszczarz T.M. Mankiewich P.M. Shaver D.C. Wind S. Abusch-Magder D. Goldhaber-Gordon D.J. Morgan N.Y. Artificial Atoms Technical Report Research Laboratory of Electronics (RLE) at the Massachusetts Institute of Technology (MIT) Cambridge, MA, USA 1994 14. Goldhaber-Gordon D. Shtrikman H. Mahalu D. Abusch-Magder D. Meirav U. Kastner M.A. Kondo effect in a single-electron transistor Nature 1998 391 156 159 10.1038/34373 15. Cronenwett S.M. Oosterkamp T.H. Kouwenhoven L.P. A tunable Kondo effect in quantum dots Science 1998 281 540 544 10.1126/science.281.5376.540 9677192 16. Van der Wiel W. Franceschi S.D. Fujisawa T. Elzerman J. Tarucha S. Kouwenhoven L. The Kondo effect in the unitary limit Science 2000 289 2105 2108 10.1126/science.289.5487.2105 11000108 17. Mitchell A.K. Becker M. Bulla R. Real-space renormalization group flow in quantum impurity systems: Local moment formation and the Kondo screening cloud Phys. Rev. B 2011 84 115120 10.1103/PhysRevB.84.115120 18. Yoo G. Lee S.S. Sim H.S. Detecting Kondo entanglement by electron conductance Phys. Rev. Lett. 2018 120 146801 10.1103/PhysRevLett.120.146801 29694152 19. Pustilnik M. Glazman L. Kondo effect in quantum dots J. Phys. Condens. Matter 2004 16 R513 10.1088/0953-8984/16/16/R01 20. Vojta M. Impurity quantum phase transitions Philos. Mag. 2006 86 1807 1846 10.1080/14786430500070396 21. Mitchell A.K. Jarrold T.F. Logan D.E. Quantum phase transition in quantum dot trimers Phys. Rev. B 2009 79 085124 10.1103/PhysRevB.79.085124 22. Keller A. Amasha S. Weymann I. Moca C. Rau I. Katine J. Shtrikman H. Zaránd G. Goldhaber-Gordon D. Emergent SU (4) Kondo physics in a spin–charge-entangled double quantum dot Nat. Phys. 2014 10 145 150 10.1038/nphys2844 23. Mitchell A.K. Liberman A. Sela E. Affleck I. SO (5) non-Fermi liquid in a Coulomb box device Phys. Rev. Lett. 2021 126 147702 10.1103/PhysRevLett.126.147702 33891454 24. Potok R. Rau I. Shtrikman H. Oreg Y. Goldhaber-Gordon D. Observation of the two-channel Kondo effect Nature 2007 446 167 171 10.1038/nature05556 17344849 25. Keller A. Peeters L. Moca C. Weymann I. Mahalu D. Umansky V. Zaránd G. Goldhaber-Gordon D. Universal Fermi liquid crossover and quantum criticality in a mesoscopic system Nature 2015 526 237 240 10.1038/nature15261 26450057 26. Mitchell A.K. Sela E. Universal low-temperature crossover in two-channel Kondo models Phys. Rev. B 2012 85 235127 10.1103/PhysRevB.85.235127 27. Mitchell A.K. Sela E. Logan D.E. Two-channel Kondo physics in two-impurity Kondo models Phys. Rev. Lett. 2012 108 086405 10.1103/PhysRevLett.108.086405 22463550 28. Nozieres P. Blandin A. Kondo effect in real metals J. Phys. 1980 41 193 211 10.1051/jphys:01980004103019300 29. Affleck I. Ludwig A.W. Critical theory of overscreened Kondo fixed points Nucl. Phys. B 1991 360 641 696 10.1016/0550-3213(91)90419-X 30. Iftikhar Z. Jezouin S. Anthore A. Gennser U. Parmentier F. Cavanna A. Pierre F. Two-channel Kondo effect and renormalization flow with macroscopic quantum charge states Nature 2015 526 233 236 10.1038/nature15384 26450056 31. Mitchell A.K. Landau L. Fritz L. Sela E. Universality and scaling in a charge two-channel Kondo device Phys. Rev. Lett. 2016 116 157202 10.1103/PhysRevLett.116.157202 27127983 32. Iftikhar Z. Anthore A. Mitchell A. Parmentier F. Gennser U. Ouerghi A. Cavanna A. Mora C. Simon P. Pierre F. Tunable quantum criticality and super-ballistic transport in a “charge” Kondo circuit Science 2018 360 1315 1320 10.1126/science.aan5592 29724906 33. Han C. Iftikhar Z. Kleeorin Y. Anthore A. Pierre F. Meir Y. Mitchell A.K. Sela E. Fractional entropy of multichannel Kondo systems from conductance-charge relations arXiv 2021 2108.12878 10.1103/PhysRevLett.128.146803 35476492 34. Pouse W. Peeters L. Hsueh C.L. Gennser U. Cavanna A. Kastner M.A. Mitchell A.K. Goldhaber-Gordon D. Exotic quantum critical point in a two-site charge Kondo circuit arXiv 2021 2108.12691 35. Matveev K. Coulomb blockade at almost perfect transmission Phys. Rev. B 1995 51 1743 10.1103/PhysRevB.51.1743 36. Furusaki A. Matveev K. Theory of strong inelastic cotunnelling Phys. Rev. B 1995 52 16676 10.1103/PhysRevB.52.16676 9981071 37. Schneider I. Fritz L. Anders F.B. Benlagra A. Vojta M. Two-channel pseudogap Kondo and Anderson models: Quantum phase transitions and non-Fermi liquids Phys. Rev. B 2011 84 125139 10.1103/PhysRevB.84.125139 38. Bacon M. Bradley S.J. Nann T. Graphene quantum dots Part. Part. Syst. Charact. 2014 31 415 428 10.1002/ppsc.201300252 39. Yan Y. Gong J. Chen J. Zeng Z. Huang W. Pu K. Liu J. Chen P. Recent advances on graphene quantum dots: From chemistry and physics to applications Adv. Mater. 2019 31 1808283 10.1002/adma.201808283 40. Cai L. Zhang Z. Xiao H. Chen S. Fu J. An eco-friendly imprinted polymer based on graphene quantum dots for fluorescent detection of p-nitroaniline RSC Adv. 2019 9 41383 41391 10.1039/C9RA08726E 35541579 41. Lebanon E. Schiller A. Anders F.B. Coulomb blockade in quantum boxes Phys. Rev. B 2003 68 041311 10.1103/PhysRevB.68.041311 42. Kogan E. Poor man’s scaling: Anisotropic Kondo and Coqblin–Schrieffer models J. Phys. Commun. 2018 2 085001 10.1088/2399-6528/aad484 43. Fritz L. Vojta M. Phase transitions in the pseudogap Anderson and Kondo models: Critical dimensions, renormalization group, and local-moment criticality Phys. Rev. B 2004 70 214427 10.1103/PhysRevB.70.214427 44. Izumida W. Sakai O. Shimizu Y. Many body effects on electron tunnelling through quantum dots in an Aharonov-Bohm circuit J. Phys. Soc. Jpn. 1997 66 717 726 10.1143/JPSJ.66.717 45. Minarelli E.L. Rigo J.B. Mitchell A.K. Linear response quantum transport through interacting multi-orbital nanostructures 2022 in preparation 46. Meir Y. Wingreen N.S. Landauer formula for the current through an interacting electron region Phys. Rev. Lett. 1992 68 2512 10.1103/PhysRevLett.68.2512 10045416 47. Child T. Sheekey O. Lüscher S. Fallahi S. Gardner G.C. Manfra M. Kleeorin Y. Meir Y. Folk J. Entropy measurement of a strongly correlated quantum dot arXiv 2021 2110.14158 48. Bulla R. Costi T.A. Pruschke T. Numerical renormalization group method for quantum impurity systems Rev. Mod. Phys. 2008 80 395 450 10.1103/RevModPhys.80.395 49. Weichselbaum A. Von Delft J. Sum-rule conserving spectral functions from the numerical renormalization group Phys. Rev. Lett. 2007 99 076402 10.1103/PhysRevLett.99.076402 17930909 50. Bulla R. Pruschke T. Hewson A. Anderson impurity in pseudo-gap Fermi systems J. Phys. Condens. Matter 1997 9 10463 10.1088/0953-8984/9/47/014 51. Peters R. Pruschke T. Anders F.B. Numerical renormalization group approach to Greenâ’s functions for quantum impurity models Phys. Rev. B 2006 74 245114 10.1103/PhysRevB.74.245114 52. Anders F.B. Schiller A. Real-time dynamics in quantum-impurity systems: A time-dependent numerical renormalization-group approach Phys. Rev. Lett. 2005 95 196801 10.1103/PhysRevLett.95.196801 16384007 53. Gonzalez-Buxton C. Ingersent K. Renormalization-group study of Anderson and Kondo impurities in gapless Fermi systems Phys. Rev. B 1998 57 14254 10.1103/PhysRevB.57.14254 54. Logan D.E. Glossop M.T. A local moment approach to magnetic impurities in gapless Fermi systems J. Phys. Condens. Matter 2000 12 985 10.1088/0953-8984/12/6/320 55. Vojta M. Fritz L. Upper critical dimension in a quantum impurity model: Critical theory of the asymmetric pseudogap Kondo problem Phys. Rev. B 2004 70 094502 10.1103/PhysRevB.70.094502 56. Vojta M. Fritz L. Bulla R. Gate-controlled Kondo screening in graphene: Quantum criticality and electron-hole asymmetry EPL (Europhys. Lett.) 2010 90 27006 10.1209/0295-5075/90/27006 57. Mitchell A.K. Vojta M. Bulla R. Fritz L. Quantum phase transitions and thermodynamics of the power-law Kondo model Phys. Rev. B 2013 88 195119 10.1103/PhysRevB.88.195119 58. Vojta M. Bulla R. Kondo effect of impurity moments in d-wave superconductors: Quantum phase transition and spectral properties Phys. Rev. B 2001 65 014511 10.1103/PhysRevB.65.014511

PMC009xxxxxx/PMC9099600.txt

==== Front Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods11091368 foods-11-01368 Article Comparison of Effects from Ultrasound Thawing, Vacuum Thawing and Microwave Thawing on the Quality Properties and Oxidation of Porcine Longissimus Lumborum Wang Bo Bai Xue Du Xin Pan Nan Shi Shuo Xia Xiufang * Panea Begoña Academic Editor College of Food Science, Northeast Agricultural University, Harbin 150030, China; wangbo9214@163.com (B.W.); snowbx1029@163.com (X.B.); dbnydxdx@163.com (X.D.); pannan36@163.com (N.P.); shishuo0902@163.com (S.S.) * Correspondence: xiaxiufang@neau.edu.cn; Tel.: +86-451-55191289 09 5 2022 5 2022 11 9 136806 3 2022 29 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). The effects of vacuum thawing (VT), ultrasound thawing (UT) and microwave thawing (MT) on the quality, protein and lipid oxidation, internal temperature distribution and microstructure of porcine longissimus lumborum were compared. The results showed that a significant decrease (p < 0.05) in quality compared with those of fresh meat (FM) occurred for all of the thawing samples, especially for the MT samples. Changes in quality of the VT and UT samples were less significant than those of the MT samples. The increases in carbonyl content and TBARS value indicated that proteins and lipids in the thawing samples were oxidized. The decreases in uniform degrees of internal temperature distributions of muscles from the thawing samples were analysed by infrared thermography. Scanning electron microscopy images showed that the myofibril arrangements of thawing samples were looser than those of the FM samples with compact and ordered structure, which was proven by the obvious increase in the myofibril gap value of the thawing samples. thawing method porcine longissimus lumborum texture fluid losses oxidation National Natural Science Foundation of China32172273 This study was supported by the National Natural Science Foundation of China (Grant No. 32172273). ==== Body pmc1. Introduction The thawing processes represent an important step performed prior to the further processing of frozen meat, such as slicing, chopping and subsequent cooking [1]. Generally, the time of the thawing process is longer than that of freezing process, and long thawing times can potentially promote physicochemical changes. In the thawing process, some undesirable effects on muscle quality include those related to flavour [2], texture [3] and colour [4], while denaturation and aggregation [5,6] of protein can also occur through some physical and chemical changes. The quality decrease in meat during thawing depends on many factors, such as thawing methods [7], thawing time [8] and thawing temperature [9]. Compared to common thawing methods (refrigerator [10], water-assisted [11] and air thawing [9]), some efficient and timesaving thawing methods, including high-voltage electrostatic field [12], vacuum [3], microwave [13] and ultrasonic thawing [14], have been widely applied to frozen foods such as fish, beef, edamame and mango. Cai et al. [5] noted that low temperature, fast thawing methods offer many advantages, such as shortened thawing time, reduced oxidation reaction and promoted quality. Ultrasound has been verified to function effectively for frozen foods, saving thawing time and improving the qualities of the meat [15]. Previous studies have reported that the melting speed of ice crystals is faster than that of most common thawing methods, shortening the thawing times of samples [16]. For example, the thawing times of mango [8], edamame [17] and frozen fish [18] induced by ultrasound thawing were reduced by 51–73%, 54.39% and 71%, respectively. In addition, the insignificant negative impacts on pH and microbial growth in pork [19] and sensory properties of the blocks of Pacific cod [18] were compared between ultrasound thawing and common thawing. For vacuum thawing, the frozen samples were thawed in a low oxygen environment which exerts reduced effects on microbial reproduction and oxidation reactions [20]. The insignificant changes in gel qualities and protein structure induced by vacuum as well as vacuum-assisted thawing were observed compared with control group [3,20]. Microwave thawing can control bacterial growth and the quality reduction because of shorter thawing time [21]. Some studies suggested that the thawing rate of frozen pork thawed by microwave was 100 times faster than that of air thawing [22], and the quality of native starch sauces thawed by microwave was improved by inhibiting additional amylose degradation compared with water bath [23]. In our previous studies, the effects of freezing storage on the quality [24,25,26,27,28] and thawing methods on the structure [7] and gelling properties [1] of proteins have been evaluated. Most other research concerning thawing was related to the changes in muscle quality characters and oxidation induced by single thawing methods [19,29]. Nevertheless, less attention was received with respect to comparison of the effects between novel thawing methods and conventional thawing methods on pork quality traits. Therefore, in this experiment, the comparison of vacuum thawing (VT), ultrasound thawing (UT) and microwave thawing (MT) on the quality traits, protein and lipid oxidation, internal temperature distribution and microstructure of porcine longissimus lumborum was evaluated. 2. Materials and Methods 2.1. Materials Longissimus lumborum (24 h post-mortem) purchased from local commercial abattoir (Harbin, China) were obtained from eight pork carcasses (total pork carcasses number in three independent batches) of similar age. All chemical reagents used were of analytical grade. 2.2. Sample Preparation A total of 120 chop samples (100 ± 0.1 g, total samples number in each independent batches) were divided into four groups, and each group was subjected to the freeze process (−18 °C, 7 days). Thirty chop samples were randomly picked for each experimental groups and then thawing operation was performed. Due to the space limitation of thawing room, five chop samples were thawed each time. The frozen muscle chops were thawed using four different methods: VT, vacuum thawing (25 °C, 30 min), UT, ultrasonic thawing (20 °C, 20 min), MT, microwave thawing (5 min) and WT, water immersion thawing (14 °C, 55 min). To monitor the temperature changes, a temperature recorder (Applent Precision Instrument Co., Ltd., Changzhou, China) was inserted at the centre of the samples. The endpoint temperature of thawing was set to approximately 4 °C. VT was carried out in a vacuum chamber including a water tank (9 KPa). UT was carried out in an ultrasonic chamber (500 W, Nanjing Xianou Co., Ltd., Nanjing, China). MT was carried out in a microwave oven (800 W, BE525LMS0W, Siemens, 594 × 382 × 317 mm, Guangdong, China). 2.3. MP Extraction The thawing muscles were chopped into small pieces for the myofibrillar protein (MP) extraction based on the method described by Li et al. [30] and Du et al. [31]. The measurement of protein concentration was carried out using bovine serum albumin as a standard. 2.4. Thawing Loss of Muscle Thawing loss (%)=W1−W2W1×100% where W1 represents the weight of porcine samples before thawing, and W2 represents the weight of porcine samples after thawing. 2.5. Cooking Loss of Muscle The cooking loss from thawed samples was measured according to the methods described by Jin et al. [32] with slight modification. The thawed porcine chops were weighed (W1) and cooked in an 80 °C water bath until the centre temperature reached 75 °C. The samples were weighed (W2) after cooking. The cooking loss is defined as follows:Cooking loss (%)=W1−W2W1×100%  2.6. Colour of Muscle The measurement of muscle colour was carried out based on the method described by Li et al. [33] with a colourimeter (ZE-6000, Illuminant D65; Nippon Denshoku, Tokyo, Japan). The thawed samples were balanced to room temperature and cut into slices with a diameter of 18 mm. The instrument was calibrated using a white standard plate (L* = 90.26, a* = −1.29, b* = 5.18). The values, expressed as L* (lightness), a* (redness), b* (yellowness) and ΔE units, were obtained from four different areas on the surface of each chop, and a minimum of 3 chops per treatment block was analysed to obtain an average value. The values of chroma (chroma = [a*2 × b*2] × 0.5) and hue angle (hue angle = arctan [b*/a*]) were calculated. 2.7. Texture Analysis of Muscle The texture of muscle was analysed according to the determination of shear force, hardness, springiness, cohesiveness and chewiness of muscle. The muscle chops were cooled to room temperature (18 °C) after cooking. The shear force was determined by using a tenderization analysis (C-LM3B tenderization instruments, Northeast Agricultural University, Harbin, China) with a 25 kg load transducer according to the research by Wang et al. [34] and Holman et al. [35] with slight modifications. The cooled muscle chops (18 °C) were cut into strips (1.5 × 1.5 × 5.0 cm) parallel to the muscle fibre orientation for the test, and a crosshead speed of 200 mm/min. The greatest force value was recorded as the shear force when the samples were cut. The hardness, springiness, cohesiveness and chewiness of muscle was determined by Texture profile analysis (TPA) according to the method described by Pan et al. [36] at room temperature using a TA-XT plus Texture Analyser (Stable Micro System, Surrey, UK). The cooled muscle chops were cut into uniform cubic size (2 × 2 × 2 cm3) and subjected to two-cycle compression with a P/50 flat-surface cylindrical probe, 5 cm diameter cylindrical probe and 5 mm/s cross-head speed. Prior to analysis, samples were placed on the centre of the TPA platform and compressed to 40% of their original height. The conditions were as follows: pre-test speed 1.0 mm/s, test speed 1.0 mm/s, post-test speed 5.0 mm/s and the testing interval 5 s. 2.8. Moisture Mobility and Distribution of Muscle The moisture mobility and distribution from thawed samples were evaluated using transverse relaxation time (T2). The low-field nuclear magnetic resonance (LF-NMR) relaxation time was measured according to the method of Jin et al. [37] with a LF-NMR analyser minispec mq 20 (Bruker Optik GmbH, Ettlingen, Germany) with a magnetic field strength of 0.47 T corresponding to a proton resonance frequency of 20 MHz. The thawed samples were cut into parallelepipeds (1 × 1 × 2 cm3) and placed in NMR tubes (18 mm of diameter). The transverse relaxation time (T2) was measured using the Carr–Purcell–Meiboom–Gill pulse sequence. For each sample, 16 scans were obtained at 2 s intervals with 3000 echoes in total. The continuous distribution of exponentials related to water located in different muscle compartments was fitted for all Carr–Purcell–Meiboom–Gill curves using the CONTIN algorithm after normalizing the raw data. Three relaxation times (T2b, T21 and T22) and their corresponding area fractions (P2b, P21 and P22) were recorded as outputs. 2.9. Oxidation Reaction Protein oxidation was evaluated according to the content of carbonyl groups based on Jin et al. [37]. Carbonyl groups were detected by reaction with DNPH to form protein hydrazones. Myofibrillar protein solution was precipitated with 100 g/L TCA (w/v; final concentration). After centrifugation (2000× g, 10 min), pellet was treated with 2 g/L DNPH in 2 mol/L HCl with agitation for 1 h at room temperature. The fractions were then precipitated with 100 g/L TCA (final concentration) and centrifuged. The pellets were washed twice with 1 mL of ethanol:ethyl acetate (1:1 v/v), and the solution was precipitated with 100 g/L TCA (final concentration) and centrifuged. Proteins were then dissolved in 2 mL of 6 mol/L guanidine with 20 mmol/L sodium phosphate buffer at pH 6.5. Absorbance was measured at 365 nm for the DNPH-treated sample against an HCl control. The amount of carbonyl was expressed as nmol of DNPH fixed/mg of protein using an absorption coefficient of 22,000 (mol/L)−1 cm−1 for protein hydrazones. Lipid oxidation was evaluated according to Wang and Xiong [38] with slight modifications. An accurately weighed finely chopped meat sample (ca. 0.4× g) was placed in a 25 mL screw cap test tube, and three drops of antioxidant solution (BHA), 3 mL of TBA solution and 17 mL of TCA-HCI solution were subsequently added. The mixture was vortexed, flushed with nitrogen gas and then heated in boiling water (100 °C) for 30 min. After being cooled to room temperature, a 5 mL aliquot of the suspension was mixed by vortexing with 5 mL of chloroform for 1 min, followed by centrifugation at 1800× g for 10 min. The upper phase (aqueous) was centrifuged again for 10 min under the same condition, and the absorbance of the supernatant was read at 532 nm with Ultraviolet spectrophotometer. The TBARS value, expressed as mg of malonaldehyde/kg of muscle sample, was calculated by using the following equation:TBARS (mg/kg)=A532Ws×9.48  here A532 is the absorbance (532 nm) of the assay solution, Ws is assigned to the weight of meat sample (g); ‘9.48’ corresponds to a derived constant. 2.10. Internal Temperature Distribution of Muscle The measurement of internal temperature distribution of muscle was performed according to the procedure reported by Choi et al. [22] with a slight modification. The sample was immediately bisected vertically when the central temperature of thawed muscle reached approximately 4 °C, and the thermographic image of muscle was acquired by an infrared camera (GTC40, Bosch Co., Stuttgart, Germany) within 20 s. 2.11. Microstructure of Muscle Images of microstructure from muscle were acquired by a scanning electron microscope (SEM) (S-3400N; Hitachi, Tokyo, Japan) according to the method reported by Wang et al. [39]. The samples (0.5 × 0.5 × 0.25 cm3, cut with a sharp-edged razor) were fixed with 25 mL/L glutaraldehyde in 0.2 mol/L phosphate buffer (pH 7.2) for 2 h. The samples were then rinsed for 1 h with distilled water before being dehydrated in ethanol with serial concentrations of 500, 700, 800, 900 and 1000 mL/L. Dried samples were mounted on a bronze stub and sputter-coated with gold (Sputter coater SPI-Module, West Chester, PA, USA). The microstructures of the samples were examined using a SEM at an operating voltage of 3.0 kV. 2.12. Sensory Evaluation and Consumer Testing Once each experiment for a given thawing process was complete, 15 chop samples were randomly picked and cooked at 80 °C until the centre temperature reached 75 °C. The cooled muscle chops (18 °C) were cut into strips (1.0 × 1.0 × 2.0 cm3) parallel to the muscle fibre orientation for the sensory evaluation and consumer testing. Each chop samples could be divided into 18 strips to ensure that each sample could be tasted at least twice. Panellists were asked to evaluate the appearance, tenderness, juiciness, flavour, overall acceptability of the cooked muscles. All panellists received samples in the same order with a 20-min rest period between samples. During the evaluation, the panellists were situated in a private booth under incandescent light. In order to clean the palate between the experiments, room temperature water was used. All samples were coded with three-digit random numbers and served in a random order to the panellists. Twenty panellists who had prior experience with meat product evaluation assessed the sensory attributes of cooked muscles using a 9-point scale. Scores were recorded on a scale of 1–9 (1 = extreme dislike, 9 = extreme like). Scores with means from 5 to 9 were considered acceptable. The mean scores from panellists for each sample and session were calculated and analysed. Consumer testing: Consumer testing was analysed by the method of Heck et al. [40] and Ares and Jaeger [41]. One hundred and five consumers (46% male, 54% female, aged 18–55 years) were asked to complete a check-all-that-apply (CATA) questionnaire with four attributes defined by the trained panel as described above. The consumers were asked to check all the terms that they considered appropriate to describe each sample. Filtered water were provided to the consumers for mouth cleansing between samples. The sensory terms listed were randomized within and across consumers, meaning that each consumer received the CATA question with the terms in different order and this order was modified from sample to sample within the test. Analysis of the CATA questionnaire: Correspondence Analysis (CA) was used to analyse data from the CATA questionnaire, considering the chi-square distance [42], calculated on the matrix containing the use frequency of attributes (Brightness, Mouthfeel, Juiciness, Pleasant) for each sample. 2.13. Statistical Analysis All experiments were conducted three times (three independent batches), and each sample was evaluated in triplicate to evaluate the changes in quality properties and oxidation of pork from different thawing samples. The results are expressed as means values ± standard deviation (SD). One-way analysis of variance (ANOVA) was carried out followed by Duncan’s test using SPSS 22.0 (SPSS Inc., Chicago, IL, USA). The level of significance was set to p < 0.05. All graphs were generated using SigmaPlot 12.5. 3. Results and Discussion 3.1. Quality Traits 3.1.1. Fluid Losses Thawing loss is a vital quality indicator for frozen muscle because it would result in not only the change in weight but also decrease in quality, ultimately causing economic loss [43]. Table 1 showed that the thawing losses from different thawing samples increased obviously (p < 0.05). During the thawing process, the water melted by ice crystals in/outside the cell encountered difficulty in returning to the initial position in the fresh meat, which caused the decrease in the water-holding capacity [44]. The thawing loss of the MT samples was the highest (4.71%) among all thawing samples. Boonsumrej et al. [45] found that the high thawing loss of tiger shrimp was mainly induced by evaporated water upon instantaneous high temperature during microwave thawing. In addition, there was significantly low (p < 0.05) thawing loss of the samples treated with VT (2.8%) and UT (3.0%) compared with WT (3.7%). The low thawing loss primarily occurred on account of the low oxygen environment during VT, which slowed oxidation and decreased the degree of protein degeneration. Sun et al. [44] asserted that protein oxidation may result in degeneration and changes in the structure of protein, which caused the reduction in the water-binding ability of protein. The cooking loss of FM was 18.70%, and there were increments of 10.7%, 3.6%, 23.4% and 27.1% after UT, VT, MT and WT, respectively (Table 1). Choi et al. [22] also showed the increased cooking loss of porcine longissimus dorsi after thawing was observed. The cooking loss from VT samples showed an insignificant increase (p > 0.05), and UT, MT and WT samples showed significant increases (p < 0.05) as compared with FM samples. The result in cooking loss could resulted in the decrease in water-holding capacity of muscle tissue and the degradation of myofibrillar protein, especially myosin [46]. Wang et al. [7] reported that the decrease in Ca2+-ATPase activity of MP from VT and UT samples was insignificant compared with that from FM (p > 0.05), while the significant decrease in that from MT samples was observed (p < 0.05), suggesting that the protein integrity of MT samples was severely destroyed and degraded. Combined with the results of thawing loss and cooking loss, VT and UT were more conducive to maintaining water-holding capacity of muscle than MT. 3.1.2. Colour of Muscle Colour is an important sensory attribute of meat because it determines the purchasing desire of consumers [47]. Table 2 shows the changes in colour of different thawing samples. Compared with FM samples, the a* value and chroma of different thawing samples decreased significantly (p < 0.05), while L*, ΔE and hue value increased significantly (p < 0.05). The L* values can be used as a marker to judge the degree of muscle loss. The higher the muscle quality, the lower the L* values. Decreased water content may lead to increased reflection of light [48]. Table 2 showed that L* value from different thawing samples increased obviously (p < 0.05). The L* value from UT, VT, MT and WT samples were 2.44%, 1.20%, 8.07% and 10.07% higher than that from FM samples respectively (p < 0.05). Compared with UT and VT samples, significantly higher L* value from MT samples was obtained (p < 0.05). The difference in L* value among thawing samples may be related to the moisture content, state and distribution of thawing samples. Zhang et al. [49] demonstrated that samples with lower water content possessed higher L* values. The high thawing loss from MT samples led to stronger light reflection and lighter colour. The increase of L* value from UT samples may be attributed to the thawing environment. The UT samples were immersed in water during thawing, and part of water could be infiltrated into the poly nylon pouch packed with muscle, increasing the light reflection and L* value. The lower L* value from VT samples was not only due to the less thawing loss, but also due to the minimum moisture migration proved by LF-NMR analysis, which resulted in less thawed water attached to the meat surface and reduced the light reflection intensity [50]. The a* value from UT, VT, MT and WT samples decreased by 12.84%, 19.05%, 24.03% and 13.49% compared with that from FM samples, respectively (p < 0.05). Compared with VT and MT samples, a* value from UT samples decreased significantly (p < 0.05). The decrease in a* value during thawing was mainly due to the decrease in the amount of myoglobin, the main pigment in the meat, and the change in its chemical state [51]. The lower a* value from MT samples may be attributed to the loss of methemoglobin reducing enzyme with the exudation. The higher thawing loss from MT samples has been showed in Table 1. The metmyoglobin reducing enzyme is very active in fresh muscle and the metmyoglobin formed is quickly reduced to deoxymyoglobin and oxygenated back to oxymyoglobin, thereby retaining the colour. However, the metmyoglobin reducing enzyme could be lost from the sarcoplasmic environment as exudate during thawing, leading to the accumulate of metmyoglobin on the muscle surface and accelerating the loss of redness [52]. Muela et al. [53] also found that the decrease in redness of frozen muscle was related to the decrease in metmyoglobin reducing enzyme activity. In addition, the protein and lipid oxidation from MT samples caused by local overheating could increase the quantity of free radicals, leading to increased rates of myoglobin oxidation and metmyoglobin formation [54]. Choi et al. [22] reported that the a* value of pork loin from microwave thawing was lower than those from other thawing methods (radio frequency, water immersion and forced-air convection thawing). Although the less thawing loss and lower protein and lipid oxidation occurred, the low a* value from VT was obtained, which may be explained by the fact that myoglobin could be oxidized into metmyoglobin in a low oxygen partial pressure environment. In addition, the decrease in a* values from the UT samples resulted from the loss of water-soluble myoglobin with the increase in thawing loss during thawing [55]. The change in b* value from thawing samples was similar to that in L* value. The b* value from MT samples was higher than that from UT and VT samples (p < 0.05). Meanwhile, the change in b* value can also be demonstrated from the increase in hue angle from thawing samples. The hue angle was used to illustrate the colour change from red to yellow, and the increased hue angle indicated that the muscle subjected to thawing exhibited a more yellowish hue [56]. The higher b* value and hue angle from MT samples may be attributed to the high degree of protein oxidation. Coombs et al. [57] revealed that protein oxidation could increase b* values and hue angle in the process of freezing and thawing. This result could be related to the increase in TBARS during thawing. However, the hue angle from VT samples was lower than that form UT samples, which may be due to the low a* value induced by low oxygen partial pressure environment. Chroma value represents colour intensity and has been considered a good indicator of meat stability [58]. In this study, the chroma value from thawing samples decreased significantly (p < 0.05), suggesting that thawing process was not conductive to maintaining the stability of muscle colour [59]. The chroma values from VT and MT samples were significantly lower than those from UT samples (p < 0.05), which illustrated that VT and MT were not conducive to maintaining the stability of muscle colour. 3.1.3. Texture of Muscle The shear force and texture (hardness, cohesiveness, chewiness and springiness) was used to analyse the texture of muscle. The measurement of shear force from muscle is carried out to assess tenderness. The higher tenderness in muscle, the lower shear force [60]. As shown in Table 3, the shear force value from VT, UT, MT and WT samples was obviously increased by 10.8%, 17.8%, 36.9 and 34.5%, respectively, compared with FM (28.89 N) (p < 0.05), which illustrated that the tenderness of different thawing samples decreased. The decrease in tenderness from frozen and thawed samples may be caused by the high thawing and cooking loss [25] and the shrinkage and cracking of muscle fibres [61]. The results of thawing and cooking loss during thawing have been demonstrated in Table 1. Lagersted et al. [62] reported that decreased tenderness value was due to the loss of fluid during thawing, which led to the decrease of water available for hydration of muscle fibres. Furthermore, toughening is caused by sarcomere shortening in the process of thawing. Dransfield et al. [63] found that the muscle length of beef thawed at room temperature was shortened by roughly 40% and endowed the meat with greater hardness. Lagerstedt et al. [62] also reported significantly reduced tenderness from thawed beef meat. There was no significant difference in shear force value between VT and UT samples (p > 0.05). However, the shear force value from MT samples was significantly lower than that from VT and UT samples (p < 0.05). The decreased tenderness during thawing may also be due to structural changes in proteins induced by oxidation [45]. The low tenderness of muscle treated with MT may be linked to protein oxidation of porcine longissimus dorsi induced by instantaneous high temperature during thawing. Xia et al. [25] verified that protein oxidation of pork during MT process might be due to the release of pro-oxidant and oxidative enzymes from ruptured cellular organs. However, protein oxidation and intra/inter-molecular cross-linking were inhibited during VT process because of the low oxygen environment, which contributed to maintain muscle tenderness [64]. The high tenderness of muscle treated with UT might be attributed to the quick thawing rate [7] and the high physical forces associated with shear force and turbulence from cavitation [7]. Thawing rate is one of the most important factors during thawing, because it controls the time at which proteins are oxidized [65]. A quick thawing rate during UT was due to the quick phase transition speed from ice to water in frozen samples. Ultrasound treatment could convert acoustic energy into heat energy by the collapse of cavitation bubbles causing a temperature rise in the frozen samples [66]. Jayasooriya et al. [67] found that during the ultrasound process, a large number of lysosomes were released into the sarcoplasm due to the shear force and turbulence caused by cavitation, thus accelerating the enzymatic hydrolysis reaction and increasing the muscle tenderness. Dickens et al. [68] also reported the tenderization effect of ultrasonic treatment on beef and poultry muscle. Overall, the negative effects of VT and UT on tenderness were lower than those of MT. The measurement of texture is a direct and important method to assess the meat quality, which condition its palatability for consumers. Table 3 shows the changes in texture (hardness, cohesiveness, chewiness and springiness) from different thawing samples. Compared with FM samples, increased hardness value and decreased cohesiveness, chewiness and springiness values of muscle treated with different thawing methods were observed. It is noted that insignificant difference (p > 0.05) in hardness value from VT samples as well as cohesiveness and chewiness values from UT and VT samples were observed compared with the FM samples. Nevertheless, significantly higher (p < 0.05) hardness value and significantly lower (p < 0.05) cohesiveness, chewiness and springiness values were obtained in WT and MT samples with respect to other thawing samples. Mousakhani-Ganjeh et al. [69] also pointed out that the hardness value was increased, and the cohesiveness and springiness values were decreased during the thawing process, which might be attributed to the protein denaturation occurring in the proteins of fish during thawing. 3.2. Moisture Mobility and Distribution of Muscle The mobility and distribution of three kinds of moisture in porcine longissimus lumborum could be analysed by the measurement of T2 (T2b, T21 and T22) and corresponding percentages (P2b, P21 and P22) of peak areas based on LF-NMR [44]. As shown in Figure 1, three peaks (T2b, T21 and T22) were showed and represented three distinct water populations. Table 4 shows the changes in T2b, T21 and T22 relaxation times of different thawing muscle. Insignificant difference (p > 0.05) in T2b relaxation times of thawing samples was observed because bound water was very resistant to freezing or thawing [70]. Zhang et al. [71] also found that the change in T2b relaxation time of porcine samples after thawing was not obvious compared with fresh muscle (p > 0.05), which was primarily attributed to the fact that bound water is not affected by any change in mechanical stress and microstructure of muscle. The significant increase (p < 0.05) in T21 relaxation times of thawing samples was observed. The prolonged relaxation times of water reflected the decrease in combination ability between water and protein and the increase in mobility of water [72]. It was worth noting that T21 relaxation times from the MT samples were significantly longer (p < 0.05) compared with other thawing samples, which signified that the tightness between immobilized water and protein molecules was weakened and that the mobility of immobilized water from MT was high. For T22, the changes in relaxation time from thawing samples were similar to those of T21. This result indicated that the free water of thawing samples encountered difficulty returning to the position as in the fresh meat because of the high mobility of free water in the destructed myofibrils. Table 4 shows the percentages of peak areas (P2b, P21 and P22) corresponding to T2 relaxation time. Compared with the P21 values from FM samples (90.55%), the values of the VT, UT, MT and WT samples decreased by 4.61%, 4.10%, 8.71% and 6.70%, and the significant difference among thawing samples and the FM samples was observed (p < 0.05). The changing trend of P21 values from different thawing samples was consistent with the result that the WHC of porcine longissimus lumborum was decreased (Table 1). P22 values of the UT, VT, MT and WT samples increased by 44.23%, 39.26%, 83.49% and 64.13% compared with that of the FM samples (9.97%), respectively. The changes in P21 and P22 indicated that the immobilized water entrapped within the space between the thick and thin filaments or in the myofibrillar network was transformed into free water existed in the intercellular space during the thawing process. The result was consistent with the study of Han et al. [73], who reported that immobilized water was redistributed into the space between fibres, which led to the decline in the water-holding capacity of samples after the thawing process. 3.3. Protein and Lipid Oxidation Protein and lipid oxidation were monitored by carbonyl content and TBARS, respectively (Table 5). The carbonyl content and TBARS from the VT, UT, MT and WT samples were increased by 2.9%, 3.8%, 8.6% and 5.7%, and 13.3%, 6.7%, 53.3% and 20.0% compared with FM samples, respectively. Kim et al. [74] confirmed that the post mortem processing of muscles, such as frozen storage and freezing–thawing cycles, could accelerate protein and lipid oxidation. The increases in carbonyl content and TBARS of muscle during processing were associated with the release of pro-oxidative factors involved with free radicals and oxidative enzymes [2]. Boonsumrej et al. [45] also pointed out that the disruption of muscle ultrastructure during freezing and thawing could release haem iron, oxidative enzymes and other pro-oxidants which accelerate the oxidation of muscle. The differences in oxidation induced by UT and VT were not significant (p > 0.05), and their oxidation degrees were lower than that of MT. Xia et al. [25] reported that the TBARS of porcine longissimus muscle during microwave thawing was higher than during refrigerator thawing, which could be explained by the fact that the instantaneous high temperature induced by microwave thawing triggered the protein and lipid oxidation of muscle. The low carbonyl content and TBARS of samples might be attributed to the low speed of oxidation under the low oxygen environment during VT. Oxidation happening in the thawing process may result in the decrease in eating quality referred to with respect to tenderness, WHC and colour (Table 1). 3.4. Internal Temperature Distribution The changes in internal temperature distribution of muscle treated with different thawing methods were observed by thermographic imaging using infrared thermography (Figure 2). The coloration from blue (−2 °C, lower temperatures) to red (30 °C, higher temperatures) indicates the temperature change of muscle during thawing. A clear difference in colour distribution of muscle between FM and thawing samples was observed. The colour distribution from FM samples was symmetric, which showed the uniform internal temperature distribution within muscle. The internal temperature distribution from VT samples during thawing exhibited the form of concentric circles transferring inwards from high to low temperature, which reflected the uniformity of thawing: at all points equidistant from the middle of the sample (the coldest point, 4 °C), the heat penetration is equivalent. The thawing patterns of muscle from the UT, MT and WT samples did not exhibit the concentric circle shape, which illustrated non-homogenous thawing patterns during UT, MT and WT processes. The inhomogeneous internal temperature distribution within muscle was mainly attributed to the local overheating phenomenon during UT and MT processes. This finding was consistent with previous studies representing local overheating at the edge region of shrimp [45] and beef [75] induced by microwave thawing. Li and Sun [76] also pointed out that some disadvantages such as localized heating could occur during ultrasound thawing. For the WT process, the non-uniform internal temperature distribution could be attributed to the low heat conductivity and heat diffusivity, which led to higher temperatures at the top than within the centre of the samples centre [22]. 3.5. Microstructure of Muscle The changes in microstructure of different thawing samples are displayed in Figure 3. The intact and tight microstructure of the FM samples was destroyed, and the intermuscular gap was increased significantly (p < 0.05) after thawing. The intermuscular space of the VT samples (0.12 mm2) was smaller than for other thawing methods, which illustrated the destruction of VT to the microstructure of samples. The myofibre of the MT samples was separated, and the intermuscular gap was large compared with other thawing methods, which indicated that the negative effect of MT on the microstructure of samples was the greatest. The results were consistent with the study of Xia et al. [25], who reported that the destruction from MT samples on the microstructure was greatest among all thawing methods (refrigerator, ambient temperature, water immersion and lotic water thawing). Cai et al. [13] also suggested that the obvious shrinkage of muscle was induced by the overheating phenomenon during microwave thawing. The increase in intermuscular gap during thawing may be attributed to the excessive dehydration of muscle induced by protein denaturation and the breakage of epimysium, perimysium and endomysium [77]. Xia et al. [25] also reported that protein denaturation and cell disruption could be induced by improper thawing processes. Choi [22] reported that the looser structure of muscle fibre was associated with low water-holding capacity and tenderness of pork loin during freezing and thawing. The obtained results of low water-holding capacity and high shear force were coincidental with the disrupted microstructure (Table 1). 3.6. Sensory Evaluation and Consumer Testing The sensory properties of food products play a decisive role on the food choices of consumers. Table 6 shows the sensory attributes from different thawing samples. As shown in Table 6, differences among different thawing samples were found for tenderness, juiciness and overall acceptability (p < 0.05), while no difference were registered for appearance and flavour (p > 0.05). There was no significant difference in the appearance from different thawing samples, which may be due to the fact that the colour differences among cooked thawing samples were not easy to identify. The significantly lower score of tenderness from MT samples was obtained (p < 0.05). Lagerstedt et al. [62] showed that a trained sensory panel rated the freeze/thawed meat significantly less tender than the chilled meat. This sensory result was attributed to the loss of fluid during thawing that resulted in less water available to hydrate the muscle fibres; thus, a greater quantity of fibres per surface area seemed to increase the toughness as perceived by the sensory panel [78]. The high thawing and cooking loss and protein oxidation from MT samples induced by instantaneous high temperature during thawing led to the low tenderness scores. However, protein oxidation was inhibited during the VT process because of the low oxygen environment, which contributed to maintaining muscle tenderness [64]. Meanwhile, the higher tenderness scores from UT samples could be linked to the quick thawing rate and high physical force associated with shear force and turbulence from cavitation [7]. The variation trend of juiciness scores from thawing samples was consistent with tenderness. The sensation of juiciness upon the mastication of a red meat product is a function of the water content of the meat. The increased juiciness in UT and VT samples could be linked with the quick thawing rate and low degree of oxidation, which could enhance water holding capacity. Meanwhile, the results that UT and VT samples were perceived by the tasters as more tender and juicier than MT samples influenced the overall acceptability. Table 7 shows the results of the consumer preference test. CATA was applied using four sensory descriptors. The results of brightness, mouthfeel and juiciness from thawing samples were similar to that of sensory attributes. The brightness from thawing samples was also not easily identified. A bad mouthfeel from MT samples was obtained. Consumer satisfaction with the tenderness of a red meat product is based on the interaction between the textural properties of the meat and ‘mouthfeel’—this being an experience that includes mastication and biting properties (e.g., chewiness, hardness, firmness and softness). The texture results from MT samples showed higher hardness and lower cohesiveness, chewiness and springiness. The results of mouthfeel and juiciness from UT and VT samples led to the fact that more muscles were marked as ‘pleasant’ which indicated that the muscles were accepted. However, fewer muscles were marked as ‘pleasant’ which indicated that the muscles were less accepted. 4. Conclusions The effects of thawing methods (UT, VT, MT) on quality properties of porcine longissimus lumborum were confirmed by the increases in thawing loss, cooking loss, shear force, L*, b*, ΔE, hue, hardness, relaxation time, the ratios of free water (P22), carbonyl contents and TBARS and the decreases in a*, chroma, cohesiveness, chewiness, springiness and the ratios of immobilized water (P21). The internal temperature distribution of muscle samples after thawing was nonuniform, except for in the case of VT. Based on the observation of SEM, the tight and intact myofibril structure of muscle was destroyed during thawing. The effects of MT on the quality properties of samples were high, and the sensory evaluation and consumer testing showed that the MT samples were less accepted. The UT and VT could maintain the quality better. However, the low a* value from VT samples caused by low oxygen environment and the inhomogeneous temperature distribution within muscle from UT samples caused by local overheating could not be neglected, which needs to be solved in further research and practical application. Author Contributions Methodology, software, validation, formal analysis, investigation and writing, B.W.; methodology and investigation, X.B.; formal analysis, X.D.; methodology and investigation, N.P.; software, S.S.; conceptualization, resources, writing—original draft, supervision and funding acquisition, X.X. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement The data presented in this study are available in the article. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Change in distribution of the LF-NMR relaxation times of porcine longissimus dorsi induced by thawing methods. FM, fresh meat; UT, ultrasonic thawing (20 °C); VT, vacuum thawing (25 °C); MT, microwave thawing; WT, water immersion thawing (14 °C). Figure 2 Change in internal temperature distribution of porcine longissimus dorsi induced by thawing methods. FM, fresh meat; UT, ultrasonic thawing (20 °C); VT, vacuum thawing (25 °C); MT, microwave thawing; WT, water immersion thawing (14 °C). Figure 3 Change in microstructure of porcine longissimus dorsi induced by thawing methods. FM, fresh meat; UT, ultrasonic thawing (20 °C); VT, vacuum thawing (25 °C); MT, microwave thawing; WT, water immersion thawing (14 °C). The gap areas (mm2) are given as the means ± SD, with different lowercase letters (a–e) indicating significant differences (p < 0.05). foods-11-01368-t001_Table 1 Table 1 Change in fluid losses of porcine longissimus dorsi induced by thawing methods. Thawing Methods Fluid Losses (%) Thawing Loss Cooking Loss FM - 18.70 ± 0.33 c UT 2.98 ± 0.29 c 20.71 ± 0.72 b VT 2.78 ± 0.27 c 19.37 ± 0.45 c MT 4.71 ± 0.15 a 23.07 ± 0.24 a WT 3.70 ± 0.23 b 23.77 ± 0.57 a The means in the same column with different lowercase letters (a–cdiffer significantly (p < 0.05). The results are mean ± SD (n = 3 × 3). FM, fresh meat; UT, ultrasonic thawing (20 °C); VT, vacuum thawing (25 °C); MT, microwave thawing; WT, water immersion thawing (14 °C). foods-11-01368-t002_Table 2 Table 2 Change in colour of porcine longissimus dorsi induced by thawing methods. Thawing Methods L* a* b* ΔE Chroma Hue (°) FM 36.44 ± 0.12 e 13.86 ± 0.23 a 8.88 ± 0.06 d - 16.46 ± 0.43 a 32.65 ± 0.16 d UT 37.33 ± 0.16 c 12.08 ± 0.35 b 9.08 ± 0.42 d 2.00 ± 0.04 d 15.11 ± 0.21 bc 36.93 ± 0.25 c VT 36.88 ± 0.07 d 11.22 ± 0.31 c 9.17 ± 0.32 c 2.69 ± 0.01 c 14.49 ± 0.46 c 39.26 ± 0.44 b MT 39.38 ± 0.14 b 10.53 ± 0.19 c 10.55 ± 0.32 a 4.75 ± 0.06 a 14.90 ± 0.24 c 45.05 ± 0.26 a WT 40.11 ± 0.19 a 11.99 ± 0.24 b 10.01 ± 0.18 b 4.27 ± 0.03 b 15.62 ± 0.31 b 39.85 ± 0.31 b The means in the same column with different lowercase letters (a–e) differ significantly (p < 0.05). FM, fresh meat; UT, ultrasonic thawing (20 °C); VT, vacuum thawing (25 °C); MT, microwave thawing; WT, water immersion thawing (14 °C). foods-11-01368-t003_Table 3 Table 3 Change in shear force and texture (hardness, cohesiveness, chewiness and springiness) of porcine longissimus dorsi induced by thawing methods. Thawing Methods Shear Force (N) Hardness Cohesiveness Chewiness Springiness FM 28.89 ± 0.49 c 34.57 ± 0.52 d 0.54 ± 0.01 a 22.37 ± 0.43 a 1.71 ± 0.02 a UT 34.03 ± 0.78 b 37.09 ± 0.98 bc 0.51 ± 0.01 ab 21.07 ± 0.76 a 1.56 ± 0.03 b VT 32.02 ± 0.84 b 35.77 ± 0.27 cd 0.52 ± 0.01 a 21.45 ± 0.49 a 1.60 ± 0.01 b MT 39.54 ± 0.98 a 39.61 ± 0.54 a 0.46 ± 0.02 c 18.42 ± 0.41 b 1.35 ± 0.03 d WT 38.86 ± 0.31 a 38.39 ± 0.54 ab 0.48 ± 0.01 bc 19.43 ± 0.46 b 1.42 ± 0.02 c Means in shear force, hardness, cohesiveness, chewiness and springiness with different lowercase letters (a–d) differ significantly (p < 0.05). FM, fresh meat; UT, ultrasonic thawing (20 °C); VT, vacuum thawing (25 °C); MT, microwave thawing; WT, water immersion thawing (14 °C). foods-11-01368-t004_Table 4 Table 4 Change in distribution of the T2 relaxation time and the P2 (T2 peak ratio) of porcine longissimus dorsi induced by thawing methods. Thawing Methods T2 (ms) P2 (%) T 2b T 21 T 22 P 2b P 21 P 22 FM 1.52 ± 0.04 a 26.23 ± 0.31 a 114.33 ± 2.52 c 0.14 ± 0.01 a 90.55 ± 0.59 a 9.45 ± 0.96 c UT 1.61 ± 0.02 a 43.30 ± 0.40 c 118.33 ± 2.08 cd 0.11 ± 0.01 ab 86.37 ± 1.08 b 13.63 ± 1.76 b VT 1.57 ± 0.04 a 37.43 ± 0.31 d 116.67 ± 2.08 cd 0.11 ± 0.02 ab 86.84 ± 0.53 b 13.16 ± 1.18 b MT 1.73 ± 0.02 a 49.30 ± 0.56 a 137.33 ± 3.21 a 0.08 ± 0.01 b 82.66 ± 0.68 c 17.34 ± 1.73 a WT 1.62 ± 0.02 a 45.53 ± 0.40 b 126.67 ± 2.08 b 0.09 ± 0.02 b 84.49 ± 0.98 c 15.51 ± 0.51 ab The means in the same T2 with different lowercase letters (a–d) differ significantly (p < 0.05). The means in the same P2 with different lowercase letters (a–d) differ significantly (p < 0.05). FM, fresh meat; UT, ultrasonic thawing (20 °C); VT, vacuum thawing (25 °C); MT, microwave thawing; WT, water immersion thawing (14 °C). foods-11-01368-t005_Table 5 Table 5 Change in carbonyl content and TBARS of porcine longissimus dorsi induced by thawing methods. Thawing Methods Carbonyl Content nmol/mg MP TBARS mg/kg MP FM 1.05 ± 0.02 c 0.15 ± 0.01 c UT 1.09 ± 0.02 b 0.18 ± 0.01 b VT 1.09 ± 0.01 b 0.16 ± 0.01 bc MT 1.14 ± 0.01 a 0.23 ± 0.02 a WT 1.11 ± 0.02 ab 0.18 ± 0.01 b FM, fresh meat; UT, ultrasonic thawing (20 °C); VT, vacuum thawing (25 °C); MT, microwave thawing; WT, water immersion thawing (14 °C). Means in carbonyl content with different lowercase letters (a,b) differ significantly (p < 0.05); Means in TBARS with different lowercase letters (a–c) differ significantly (p < 0.05). foods-11-01368-t006_Table 6 Table 6 Sensory characteristics of cooked porcine longissimus dorsi induced by thawing methods. Thawing Methods Sensory Attributes Appearance Tenderness Juiciness Flavor Overall Acceptability FM 7.67 ± 0.25 a 8.02 ± 0.17 a 7.06 ± 0.13 a 6.83 ± 0.15 a 7.26 ± 0.25 a UT 7.52 ± 0.27 a 7.79 ± 0.35 a 6.69 ± 0.15 a 6.77 ± 0.12 a 7.04 ± 0.11 a VT 7.57 ± 0.16 a 7.93 ± 0.24 a 6.78 ± 0.11 a 6.81 ± 0.26 a 7.12 ± 0.14 a MT 7.38 ± 0.22 a 7.04 ± 0.19 b 6.15 ± 0.22 b 6.58 ± 0.13 a 6.33 ± 0.25 b WT 7.47 ± 0.21 a 7.13 ± 0.27 b 6.22 ± 0.10 b 6.61 ± 0.33 a 6.51 ± 0.11 b The means in the same column with different lowercase letters (a,b) differ significantly (p < 0.05). FM, fresh meat; UT, ultrasonic thawing (20 °C); VT, vacuum thawing (25 °C); MT, microwave thawing; WT, water immersion thawing (14 °C). foods-11-01368-t007_Table 7 Table 7 Frequency mentioned attributes by porcine longissimus dorsi CATA for each sample. Thawing Methods Attributes Brightness Mouthfeel Juiciness Pleasant FM 25 45 51 32 UT 23 43 47 28 VT 20 44 49 28 MT 20 37 40 23 WT 22 39 38 25 FM, fresh meat; UT, ultrasonic thawing (20 °C); VT, vacuum thawing (25 °C); MT, microwave thawing; WT, water immersion thawing (14 °C). Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Wang B. Du X. Kong B. Liu Q. Li F. Pan N. Xia X. Zhang D. Effect of ultrasound thawing, vacuum thawing, and microwave thawing on gelling properties of protein from porcine longissimus dorsi Ultrason. Sonochem. 2020 64 104860 10.1016/j.ultsonch.2019.104860 31948851 2. Leygonie C. Britz T. Hoffman L. Impact of freezing and thawing on the quality of meat: A review Meat Sci. 2012 91 93 98 10.1016/j.meatsci.2012.01.013 22326063 3. Li F. Wang B. Liu Q. Chen Q. Zhang H. Xia X. Kong B. Changes in myofibrillar protein gel quality of porcine longissimus muscle induced by its stuctural modification under different thawing methods Meat Sci. 2019 147 108 115 10.1016/j.meatsci.2018.09.003 30219362 4. Jia G. Sha K. Meng J. Liu H. Effect of high voltage electrostatic field treatment on thawing characteristics and post-thawing quality of lightly salted, frozen pork tenderloin LWT-Food Sci. Technol. 2019 99 268 275 10.1016/j.lwt.2018.09.064 5. Cai L. Zhang W. Cao A. Cao M. Effects of different thawing methods on the quality of largemouth bass (Micropterus salmonides) LWT-Food Sci. Technol. 2020 120 108908 10.1016/j.lwt.2019.108908 6. Li F. Wang B. Kong B. Shi S. Xia X. Decreased gelling properties of protein in mirror carp (Cyprinus carpio) are due to protein aggregation and structure deterioration when subjected to freeze-thaw cycles Food Hydrocoll. 2019 97 105223 10.1016/j.foodhyd.2019.105223 7. Wang B. Kong B. Li F. Liu Q. Zhang H. Xia X. Changes in the thermal stability and structure of protein from porcine longissimus dorsi induced by different thawing methods Food Chem. 2020 316 126375 10.1016/j.foodchem.2020.126375 32062575 8. Liu Y. Chen S. Pu Y. Muhammad A. Hang M. Ultrasound-assisted thawing of mango pulp: Effect on thawing rate, sensory, and nutritional properties Food Chem. 2019 286 576 583 10.1016/j.foodchem.2019.02.059 30827649 9. Lee S. Kim E. Park D. Choi M. Two-stage air thawing as an effective method for controlling thawing temperature and improving the freshness of frozen pork loin LWT-Food Sci. Technol. 2021 140 110668 10.1016/j.lwt.2020.110668 10. Anderson B. Sun S. Erdogdu F. Singh R. Thawing and freezing of selected meat products in household refrigerators Int. J. Refrig. 2004 27 63 72 10.1016/S0140-7007(03)00093-8 11. Li D. Zhao H. Muhammad A. Song L. Liu D. The comparison of ultrasound-assisted thawing, air thawing and water immersion thawing on the quality of slow/fast freezing bighead carp (Aristichthys nobilis) fillets Food Chem. 2020 320 126614 10.1016/j.foodchem.2020.126614 32208187 12. Mousakhani-Ganjeh A. Hamdami N. Soltanizadeh N. Effect of high voltage electrostatic field thawing on the lipid oxidation of frozen tuna fish (Thunnus albacares) Innov. Food Sci. Emerg. Technol. 2016 36 42 47 10.1016/j.ifset.2016.05.017 13. Cai L. Zhang W. Cao A. Cao M. Li J. Effects of ultrasonics combined with far infrared or microwave thawing on protein denaturation and moisture migration of sciaenops ocellatus (red drum) Ultrason. Sonochem. 2019 55 96 104 10.1016/j.ultsonch.2019.03.017 31084796 14. Sun Q. Kong B. Liu S. Zheng O. Zhang C. Ultrasound-assisted thawing accelerates the thawing of common carp (Cyprinus carpio) and improves its muscle quality LWT-Food Sci. Technol. 2021 141 111080 10.1016/j.lwt.2021.111080 15. Chandrapala J. Oliver C. Kentish S. Ashokkumar M. Use of power ultrasound to improve extraction and modify phase transitions in food processing Food Res. Int. 2012 29 67 91 10.1080/87559129.2012.692140 16. Wu X. Zhang M. Adhikari B. Sun J. Recent developments in novel freezing and thawing technologies applied to foods Food Sci. Nutr. 2016 57 3620 3631 10.1080/10408398.2015.1132670 17. Cheng X. Zhang M. Adhikari B. Effects of ultrasound-assisted thawing on the quality of edamames [Glycine max (L.) Merrill] frozen using different freezing methods Food Sci. Biotechnol. 2014 23 1095 1102 10.1007/s10068-014-0150-0 18. Kissam A. Nelson R. Ngao J. Hunter P. Water-thawing of fish using low frequency acoustics J. Food Sci. 1982 47 71 75 10.1111/j.1365-2621.1982.tb11029.x 19. Gambuteanu C. Alexe P. Effects of ultrasound assisted thawing on microbiological, chemical and technological properties of unpackaged pork longissimus dorsi Food Technol. 2013 37 98 107 20. Cai L. Cao M. Cao A. Regenstein J. Li J. Guan R. Ultrasound or microwave vacuum thawing of red seabream (Pagrus major) fillets Ultrason. Sonochem. 2018 47 122 132 10.1016/j.ultsonch.2018.05.001 29908601 21. Taher B. Farid M. Cyclic microwave thawing of frozen meat: Experimental and theoretical investigation Chem. Eng. Process. 2001 40 379 389 10.1016/S0255-2701(01)00118-0 22. Choi E. Park H. Chung Y. Park S. Jin S. Chun H. Effect of tempering methods on quality changes of pork loin frozen by cryogenic immersion Meat Sci. 2017 124 69 76 10.1016/j.meatsci.2016.11.003 27838574 23. Arocas A. Sanz T. Hernando M. Fiszman S. Comparing microwave- and water bath-thawed starch-based sauces: Infrared thermography, rheology and microstructure Food Hydrocoll. 2011 25 1554 1562 10.1016/j.foodhyd.2011.01.013 24. Xia X. Kong B. Liu Q. Liu J. Physicochemical change and protein oxidation in porcine longissimus dorsi as influenced by different freeze-thaw cycles Meat Sci. 2009 83 239 245 10.1016/j.meatsci.2009.05.003 20416749 25. Xia X. Kong B. Liu J. Diao X. Liu Q. Influence of different thawing methods on physicochemical changes and protein oxidation of porcine longissimus muscle LWT-Food Sci. Technol. 2012 46 280 286 10.1016/j.lwt.2011.09.018 26. Pan N. Wan W. Du X. Kong B. Liu Q. Lv H. Xia X. Li F. Mechanisms of change in emulsifying capacity induced by protein denaturation and aggregation in quick-frozen pork patties with different fat levels and freeze-thaw cycles Foods 2022 11 44 10.3390/foods11010044 27. Du X. Chang P. Tian J. Kong B. Sun F. Xia X. Effect of ice structuring protein on the quality, thermal stability and oxidation of mirror carp (Cyprinus carpio L.) induced by freeze-thaw cycles LWT-Food Sci. Technol. 2020 124 109140 10.1016/j.lwt.2020.109140 28. Du X. Li H. Dong C. Ren Y. Pan N. Kong B. Liu H. Xia X. Effect of ice structuring protein on the microstructure and myofibrillar protein structure of mirror carp (Cyprinus carpio L.) induced by freeze-thaw processes LWT-Food Sci. Technol. 2021 139 110570 10.1016/j.lwt.2020.110570 29. Amiri A. Mousakhani-Ganjeha A. Shafiekhanib S. Mandalc R. Singhc A. Kenarid R. Effect of high voltage electrostatic field thawing on the functional and physicochemical properties of myofibrillar proteins Innov. Food Sci. Emerg. Technol. 2019 56 102191 10.1016/j.ifset.2019.102191 30. Li F. Du X. Ren Y. Kong B. Wang B. Xia X. Bao Y. Impact of ice structuring protein on myofibrillar protein aggregation behaviour and structural property of quick-frozen patty during frozen storage Int. J. Biol. Macromol. 2021 178 136 142 10.1016/j.ijbiomac.2021.02.158 33636271 31. Du X. Zhao M. Pan N. Wang C. Xia X. Zhang D. Tracking aggregation behaviour and gel properties induced by structural alterations in myofibrillar protein in mirror carp (cyprinus carpio) under the synergistic effects of pH and heating Food Chem. 2021 362 130222 10.1016/j.foodchem.2021.130222 34090040 32. Jin S. Yang H. Liu F. Diao X. Feng X. Effect of Dietary Curcumin on the Growth Performance, Serum Antioxidation and Meat Quality of Ducks (Anas Platyrhynchos) Foods 2021 10 2981 10.3390/foods10122981 34945532 33. Li F. Zhong Q. Kong B. Wang B. Pan N. Xia X. Deterioration in quality of quick-frozen pork patties induced by changes in protein structure and lipid and protein oxidation during frozen storage Food Res. Int. 2020 133 109142 10.1016/j.foodres.2020.109142 32466938 34. Wang B. Li H. Huang Z. Kong B. Liu Q. Wang H. Xu M. Xia X. Dynamic changes in the qualities and heterocyclic aromatic amines of roasted pork induced by frying temperature and time Meat Sci. 2021 176 108457 10.1016/j.meatsci.2021.108457 33618127 35. Holman B.W.B. Alvarenga T. Vandeven R. Hopkins D. A comparison of technical replicate (cuts) effect on lamb Warner-Bratzler shear force measurement precision Meat Sci. 2015 105 93 95 10.1016/j.meatsci.2015.02.013 25828163 36. Pan N. Dong C. Du X. Kong B. Xia X. Effect of freeze-thaw cycles on the quality of quick-frozen pork patty with different fat content by consumer assessment and instrument-based detection Meat Sci. 2020 172 108313 10.1016/j.meatsci.2020.108313 32980723 37. Jin S. Pang Q. Yang H. Diao X. Feng X. Effects of dietary resveratrol supplementation on the chemical composition, oxidative stability and meat quality of ducks (Anas platyrhynchos) Food Chem. 2021 363 130263 10.1016/j.foodchem.2021.130263 34116495 38. Wang L. Xiong Y. Inhibition of lipid oxidation in cooked beef patties byhydrolyzed potato protein is related to its reducing andradical scavenging ability J. Agric. Food Chem. 2005 53 9186 9192 10.1021/jf051213g 16277421 39. Wang B. Li F. Pan N. Kong B. Xia X. Effect of ice structuring protein on the quality of quick-frozen patties subjected to multiple freeze-thaw cycles Meat Sci. 2021 172 108335 10.1016/j.meatsci.2020.108335 33059179 40. Heck R.T. Fagundes M.B. Cichoski A.J. De Menezes C.R. Barin J.S. Lorenzo J.M. Campagnol P.C. Volatile compounds and sensory profile of burgers with 50% fat replacement by microparticles of chia oil enriched with rosemary Meat Sci. 2018 148 164 170 10.1016/j.meatsci.2018.10.017 30388481 41. Ares G. Jaeger S. Check-all-that-apply questions: Influence of attribute order on sensory product characterization Food Qual. Prefer. 2013 28 141 153 10.1016/j.foodqual.2012.08.016 42. Vidal L. Tárrega A. Antúnez L. Ares G. Jaeger S.R. Comparison of correspondence analysis based on Hellinger and chi-square distances to obtain sensory spaces from check-all-that-apply (CATA) questions Food Qual. Prefer. 2015 43 106 112 10.1016/j.foodqual.2015.03.003 43. Hong H. Luo Y. Zhou Z. Bao Y. Lu H. Shen H. Effects of different freezing treatments on the biogenic amine and quality changes of bighead carp (aristichthys nobilis) heads during ice storage Food Chem. 2013 138 1476 1482 10.1016/j.foodchem.2012.11.031 23411270 44. Sun Q. Sun F. Xia X. Xu H. Kong B. The comparison of ultrasound-assisted immersion freezing, air freezing and immersion freezing on the muscle quality and physicochemical properties of common carp (cyprinus carpio) during freezing storage Ultrason. Sonochem. 2019 51 281 291 10.1016/j.ultsonch.2018.10.006 30337027 45. Boonsumrej S. Chaiwanichsiri S. Tantratian S. Suzuki T. Takai R. Effects of freezing and thawing on the quality changes of tiger shrimp (Penaeus monodon) frozen by air-blast and cryogenic freezing J. Food Eng. 2007 80 292 299 10.1016/j.jfoodeng.2006.04.059 46. Ali S. Zhang W. Rajput N. Khan M.A. Li C. Zhou G. Effect of multiple freeze-thaw cycles on the quality of chicken breast meat Food Chem. 2015 173 808 814 10.1016/j.foodchem.2014.09.095 25466093 47. Faustman C. Cassens R. The biochemical basis for discoloration in fresh meat: A review J. Muscle Foods 1990 1 217 243 10.1111/j.1745-4573.1990.tb00366.x 48. Hong G.P. Park S.H. Kim J.Y. Min S.G. The effects of high pressure and various binders on the physico-chemical properties of restructured pork meat Asian-Australas. J. Anim. Sci. 2006 19 1484 1489 10.5713/ajas.2006.1484 49. Zhang L. Du H. Zhang P. Kong B. Liu Q. Heterocyclic aromatic amine concentrations and quality characteristics of traditional smoked and roasted poultry products on the northern Chinese market Food Chem. Toxicol. 2020 135 110931 10.1016/j.fct.2019.110931 31678486 50. Farouk M.M. Farouk K.J. Wieliczko I. Ultra-fast freezing and low storage temperatures are not necessary to maintain the functional properties of manufacturing beef Meat Sci. 2004 66 171 179 10.1016/S0309-1740(03)00081-0 22063945 51. Beltrán J. Bellés M. Effect of freezing on the quality of meat Ref. Modul. Food Sci. 2019 2 493 497 10.1016/B978-0-08-100596-5.22461-X 52. Abdallah M.B. Marchello J.A. Ahmad H.A. Effect of freezing and microbial growth on myoglobin derivatives of beef J. Agric. Food Chem. 1999 47 4093 4099 10.1021/jf9809434 10552772 53. Muela E. Sañudo C. Campo M.M. Medel I. Beltrán J.A. Effect of freezing method and frozen storage duration on instrumental quality of lamb throughout display Meat Sci. 2010 84 662 669 10.1016/j.meatsci.2009.10.028 20374840 54. Kang H. Lee H. Park J. Kum J. Effect of microwave treatment on the physicochemical and microbiological characteristics of beef loin during storage at 4 °C Korean J. Food Sci. Technol. 2013 45 161 166 10.9721/KJFST.2013.45.2.161 55. Añón M. Calvelo A. Freezing rate effects on the drip loss of frozen beef Meat Sci. 1980 4 1 14 10.1016/0309-1740(80)90018-2 22055482 56. Ozlu N. Akten B. Timm W. Haseley N. Steen H. Steen J. Phosphoproteomics Wsbm 2010 2 255 276 10.1002/wsbm.41 57. Coombs C.E.O. Holman B.W.B. Friend M.A. Long-term red meat preservation using chilled and frozen storage combinations: A review Meat Sci. 2017 125 84 94 10.1016/j.meatsci.2016.11.025 27918929 58. Seyfert M. Mancini R.A. Hunt M.C. Tang J. Faustman C. Garcia M. Color stability, reducing activity, and cytochrome c oxidase activity of five bovine muscles J. Agric. Food Chem. 2006 54 8919 8925 10.1021/jf061657s 17090142 59. Washington T. Reecy J. Thompson R. Lowe L. McClung J. Carson J. Lactate dehydrogenase expression at the onset of altered loading in rat soleus muscle J. Appl. Physiol. 2004 97 1424 1430 10.1152/japplphysiol.00222.2004 15358753 60. Guro I. Tveit G. Backi C. Jónsson Á. Karlsdóttir M. Lunestad B. Effects of controlled thawing media temperatures on quality and safety of pre-rigor frozen Atlantic cod (Gadus morhua) LWT-Food Sci. Technol. 2018 90 138 144 10.1016/j.lwt.2017.12.030 61. Pan B. Yeh W. Biochemical and morphological changes in grasss hrimp (Penaeus monodon) muscle following freezing by air blast and liquid nitrogen methods J. Food Biochem. 1993 17 147 160 10.1111/j.1745-4514.1993.tb00464.x 62. Lagerstedt Å. Lundström K. Enfält L. Johansson L. Effect of freezing on sensory quality, shear force and water loss in beef M. longissimus dorsi Meat Sci. 2008 80 457 461 10.1016/j.meatsci.2008.01.009 22063353 63. Dransfield E. Calpains from thaw rigor muscle Meat Sci. 1996 43 311 320 10.1016/S0309-1740(96)00022-8 22060598 64. Zainudin M. Poojary M. Jongberg S. Lund M. Light exposure accelerates oxidative protein polymerization in beef stored in high oxygen atmosphere Food Chem. 2019 299 125 132 10.1016/j.foodchem.2019.125132 31299519 65. Jia G. Nirasawa S. Ji X. Luo Y. Liu H. Physicochemical changes in myofibrillar proteins extracted from pork tenderloin thawed by a high-voltage electrostatic field Food Chem. 2018 240 910 916 10.1016/j.foodchem.2017.07.138 28946361 66. Jambrak A. Mason T. Lelas V. Herceg Z. Herceg I. Effect of ultrasound treatment on solubility and foaming properties of whey protein suspensions Food Eng. 2008 86 281 287 10.1016/j.jfoodeng.2007.10.004 67. Jayasooriya S.D. Torley P.J. D’arcy B.R. Bhandari B.R. Effect of high power ultrasound and ageing on the physical properties of bovine Semitendinosus and Longissimus muscles Meat Sci. 2007 75 628 639 10.1016/j.meatsci.2006.09.010 22064027 68. Dickens J. Lyon C. Wilson R. Effect of ultrasonic radiation on some physical characteristics of broiler breast muscle and cooked meat Poult. Sci. 1991 70 389 396 10.3382/ps.0700389 69. Mousakhani-Ganjeh A. Hamdami N. Soltanizadeh N. Impact of high voltage electric field thawing on the quality of frozen tuna fish (Thunnus albacares) J. Food Eng. 2015 156 39 44 10.1016/j.jfoodeng.2015.02.004 70. Mcdonnell C. Allen P. Duggan E. Arimi J. Casey E. Duane G. James G. The effect of salt and fibre direction on water dynamics, distribution and mobility in pork muscle: A low field NMR study Meat Sci. 2013 95 51 58 10.1016/j.meatsci.2013.04.012 23648432 71. Zhang M. Li F. Diao X. Kong B. Xia X. Moisture migration, microstructure damage and protein structure changes in porcine longissimus muscle as influenced by multiple freeze-thaw cycles Meat Sci. 2017 133 10 18 10.1016/j.meatsci.2017.05.019 28577375 72. Shao J. Deng Y. Song L. Batur A. Jia N. Liu D. Investigation the effects of protein hydration states on the mobility water and fat in meat batters by LF-NMR technique LWT-Food Sci. Technol. 2016 66 1 6 10.1016/j.lwt.2015.10.008 73. Han M. Wang P. Xu X. Zhou G. Low-field nmr study of heat-induced gelation of pork myofibrillar proteins and its relationship with microstructural characteristics Food Res. Int. 2014 62 1175 1182 10.1016/j.foodres.2014.05.062 74. Kim H. Kim J. Seo J. Setyabrata D. Kim Y. Effects of aging/freezing sequence and freezing rate on meat quality and oxidative stability of pork loins Meat Sci. 2018 139 162 170 10.1016/j.meatsci.2018.01.024 29427852 75. Farag K. Lyng J. Morgan J. Cronin D. A comparison of conventional and radio frequency thawing of beef meats: Effects on product temperature distribution Food Bioprocess Technol. 2011 4 1128 1136 10.1007/s11947-009-0205-z 76. Li B. Sun D. Effect of power ultrasound on freezing rate during immersion freezing of potatoes J. Food Eng. 2002 55 77 282 10.1016/S0260-8774(02)00102-4 77. Cao M. Cao A. Wang J. Cai L. Regenstein J. Ruan Y. Effect of magnetic nanoparticles plus microwave or far-infrared thawing on protein conformation changes and moisture migration of red seabream (Pagrus major) fillets Food Chem. 2018 266 498 507 10.1016/j.foodchem.2018.06.057 30381218 78. Lui Z. Xiong Y. Chen J. Protein oxidation enhances hydration but suppresses water-holding capacity in Porcine Longissimus muscle J. Agric. Food Chem. 2010 58 10697 10704 10.3168/jds.2007-0179 20806938

PMC009xxxxxx/PMC9099601.txt

==== Front Plants (Basel) Plants (Basel) plants Plants 2223-7747 MDPI 10.3390/plants11091125 plants-11-01125 Article Diazotrophic Bacteria Is an Alternative Strategy for Increasing Grain Biofortification, Yield and Zinc Use Efficiency of Maize https://orcid.org/0000-0002-9451-0508 Jalal Arshad 1 https://orcid.org/0000-0002-3894-9559 Oliveira Carlos Eduardo da Silva 1 Fernandes Henrique Benetasse 1 https://orcid.org/0000-0001-5118-7459 Galindo Fernando Shintate 2 https://orcid.org/0000-0002-1813-490X da Silva Edson Cabral 3 Fernandes Guilherme Carlos 1 https://orcid.org/0000-0002-1783-3311 Nogueira Thiago Assis Rodrigues 1 de Carvalho Pedro Henrique Gomes 1 Balbino Vinícius Rodrigues 1 de Lima Bruno Horschut 1 https://orcid.org/0000-0003-2303-3465 Teixeira Filho Marcelo Carvalho Minhoto 1* Reboredo Fernando Henrique Academic Editor 1 Department of Plant Protection, São Paulo State University (UNESP), Rural Engineering and Soils (DEFERS), Ilha Solteira 15385-000, SP, Brazil; arshad.jalal@unesp.br (A.J.); ces.oliveira@unesp.br (C.E.d.S.O.); henrique.b.fernandes@unesp.br (H.B.F.); guilherme.carlos.fernandes@gmail.com (G.C.F.); tar.nogueira@unesp.br (T.A.R.N.); pedro.goomes04@gmail.com (P.H.G.d.C.); viniciusrbalbinoega@gmail.com (V.R.B.); bruno.horschut@unesp.br (B.H.d.L.) 2 Center for Nuclear Energy in Agriculture (CENA), University of São Paulo (USP), Piracicaba 13416-000, SP, Brazil; fs.galindo@yahoo.com.br 3 Rio Verde Campus, Goiano Federal Institute, Rio Verde 75901-970, GO, Brazil; edsoncabralsilva@gmail.com * Correspondence: mcm.teixeira-filho@unesp.br 21 4 2022 5 2022 11 9 112530 3 2022 18 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Biofortification of cereal crops with zinc and diazotrophic bacteria is a sustainable solution to nutrient deficiency and hidden hunger. The inoculation of staple grain crops such as maize is increased with reducing productivity losses while improving nutrition and use efficiency under climatic extremes and weathered soils of tropical savannah. Therefore, objectives of our study were to evaluate the influence of seed inoculation with diazotrophic bacteria (No inoculation–Control, Azospirillum brasilense, Bacillus subtilis, and Pseudomonas fluorescens) together with residual effect of soil Zn (absence and presence) on growth, yield, Zn nutrition, Zn use efficiencies, and intake of maize in 2019 and 2020 cropping seasons. The inoculation of B. subtilis increased hundred grain mass and yield (14.5 and 17%), while P. fluorescens under residual Zn fertilization has improved shoot and grain Zn concentration in shoot (29.5 and 30.5%). and grain (25.5 and 26.2%), while improving Zn accumulation in shoot (33.8 and 35%) and grain (37.2 and 42%) of maize. The estimated Zn intake in maize was also increased with A. brasilense inoculation and residual Zn application. The Zn use efficiencies including Zn use efficiency, agro-physiological, and utilization efficiency was increased with B. subtilis, while applied Zn recovery was increased with A. brasilense inoculations under residual Zn fertilization. Zinc use efficiency was increased by 93.3 and 397% with inoculation of B. subtilis regardless of Zn application. Therefore, inoculation with B. subtilis and P. fluorescens along residual Zn fertilization is considered the most effective and sustainable strategy for agronomic biofortification of maize under harsh tropical conditions of Brazil. Zea mays L. agronomic biofortification zinc uptake productivity zinc efficiencies zinc nutrition inoculation ==== Body pmc1. Introduction Maize (Zea mays L.) is a crop of social and economic importance, and feeds more than 65% of the global population with a sustainable intake of proteins and calories [1]. It is a versatile cereal cultivated in diversified environments due to its changing food habits and increasing consumption by non-vegetarians [2]. Brazil is the 3rd largest producer of maize around the world with a production of 95 million tons from 17.5 million hectares [3], but still low in average yield production as compared to American and European regions [4]. Maize is inherently poor in minerals’ concentration which is usually plagued by a widespread zinc (Zn) deficiency in tropical regions and ultimately confronts plants’ nutrient acquisition, productivity, and food quality as well as human nutrition and health [5]. Zinc deficiency is a global threat, affecting one third of agricultural soils, and leading to poor production and nutritional quality of cereal crops [6]. Zinc soil deficiency is mainly caused by its abundant soil silicate, oxide, phosphate, and carbonates in soil as well as extensive farming and chemical fertilization, and inadequate irrigation [7]. In addition, Zn is the most transitional nutrient for plant physiological processes, protein synthesis, energy production, genes expression, photosynthesis, and enzymatic activities, as well as pollen fertility, and hormonal and carbohydrate metabolism while discouraging pathogen infestation in cereal crops [8]. Besides this, cereals based on low Zn nutritional security are mainly contributing to human Zn deficiency, and have become the challenge of the day, especially in developing countries [9] and tropical soils [10]. This has introduced the requirement of ultra-nourished strategy such as agronomic biofortification to alleviate malnutrition with effective improvement of nutrition and dietary consumption for the targeted population [11]. However, biofortification of crops with single nutrient in soil application may not be enough for better growth, nutrition, and productivity under harsh tropical conditions. Thus, microbes-mediated biofortification of field crops is an ecofriendly and sustainable strategy to better understand transport of nutrients to grains with greater productivity and nutrient use efficiency [10]. Several plant growth promoting bacteria (PGPBs) are being involved in stimulation of different direct (nutrient acquisition and growth stimulation) and indirect mechanisms (stress diminution and bio-control resistance) that can improve nutrient cycling, maintaining homeostasis, and decomposition of organic material with greater crop production under a sustainable and ecofriendly environment [12]. These rhizosphere bacteria unlocked Zn in the soil by establishing association with plant roots for better accessibility to support plant growth and production [7]. These microbes of synthesis chelating compound in the rhizosphere of roots where they form complexes with Zn and increase its availability and consequently, biofortification through production of siderophores, indole acetic acid (IAA), gibberellins and cytokinins, and reducing phytic acid [12,13]. Several genera of beneficial bacteria such as Rhizobium, Pseudomonas, Azospirillum, and Bacillus are being quoted as Zn solubilizer that facilitate translocation of Zn from soil to different plant tissues, promoting productivity and enriching grains, thus supporting ecofriendly agronomic biofortification [7,14,15]. Inclusion of Zn solubilizing bacteria is the most competent, feasible, and least expensive strategy for Zn biofortification of edible grains (especially maize) with admirable results on sustainable agriculture [16]. Maize is currently the largest cereal source in the world, therefore, it is required to determine sources and dissemination of Zn uptake in maize grains for better understanding of its performance on global Zn cycling. The literature is lacking with Zn biofortification of maize under the interaction of diazotrophic bacteria and soil applied Zn in tropical Savannah. There is also a research gap on the association of diazotrophic bacteria and residual Zn fertilization on Zn nutrition, Zn use efficiency (ZnUE), and yield of maize crop. The integrated use of diazotrophic bacteria and chemical fertilizer is an emerging alternative in the agricultural world. Therefore, it was hypothesized that inoculations of diazotrophic bacteria may have synergetic relation with residual Zn application on plant and grain concentrations, growth, yield, ZnUE, and daily intake of biofortified maize grains in tropical Savannah of Brazil. Therefore, the specific objectives of this study were to better evaluate the performing diazotrophic inoculant in the presence and absence of residual soil Zn fertilization on maize growth, leaf, and grain Zn concentration, and accumulation, yield, and Zn use efficiencies in two consecutive growing seasons. 2. Results 2.1. Plant Height, Dry Matter and Grain Yield The insertion of first productive cob, plant height, shoot dry matter, and grain yield of maize were significantly increased with diazotrophic bacterial inoculation in residual Zn applied treatments as compared to without Zn residual treatments (Table 1). The residual Zn applied treatments produced taller plants with elevated insertion of productive cob (first cob insertion) as compared to control. The plant height was increased by 1.9 and 2.2% in 2019–2020 and 2020–2021 cropping seasons, respectively. Seed inoculation with A. brasilense increased plant height by 4.5%, while height of productive cob was increased by 5.7 and 6.4% with seed inoculation of P. fluorescens as compared to control in both cropping seasons, respectively. Shoot dry matter was significantly greater with residual Zn application and bacterial inoculation in 2019–2020 and 2020–2021 cropping seasons (Table 1). Residual Zn applied treatments were observed with greater shoot dry matter (3.7 and 3.9%) as compared to control treatments in both cropping seasons. The treatments with inoculation of A. brasilense were noted with greater dry matter (5.8 and 6.9%), which were statistically at per with treatments of B. subtilis and P. fluorescens in both studied cropping seasons. The interaction of residual Zn doses and bacterial inoculations for shoot dry matter was significant only in the second cropping season (Figure 1A). Hundred grains mass and grain yield of maize were significantly increased with residual Zn application and diastrophic bacteria inoculations in 2019–2020 and 2020–2021 cropping seasons (Table 1). The single effect of residual Zn (8 kg ha−1) increased mass of 100 grains by 5.9 and 6.4% in relation to control. The inoculation with A. brasilense produced heavy 100 grains (9.5%), which was statistically similar to treatments of B. subtilis and P. fluorescens in 2019–2020 maize harvest in comparison to control. The treatments inoculated with B. subtilis increased 100 grains mass by 11.7% in 2020–2021 maize harvest which was statistically at per with treatments of A. brasilense as compared to control. The interaction for 100 grains mass was significant only in second harvest (Figure 1B). In addition, residual Zn applied treatments increased grain yield by 6.6 and 5.3% while inoculation with B. subtilis increased by 14.5 and 17.1% in 2019–2020 and 2020–2021 cropping seasons as compared to control (Table 1). The interactions for grain yield were significant in both cropping seasons (Figure 1C,D). The seed inoculation with B. subtilis was observed with greater grain yield under residual Zn application while A. brasilense treatments were noted with greater grain yield in the treatments without Zn fertilization (control) in both maize harvests. The treatments with P. fluorescens and the control were observed with lower grain yield regardless of the Zn fertilization in both years of crop harvest (Figure 1C,D). 2.2. Zinc Concentration in Leaf, Shoot, and Grains Residual Zn application and bacteria inoculation significantly increased leaf, shoot, and grain Zn concentrations of maize in 2019–2020 and 2020–2021 cropping seasons (Table 2). Leaf Zn concentration was increased by 15.2 and 17.6% under soil applied Zn doses in relation to control in the first and second cropping seasons, respectively. Inoculation with P. fluorescens was observed with higher leaf Zn concentration (23.8 and 34.1%), which were statistically at per with the treatments inoculated with B. subtilis and A. brasilense in both cropping seasons. The interaction of residual Zn and bacterial inoculations for leaf Zn concentration was significant only in the 2020–2021 cropping season (Figure 2A). The inoculations of all studied bacteria increased leaf Zn concentration in the presence of residual Zn fertilization. However, inoculation with P. fluorescens was observed for higher leaf Zn concentration in the presence of residual Zn fertilization while A. brasilense with lower in the absence of Zn fertilization (Figure 2A). Maize shoot Zn concentration was increased by 15.2 and 15.7% as a function of residual Zn fertilization in 2019–2020 and 2020–2021 harvest as compared without Zn fertilization treatments (Table 2). The inoculation with P. fluorescens increased shoot Zn concentration by 29.5 and 30.5% in the first and second harvest as compared to control, which were statistically similar to the treatments inoculated with B. subtilis and A. brasilense (Table 2). Grain Zn concentration in 2019–2020 and 2020–2021 maize harvest were significantly increased by 12.7 and 18.2% under 8 kg ha−1 soil residual Zn fertilization while inoculation with A. brasilense increased grain Zn concentration by 25.5 and 26.2% as compared to control (Table 2). The interaction for grain Zn concentration was significant only the second cropping season (Figure 2B). The inoculation with A. brasilense was noted for higher grain Zn concentration regardless of the residual Zn fertilization, while treatments with B. subtilis were observed with low grain Zn concentration in the absence of Zn fertilization (Figure 2B). 2.3. Zinc Shoot and Grain Accumulation, Partitioning, and Intake in Maize Zinc accumulation in shoot and grain, partitioning index, and estimated Zn intake in maize consumption were significantly influenced by residual Zn doses and diazotrophic bacteria inoculations in maize cropping seasons of 2019–2020 and 2020–2021 (Table 3). Shoot Zn accumulation in maize was improved by 19.2 and 20.5% with residual Zn fertilization in first and second maize harvest, respectively. Treatments with inoculation of P. fluorescens were observed with better accumulation of Zn in shoot (33.8 and 35%) in relation to control, which were statistically similar to the values obtained in treatments with A. brasilense in both maize harvest seasons (Table 3). Residual Zn fertilization in 2019–2020 and 2020–2021 maize cropping season improved grain Zn accumulation by 20.2 and 24.6% (Table 3). Inoculation with A. brasilense was noted with higher grain Zn accumulation (37 and 42%) in first and second maize cropping seasons, which was statistically at per with inoculated treatments of B. subtilis and P. fluorescens in 2019–20202 maize harvest, and with B. subtilis in 2020–2021 maize harvest. The interactions of the study factors for grain Zn accumulation were significant in both cropping seasons (Figure 2C,D). Inoculation with A. brasilense and B. subtilis tended to perform better for grain Zn accumulation under residual Zn fertilization while B. subtilis in the absence of residual Zn fertilization was observed with low grain Zn accumulation in both crop harvests (Figure 2C,D). Zinc partitioning index was not significantly influenced by residual Zn fertilization and bacteria inoculation (Table 3). The treatments with residual Zn fertilization and bacteria inoculations were not statistically different, however, inoculation with P. fluorescens performed better in partitioning Zn to grains from low Zn in soil, which was statistically similar to non-inoculated treatments. The estimated daily Zn intake in maize consumption in Brazil was significantly increased with residual Zn fertilization and diazotrophic bacteria inoculation in 2019–2020 and 2020–2021 maize harvest seasons (Table 3). The residual Zn fertilization increased daily Zn intake by 14.3 and 17.4% in the first and second cropping seasons as compared to control. The treatments with inoculation of A. brasilense tended to increase Zn intake by 26.3 and 22.7% in 2019–2020 and 2020–2021 maize harvest, which were statistically similar to the treatments with B. subtilis and P. fluorescens in first season, and with P. fluorescens in second maize copping season. The interaction for daily Zn intake was significant only in the second cropping season (Figure 2E). Inoculation with A. brasilense tended to increase Zn intake in daily maize consumption regardless of the Zn fertilization, while B. subtilis was observed with lower daily Zn intake in the absence of residual Zn fertilization (Figure 2E). 2.4. Zinc Efficiencies Zinc efficiencies such as Zn use efficiency, agro-physiological efficiency, utilization efficiency, and applied Zn recovery were positively increased by diazotrophic bacteria inoculations in residual Zn fertilization (Table 4). Zinc use efficiency (ZnUE) of maize was increased with inoculation of B. subtilis in the treatments applied with residual Zn fertilizations in the first and second maize cropping seasons (Table 4). The lower ZnUE was observed in the treatments without inoculations. Agro-physiological efficiency (APE) was statistically not significant in 2019–2020 maize cropping season. Interestingly, APE of maize was significantly increased with inoculation of B. subtilis under residual Zn fertilization in 2020–2021 cropping seasons as compared to control (Table 4). The highest APE was observed in the treatments of B. subtilis while the lowest was recorded in treatments of P. fluorescens inoculation (Table 4). Zinc utilization efficiency (UE) was increased by 77.4 and 190.8% with seed inoculation of B. subtilis in residual Zn fertilization in relation to non-inoculated treatments, which were statistically similar to the treatments of A. brasilense in the first and second maize cropping seasons, respectively (Table 4). The highest Zn utilization efficiency was observed with B. subtilis while the lowest was noted in control treatments (Table 4). Inoculation with A. brasilense under residual Zn fertilization performed better in recovery of applied Zn fertilization in 2019–2020 and 2020–2021 maize cropping seasons. Applied Zn recovery was increased by 191.6 and 213.3% in the treatments with residual Zn fertilization and A. brasilense inoculation, which were statistically similar to the treatments inoculated with B. subtilis and P. fluorescens in both crop harvests (Table 4). The lowest applied Zn recovery was observed in control (without inoculation) treatments. 2.5. Pearson’s Correlation among Evaluated Attributes of Maize There were overall positive and significant correlations among zinc concentrations in maize plants (leaf, shoot, and grains) and insertion of first productive cob, plant height, shoot dry matter shoot, and grain Zn accumulation, and negative correlation with agro-physiological efficiency, while non-significant correlations with zinc partitioning index, zinc use efficiency, applied zinc recovery, and utilization efficiency (Figure 3A). A positive correlation was observed between leaf, shoot, and grain concentration, and shoot and grain Zn accumulation, daily Zn intake, applied Zn recovery, plant height, shoot dry matter, insertion of productive cob, and grain yield. A negative correlation was noted between Zn partitioning index and plant height, shoot dry matter, insertion of first productive cob, leaf, shoot, and grain concentration, and shoot and grain Zn accumulation, and daily Zn intake. A non-significant correlation was noted between zinc utilization efficiency and Zn partitioning index, zinc use efficiency, and 100 grains mass (Figure 3B). 3. Discussion Agronomic biofortification has been recognized as the most feasible and effective mechanism for correcting zinc (Zn) deficiency in soil and plant along with better quality yield to improve human health. Single Zn fertilization could not facilitate Zn soil, plant, human, and environment at the same time, especially in tropical regions [10,17]. Therefore, integrated use of bio- and mineral fertilizers is an emerging strategy that can mediate nutrient acquisition for soil-plant-human health to facilitate millions in the population in a sustainable and ecofriendly manner. Diazotrophic bacteria are colonializing root rhizosphere to soluble mineral nutrients, stimulating plant growth with greater yield as well as improving acquisition of nutrient to edible grains [18]. The positive correlation between zinc concentrations in maize plants (leaf, shoot, and grains) and insertion of first productive cob, plant height, shoot dry matter, grain yield, and shoot and grain Zn accumulation validated the hypothesis of the current study (Figure 3). Zinc is an essential element of cell development and multiplication as well as pollen fertility for better plant establishment, growth, and reproduction, where its deficiency can plague growth and yield [19,20]. However, the integrated application of diazotrophic bacteria such as Zn solubilizing bacteria and Zn fertilization is one of the best alternative and sustainable strategies to improve Zn nutrition with greater growth and productivity [10,16]. Therefore, the current results verified that residual Zn fertilization and inoculation of A. brasilense and P. fluorescens has increased plant height, height of insertion of first productive cob, dry matter, and hundred grains and grain yield of maize (Table 1; Figure 1). Several previous studies reported that Zn solubilizing bacteria can rapidly colonialize in root rhizosphere, where they could increase Zn solubilization by producing siderophores, chelators, and several plant growth hormones such as indole acetic acid (IAA), gibberellins, and cytokinins that are immensely linked to better plant health, growth, and production [16,21]. Zinc solubilizing bacteria such A. brasilense [17] and P. fluorescens [22], together with Zn fertilization, are being reported with greater growth and yield of cereal crops. Maize grains are inherently low in Zn concentration which can further hinder nutrient acquisition and yield [5] in Zn-deficient soils. The adequate Zn concentration in maize leaf is ranging from 15–50 mg kg−1, below this is considered adequate [23]. Plants and microbes interactions in root rhizosphere stimulate nutrient cycling by solubilization, mineralization, carboxylation, and hormones synthesis [24,25], which could empower Zn concentration and uptake in cereals to support biofortification [22]. Thus, our results verified that residual Zn fertilization and inoculation with P. fluorescens and A. brasilense has increased concentration in leaf, shoot, and grains (Table 3; Figure 2A,B), and Zn accumulation in shoot and grains of maize (Table 4; Figure 2C,D). The reason might be the presence of microbes in root rhizosphere which could interact with applied inoculants to stimulate transportation of nutrients (especially Zn) to leaf, shoot, and grains by modifying root architecture, secreting phenolic acids, and reducing phytic acid supply to grains [26]. Several other studies exhibited that different strains of Azospirillum, Bacillus, and Pseudomonas sp. promote availability and solubilization of nutrients by synthesis of different plant hormones and enzymes, as well as biological nitrogen fixation [27,28]. Abadi et al. [29] exhibited that inoculants of Pseudomonas sp. could alleviate Zn deficiency by increasing root branching and proliferation for greater Zn accumulation and better plant health under harsh environmental circumstances. Zinc is an important nutrient of several biological and anabolic processes of humans, while its deficiency could lead to several disorders and hidden hunger [30]. Zinc is an indispensable element for plants and humans to perform their functions and increase productivity [31]. In addition, most of the population are consuming cereals to meet their daily food requirements and therefore, an urgent-based approach such as microbes-mediated Zn biofortification of staple crops can be the most authentic strategy to increase Zn concentration in edible crops under Zn-deficient soils [22]. In this context, the current research indicated that residual Zn fertilization and bacterial inoculation has increased estimated daily Zn intake in Brazil while Zn partitioning was not statistically different (Table 3; Figure 2E). The fact may be the activation of different mechanisms such as acidification, exchange reactions, chelation, and release of organic matter by soil microbes to solubilize nutrients (especially Zn) for better uptake in edible parts [32]. The strains of Azospirillum, Pseudomonas, and Bacillus sp. are being observed with increasing daily intake and partitioning of Zn from soil to grains of different cereal crops [22,33]. It has also been described that inoculation of wheat Zn solubilizing bacteria could increase root volume, diameter, length, and surface area that has ultimately increased Zn uptake by two folds in edible grains [34]. These microbes had increased bioavailability and transportation of Zn to edible grains by reducing phytic acids, which could substantially increase human consumption in a more green and sustainable manner [14,22]. Zinc efficiencies such as Zn use, agro-physiological, and utilization efficiency, as well as applied Zn recovery have differently responded to inoculation and residual Zn fertilization (Table 4). These efficiencies are derived from shoot and grain Zn concentration in Zn-deficient soils, where only Zn fertilization is a fraction of Zn use efficiency while increasing fertilizer dose could decrease Zn efficiency [35]. Most of the studied efficiencies were increased with inoculation of B. subtilis under residual Zn fertilization (Table 4). This increase might be due to greater growth, yield, and Zn uptake in the current experiment (Table 1, Table 2 and Table 3). Roots and soil Zn interaction are severally decreased due to low soil moisture and organic matter that can limit Zn absorption, however, greater root dry matter can scavenge and intercept nutrients into plants [36], which is the main access point to increase nutrient uptake and assimilation with higher Zn use efficiency. Nutrient use efficiency is not only dependent on nutrient uptake by plants from soil but is also dependent on growth stage, internal transportation, recycling, and mobilization. Several microbes have capability to increase use efficiency of nutrients by solubilization, where application of minerals and inoculation have promising roles to increase nutrient use efficiency in nutrient-deficient soils [37]. Several studies reported that seed inoculation with strains of Bacillus, Pseudomonas, and Azospirillum enhanced Zn translocation to grains with higher Zn use, agro-physiological, and utilization efficiency, and applied Zn recovery in cereal crops [17,22]. 4. Materials and Methods 4.1. Experimental Site and Climate Description Two years of maize (Zea mays L.) field experiments were conducted at Research and Extension Farm of School of Engineering, São Paulo State University (UNESP) in Selvíria, state of Mato Grosso do Sul, Brazil, at geographical coordinates of 20°22′ S latitude, 51°22′ W longitude, and an altitude of 335 m above sea level (Figure 4). The site has been cultivated with cereals-legumes cropping system for more than 28 years, previously cultivated with wheat, wherein the last 13 years were under a no-tillage system. The soil of the experimental site is classified as Rhodic Haplustox [38] and Red Latosol Dystrophic with a clayey texture [39]. The climate of the site is classified as humid tropical of Aw type, rainy in summer, and dry in winter according to Koppen climate classification [40]. The rainfall, minimum, maximum, and average temperatures, and air humidity of maize cultivation period is summarized in Figure 5. 4.2. Soil Analysis Twenty random soil samples were collected before experiment installation with cup auger from 0.00–0.20 m soil layer. The samples were mixed to make a homogeneous/composite sample for the determination of chemical [41] and granulometric characterization such as clay = 433, sand = 470, and silt = 90 g kg−1 following the methodology of Teixeira et al. [42]. The chemical and physical characterization of the site are summarized in Table 5. 4.3. Experimental Design and Treatments The experiments were designed in randomized complete blocks in 4 × 2 factorial scheme with four replications. The treatments consisted of four seed inoculations with diazotrophic bacteria (no inoculation, Azospirillum brasilense, Bacillus subtilis, and Pseudomonas fluorescens) and two residual zinc (Zn) applications (without 0 kg ha−1 of Zn and with 8 kg ha−1 of Zn), applied from zinc sulphate (21% Zn and 10% S). The inoculation of A. brasilense strains Ab-V5 and Ab-V6 (Ab-V5 = CNPSo 2083 and Ab-V6 = CNPSo 2084 with guarantee of 2 × 108 CFU mL−1) was conducted at a dose of 200 mL ha−1 (liquid inoculant) added in a small quantity of water to uniformly mix in around 24 kg of maize seeds sown ha−1. The B. subtilis (strain CCTB04 with guarantee of 1 × 108 CFU mL−1) and P. fluorescens (strain CCTB03 with guarantee of 2 × 108 CFU mL−1) were performed at a dose of 150 mL ha-1 (liquid) according to the recommendation of an inoculants providing company (Total Biotechnology®), Curitiba, Brazil. The inoculations were performed an hour before plantation of the crop, followed in both cropping seasons. Zinc fertilization (0 and 8 kg ha−1) was performed only in 2019 and 2020 (May to September both years) of wheat cropping seasons. Zinc sulphate was manually applied to soil surface between rows of wheat at V1/V2 stage (one to two completely unfolded leaves) and followed by 14 mm irrigation to incorporate into soil. Thus, zinc was not directly applied in maize cultivation season, residual effect of Zn applied in wheat was evaluated in current experiment. 4.4. Plant Materials The experimental site was applied with herbicides glyphosate (1800 g ha−1 of a.i.) and 2,4-D (670 g ha−1 of a.i.) 15 days before plantation. Seeds were chemically treated with Standak Top®, a mix formulation of insecticide-imidacloprid + thiodicarb (45 g + 135 g of a.i. per 100 kg seeds) and fungicide-carbendazim + thiram (45 g + 105 g of a.i. per 100 kg seeds) a day before inoculations and plantation. A simple maize hybrid (FS500PWU-Forseed, registered with National Technical Commission on Biosafety of Brazil under reference no. 1596/2008 for tropical and sub-tropical regions) was sown on 18th November, 2019 and 12 November, 2020 in a no-tillage system at 3.3 seeds m−1. All treatments were applied with 350 kg ha-1 of NPK (08-28-16, urea) on the basis of soil analysis. Each experimental unit was 6 m long with 6 rows, 0.45 m apart with total plot size of 16.2 m2. Post-emergence herbicides atrazine + tembotrione (1000 + 84 g a.i. ha−1, respectively) were applied at V3 stage to control weeds. The topdressing fertilization of nitrogen (120 kg N ha−1 from ammonium sulphate) at V6 stage was performed in all treatments to uniformly distribute on soil surface. The crop was irrigated with pivot-irrigation system at 14 mm water volume according to the need of the crop. The crop was manually harvested on 2 March 2020 and 7 March 2021. 4.5. Evaluations and Analysis 4.5.1. Growth and Yield Attributes Plant height was measured with meter-rod from ground to upper apex. Shoot dry matter was determined by harvesting four central lines, sun dried and weighed. Ten random ears were collected at harvest to count number of rows and grains ear−1plot−1. Hundred grains mass was measured with a precise scale on 13% humidity (wet basis). Ears were collected from central lines of each plot, threshed with electric thresher, and processed to calculate yield in kg ha−1 (productivity at 13% moisture content). The dried grains were then ground in a Willey mill for nutritional analysis. 4.5.2. Nutritional Analysis Twenty random leaves were collected from ear insertion at flowering stage in each plot. The plant material (shoot and grain) was collected at the time harvest. The samples were dried in an air-tight oven at 60 ± 5 °C for 72 h to attain uniform humidity. The material for each attribute was then individually grounded in a stainless-steel Willey knife mill by passing through a 10-mesh sieve in labeled plastic bags. Each sample was weighed (0.25 g), digested with nitroperchloric digestion (HNO3:HClO4 solution), and quantified by atomic absorption spectrophotometry following procedure of [23]. Zinc shoot and grains accumulation (g ha−1) were calculated from respective Zn concentration in shoot and grains, and dry matter yield ha−1, respectively. 4.5.3. Zinc Partitioning Index, Intake, and Use Efficiencies Zinc partitioning index (ZPI) was calculated from the ratio of shoot Zn concentration to total (shoot + grains) Zn concentration in percent following Rengel and Graham [43]. Estimated Zn intake in Brazil (Equation (1)) was calculated from Zn biofortified grains of present study [44]. Brazil per capita maize consumption is around 24.69 kg person−1 year−1 (67.6 g person−1 day−1). Based on this information, estimated Zn intake of biofortified grains was calculated below in Equation (1). Zn intake = [Zn grain] × C (1) where Zn intake (g person−1 day−1) is daily estimated Zn consumption person−1, [Zn grain] (g kg−1) is Zn concentration in biofortified grains, and C (g person−1 day−1) is average maize consumption per person in Brazil [45]. Zinc use efficiency (ZnUE), agro-physiological efficiency (APE), recovery applied Zn (RAZn), and utilization efficiency (UE) were derived from the fractions of Zn concentration and accumulation in shoot and grains, dry matter, and grain yield following procedures of [22,46]. ZnUE = (GYF − GYWF) ÷ Zn applied dose(2) APE = (GYF − GYWF) ÷ (ZnAF − ZnAWF) (3) RAZn (%) = (ZnAF − ZnAWF) ÷ Zn applied dose(4) UE = PE × RAZn (5) where GYF = grain yield in Zn fertilized treatments, GYWF = grain yield in without Zn fertilized treatments, ZnAF = zinc accumulation in shoot + grain within fertilized treatments, ZnAWF = zinc accumulation in shoot + grain without fertilized treatments, and PE = physiological efficiency. 4.6. Statistical Analysis The data were tested for normality with Shapiro–Wilk test which showed that data were normally distributed (W ≥ 0.90). The data were submitted to analysis of variance (F test). Zn soil application, and diazotrophic bacterial inoculations and their interactions were considered fixed effects in the model. When a main effect or interaction was observed significant by F test (p ≤ 0.05), then Tukey test (p ≤ 0.05) was used for comparison of means of residual soil Zn fertilization and diazotrophic bacterial inoculations [47]. The Pearson correlation analysis (p ≤ 0.05) was performed using R software [47]. To create a heatmap, the corrplot package was used, using the "color" and "cor.mtest" functions to calculate the coefficients and p-value matrices. 5. Conclusions Microbes-mediated Zn biofortification is one the most accessible, easy, and authentic strategies to increase Zn concentration and accumulation in the edible part of the maize crop. Our results indicated that residual Zn fertilization is a feasible and sustainable technique which has increased plant growth, yield, and Zn nutrition in both cropping seasons. The inoculation of diazotrophic bacteria along residual Zn fertilization performed better than without Zn fertilized treatments. Seed inoculation of A. brasilense and B. subtilis has increased height of insertion of first productive cob, plant height, shoot dry matter, and grain yield of maize under residual Zn fertilization. Most of the growth and yield attributes performed better with inoculation of A. brasilense in the absence of residual Zn fertilization. Zinc concentration in leaf, and accumulation in shoot and grains of maize were increased with A. brasilense and P. fluorescens under residual Zn fertilization. The highest Zn partitioning and daily Zn intake were also increased with inoculation of P. fluorescens and A. brasilense with residual Zn fertilization. All Zn efficiencies were increased with inoculation of B. subtilis except applied Zn recovery, which was greater with inoculation of A. brasilense when analyzed in residual Zn fertilized treatments. Therefore, inoculation of maize seeds with B. subtilis and P. fluorescens together with residual Zn fertilization could be an efficient alternative mechanism to improve Zn acquisition and use efficiencies, as well as productivity of maize in a sustainable manner in tropical savannahs. Prospective research aiming to improve Zn use efficiency and recovery with inoculation and co-inoculation of diazotrophic bacteria, and their influence on cereal biofortification, physiological, and molecular aspects is required to be carried out in different edaphic conditions to better understand Zn solubilizing bacteria under field conditions. Acknowledgments The authors thank São Paulo State University (UNESP) for providing technology and support. Author Contributions Conceptualization, A.J. and M.C.M.T.F.; methodology, A.J. and C.E.d.S.O.; software, C.E.d.S.O. and F.S.G.; validation, A.J., F.S.G. and E.C.d.S.; formal analysis, H.B.F. and A.J.; investigation, A.J., G.C.F. and H.B.F.; resources, M.C.M.T.F. and T.A.R.N.; data curation, A.J. and C.E.d.S.O.; writing—original draft preparation, A.J.; writing—review and editing, M.C.M.T.F., T.A.R.N., E.C.d.S. and F.S.G.; visualization, B.H.d.L., P.H.G.d.C. and V.R.B.; supervision, M.C.M.T.F.; project administration, A.J. and M.C.M.T.F.; funding acquisition, A.J. and M.C.M.T.F. All authors have read and agreed to the published version of the manuscript. Funding This research received funding from The World Academy of Science (TWAS) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), first author’s doctoral fellowship (CNPq/TWAS grant number: 166331/2018-0), and productivity research grant (award number 311308/2020-1) of corresponding author. This research was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior—Brasil (CAPES/AUXPE Award Number 88887.592666/2020-00 | 0242/2021). Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Maize shoot dry matter (A) and 100 grains mass (B) in 2020–2021 respectively, and grain yield in 2019–2020 (C) and 2020–2021 (D) as function of residual Zn doses and diazotrophic bacteria. Without = control (no inoculation). The uppercase letters are used for inoculation interactions within each level of soil applied residual Zn whereas lowercase letters are used for the interactions of Zn doses (presence and absence) within each inoculation treatment. The identical alphabetic letters do not differ from each other by Tukey test (p < 0.05) for Zn doses and inoculations in 2019–2020 and 2020–2021. Error bars indicate standard error of the mean (n = 4 replications). Figure 2 Leaf zinc concentration (A) and grain zinc concentration (B) in 2020–2021, respectively, and grain zinc accumulation in 2019–2020 (C) and 2020-2021 (D), and estimated daily zinc intake in Brazil in 2020−2021 cropping maize season (E) as function of residual Zn doses and diazotrophic bacteria. Without = control (no inoculation). The uppercase letters are used for inoculation interactions within each level of soil applied residual Zn whereas lowercase letters are used for the interactions of Zn doses (presence and absence) within each inoculation treatment. The identical alphabetic letters do not differ from each other by Tukey test (p < 0.05) for Zn doses and inoculations in 2019–2020 and 2020–2021. Error bars indicate standard error of the mean (n = 4 replications). Figure 3 Heat-map color scale indicating Pearson’s correlation among evaluated parameters of maize plants in response to residual soil Zn applications and diazotrophic bacteria inoculations in 2019–2020 (A) and 2020–2021 (B) cropping seasons. X = indicates a non-significant relationship (p ≤ 0.05). Abbreviations: CI = Insertion of first productive cob, PH = plant height, SM = shoot dry matter, GY = grain yield, LZC = Leaf Zn concentration, SZC = Shoot Zn concentration, GZC = Grain Zn concentration, SZA = Shoot Zn accumulation, GZA = Grain Zn accumulation, ZPI = Zn partitioning index, ZCB = Zn intake in Brazil, ZnUE = Zn use efficiency, APE = Agro-physiological efficiency, RAZn = Applied Zn recovery, and UTE = Utilization efficiency. Figure 4 Location of the experimental area at Extension and Research Farm, UNESP—Ilha Solteira Campus, at Selvíria, state of Mato Grosso do Sul, Brazil (20°22′ S, 51°22′ W, altitude of 335 m) in 2019–2020 and 2020–2021 crop seasons. The map was created by using geographic information system (QGIS) software and the Google Earth program. The QGIS Development Team (2021). Open Source Geospatial Foundation project. http://qgis.osgeo.org, accessed on 27 February 2022. Projection System WGS 84/UTM 200DC (EPSG: 4326). This image was taken from the Google Earth program, Google Company (2021). From map data: Google, Maxar Technologies. Figure 5 Rainfall, maximum, and minimum temperatures, and light radiation were acquired from the weather station of Extension and Research Farm of School of Engineering—UNESP during maize cultivation period from November to March 2019–2020 (A) and 2020–2021 (B). plants-11-01125-t001_Table 1 Table 1 First productive cob insertion, plant height, shoot dry matter, and grain yield of maize as influenced by diazotrophic bacteria and residual zinc doses in 2019–2020 and 2020–2021 cropping season. Treatments Plant Height First Cob Insertion Shoot Dry Matter 100 Grains Mass Grain Yield m kg ha−1 g kg ha−1 2019–2020 2020–2021 2019–2020 2020–2021 2019–2020 2020–2021 2019–2020 2020–2021 2019–2020 2020–2021 Inoculations (I) Without 2.66 b 2.67 b 1.22 b 1.25 b 11,945 b 11,832 30.6 b 31.7 7379 7307 A. brasilense 2.78 a 2.79 a 1.29 ab 1.31 a 12,642 a 12,654 33.5 a 34.4 8109 8233 B. subtilis 2.67 ab 2.72 ab 1.27 ab 1.31 a 12,381 a 12,381 32.7 ab 35.4 8449 8555 P. fluorescens 2.72 ab 2.77 a 1.29 a 1.33 a 12,355 a 12,243 31.8 ab 33.5 7911 7952 Residual Zinc (Zn) Doses (kg ha−1) 0 2.67 b 2.71 b 1.25 a 1.28 b 12,102 b 12,040 31.3 b 32.7 7709 7806 8 2.72 a 2.77 a 1.28 a 1.32 a 12,559 a 12,515 32.9 a 34.8 8215 8218 F-values I 0.004 ** 0.00 ** 0.03 * 0.00 ** 0.00 ** 0.00 ** 0.01 * 0.00 ** 0.00 ** 0.00 ** Zn 0.01 ** 0.00 ** 0.11ns 0.008 ** 0.00 ** 0.00 ** 0.01 * 0.00 ** 0.00 ** 0.00 ** I × Zn 0.63 ns 0.19 ns 0.99 ns 0.86 ns 0.36 ns 0.04 * 0.94 ns 0.02 * 0.03 * 0.02 ** CV (%) 2.1 1.7 3.9 2.4 2.3 1.8 5.2 2.8 4.0 3.6 Means in the column followed by different letters are statistically different by Tukey test, p ≤ 0.05. ** and *—significant at p < 0.01 and p < 0.05, respectively, while ns—non-significant by F-test. plants-11-01125-t002_Table 2 Table 2 Leaf, shoot, and grain zinc (Zn) concentrations of maize as function of residual Zn doses and diazotrophic bacteria in 2019–2020 and 2020–2021 cropping season. Treatments Leaf Zn Concentration Shoot Zn Concentration Grain Zn Concentration mg kg−1 2019–2020 2020–2021 2019–2020 2020–2021 2019–2020 2020–2021 Inoculations (I) Without (control) 20.6 b 21.7 29.1 b 29.5 b 28.2 b 32.5 A. brasilense 23.5 ab 27.4 35.8 ab 35.7 a 35.4 a 41.0 B. subtilis 23.9 ab 28.0 33.4 ab 33.9 ab 32.9 a 36.9 P. fluorescens 25.5 a 29.1 37.7 a 38.5 a 34.6 a 38.5 Residual Zinc (Zn) Doses (kg ha −1 ) 0 21.7 b 24.4 31.6 b 31.9 b 30.8 b 34.1 8 25.0 a 28.7 36.4 a 36.9 a 34.7 a 40.3 F-values I 0.03 * 0.00 * 0.01 * 0.002 ** 0.00 ** 0.00 ** Zn 0.008 ** 0.00 ** 0.01 * 0.002 ** 0.00 ** 0.00 ** I × Zn 0.43 ns 0.04 * 0.76 ns 0.78 ns 0.33 ns 0.03 * CV (%) 13.5 6.2 14.2 12.0 8.6 6.7 Means in the column followed by different letters are statistically different by Tukey test, p ≤ 0.05. ** and *—significant at p < 0.01 and p < 0.05, respectively, while ns—non-significant by F-test. plants-11-01125-t003_Table 3 Table 3 Shoot and grain zinc accumulation, zinc partitioning index, and estimated daily zinc intake by maize in Brazil as function of residual zinc fertilization and diazotrophic bacteria inoculations in 2019–2020 and 2020–2021 cropping seasons. Treatments Shoot Zn Accumulation Grain Zn Accumulation Zn Partitioning Index Zn Intake (Brazil) g ha−1 % g person−1 day−1 2019–2020 2020–2021 2019–2020 2020–2021 2019–2020 2020–2021 2019–2020 2020–2021 Inoculations (I) Without I 348.5 b 349.3 b 208.9 237.7 51.8 a 47.5 a 1.9 b 2.2 A. brasilense 452.8 a 453.4 a 286.7 337.6 50.2 a 46.6 a 2.4 a 2.7 B. subtilis 413.5 ab 420.0 ab 279.5 317.9 50.1 a 47.8 a 2.2 a 2.5 P. fluorescens 466.3 a 471.6 a 273.9 306.2 51.8 a 49.8 a 2.3 a 2.6 Residual Zinc (Zn) Doses (kg ha−1) 0 383.4 b 384.2 b 238.2 266.9 50.5 a 48.1 a 2.1 b 2.3 8 457.2 a 463.0 a 286.3 332.7 51.1 a 47.8 a 2.4 a 2.7 F-values I 0.002 ** 0.00 ** 0.00 ** 0.00 ** 0.86 ns 0.31 ns 0.00 ** 0.00 ** Zn 0.001 ** 0.00 ** 0.00 ** 0.00 ** 0.71 ns 0.77 ns 0.00 ** 0.00 ** I × Zn 0.63 ns 0.80 ns 0.04 * 0.00 ** 0.99 ns 0.78 ns 0.33 ns 0.02 ** CV (%) 13.7 12.5 8.9 7.3 8.6 6.8 8.6 6.6 Means in the column followed by different letters are statistically different by Tukey test, p ≤ 0.05. ** and *—significant at p < 0.01 and p < 0.05, respectively, while ns—non-significant by F-test. plants-11-01125-t004_Table 4 Table 4 Zinc efficiencies of maize as function of residual zinc fertilization and diazotrophic bacteria inoculations in 2019–2020 and 2020–2021 cropping seasons. Treatments ZnUE APE UE AZnR kg kg−1 % 2019–2020 2020–2021 2019–2020 2020–2021 2019–2020 2020–2021 2019–2020 2020–2021 Inoculations (I) Without (control) 164 c 68 c 19 a 4.5 b 297 b 131 b 12 b 15 b A. brasilense 233 b 178 b 7 a 3.8 b 483 a 353 a 35 a 47 a B. subtilis 317 a 270 a 10 a 6.5 a 509 a 381 a 32 a 42 a P. fluorescens 190 c 135 b 7 a 3.2 b 379 b 219 b 33 a 43 a F-values I 0.00 ** 0.00 ** 0.05 * 0.003 ** 0.00 ** 0.00 ** 0.008 ** 0.00 ** CV (%) 8.3 13 54 20 10 17 29 21 ZnUE = Zinc use efficiency, APE = Agro-physiological efficiency, UE = Utilization efficiency, and AZnR = Applied zinc recovery. Means in the column followed by different letters are statistically different by Tukey test, p ≤ 0.05. ** and *—significant at p < 0.01 and p < 0.05, respectively, while ns—non-significant by F-test. plants-11-01125-t005_Table 5 Table 5 Pre-maize experiments soil analysis of composite sample in a soil layer (0–0.20 m) in 2019–2020 and 2020–2021 cropping seasons. Properties Units Status 2019–2020 2020–2021 pH (CaCl2) ---- 5.2 5.3 Organic matter mg dm−3 18 23 P (resin) mg dm−3 38 40 K mmolc dm−3 1.7 1.9 Ca mmolc dm−3 21 22 Mg mmolc dm−3 15 12 B (hot water) mg dm−3 0.14 0.39 Cu (DTPA) * mg dm−3 3.4 3.7 Fe (DTPA) * mg dm−3 25 28 Mn (DTPA) * mg dm−3 38.1 37.3 S-SO4 mg dm−3 4.0 22 H + Al mmolc dm−3 34 31 CEC (pH7) * mmolc dm−3 75.7 66.9 V * % 50 54 Zn content (DTPA) Without Zn fertilization mg dm−3 0.9 1.1 Residual Zn fertilization mg dm−3 2.2 3.0 * CEC: cation exchange capacity, V: base saturation, DTPA: Diethylenetriaminepentaacetic acid. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Xue Y.F. Yue S.C. Liu D.Y. Zhang W. Chen X.P. Zou C.Q. Dynamic zinc accumulation and contributions of pre-and/or post-silking zinc uptake to grain zinc of maize as affected by nitrogen supply Front. Plant Sci. 2019 10 1203 10.3389/fpls.2019.01203 31632429 2. Ayyar S. Appavoo S. Basker M. Pandiyarajan P. Kavimani R. Effect of Zinc and microbial inoculation on soil enzyme activities for maize (Zea mays L.) in black soil Int. J. Curr. Microbiol. Appl. Sci. 2019 8 1804 1814 10.20546/ijcmas.2019.808.213 3. Conab Companhia Nacional de Abastecimento Grains Report 11/2020. Brasília: Conab Available online: https://www.conab.gov.br/info-agro/safras (accessed on 15 March 2022) (In Portuguese) 4. United States Department of Agriculture National Agricultural Statistics Service Crop Production 2020 Available online: http://www.nass.usda.gov/Publications/index.php (accessed on 25 March 2022) 5. Obaid H. Shrestha R.K. Liu D. Elsayed N.S. Ni J. Ni C. Biofortification of Maize with Zinc and Its Effect on Human Health J. Soil Sci. Plant Nutr. 2022 7 1 13 10.1007/s42729-022-00772-5 6. Cakmak I. McLaughlin M.J. White P. Zinc for better crop production and human health Plant Soil 2017 411 1 4 10.1007/s11104-016-3166-9 7. Masood F. Ahmad S. Malik A. Role of Rhizobacterial Bacilli in Zinc Solubilization Microbial Biofertilizers and Micronutrient Availability Khan S.T. Malik A. Springer Cham, Switzerland 2022 361 377 10.1007/978-3-030-76609-2_15 8. Suganya A. Saravanan A. Manivannan N. Role of zinc nutrition for increasing zinc availability, uptake, yield, and quality of maize (Zea mays L.) grains: An overview. Commun Soil Sci. Plant Anal. 2020 51 2001 2021 9. Dhaliwal S.S. Sharma V. Shukla A.K. Verma V. Kaur M. Shivay Y.S. Nisar S. Gaber A. Brestic M. Barek V. Biofortification—A Frontier Novel Approach to Enrich Micronutrients in Field Crops to Encounter the Nutritional Security Molecules 2022 27 1340 10.3390/molecules27041340 35209127 10. Jalal A. Galindo F.S. Boleta E.H.M. da Silva Oliveira C.E. dos Reis A.R. Nogueira T.A.R. Moretti Neto M.J. Mortinho E.S. Fernandes G.C. Teixeira Filho M.C.M. Common Bean Yield and Zinc Use Efficiency in Association with Diazotrophic Bacteria Co-Inoculations Agronomy 2021 11 959 10.3390/agronomy11050959 11. Ahsin M. Hussain S. Rengel Z. Amir M. Zinc status and its requirement by rural adults consuming wheat from control or zinc-treated fields Environ. Geochem. Health 2020 42 1877 1892 10.1007/s10653-019-00463-8 31696401 12. Hakim S. Naqqash T. Nawaz M.S. Laraib I. Siddique M.J. Zia R. Mirza M.S. Imran A. Rhizosphere engineering with plant growth-promoting microorganisms for agriculture and ecological sustainability Front. Sustain. Food Syst. 2021 5 16 10.3389/fsufs.2021.617157 13. Bashir S. Basit A. Abbas R.N. Naeem S. Bashir S. Ahmed N. Ahmed M.S. Ilyas M.Z. Aslam Z. Alotaibi S.S. Combined application of zinc-lysine chelate and zinc-solubilizing bacteria improves yield and grain biofortification of maize (Zea mays L.) PLoS ONE 2021 16 e0254647 10.1371/journal.pone.0254647 34255800 14. Mumtaz M.Z. Ahmad M. Jamil M. Hussain T. Zinc solubilizing Bacillus spp. potential candidates for biofortification in maize Microbiol. Res. 2017 202 51 60 10.1016/j.micres.2017.06.001 28647123 15. Rehman A. Farooq M. Naveed M. Nawaz A. Shahzad B. Seed priming of Zn with endophytic bacteria improves the productivity and grain biofortification of bread wheat Eur. J. Agron. 2018 94 98 107 10.1016/j.eja.2018.01.017 16. Upadhayay V.K. Singh A.V. Khan A. Cross Talk between Zinc-Solubilizing Bacteria and Plants: A Short Tale of Bacterial-Assisted Zinc Biofortification Front. Soil Sci. 2022 1 788170 10.3389/fsoil.2021.788170 17. Galindo F.S. Bellotte J.L. Santini J.M. Buzetti S. Rosa P.A. Jalal A. Teixeira Filho M.C.M. Zinc use efficiency of maize-wheat cropping after inoculation with Azospirillum brasilense Nutr. Cycl. Agroecosyst. 2021 120 205 221 10.1007/s10705-021-10149-2 18. Chandra A.K. Kumar A. Bharati A. Joshi R. Agrawal A. Kumar S. Microbial-assisted and genomic-assisted breeding: A two-way approach for the improvement of nutritional quality traits in agricultural crops 3 Biotech 2020 10 2 10.1007/s13205-019-1994-z 19. Noulas C. Tziouvalekas M. Karyotis T. Zinc in soils, water and food crops J. Trace Elem. Med. Biol. 2018 49 252 260 10.1016/j.jtemb.2018.02.009 29472130 20. Jalal A. Shah S. Teixeira Filho M.C.M. Khan A. Shah T. Hussain Z. Younis M. Ilyas M. Yield and phenological indices of wheat as affected by exogenous fertilization of Zinc and Iron Braz. J. Agric. Sci./Rev. Bras. Cienc. Agrar. 2020 15 e7730 10.5039/agraria.v15i1a7730 21. Marag P.S. Suman A. Growth stage and tissue specific colonization of endophytic bacteria having plant growth promoting traits in hybrid and composite maize (Zea mays L.) Microbiol. Res. 2018 214 101 113 10.1016/j.micres.2018.05.016 30031472 22. Jalal A. Oliveira C. Freitas L. Galindo F. Lima B. Boleta E. Silva E.C. Nascimento V. Nogueira T.A.R. Buzetti S. Agronomic biofortification and productivity of wheat with soil zinc and diazotrophic bacteria under tropical savannah Crop Pasture Sci. 2022 74 A N 10.1071/CP21457 23. Malavolta E. Vitti G.C. Oliveira S.A. Evaluation of the Nutritional Status of Plants: Principles and Applications 2nd ed. Potafos Piracicaba, Brazil 1997 319 (In Portuguese) 24. Rehman A. Farooq M. Naveed M. Ozturk L. Nawaz A. Pseudomonas-aided zinc application improves the productivity and biofortification of bread wheat Crop Pasture Sci. 2018 69 659 672 10.1071/CP17441 25. Mitter B. Pfaffenbichler N. Flavell R. Compant S. Antonielli L. Petric A. Berninger T. Naveed M. Sheibani-Tezerji R. von Maltzahn G. A new approach to modify plant microbiomes and traits by introducing beneficial bacteria at flowering into progeny seeds Front. Microbiol. 2017 8 11 10.3389/fmicb.2017.00011 28167932 26. Singh D. Prasanna R. Potential of microbes in the biofortification of Zn and Fe in dietary food grains. A review Agron. Sustain. Dev. 2020 40 15 10.1007/s13593-020-00619-2 27. Jalal A. Azeem K. Teixeira Filho M.C.M. Khan A. Enhancing soil properties and maize yield through organic and inorganic nitrogen and diazotrophic bacteria Sustainable Crop Production IntechOpen London, UK 2020 Volume 20 165 178 28. Pereira N.C.M. Galindo F.S. Gazola R.P.D. Dupas E. Rosa P.A.L. Mortinho E.S. Corn yield and phosphorus use efficiency response to phosphorus rates associated with plant growth promoting bacteria Front. Environ. Sci. 2020 8 40 10.3389/fenvs.2020.00040 29. Abadi V.A.J.M. Sepehri M. Khatabi B. Rezaei M. Alleviation of zinc deficiency in wheat inoculated with root endophytic fungus Piriformospora indica and rhizobacterium Pseudomonas putida Rhizosphere 2021 17 100311 10.1016/j.rhisph.2021.100311 30. Khan A. Singh J. Upadhayay V.K. Singh A.V. Shah S. Microbial biofortification: A green technology through plant growth promoting microorganisms Sustainable Green Technologies for Environmental Management Springer Singapore 2019 255 269 31. Jalal A. Shah S. Teixeira Filho M.C.M. Khan A. Shah T. Ilyas M. Rosa P.A.L. Agro-biofortification of zinc and iron in wheat grains Gesunde Pflanz. 2020 72 227 236 10.1007/s10343-020-00505-7 32. Kaur T. Rana K.L. Kour D. Sheikh I. Yadav N. Kumar V. Yadav A.N. Dhaliwal H.S. Saxena A.K. Microbe-mediated biofortification for micronutrients: Present status and future challenges New and Future Developments in Microbial Biotechnology and Bioengineering Elsevier Amsterdam, The Netherlands 2020 1 17 10.1016/B978-0-12-820528-0.00002-8 33. Singh D. Rajawat M.V.S. Kaushik R. Prasanna R. Saxena A.K. Beneficial role of endophytes in biofortification of Zn in wheat genotypes varying in nutrient use efficiency grown in soils sufficient and deficient in Zn Plant Soil 2017 416 107 116 10.1007/s11104-017-3189-x 34. Shakeel M. Rais A. Hassan M.N. Hafeez F.Y. Root associated Bacillus sp. improves growth, yield and zinc translocation for basmati rice (Oryza sativa) varieties Front. Microbiol. 2015 6 1286 10.3389/fmicb.2015.01286 26635754 35. Shivay Y.S. Prasad R. Rahal A. Relative efficiency of zinc oxide and zinc sulphate-enriched urea for spring wheat Nutr. Cycl. Agroecosyst. 2008 82 259 264 10.1007/s10705-008-9186-y 36. Galindo F.S. Buzetti S. Rodrigues W.L. Boleta E.H.M. Silva V.M. Tavanti R.F.R. Fernandes G.C. Biagini A.L.C. Rosa P.A.L. Teixeira Filho M.C.M. Inoculation of Azospirillum brasilense associated with silicon as a liming source to improve nitrogen fertilization in wheat Crops Sci. 2020 10 6160 10.1038/s41598-020-63095-4 37. Schütz L. Gattinger A. Meier M. Müller A. Boller T. Mader P. Mathimaran N. Improving Crop Yield and Nutrient Use Efficiency via Biofertilization—A Global Meta-Analysis Front. Plant Sci. 2018 8 2204 10.3389/fpls.2017.02204 29375594 38. Soil Survey Staff Keys to Soil Taxonomy 12th ed. USDA, Natural Resources Conservation Service Washington, DC, USA 2014 39. Santos H.G. Jacomina P.K.T. Anjos L.H.C. Oliveira V.A. Lumbreras J.F. Coelho M.R. Almeida J.A. Araujo Filho J.C. Oliveira J.B. Cunha T.J.F. Brazilian Soil Classification System 5th ed. Embrapa Brasília, Brazil 2018 Available online: https://www.embrapa.br/en/busca-de-publicacoes/-/publicacao/1094001/brazilian-soil-classification-system (accessed on 18 February 2022) 40. Alvares C.A. Stape J.L. Sentelhas P.C. Gonçalves J.D.M. Sparovek G. Köppen’s climate classification map for Brazil Meteorol. Z. 2013 22 711 728 10.1127/0941-2948/2013/0507 41. Raij B. van Andrade J.C. Cantarella H. Quaggio J.A. Chemical Analysis for Fertility Evaluation of Tropical Soils IAC Campinas, Brazil 2001 285 (In Portuguese) 42. Teixeira P.C. Donagemma G.K. Fontana A. Teixeira W.G. Manual of Soil Analysis Methods Centro Nacional de Pesquisa de Solos, Embrapa Rio de Janeiro, Brazil 2017 (In Portuguese) 43. Rengel Z. Graham R.D. Uptake of zinc from chelate-buffered nutrient solutions by wheat genotypes differing in zinc efficiency J. Exp. Bot. 1996 47 217 226 10.1093/jxb/47.2.217 44. Lessa J.H.D.L. Araujo A.M. Ferreira L.A. da Silva E.C. Jr. De Oliveira C. Corguinha A.P.B. Martins F.A.D. De Carvalho H.W.P. Guilherme L.R.G. Lopes G. Agronomic biofortification of rice (Oryza sativa L.) with selenium and its effect on element distributions in biofortified grains Plant Soil 2019 444 331 342 10.1007/s11104-019-04275-8 45. Statistica Research Department Brazil Per Capita Consumption of Selected Cereals in Brazil 2022 Available online: https://www.statista.com/statistics/1121036/brazil-per-capita-consumption-cereals-type/ (accessed on 24 March 2022) 46. Fageria N.K. Dos Santos A.B. Cobucci T. Zinc Nutrition of Lowland Rice Commun. Soil Sci. Plant Anal. 2011 42 1719 1727 10.1080/00103624.2011.584591 47. R Core Team R: A Language and Environment for Statistical Computing R Foundation for Statistical Computing Vienna, Austria 2015 Available online: https://www.R-project.org/ (accessed on 9 February 2022)

PMC009xxxxxx/PMC9099602.txt

==== Front Int J Environ Res Public Health Int J Environ Res Public Health ijerph International Journal of Environmental Research and Public Health 1661-7827 1660-4601 MDPI 10.3390/ijerph19095475 ijerph-19-05475 Article Emotional Self-Regulation in Primary Education: A Heart Rate-Variability Biofeedback Intervention Programme Aritzeta Aitor 1* Aranberri-Ruiz Ainara 1 Soroa Goretti 2 Mindeguia Rosa 1 https://orcid.org/0000-0001-8984-5429 Olarza Amaiur 1 Herr Raphael M. Academic Editor 1 Department of Basic Psychological Process and Development, University of the Basque Country (UPV/EHU), 20018 San Sebastian, Spain; ainara.aranberri@ehu.eus (A.A.-R.); rosa.mindeguia@ehu.eus (R.M.); amaiur.olarza@ehu.eus (A.O.) 2 Department of Clinical and Health Psychology and Research Methodology, University of the Basque Country (UPV/EHU), 20018 San Sebastian, Spain; goretti.soroa@ehu.eus * Correspondence: aitor.aritzeta@ehu.eus; Tel.: +34-943-01-8332 30 4 2022 5 2022 19 9 547504 3 2022 28 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). This study investigated the benefits of using a biofeedback intervention programme to train children in controlling their heart rate variability (HRV) through slow-paced breathing in real time. HRV biofeedback interventions focused on showing subjects to breathe such that their HRV numbers rise, improving their self-regulation. The HRV biofeedback intervention, focused on breathing, was conducted with primary education students aged between 7 and 11 years. The programme consisted of five biofeedback sessions, where students were taught to breathe six long and slow pairs of breaths per minute, to increase their HRV. After participation in the programme, students, regardless of gender, increased their HRV in a statistically significant fashion with a large effect, but this effect was not the same for all ages. HRV biofeedback interventions are rarely applied in schools and given the effectiveness of the intervention to improve HRV in children, the applied implications of our results in educational settings are discussed, especially taking into account the children’s ages. biofeedback intervention heart rate variability children polyvagal theory HeartMath emWave software Basque GovernmentBasque Government, Research Groups 2022. ==== Body pmc1. Introduction In childhood, emotions are experienced with great intensity and with low capacity for emotional regulation [1] which may have a negative impact on psychosocial skills, school performance and well-being [2,3]. During primary education, the capacity for emotional self-regulation is still under development which offers an opportunity to implement interventions with the goal of improving emotional self-regulation abilities [4]. Specifically, interventions based on biofeedback of the HRV and focused on breathing have been shown to be effective in improving emotional self-regulation capacity in children [5]. Biofeedback is a broadly used method to train people to voluntarily control certain physiological functions, such as breathing, which consists of providing users with instantaneous information on variations that occur in their own physiological activity [6]. Using heart rate variability-based biofeedback programmes, by means of breathing practise, the subjects learn to breath such that heart rate variability (HRV) increases [7,8]. In this regard, it was observed that adequate HRV biofeedback teaches people to breathe at a frequency of approximately six breaths per minute [9,10]. As far as age is concerned, it has been observed that in children the degree of HRV and self-regulation capacity are related. Different studies have shown that greater HRV reflects better psychosocial adjustment in the childhood population [11,12]. Children with behaviour problems display lower HRV [13,14], however, we do not have enough evidence about the differential effectiveness that, regarding age, a biofeedback training programme based in breath pacing, may have on children from 7 to 11 years old in Primary Education. In order to offer a comprehensive theory on children’s capacity for self-regulation through biofeedback programmes, the Polyvagal Theory should be mentioned. This theory addresses the interrelation between the vagus nerve and the emotional experience allowing us to associate self-regulation abilities with values of HRV. The vagus nerve is the tenth cranial nerve and is the main nerve in the parasympathetic division determining emotional self-regulation [15]. Primary education children completely possess the vagus nerve system [16,17] and the ventral-vagus branch and the dorsal-vagus branch of the vagus nerve, provide an inhibitory entry to the heart through the parasympathetic nervous system (reducing heart rate) and influence HRV outcomes. The vagus nerve system is hierarchically structured into three sequential functional sub-systems: (a) The ventral vagal complex or myelinated vagus, (b) The sympathoadrenal system, and (c) The dorsal vagal complex or non-myelinated vagus [18,19]. When the environment is perceived to be safe, the ventral vagal complex is activated, leading to an increased influence of the myelinated vagal channels that slow down heartbeat frequency, increase HRV, and inhibit fight-or-flight mechanisms of the sympathetic nervous system. This physiological state makes it possible for the prefrontal cortex’s structures to work properly, which are in charge of attention and self-regulation, both of which are differentially developed for a child of 7 years old and one of 11 years old. However, and despite primary education children completely possessing the vagus nerve system, we still do not have enough evidence to know how effective an HRV and breathing based biofeedback programme can be dependent on the age of the children The aforementioned HRV can be assessed with several analytical foci, although the most commonly used are frequency domain analysis (power spectral density) and time domain analysis [20,21]. As it has consistently been shown in the literature, HRV is considered a measurement of physiological emotional self-regulation [22,23,24,25,26] and a biomarker for psychopathology [27]. High HRV also has been related with lower levels of frustration and higher performance levels [28] and positive psychological adjustment in children, adolescents, and the adult population [27,29], including, for example, empathetic responses to other people in danger [30,31], social competence [32], and positive social interactions amongst equals [33]. Low HRV is the result of supressing the vagal brake. Suppression of the vagal brake entails disappearance of the parasympathetic influence wielded by the ventral vagus, and increased cardiovascular pace, which reduces HRV. The subjective emotional experience under this neurophysiological state is a variable degree of tension, malaise, and anxiety [34,35]. In terms of psychological regulation, reduced HRV has been associated with reduced capacity for self-regulation and cognitive functions that are primarily located in the prefrontal cortex [26,36,37,38]. Through breath control, healthy people can increase their own HRV by modifying respiratory frequency, through slower and steadier breathing [39,40] by means of the vagus nerve’s influence on the sinoatrial node of the heart, slowing the heart rate, and increasing HRV [36,41,42]. The high HRV generated through voluntary breathing is a suitable emotional self-regulation measure [43,44]. HRV biofeedback can be conducted by placing a device on a person that connects to a computer and provides real-time information on their HRV. By observing the impact that breathing has on HRV in real time, one learns to breathe (through trial and error and feedback), and to improve their emotional self-regulation capacity [7]. In this sense, scientific literature has shown that biofeedback interventions focused on breathing, increase HRV both in the adult population [45,46,47,48,49,50] and in the childhood population [50,51]. For example, Van der Zwan et al. [52] conducted an HRV biofeedback intervention focused on breathing for 5 weeks, analysing the programme’s effect on the well-being of pregnant and non-pregnant women. After the intervention, both groups had reduced anxiety and increased well-being. In order to better tackle stress during exams, Deschodt-Arsac et al. [53] conducted a 5-week intervention on HRV biofeedback focused on breathing, with university athletes. After the intervention, the experimental group significantly improved HRV. This improvement was maintained for 12 weeks. Poskotinova et al. [54] conducted an HRV biofeedback intervention in one single session with 20 Secondary School students (15 and 16 years old), in order to increase the participants’ HRV as a protective resource against internet addiction disorders. After the intervention, participants’ HRV was significantly increased, and correlation analyses showed a significant negative correlation between increased HRV and scores obtained in measuring internet addiction. Mei et al. [9] conducted an intervention with 32 adolescents between 13 and 18 years of age, also in one single HRV biofeedback session. After the intervention, a significant increase in HRV was observed, and more significant than another group of students who received alternative treatment in an autogenous training session. In this realm, other interventions in HRV biofeedback were also proven effective. For example, in terms of the university population, HRV biofeedback interventions focused on breathing, improved participants’ HRV [45,49], and in terms of the school-aged population, students participating in HRV biofeedback interventions also improved HRV [47,50,51,55]. Therefore, literature consistently shows biofeedback interventions have a positive impact on participants HRV, especially in adolescents and adults. Less evidence is gathered on the impact of these programs in children from 7 to 11 years old. Therefore, the main goal of this research was to examine the effectiveness of a biofeedback breath pacing training programme on HRV in children at primary school, taking into account the age of the participants. Based on previous results, we expect that; compared to a control group, the values of the HRV of the experimental group will be higher than the control group values after the intervention. Regarding the above-mentioned neuro-anatomical development processes, we will explore if the lower neuro-anatomical maturity progress at lower ages might also affect the effectiveness of the intervention programme, so as to observe better HRV results as the age of the participants increases. Finally, in terms of gender, we did not find similar research conducted on school-aged populations to examine differences based on gender, which is possibly explained by limited sample sizes. In other age ranges, Hill et al. [56] examined 172 studies mainly focused on the adult population. The results showed that, in comparison with men, women displayed greater amplitude in HRV. Brunetto et al. [57] conducted a study with 41 adolescents (20 boys and 21 girls), aged 12–17 years, where no statistically significant gender differences were found in HRV scores. Finally, in the study conducted by Aziz et al. [58], upon comparing HRV parameters between 33 girls and 37 boys, they observed that only boys had lower scores than girls in only three out of 11 HRV parameters (SDNN, LF, and SD2). Therefore, we also aim to investigate if there were gender differences in HRV values before and after the intervention programme. 2. Materials and Methods 2.1. Sample The experimental and control groups were divided into the three cycles: 29.2% belonged to the first cycle (2nd grade, M = 7.58 years and SD = 0.38), 57.7% to the second cycle (3rd and 4th grades, M = 8.93 years old and SD = 0.85), and the remaining 13.1% were third-cycle students (5th and 6th grades, M = 11.04 years old and SD = 0.91). Participants were studying in primary education in two different public schools using the Amara Berri Education System which seeks to educate students using everyday problems and organizing their everyday learning process in thematic spaces (Murumendi and Larrea public schools). The experimental and control groups were divided in the three cycles: 29.2% belonged to the first cycle (2nd grade, 7 years old), 57.7% to the second cycle (3rd and 4th grades, 8–9 years old), and the remaining 13.1% were third-cycle students (5th and 6th grades, 10–11 years old). Student participation was voluntary and consented to by the school board, parents, and guardians. Classrooms were randomly assigned to experimental and control groups, however, all children, except those with a clinical diagnosis (i.e., hyperactivity, depression or anxiety disorders) belonging to those classrooms participated in the study. So, children were not randomly assigned to the treatment/non-treatment conditions. The study had a favourable report from the ethics committee for research with humans, their samples, and their data (M10-2020-318) from the University of the Basque Country/Euskal Herriko Unibertsitatea (UPV/EHU). Ethical aspects required for research with humans were scrupulously followed (informed consent, right to information, personal data protection, confidentiality guarantees, non-discrimination, no cost, and possibility to leave the study during any of its phases). 2.2. Design Given the inherent characteristics of the natural groups (primary education classrooms) and the natural context (intervention takes place at the school), a quasi-experimental intervention was designed with a control group. Five HRV scores were taken from all experimental participants. The first measurement of session one (S1) is the pre-treatment measurement and informs about the baseline HRV numbers. The fifth measurement of session five (S5) is the post-treatment measurement. For the control group only two measures were taken in session one (S1) and session five (S5) and at the same period of time as the experimental group. 2.3. Instruments and Materials As follows, a description of the instruments and materials used are described. Recording sheet (ad hoc): A document was specifically prepared to record the data obtained from each student in each session throughout the intervention. It contains the HRV scores from the five sessions. HRV: HeartMath EmWave software was selected for this study [59], whose efficacy has been proven in different studies [5,50,55]. This is a non-invasive instrument that measures HRV in real time through a sensor placed on the earlobe. Using a mathematical algorithm, the EmWave software transforms the collected data (e.g., RMSSD, SDNN...) into low, medium, and high coherence ratios. These low, medium and high coherence ratios are directly related to HRV amplitude rates. Thus, a low coherence level matches low HRV, a medium coherence level matches medium HRV, and a high coherence level matches high HRV. After each session, the application shows the HRV scores by distributing 100 points in three levels (1) A low coherence level, or low HRV, shown in red; (2) An intermediate coherence level, or intermediate HRV, shown in blue; and (3) A high coherence level, or high HRV, shown in green. These scores are set forth on the record sheet and used for later statistical analyses. Two EmWave software applications were used [59]: Coherence Coach and Balloon Game. By applying Coherence Coach, children learn to breathe following the rhythm of a moving ball on the computer screen, learning to breathe six breaths per minute in the following fashion: when the ball goes upward, participants must breathe in through the nose, and when the ball goes down, they must also breathe out through the nose. The application named Balloon Game is an interactive game where students practise the type of breathing learned in the Coherence Coach application, but without following the rhythm of the ball. Specifically, in the Balloon Game application, a hot-air balloon appears, flying over different scenarios with variable degrees of speed, depending on the participants’ pace of breathing. If they breathe under parameters similar to those learned in the Coherence Coach application, the journey will be faster, while if they do not breathe under these parameters, the journey will be slower. 2.4. Procedure The programme was implemented in two phases. First, since the intervention was designed for the tutor to train students individually, the pre-intervention phase was designed, where teachers were trained to learn to use the EmWave software’s Coherence Coach and Balloon Game applications [59] so that, later on, they could apply this intervention to students. This training’s goal was two-fold. On one hand, the objective was to teach how to use the programme and the instruments necessary to use it in the classroom in order to take individual HRV measurements throughout the intervention. On the other hand, to help teachers to understand the relationship between steady, calm breathing, and HRV. Teacher training lasted 8 h. Secondly, for the experimental group, the application phase for the intervention programme consisted of five sessions. One session per week was held individually with each student. Each session lasted approximately 15 min and was conducted in the same physical space (a classroom prepared to this end). The idea is for each girl and boy to conduct their session the same days and times of the week. The different activities proposed were directed by the usual teachers who had been previously trained and were supervised by a member of the research team. Moreover, the HRV scores obtained on each record for each session were collected on a record sheet made to this end. A detailed step by step description of the biofeedback training programme can be obtained by the authors. Table 1, summarizes the intervention protocol both in intervention and control groups. 3. Statistical Analysis In order to observe if HRV scores were significantly different upon completion of training, we conducted the Wilcoxon signed-rank test comparing averages before and after the training sessions. Later, a Student t-test analysis was performed to examine if averages between pre-test and post-test low HRV and high HRV by cycles were different. We also analysed the gender differences in HRV by using a variance analysis of HRV results in S1 and S5 based on gender and finally, to examine if the effectiveness of the programme was stable from session to session and taking into account the age of the participants, we used an ANOVA analysis using a MIXED method. 4. Results Once the intervention was completed, the five HRV scores collected on the record sheets were processed and analysed. Below we present the results of the Wilcoxon signed-rank test comparing averages before and after the training sessions. We deemed it essential to observe whether HRV scores were significantly different upon completion of training. In addition, scores from the baseline (S1) were compared with those from session five (S5). It should be mentioned that the criteria used by Kelley and Preacher [60] were followed to interpret the effect size. In terms of low HRV, for the experimental group the average levels were significantly higher at the baseline (S1; M = 62.70) than after the intervention (S5; M = 31.80), z = −10.26, p < 0.05, r = −0.53. For medium HRV, the average levels were lower on the baseline session (S1; M = 14.50) than after the intervention (M = 23.08) z = −5.98, p < 0.05, r =−0.34. Finally, for high HRV, the average levels were lower on the baseline session (S1; M = 23.04) than after the intervention (M = 79.52), z = −12.37, p < 0.05, r = −0.71. Comparing these same values for the control group, no statistical differences could be observed in any of the three levels of HRV. Low HRV values (S1; M = 60.21); (S5; M = 62.70), z = 0.87, p > 0.05, r = 0.13. For medium HRV, (S1; M = 16.76) and S5 (M = 17.09) z = 0.73, p > 0.05, r = 0.12. Finally, for high HRV, (S1; M = 24.17) and S5 (M = 24.98) z = 0.44, p > 0.05, r = 0.04. As there were no statistical differences for the control group, we examined differences just for the experimental groups and taking into account the main goal of this investigation, that is to examine if there were differences between age ranges. As follows, with the analysis by age ranges (first, second and third cycle), in Table 2, it can be observed that there is a differential effect of age in relation to high HRV. With changes being statistically significant in all cycles, a lower effect size is observed in the first cycle (7 years old), being a moderate effect size. In the second cycle, a large effect size is observed and in the third cycle, the biggest effect size is observed. In relation to results by ages of low HRV in Table 3, we also observe the programme’s differential impact. In this case, with the change effect being statistically significant in all cycles, the effect’s lesser size occurs again in the first cycle, this being a moderate effect size. In the second cycle, we observe a large effect size. And in the third cycle, the greatest effect sizes of all cycles occur. In other words, as the children are older, the greater the impact of the programme (measured through the increased high HRV and reduced low HRV). In order to give an answer to the third exploratory hypothesis, we analysed the gender differences in HRV. Table 4 shows the results. None of the HRV scores compared between girls and boys in sessions one and five showed statistically significant results. Thus, there are no differences between both genders in HRV values. At this point of the study, we also asked if the effectiveness of the programme is stable from session to session across the five sessions and also if there are any differences in each session regarding the age of the participants. In order to answer to these questions, the evolution of high HRV values from session to session was examined by an ANOVA analysis using a MIXED method. The fixed effect estimation showed that the increase of the high HRV observed in each session was statistically significant. See Table 5. In Figure 1 the graphical evolution of the high HRV of all samples can be seen. Examining the results, we could observe that the pattern of the first cycle was a little different from the total samples. The fixed effect estimation indicates that there is an increment of the high HRV from S1 to S2, however there a decrease from S2 to S3. This reduction only happens in children of the first cycle, but not in the second and third cycle. See Table 6 and Figure 2. In the following the graphical evolution of the high HRV of students of cycle 1 can be seen. 5. Discussion As set forth in the study’s hypotheses, it was expected that the breathing-based biofeedback programme would increase HRV. In other words, that the HRV biofeedback intervention programme would be effective if, upon its completion, participants were able to increase their high HRV. It was also expected that such values will be higher for the experimental group compared to the control group. We also intended to explore if there were differences between boys and girls and, most importantly, if the age of the participants had an effect on the results. Upon analysing the comparison of pre-test (S1) and post-test (S5) high HRV averages from the experimental group, we observe that intervention programme improved high levels of HRV. We also observed that comparing the values of the HRV for the control group, the pre-test and post-test results did show statistically significant differences, showing for this group the same values in both measures. Therefore, we consider that the intervention programme increased the HRV of participants, which can be considered as an indicator of their improving capacity for self-regulation. Emotional self-regulation is a fundamental tool to develop students′ effort, motivation, and personal responsibility about learning and to guarantee scholastic adaptation [14,61,62]. Throughout the primary education period, children experience complex feelings without having internalized the ability to communicate and self-regulate their emotions effectively yet. Therefore, this kind of programme can be essential for children′s well-being by promoting the ability to self-regulate their own emotions, attention and behaviour. The improvement of the treatment group in HRV observed in our research, matches results observed in other interventions with other age groups. For example, Kuppusamy, et al. [63] carried out an intervention with 520 adolescents (13–18 years) to analyse the impact of a 6-month programme focused on voluntary breathing control on domain parameters of frequency and HRV. After the intervention, it was observed that the experimental group obtained a statistically significant improvement in comparison with the control group in the indicator for HRV’s parasympathetic influence. Mei et al. [9] conducted an intervention with 60 subjects, where they analysed the differences in HRV of participants in two groups treated differently: one group with an HRV biofeedback session, and a group with an autogenous training session. The results revealed significant effects in both interventions on HRV values. We also found research conducted with the same HeartMath technology used in our research [59]. Field et al. [64] conducted an intervention with 13 participants aged between 26 and 62 years. After the intervention, the experimental group showed greater HRV amplitude than the control group in a statistically significant fashion and with a great effect size (d = 1.97). Aritzeta et al. [45] conducted an HRV biofeedback intervention programme with 152 university students (average age = 19.6), and a control group with 81 university students (average age = 19.4), in order to reduce anxiety levels before exams, and thus improve academic performance. The results indicate a significant improvement in HRV scores, with an effect size of η2 = 0.77. Regarding the gender variable, we observed significant differences between boys and girls in HRV values. The effect size was low in all cases (d =< 0.3). In this regard, it should be mentioned that we did not find similar research conducted on the schoolchildren population to examine differences based on gender, which is possibly explained by limited sample sizes. Thus, having observed our results and those shown in the scientific literature [57,58], we cannot conclude that there are evident differences between boys and girls in HRV. Just like Hill et al. [56], we believe that it is necessary to research further into the impact of the gender variable on HRV to observe more conclusive results. One of the goals of this article was also to observe the effect of the programme depending the age of the participants. We should remember that, given the curricular configuration of primary education, the intervention programme in this research was applied to three different age tranches: cycle one (7 years, n = 76), cycle two (8–9 years, n = 240), and cycle three (10–11 years, n = 47). The results showed that the differences in the pre-test and post-test averages in high HRV were statistically significant in all cycles. However, it is important to note that, in terms of the effect size, there were differences between the cycles: in cycle one (d = −0.28), being low to moderate, while in cycle two (d = −0.83) and in cycle three (d = −0.79), the effect size was large. In other words, in cycle one, unlike cycle two and cycle three, we observe a moderate size effect both in high HRV and in low HRV. As such, we may suppose that these differences between cycles are associated with shared progression factors inherent to students in cycle one. Differences may be justified by maturity factors related to the age of 7 years. Therefore, specifically heeding to the neuro-anatomical cerebral zones for emotional self-regulation (for example, the anterior ventral cingulate cortex and the prefrontal ventromedial cortex) [65], and being aware of their developmental process between childhood and adolescence [66,67] we can infer that this lower neuro-anatomical maturity progress at 7 years makes it difficult to acquire the emotional self-regulation skill measured in this case through increased high HRV. As such, it should be mentioned that the results observed in our study are also corroborated by those observed in other pieces of research with similar methodologies [64,68,69]. 6. Conclusions Based on the results observed in this study, and through the statistical analyses conducted after applying our intervention programme, we can affirm that students who participated in our programme learned to increase their HRV by practising long, steady breathing, at a pace of six pairs of breaths per minute. These results fall in line with other similar studies [23,67], and they are in accordance with the Polyvagal Theory [19,70,71], which affirms that biofeedback procedures based on breathing strategies have an influence on the regulation and improvement of HRV. Moreover, HRV biofeedback has proven to be an effective strategy for children to learn to self-regulate and bolster voluntary breathing learning. A growing body of psychological research supports the idea that HRV could be an objective physiological measurement to assess emotional responses [22,23,72], including those related to emotional regulation [15,67,73]. High HRV levels correlate to positive results in psychological adjustment in children [74,75,76], adolescents, and adults [27]. It has also been directly linked to self-regulatory capacities in children [77,78,79]. The improvement of emotional self-regulation capacities is very important in Primary Education because at this period there is an increase in academic and social demands that can negatively affect emotional stress [80]. High HRV results are positively associated with emotional self-regulation abilities, which, in turn, positively influences psychological well-being [81] and academic performance [2]. Given that, in childhood, emotional reactivity is highly intense [82] and emotional self-regulation capacity is under development [83,84,85]. Given that childhood is a key period for learning to emotionally self-regulate [86,87,88], it is suitable to use resources to make emotional reactivity regulation possible. School is an important socialising agent that, along with families, plays a vital role in the promotion of positive mental health in children [89,90]. Fostering healthy relationships during primary education stage is essential to children´s positive experience of school, to promote their well-being and their cognitive and emotional development. For all this, the need to implement programmes based on emotional regulation abilities in academic contexts is justified, and it should also be mentioned that these programmes are especially relevant for children from 8 years old. Difficulties, Limitations, and Future Lines of Research Given the design of the natural groups, it was not possible to maintain equivalent sample proportionality, nor to maintain homogeneous sample sizes throughout cycles. The samples are not representative, and the generalisation of results must be taken with caution. Regarding collection of HRV data, two kinds of data contamination may have occurred: “the examiner effect”, which refers to the influence of the examiner and of the examiner’s interaction in examining data collection; and “situation effects”, which refer to the influence of different factors on the results of the execution [91]. To attempt to reduce both effects, teachers were selected as “examiners”, and school classrooms were selected as a suitable location for the intervention. HRV was exclusively analysed in this intervention. It would be advisable to conduct an analysis on the impact of this type of programme on other variables, such as on academic performance. It would also be suitable to analyse the impact the intervention programme had on social relationships. For example, one might establish the presumption that students who participate in the intervention will display more social skills than before participating in the programme, and that participants will also display better social skills than those who do not participate in the programme. After measuring the efficacy of our intervention, it would be suitable to study the longitudinal effect of the programme and it would be suitable to conduct further HRV biofeedback interventions with similar procedures, samples, and methods, so as to provide greater robustness to the already-proven effectiveness of these interventions. In closing, we should mention that, although HRV biofeedback interventions are recently and rarely applied in schools, given the high effectiveness of the intervention conducted in this study; and the relevance and need for emotional self-regulation in healthy psychosocial development, we believe that this new field has great potential to boost an emotional self-regulation resource, which is prolonged and steady breathing, in the classroom and in universal fashion, free of cost, for all Primary Education students. Acknowledgments We are grateful to Murumendi and Larrea public schools for voluntarily participating in this study. Author Contributions Conceptualization, A.A.; methodology, G.S. and R.M.; formal analysis, R.M.; data curation, G.S.; writing—original draft preparation, A.A.; writing—review and editing, A.A.-R.; visualization, A.O.; supervision, A.O. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki, and approved by the Committee for Research Involving Humans of the University of the Basque Country whit number CEISH/269 1-2-3-4-/2014. Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the patient(s) to publish this paper. Data Availability Statement Data is available on request from corresponding author. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Evolution of the high HRV in all samples. Figure 2 Evolution of the high HRV in cycle 1. ijerph-19-05475-t001_Table 1 Table 1 Intervention processes for intervention and control groups. Groups S1 S2 S3 S4 S5 Intervention HRV- Base line Breath Training + HRV-Balloon Game Connexion Breath Training + HRV-Balloon Game Connexion HRV-Balloon Game Connexion HRV-Balloon Game Connexion Control HRV- Base line 1 No training No training No training HRV-Base line 2 ijerph-19-05475-t002_Table 2 Table 2 Comparison of S1* and S5* averages by high HRV and cycles. Pre-Test Post-Test Student’s t Cohen’s d N M SD M SD p d Total 300 23.26 29.56 79.52 60.47 0.000 −1.250 Cycle 1 87 26.61 31.93 49.77 51.10 0.000 −0.558 Cycle 2 175 23.78 29.75 94.99 60.56 0.000 −1.577 Cycle 3 38 13.16 19.88 76.39 54.84 0.000 −1.692 * S1 and S5 numbers refer to sessions 1 and 5. ijerph-19-05475-t003_Table 3 Table 3 Comparison of S1–S5 averages by low HRV and cycles. Pre-Test Post-Test Student’s t Cohen’s d N M SD M SD p d Total 299 62.63 30.78 32 30.07 0.000 −1.007 Cycle 1 87 62.16 33.09 31.45 37.23 0.000 −0.506 Cycle 2 174 61.39 31.01 25.32 27.26 0.000 −1.238 Cycle 3 38 69.37 23.18 30.95 29.12 0.000 −1.469 ijerph-19-05475-t004_Table 4 Table 4 Variance analysis in HRV results in S1* and S5* based on gender. Average SD F HRV N Girls N Boys Girls Boys F p Cohen’s d Low HRV S1 138 61.80 163 63.21 31.35 30.18 0.16 0.69 −0.046 Low HRV S5 138 33.82 163 30.45 31.26 29.12 0.93 0.34 0.112 Medium HRV S1 138 14.04 163 14.93 9.10 10.13 0.63 0.43 −0.093 Medium HRV S5 138 23.32 163 22.96 23.03 19.74 0.02 0.88 0.017 High HRV S1 138 24.14 163 22.35 30.09 29.10 0.28 0.60 0.061 High HRV S5 138 80.22 163 78.45 63.11 58.01 0.06 0.80 0.029 * S1 and S5 numbers refer to sessions 1 and 5. ijerph-19-05475-t005_Table 5 Table 5 Evolution of high HRV values from session to session in all samples. Predictor Sum of Squares df Mean Square F p Session 1 high HVR 4616.81 2 2308.41 2.70 0.06 Session 2 high HVR 24578.54 2 12289.27 11.09 0.001 Session 3 high HVR 102411.57 2 51205.78 24.52 0.001 Session 4 high HVR 79778.77 2 39889.38 15.61 0.001 Session 5 high HVR 119275.36 2 59637.69 18.18 0.001 ijerph-19-05475-t006_Table 6 Table 6 Evolution of high HRV values from session to session in cycle 1. Predictor Estimation SE df t p IC95% [LL, UL] Session 1 high HVR 26.00 4.04 291.64 6.43 0.001 [18.04, 33.95] Session 2 high HVR 42.66 4.04 291.64 10.56 0.001 [34.71, 50.61] Session 3 high HVR 36.43 4.04 291.64 9.01 0.001 [28.48, 44.38] Session 4 high HVR 45.80 4.15 291.64 11.02 0.001 [18.04, 33.95] Session 5 high HVR 49.77 4.04 291.64 12.31 0.001 [18.04, 33.95] Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Silvers J.S. Insel C. Powers A. Franz P. Helion C. Martin R. Weber J. Mischel W. Casey B.J. Ochsner J.N. The transition from childhood to adolescence is marked by a general decrease in amygdala reactivity and an affect-specific ventral-to-dorsal shift in medial prefrontal recruitment Dev. Cogn. Neurosci. 2017 25 128 137 10.1016/j.dcn.2016.06.005 27445112 2. Datu J.A.D. King R.B. Subjective well-being is reciprocally associated with academic engagement: A two-wave longitudinal study J. Sch. Psychol. 2018 69 100 110 10.1016/j.jsp.2018.05.007 30558746 3. Goldin P.R. Moodie C.A. Gross J.J. Acceptance versus reappraisal: Behavioral, autonomic, and neural effects Cogn. Affect. Behav. Neurosci. 2019 19 927 944 10.3758/s13415-019-00690-7 30656602 4. Thomas J.C. APrON Study Team Letourneau N. Campbell T.S. Tomfohr-Madsen L. Giesbrecht G.F. Developmental origins of infant emotion regulation: Mediation by temperamental negativity and moderation by maternal sensitivity Dev. Psychol. 2017 53 611 628 10.1037/dev0000279 28333524 5. Cruz A. Biofeedback as an Intervention to Increase Self-Regulation in School-Aged Children in an Urban Charter School Doctoral Dissertation Widener University Chester, PA, USA 2019 Available online: https://search.proquest.com/docview/2284756111/fulltextPDF/20682CE768684BC1PQ/1?accountid=17248 (accessed on 12 December 2021) 6. Schwartz M.S. Andrasik F. Definitions of Biofeedback and Applied Psychophysiology Biofeedback: A Practitioner’s Guide Guilford Press New York, NY, USA 2003 7. Kiselev A.R. Karavaev A.S. Gridnev V. Prokhorov M.D. Ponomarenko V. Borovkova E.I. Shvartz V. Ishbulatov Y. Posnenkova O.M. Bezruchko B.P. Method of estimation of synchronization strength between low-frequency oscillations in heart rate variability and photoplethysmographic waveform variability Russ. Open Med J. 2016 5 e0101 10.15275/rusomj.2016.0101 8. Lehrer P. Eddie D. Dynamic Processes in Regulation and Some Implications for Biofeedback and Biobehavioral Interventions Appl. Psychophysiol. Biofeedback 2013 38 143 155 10.1007/s10484-013-9217-6 23572244 9. Lin I.-M. Wang S.-Y. Fan S.-Y. Peper E. Chen S.-P. Huang C.-Y. A Single Session of Heart Rate Variability Biofeedback Produced Greater Increases in Heart Rate Variability Than Autogenic Training Appl. Psychophysiol. Biofeedback 2020 45 343 350 10.1007/s10484-020-09483-y 32767160 10. Karavaev A.S. Kiselev A. Gridnev V. Borovkova E.I. Prokhorov M.D. Posnenkova O.M. Ponomarenko V. Bezruchko B.P. Shvartz V. Phase and frequency locking of 0.1-Hz oscillations in heart rate and baroreflex control of blood pressure by breathing of linearly varying frequency as determined in healthy subjects Hum. Physiol. 2013 39 416 425 10.1134/S0362119713010040 11. Eisenberg N. Spinrad T.L. Eggum N.D. Emotion-Related Self-Regulation and Its Relation to Children’s Maladjustment Annu. Rev. Clin. Psychol. 2010 6 495 525 10.1146/annurev.clinpsy.121208.131208 20192797 12. Gentzler A.L. Santucci A.K. Kovacs M. Fox N.A. Respiratory sinus arrhythmia reactivity predicts emotion regulation and depressive symptoms in at-risk and control children Biol. Psychol. 2009 82 156 163 10.1016/j.biopsycho.2009.07.002 19596044 13. Aimie-Salleh N. Malarvili M.B. Whittaker A.C. Fusion of heart rate variability and salivary cortisol for stress response identification based on adverse childhood experience Med. Biol. Eng. Comput. 2019 57 1229 1245 10.1007/s11517-019-01958-3 30734153 14. Calkins S.D. The emergence of self- regulation: Biological and behavioral control mechanisms supporting toddler compe-tencies Socioemotional Development in the Toddler Years: Transitions and Transformations Brownell C.A. Kopp C.B. Guilford Press New York, NY, USA 2007 261 284 15. Porges S.W. Orienting in a defensive world: Mammalian modifications of our evolutionary heritage. A Polyvagal Theory Psychophysiology 1995 32 301 318 10.1111/j.1469-8986.1995.tb01213.x 7652107 16. Sachis P.N. Armstrong D.L. Becker L.E. Bryan A.C. Myelination of the Human Vagus Nerve from 24 Weeks Postconceptional Age to Adolescence J. Neuropathol. Exp. Neurol. 1982 41 466 472 10.1097/00005072-198207000-00009 7086467 17. Wenzel A. Monk C. Hane A. Fetal and infant neurobehavioral development: Basic processes and environmental influences Oxford library of Psychology. The Oxford Handbook of Perinatal Psychology Wenzel A. Oxford University Press Oxford, UK 2016 53 86 18. Porges S.W. Emotion: An evolutionary by-product of the neural regulation of the autonomic nervous system Ann. New York Acad. Sci. 1997 807 62 77 10.1111/j.1749-6632.1997.tb51913.x 9071344 19. Porges S.W. The polyvagal theory: Phylogenetic substrates of a social nervous system Int. J. Psychophysiol. 2001 42 123 146 10.1016/S0167-8760(01)00162-3 11587772 20. Hirsch J.A. Bishop B. Respiratory sinus arrhythmia in humans: How breathing pattern modulates heart rate Am. J. Physiol. Circ. Physiol. 1981 241 H620 H629 10.1152/ajpheart.1981.241.4.H620 21. McCraty R. Atkinson M. Tomasino D. Bradley R.T. The coherent heart: Heart brain interactions, psychophysiological coherence, and the emergence of system-wide order Integral Rev. 2009 5 10 115 22. Appelhans B.M. Luecken L.J. Heart rate variability as an index of regulated emotional responding Rev. Gen. Psychol. 2006 10 229 240 10.1037/1089-2680.10.3.229 23. Balzarotti S. Biassoni F. Colombo B. Ciceri M.R. Cardiac vagal control as a marker of emotion regulation in healthy adults: A review Biol. Psychol. 2017 130 54 66 10.1016/j.biopsycho.2017.10.008 29079304 24. Graziano P. Derefinko K. Cardiac vagal control and children’s adaptive functioning: A meta-analysis Biol. Psychol. 2013 94 22 37 10.1016/j.biopsycho.2013.04.011 23648264 25. Park G. Thayer J.F. From the heart to the mind: Cardiac vagal tone modulates top-down and bottom-up visual perception and attention to emotional stimuli Front. Psychol. 2014 5 278 10.3389/fpsyg.2014.00278 24817853 26. Thayer J.F. Lane R.D. Claude Bernard and the heart—brain connection: Further elaboration of a model of neurovisceral integration Neurosci. Biobehav. Rev. 2009 33 81 88 10.1016/j.neubiorev.2008.08.004 18771686 27. Beauchaine T.P. Thayer J.F. Heart rate variability as a transdiagnostic biomarker of psychopathology Int. J. Psychophysiol. 2015 98 338 350 10.1016/j.ijpsycho.2015.08.004 26272488 28. Li W.-C. Chiu F.-C. Kuo Y.S. Wu K.-J. The Investigation of Visual Attention and Workload by Experts and Novices in the Cockpit Engineering Psychology and Cognitive Ergonomics. Applications and Services Harris D. Springer Berlin/Heidelberg, Germany 2013 167 176 10.1007/978-3-642-39354-9_19 29. Park G. Vasey M. Van Bavel J.J. Thayer J.F. When tonic cardiac vagal tone predicts changes in phasic vagal tone: The role of fear and perceptual load Psychophysiology 2014 51 419 426 10.1111/psyp.12186 24571084 30. Fabes R.A. Eisenberg N. Eisenbud L. Behavioral and physiological correlates of children’s reactions to others in distress Dev. Psychol. 1993 29 655 663 10.1037/0012-1649.29.4.655 31. Liew J. Eisenberg N. Spinrad T.L. Eggum N.D. Haugen R.G. Kupfer A. Reiser M.R. Smith C.L. Lemery-Chalfant K. Baham M.E. Physiological Regulation and Fearfulness as Predictors of Young Children’s Empathy-related Reactions Soc. Dev. 2011 20 111 134 10.1111/j.1467-9507.2010.00575.x 22573929 32. Eisenberg N. Spinrad T.L. Valiente C. Emotion-related self-regulation, and children’s social, psychological, and academic functioning Child Psychology: A Handbook of Contemporary Issues Balter L. Tamis-LeMonda C. Taylor and Francis Inc London, UK 2016 219 244 33. Diamond A. Lee K. Interventions Shown to Aid Executive Function Development in Children 4 to 12 Years Old Science 2011 333 959 964 10.1126/science.1204529 21852486 34. Chalmers J.A. Quintana D.S. Abbott M.J.-A. Kemp A.H. Anxiety Disorders are Associated with Reduced Heart Rate Variability: A Meta-Analysis Front. Psychiatry 2014 5 80 10.3389/fpsyt.2014.00080 25071612 35. McCraty R. New Frontiers in Heart Rate Variability and Social Coherence Research: Techniques, Technologies, and Implications for Improving Group Dynamics and Outcomes Front. Public Health. 2017 5 267 10.3389/fpubh.2017.00267 29075623 36. Hildebrandt L.K. McCall C. Engen H.G. Singer T. Cognitive flexibility, heart rate variability, and resilience predict fine-grained regulation of arousal during prolonged threat Psychophysiology 2016 53 880 890 10.1111/psyp.12632 26899260 37. Thayer J.F. Hansen A.L. Saus-Rose E. Johnsen B.H. Heart Rate Variability, Prefrontal Neural Function, and Cognitive Performance: The Neurovisceral Integration Perspective on Self-regulation, Adaptation, and Health Ann. Behav. Med. 2009 37 141 153 10.1007/s12160-009-9101-z 19424767 38. Williams P.G. Cribbet M.R. Tinajero R. Rau H.K. Thayer J.F. Suchy Y. The association between individual differences in executive functioning and resting high-frequency heart rate variability Biol. Psychol. 2019 148 107772 10.1016/j.biopsycho.2019.107772 31577925 39. Bhagat V. Izad Y. Jayaraj J. Husain R. Mat K.C.M. Aung M.M.T. Emotional maturity among medical students and its impact on their academic performance Trans. Sci. Technol. 2017 4 48 54 40. Lehrer P.M. Gevirtz R. Heart rate variability biofeedback: How and why does it work? Front. Psychol. 2014 5 756 10.3389/fpsyg.2014.00756 25101026 41. Porges S.W. The Polyvagal Theory: Neurophysiological Foundations of Emotions, Attachment, Communication, and Self-Regulation W.W. Norton & Co New York, NY, USA 2011 42. Walker L.S. Stone A.L. Smith C.A. Bruehl S. Garber J. Puzanovova M. Diedrich A. Interacting influences of gender and chronic pain status on parasympathetically mediated heart rate variability in adolescents and young adults Pain 2017 158 1509 1516 10.1097/j.pain.0000000000000942 28715351 43. Blum J. Rockstroh C. Göritz A.S. Heart Rate Variability Biofeedback Based on Slow-Paced Breathing with Immersive Virtual Reality Nature Scenery Front. Psychol. 2019 10 2172 10.3389/fpsyg.2019.02172 31616353 44. Cheng K.S. Croarkin P.E. Lee P.F. Heart Rate Variability of Various Video-Aided Mindful Deep Breathing Durations and Its Impact on Depression, Anxiety, and Stress Symptom Severity Mindfulness 2019 10 2082 2094 10.1007/s12671-019-01178-8 45. Aritzeta A. Soroa G. Balluerka N. Muela A. Gorostiaga A. Aliri J. Reducing Anxiety and Improving Academic Performance Through a Biofeedback Relaxation Training Program Appl. Psychophysiol. Biofeedback 2017 42 193 202 10.1007/s10484-017-9367-z 28623467 46. Goessl V.C. Curtiss J.E. Hofmann S.G. The effect of heart rate variability biofeedback training on stress and anxiety: A meta-analysis Psychol. Med. 2017 47 2578 2586 10.1017/S0033291717001003 28478782 47. Henriques G. Keffer S. Abrahamson C. Horst S.J. Exploring the Effectiveness of a Computer-Based Heart Rate Variability Biofeedback Program in Reducing Anxiety in College Students Appl. Psychophysiol. Biofeedback 2011 36 101 112 10.1007/s10484-011-9151-4 21533678 48. Lantyer A.D.S. Viana M.D.B. Padovani R.D.C. Biofeedback in the treatment of stress and anxiety-related disorders: A critical review Psico-USF 2013 18 131 140 10.1590/S1413-82712013000100014 49. Prato C.A. Yucha C.B. Biofeedback-Assisted Relaxation Training to Decrease Test Anxiety in Nursing Students Nurs. Educ. Perspect. 2013 34 76 81 10.5480/1536-5026-34.2.76 23763019 50. Rush K.S. Golden M.E. Mortenson B.P. Albohn D. Horger M. The Effects of a Mindfulness and Biofeedback Program on the On- and Off-Task Behaviors of Students with Emotional Behavioral Disorders Contemp. Sch. Psychol. 2017 21 347 357 10.1007/s40688-017-0140-3 51. Jones A.M. West K.B. Suveg C. Anxiety in the School Setting: A Framework for Evidence-Based Practice Sch. Ment. Health 2017 11 4 14 10.1007/s12310-017-9235-2 52. van der Zwan J.E. Huizink A.C. Lehrer P.M. Koot H.M. de Vente W. The Effect of Heart Rate Variability Biofeedback Training on Mental Health of Pregnant and Non-Pregnant Women: A Randomized Controlled Trial Int. J. Environ. Res. Public Health 2019 16 1051 10.3390/ijerph16061051 30909539 53. Deschodt-Arsac V. Lalanne R. Spiluttini B. Bertin C. Arsac L.M. Effects of heart rate variability biofeedback training in athletes exposed to stress of university examinations PLoS ONE 2018 13 e0201388 10.1371/journal.pone.0201388 30048519 54. Poskotinova L. Krivonogova O. Zaborsky O. Effectiveness of Short-Term Heart Rate Variability Biofeedback Training and the Risk of Internet Addiction in Adolescents 15-16 Years of Age Int. J. Biomed. 2020 10 153 156 10.21103/Article10(2)_OA13 55. Bothe D.A. Grignon J.B. Olness K.N. The Effects of a Stress Management Intervention in Elementary School Children J. Dev. Behav. Pediatr. 2014 35 62 67 10.1097/DBP.0000000000000016 24336090 56. Hill L.K. Hu D.D. Koenig J. Sollers J.J. Kapuku G. Wang X. Snieder H. Thayer J.F. Ethnic Differences in Resting Heart Rate Variability: A systematic review and meta-analysis Psychosom. Med. 2015 77 16 25 10.1097/PSY.0000000000000133 25551201 57. Brunetto A.F. Roseguini B.T. Silva B.M. Hirai D.M. Guedes D.P. Effects of Gender and Aerobic Fitness on Cardiac Autonomic Responses to Head-Up Tilt in Healthy Adolescents Pediatr. Cardiol. 2005 26 418 424 10.1007/s00246-004-0808-0 16374693 58. Aziz W. Schlindwein F.S. Wailoo M. Biala T. Rocha F.C. Heart rate variability analysis of normal and growth restricted children Clin. Auton. Res. 2011 22 91 97 10.1007/s10286-011-0149-z 22045365 59. Institute of HeartMath EmWave Desktop© HeartMath Boulder Creek, CA, USA 2012 60. Kelley K. Preacher K.J. On effect size Psychol. Methods 2012 17 137 152 10.1037/a0028086 22545595 61. Gross J.J. Emotion regulation: Taking stock and moving forward Emotion 2013 13 359 365 10.1037/a0032135 23527510 62. Morrison F.J. Ponitz C.C. McClelland M.M. Self-regulation and academic achievement in the transition to school Human Brain Development. Child Development at the Intersection of Emotion and Cognition Calkins S.D. Bell M.A. American Psychological Association Washington, DC, USA 2010 203 224 10.1037/12059-011 63. Kuppusamy M. Kamaldeen D. Pitani R. Amaldas J. Ramasamy P. Shanmugam P. Vijayakumar V. Effects of yoga breathing practice on heart rate variability in healthy adolescents: A randomized controlled trial Integr. Med. Res. 2020 9 28 32 10.1016/j.imr.2020.01.006 32025489 64. Field L.H. Edwards S.D. Edwards D.J. Dean S.E. Influence of HeartMath Training Programme on Physiological and Psychological Variables Glob. J. Health Sci. 2018 10 126 133 10.5539/gjhs.v10n2p126 65. Etkin A. Büchel C. Gross J.J. The neural bases of emotion regulation Nat. Rev. Neurosci. 2015 16 693 700 10.1038/nrn4044 26481098 66. Michalska K.J. Kinzler K.D. Decety J. Age-related sex differences in explicit measures of empathy do not predict brain responses across childhood and adolescence Dev. Cogn. Neurosci. 2013 3 22 32 10.1016/j.dcn.2012.08.001 23245217 67. Michalska K.J. Davis E.L. The psychobiology of emotional development: The case for examining sociocultural processes Dev. Psychobiol. 2019 61 416 429 10.1002/dev.21795 30592032 68. Berry M.E. Chapple I.T. Ginsberg J.P. Gleichauf K.J. Meyer J.A. Nagpal M.L. Non-pharmacological Intervention for Chronic Pain in Veterans: A Pilot Study of Heart Rate Variability Biofeedback Glob. Adv. HealthMed. 2014 3 28 33 10.7453/gahmj.2013.075 69. Thurber M.R. Bodenhamer-Davis E. Johnson M. Chesky K. Chandler C.K. Effects of Heart Rate Variability Coherence Biofeedback Training and Emotional Management Techniques to Decrease Music Performance Anxiety Biofeedback 2010 38 28 40 10.5298/1081-5937-38.1.28 70. Porges S.W. The polyvagal perspective Biol. Psychol. 2007 74 116 143 10.1016/j.biopsycho.2006.06.009 17049418 71. Porges S.W. Norton series on Interpersonal Neurobiology The Pocket Guide to the Polyvagal Theory: The Transformative Power of Feeling Safe W.W. Norton & Co New York, NY, USA 2017 72. Porges S.W. Vagal Mediation of Respiratory Sinus Arrhythmia: Implications for drug delivery Temporal Control of Drug Delivery Hrushesky J.M. Langer R. Theeuwes F. New York Academy of Sciences New York, NY, USA 1991 57 66 10.1111/j.1749-6632.1991.tb27237.x 73. Porges S.W. Vagal tone: An autonomic mediator of affect Cambridge Studies in Social and Emotional Development. The Development of Emotion Regulation and Dysregulation Garber J. Dodge K.A. Cambridge University Press (CUP) Cambridge, UK 1991 111 128 10.1017/cbo9780511663963.007 74. Fox N.A. Calkins S.D. The Development of Self-Control of Emotion: Intrinsic and Extrinsic Influences Motiv. Emot. 2003 27 7 26 10.1023/A:1023622324898 75. Hastings P.D. Nuselovici J.N. Utendale W.T. Coutya J. McShane K.E. Sullivan C. Applying the polyvagal theory to children’s emotion regulation: Social context, socialization, and adjustment Biol. Psychol. 2008 79 299 306 10.1016/j.biopsycho.2008.07.005 18722499 76. Smith M.S. Womack W.M. Stress Management Techniques in Childhood and Adolescence: Relaxation training, meditation, hypnosis, and biofeedback. Appropriate clinical applications Clin. Pediatr. 1987 26 581 585 10.1177/000992288702601105 77. Graziano P.A. Reavis R.D. Keane S.P. Calkins S.D. The role of emotion regulation in children’s early academic success J. Sch. Psychol. 2007 45 3 19 10.1016/j.jsp.2006.09.002 21179384 78. Kahle S. Utendale W.T. Widaman K.F. Hastings P.D. Parasympathetic Regulation and Inhibitory Control Predict the Development of Externalizing Problems in Early Childhood J. Abnorm. Child Psychol. 2018 46 237 249 10.1007/s10802-017-0305-6 28493111 79. Marcovitch S. Leigh J. Calkins S.D. Leerks E.M. O’Brien M. Blankson A.N. Moderate vagal withdrawal in 3.5-year-old children is associated with optimal performance on executive function tasks Dev. Psychobiol. 2010 52 603 608 10.1002/dev.20462 20806334 80. Pirskanen H. Jokinen K. Karhinen-Soppi A. Notko M. Lämsä T. Otani M. Meil G. Romero-Balsas P. Rogero-García J. Children’s Emotions in Educational Settings: Teacher Perceptions from Australia, China, Finland, Japan and Spain Early Child. Educ. J. 2019 47 417 426 10.1007/s10643-019-00944-6 81. Kim D.H. Bassett S.M. Takahashi L. Voisin D.R. What does self-esteem have to do with behavioral health among low-income youth in Chicago? J. Youth Stud. 2018 21 999 1010 10.1080/13676261.2018.1441982 82. Jiang Y. Tian Y. Wang Z. Causal Interactions in Human Amygdala Cortical Networks across the Lifespan Sci. Rep. 2019 9 1 11 10.1038/s41598-019-42361-0 30626917 83. Dougherty L.R. Blankenship S.L. Spechler P.A. Padmala S. Pessoa L. An fMRI Pilot Study of Cognitive Reappraisal in Children: Divergent Effects on Brain and Behavior J. Psychopathol. Behav. Assess. 2015 37 634 644 10.1007/s10862-015-9492-z 26692636 84. McRae K. Gross J.J. Weber J. Robertson E.R. Sokol-Hessner P. Ray R.D. Gabrieli J.D. Ochsner K.N. The development of emotion regulation: An fMRI study of cognitive reappraisal in children, adolescents and young adults Soc. Cogn. Affect. Neurosci. 2012 7 11 22 10.1093/scan/nsr093 22228751 85. Silvers J.A. McRae K. Gabrieli J.D.E. Gross J.J. Remy K.A. Ochsner K.N. Age-related differences in emotional reactivity, regulation, and rejection sensitivity in adolescence Emotion 2012 12 1235 1247 10.1037/a0028297 22642356 86. Babkirk S. Rios V. Dennis T.A. The late positive potential predicts emotion regulation strategy use in school-aged children concurrently and two years later Dev. Sci. 2014 18 832 841 10.1111/desc.12258 25438825 87. Jones S.M. Doolittle E.J. Social and Emotional Learning: Introducing the Issue Futur. Child. 2017 27 3 11 10.1353/foc.2017.0000 88. Wang C. Hu Y. Weng J. Chen F. Liu H. Modular segregation of task-dependent brain networks contributes to the development of executive function in children NeuroImage 2020 206 116334 10.1016/j.neuroimage.2019.116334 31704295 89. Darling N. Steinberg L. Parenting style as context: An integrative model Psychol. Bull. 1993 113 487 496 10.1037/0033-2909.113.3.487 90. Keppens G.A. Spruyt B. The School as a Socialization Context: Understanding the influence of school bonding and an author-itative school climate on class skipping Youth Soc. 2017 51 1145 1166 10.1177/0044118X17722305 91. Fernández-Ballesteros R. Introducción a la Evaluación Psicológica [Introduction to Psychological Evaluation] Pirámide Madrid, Spain 2005

PMC009xxxxxx/PMC9099603.txt

==== Front Animals (Basel) Animals (Basel) animals Animals : an Open Access Journal from MDPI 2076-2615 MDPI 10.3390/ani12091159 animals-12-01159 Article A Refined Method for Studying Foraging Behaviour and Body Mass in Group-Housed European Starlings https://orcid.org/0000-0002-0861-0191 Bateson Melissa * https://orcid.org/0000-0003-4492-0629 Nolan Ryan Prescott Mark J. Academic Editor Graham Melanie L. Academic Editor Biosciences Institute, Newcastle University, Newcastle upon Tyne NE2 4HH, UK; ryan.nolan@liverpool.ac.uk * Correspondence: melissa.bateson@newcastle.ac.uk 29 4 2022 5 2022 12 9 115919 3 2022 26 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Simple Summary Small birds such as European starlings respond rapidly to environmental challenges by losing or gaining weight. Laboratory studies of these birds are therefore useful for understanding how the environment affects body weight. However, practical constraints including the need to catch birds frequently for weighing has meant that birds are often housed alone in small cages for such studies. Such conditions are unnatural and are likely to cause stress. Consequently, the data obtained from these studies are unrepresentative of wild birds. Here, we describe a novel technology based on smart feeders that permits continuous recording of foraging behaviour and body masses from starlings housed in groups in large indoor aviaries that permit more natural behaviour. We show that the birds quickly learn to use the feeders and that the system delivers detailed real-time data on foraging behaviour and body mass, without the need for frequent catching. The data obtained allowed us to study how the foraging decisions that a bird makes within a single day affect its body weight that day. These improvements in the quality of the data that we are able to collect will help inform our understanding of the environmental causes of weight gain and obesity. Abstract Laboratory experiments on passerine birds have been important for testing hypotheses regarding the effects of environmental variables on the adaptive regulation of body mass. However, previous work in this area has suffered from poor ecological validity and animal welfare due to the requirement to house birds individually in small cages to facilitate behavioural measurement and frequent catching for weighing. Here, we describe the social foraging system, a novel technology that permits continuous collection of individual-level data on operant foraging behaviour and body mass from group-housed European starlings (Sturnus vulgaris). We report on the rapid acquisition of operant key pecking, followed by foraging and body mass data from two groups of six birds maintained on a fixed-ratio operant schedule under closed economy for 11 consecutive days. Birds gained 6.0 ± 1.2 g (mean ± sd) between dawn and dusk each day and lost an equal amount overnight. Individual daily mass gain trajectories were non-linear, with the rate of gain decelerating between dawn and dusk. Within-bird variation in daily foraging effort (key pecks) positively predicted within-bird variation in dusk mass. However, between-bird variation in mean foraging effort was uncorrelated with between-bird variation in mean mass, potentially indicative of individual differences in daily energy requirements. We conclude that the social foraging system delivers refined data collection and offers potential for improving our understanding of mass regulation in starlings and other species. passerine bird foraging fat body mass mass gain energy requirement adaptive regulation UK Biotechnology and Biological Sciences Research CouncilBB/J016446/1 European Union666669 This research was funded by a UK Biotechnology and Biological Sciences Research Council project grant to Melissa Bateson and Daniel Nettle (BB/J016446/1) and a European Research Council Advanced Grant to Daniel Nettle under the European Union’s Horizon 2020 research and innovation programme (AdG 666669, COMSTAR). ==== Body pmc1. Introduction Wild bird species such as the European starling (Sturnus vulgaris) provide valuable animal models for studying the biology of foraging behaviour and the regulation of fat reserves and body mass [1,2]. A well-developed theoretical literature in evolutionary ecology argues that birds should adaptively regulate their fat reserves to optimize the trade-off between the increased starvation risk caused by having too little fat and the increased predation risk caused by having too much [3,4,5,6,7]. These optimality models predict how birds should strategically regulate their body mass and the shape of their mass gain trajectory over the day in response to environmental challenges including perceived predation risk and unpredictable access to food. Laboratory experiments on birds have been central in testing these models of adaptive mass regulation. For example, numerous experimental studies on starlings have confirmed the theoretical prediction that limited and unpredictable access to food should cause birds to gain mass as insurance against starvation [2,8,9,10,11]. This ‘insurance hypothesis’ has recently been suggested as a possible evolutionary explanation for the increased odds of overweight observed in food insecure humans, suggesting that studies of birds could provide new insights into the environmental causes of the human obesity epidemic [12,13,14,15]. Despite the above achievements, progress in understanding the behavioural and physiological mechanisms underlying body mass regulation in birds has been limited by the practical difficulties of obtaining high quality individual-level data on body mass, foraging behaviour, and physiology from birds housed in ecologically valid conditions in the laboratory. Moreover, the common practice of individually housing birds in small cages that facilitate behavioural measurements and frequent capture for manual weighing is likely to be chronically stressful for social species such as the starling. We therefore set out to demonstrate that high quality individual-level data on foraging behaviour and body mass, which are necessary for testing mechanistic hypotheses, can be obtained from group-housed starlings living in large aviaries that simultaneously provide greater ecological validity and improved animal welfare compared with previous methods. Laboratory studies in birds and mammals have identified a number of methodological variables that are likely to affect the ecological validity of studies of foraging behaviour and mass regulation. The first of these is cage size, which is potentially important because it affects the distances that animals can walk, run, or fly and hence their potential energy expenditure through physical activity. While it is generally assumed that increased body mass under conditions of unpredictable food results from increased energy intake [16,17,18], reduced energy expenditure is also likely to be important; studies of both starlings and zebra finches (Taeniopygia guttata) showed that birds respond to unpredictable food by reducing their physical activity [2,19]. Cage furnishing also affects the range of behaviour patterns that animals can perform in the laboratory and hence their energy expenditure. In rats, the offspring of mothers that are undernourished during pregnancy become obese when housed in standard laboratory cages without the opportunity for exercise, but when provided with daily access to a running wheel, they engaged in voluntary exercise and maintained normal body mass [20,21]. Thus, housing animals in small barren cages might constrain them to be sedentary and remove a potentially important mechanism that is present in the natural environment for regulating energy balance. Housing social animals in groups is likely to be important for body mass for two reasons. First, group housing permits social behaviour, which can form an important contribution to an individual’s time budget and contribute to energy expenditure. In songbirds such as the starling, song is primarily used in intra-specific communication and is energetically costly [22]. Second, group housing can affect body mass by increasing competition for food. Dominant individuals in a group can displace subordinate animals from a monopolizable food source, creating inequalities within the group in the predictability of food access. In starlings, as predicted by the insurance hypothesis, the effects of unpredictable food access on body mass are greater when there is more competition between birds for access to feeders [2]. Furthermore, experimentally increasing a subordinate bird’s rank in a group (by removing the most dominant individual) causes a decrease in its body mass [23]. Thus, housing animals individually removes opportunities for energetically expensive social behaviour, and also removes competition for food, reducing unpredictability in food and thereby eliminating an important source of individual variability in body mass that is present in the natural environment. All experimental studies of foraging behaviour can be classified as either open or closed-economy designs [24]. In an open economy experiment, a subject is placed in the experimental manipulation for part of the day and its total daily food intake is determined by the amount of food provided by the experimenter outside of this period. For example, in many psychological studies, subjects are maintained at a fixed percentage of their free-feeding mass, regardless of the amount of food eaten in the experimental session. In contrast, in a closed economy experiment, the manipulation is typically in place for the entire day, and the subject’s response to the manipulation completely determines its daily intake. Theoretical models show that the optimal proportion of time a subject should spend foraging depends on the rate at which it can obtain food, but that the direction of this relationship depends on whether the subject is in open or closed economy [25]. Thus, the type of economy potentially influences the effects that will be observed in experiments. Closed economy experiments provide a better simulation of the situation present in natural environments. In summary, to perform ecologically valid studies of body mass regulation and understand the behavioural, psychological, and physiological mechanisms underlying this, we need experimental laboratory studies that incorporate the following features. First, subjects must be housed in enclosures that are large enough and contain the furnishings necessary to allow performance of a natural range of behaviour. Second, subjects of social species should be housed in groups that permit social interactions and the formation of dominance hierarchies. Third, studies must be performed in closed economy, whereby the subjects’ adjustments to the experimental conditions they are placed under directly determine their daily food intake. Finally, human intervention, especially catching and restraint, which are stressful and likely to be perceived by subjects as predation attempts, should be limited as far as possible. Here, we describe the social foraging system (SFS), a novel technology that allows remotely operated experimental manipulation of food access and permits continuous collection of individual-level data on operant foraging behaviour. We present observational data on the patterns of foraging behaviour and mass gain that can be recorded using the SFS as well as the methods that we have developed to train the birds. The SFS is based around smart feeding stations that identify individual birds with radio frequency identification (RFID), weigh them with an integrated electronic balance, record foraging behaviour using pecks on an illuminable pecking key, and control food access using a retractable food hopper. The feeding stations are connected to a computer that controls the operant schedule in place on the pecking key and access to the food hopper and collects data on bird identities, body masses, and foraging behaviour continuously between dawn and dusk each day (starlings do not feed at night). A potential challenge raised by the SFS is the training of the birds. In the past, we have conducted operant training in individual cages where individual experience can be controlled and progression based on individual performance criteria (e.g., [26]). Starlings are social foragers and we speculated that group training might increase the willingness of more neophobic individuals to approach the SFS and feed from the hopper when they saw bolder birds foraging there. However, we were also concerned that during operant training, some birds might learn to scrounge food rewards earned by other individuals, rather than learning to peck the keys for food themselves. Over a series of pilot studies not described here, we developed a protocol for training starlings to forage from the SFS in groups of three birds. Six of the birds used in the current study were involved in these earlier trials and were trained to forage from the SFS prior to the start of the current study. However, the other six birds were experimentally naïve and we used this study as an opportunity to test a new social training protocol for the SFS. We start by describing the acquisition of operant key pecking in naïve group-housed birds. Our aim was to discover whether starlings can learn key pecking without the need for social isolation. Following successful operant training, we report the observational foraging and body mass data from two groups of six birds maintained on a fixed-ratio operant schedule under closed economy for 11 consecutive days. Our aims were to describe the patterns of foraging behaviour and mass gain within the day within birds; to establish the repeatability of these measurements between days; and, to explore how variation in foraging behaviour is related to variation in body mass both within and between birds. We discuss the strengths and limitations of the SFS, focusing on both the quality of the scientific data obtained and the likely welfare of the birds. 2. Materials and Methods 2.1. Ethical Statement The study adhered to ASAB/ABS guidelines for the use of animals in research. Birds were taken from the wild under Natural England permit 20121066 and the research was completed under UK Home Office licence PPL P038AB1D3 with the approval of the Animal Welfare and Ethical Review Body at Newcastle University. Reporting followed the ARRIVE 2.0 guidelines [27]. 2.2. Subjects and Basic Husbandry Subjects were 12 adult European starlings (Sturnus vulgaris), six males and six females, originally caught from the wild in March 2019 at Seaton Sluice, Northumberland, UK. Birds were sexed using bill and iris colour [28]. Tarsus length was measured twice for the right and left leg using digital callipers and a mean tarsus length calculated for each bird as a measure of skeletal size [29]. Prior to the study described in the current paper, which took place in March 2020, the birds were group-housed in a single indoor ‘home’ aviary (280 cm wide × 300 cm deep × 255 cm high). The aviary was without windows and was artificially lit and ventilated. The temperature and humidity were maintained at ~18–20 °C and ~40% respectively. The lights were on between 0800–1715, providing 9.25 h daylight. Between 0800 and 0815, the lights gradually increased in intensity to simulate dawn and between 1700–1715, the lights dimmed gradually to simulate dusk and allow birds to settle for the night. The aviary was furnished with rope perches, wood chippings on the floor, a water bath, and a drinker providing ad libitum clean water supplemented with vitamins. Birds were fed ad libitum on commercial poultry starter crumb (Special Diets Services Poultry Starter; henceforth ‘crumb’)—a homogeneous complete diet for starlings—supplemented with softbill mix (Orlux Universal Softbill Food), fresh fruit, and live mealworms. For the study described in the current paper, the birds were caught and transferred to experimental aviaries described below. After completion of the data collection, the birds were retained in the experimental aviaries for an experimental study not reported here. In April 2020, following completion of the latter experiment and inspection by a veterinary surgeon, the birds were released to the wild at the site of original capture. 2.3. Experimental Aviaries and the Social Foraging System Experimental aviaries were identical to the home aviary described above with the exception that each aviary was equipped with one SFS station for every three birds present in the aviary. The SFS was built to our specifications by Campden Instruments, Loughborough, UK. A single SFS station (Figure 1) comprised a motorized retractable food hopper filled with crumb and an illuminable pecking key, both of which could only be accessed via a wooden perch designed to accommodate a single bird. The perch was mounted at the apex of a smooth plastic pyramid designed to prevent other birds from perching on the station and protect the balance and RFID aerial that were located beneath it. The pyramid was mounted on an electronic balance that measured to a resolution of 0.01 g. The RFID aerial was tuned to read microchips that were glued to plastic leg rings worn by a starling standing on the perch. The SFS stations were connected to a single computer in an adjacent room running Whisker experimental control software [30], a custom-written programme (Starfeeder) that managed the RFID and mass data, and additional custom-written Whisker ‘client’ programmes specific to different phases of the study. The computer controlled the operant schedule in place on each station and collected continuous data on the identity and masses of birds visiting each station, plus any key pecks and food rewards delivered. The data files written by the Whisker client programmes could be accessed in real-time by custom-written R scripts [31] that produced summaries of the birds’ masses and key pecking behaviour for the current day. These latter summaries were checked a minimum of three times daily and were used for welfare monitoring purposes. Husbandry in the experimental aviaries took place between 1600 and 1700 daily throughout the study. Birds were fed on crumb that was available either from the SFS hopper or from ad libitum bowls that were provided following operant training sessions. The diet was supplemented with four live mealworms per bird given during daily husbandry and supplied in spatially separated bowls to prevent one bird from monopolizing the worms. 2.4. Welfare Monitoring A welfare monitoring protocol was designed that takes advantage of the mass data provided by the SFS and hence avoids unnecessary catching of the birds (Supplementary Materials Document: Welfare monitoring in European starlings). Catching for manual weighing was only required by the protocol in the period of habituation, before birds started perching on the SFS and recording automatic body masses. 2.5. Study Design The study was observational with no experimental manipulation. Data on foraging behaviour and body mass are described from two groups of six birds housed in separate experimental aviaries and maintained under closed economy for 11 consecutive days. The design is summarized in Figure 2. 2.6. Operant Training Three days prior to being transferred to the experimental aviaries for the start of operant training, the diet of the birds in the home aviary was restricted to ad libitum crumb in order to habituate them to the diet that would be available from the SFS. At ~1200 on the third day, 12 birds were caught from the home aviary, manually weighed, and fitted with two coloured plastic leg rings, each of which had a unique microchip attached. Birds wore two microchips to guard against identification failure in the event that one microchip fell off, broke, or was not read due to poor alignment with the aerial in the SFS. Six of the birds had been trained to forage from the SFS ~6 months previously and were released directly into a single experimental aviary equipped with two SFS stations running the continuous foraging operant schedule (see below). The other six birds were experimentally-naïve and required operant training. Training was conducted in groups of three to facilitate monitoring of individual performance during the acquisition of key pecking. Three birds were released into two experimental aviaries (three males in one and three females in the other), each equipped with a single SFS station. The first phase of operant training was to habituate the birds to the SFS stations and to feeding from the SFS hopper. Throughout this phase, crumb was provided ad libitum by permanently raising the hoppers of the SFS stations between 0800–1645. Birds were initially encouraged to approach the SFS stations by placing two bowls of crumb and two bowls of mealworms (highly attractive to starlings) on the base of each station. Feeding was monitored by watching the birds via a live CCTV camera and weighing the bowls and SFS hopper at the end of each day. When the birds started feeding from the bowls, they were removed, and dried mealworms were manually placed in the SFS food hopper entrance to attract the birds to start feeding from the hopper. When the birds started feeding from the hopper, the dried mealworms were discontinued, and the birds were henceforth restricted to feeding on the crumb in the hopper. Throughout habituation, the number of stable masses recorded per bird each day was used to monitor the foraging effort and body mass of each bird and the daily decrease in the mass of the food hopper to monitor crumb consumption by each trio of birds. Until individual birds started perching on the balance, they were caught and manually weighed every second day. As soon as all the birds in an aviary were perching on the SFS and feeding from the hopper, the aviary progressed to the next phase of training. The second phase of operant training was to teach the birds to peck the lit key on the SFS to raise the hopper and hence obtain access to food. During this phase, a daily operant training session ran from 0830 until ~1400 and was followed by the provision of bowls of ad libitum crumb placed on the base of the SFS between the end of the session and 1645 when birds were food deprived until the start of the session the following morning. The SFS hoppers remained in the inaccessible retracted position outside of the training sessions. Operant training began with daily sessions of auto-shaping, whereby illumination of the pecking key on the SFS for 15 s predicted unconditional raising of the food hopper for 15 s, starting when the key light extinguished, followed by a 200-s inter-trial interval (ITI) during which the hopper was retracted and the key light was unlit. The aim of this schedule was to set up a Pavlovian association between the lit key and food reward. The acquisition of this association typically results in a conditioned response, whereby the birds spontaneously start to direct appetitive pecks at the lit key (auto-shaping). A peck to the lit key was reinforced by immediate hopper raising, thus additionally establishing an instrumental association between pecking the lit key, and immediate food reward. Daily sessions terminated after 90 trials or 6 h. This unconditional auto-shaping training continued daily until all the birds in an aviary had started pecking the lit key—each bird was required to make a minimum of three pecks in a session before the entire aviary could progress to the next phase of training. When this criterion was met, hopper raising was made conditional on a bird pecking the illuminated key. The stimulus time was maintained at 15 s, but the feeding time was reduced to 5 s and the ITI to 66 s. The shorter feeding time was designed to reduce the potential for birds to scrounge food produced by others, which might impair them from learning the association between pecking and food. The session terminated after 270 trials or 6 h. This conditional training continued daily until all the birds in an aviary were pecking the lit key a minimum of three times in a session. When this criterion was met, the aviary progressed to continuous foraging (see below) and as soon as both aviaries of naïve birds had reached this point, the six trained birds were united in a single experimental aviary equipped with two SFS stations running the continuous foraging schedule. 2.7. Continuous Foraging under Closed Economy Following operant training, the birds were maintained in a closed economy, obtaining all of their daily food from the SFS (with the exception of the four mealworms given during husbandry). The default position of the food hopper was lowered so that food was unavailable and the key light on the SFS was illuminated to indicate that the SFS station was available for foraging. Food access was delivered on a ratio schedule, whereby a single peck at the lit key caused the key light to extinguish and the hopper to raise for 5 s (henceforth referred to as ‘a reinforcement’). At the end of the food access period, the hopper lowered and there was a 2-s ITI before the key re-lit, signalling the start of the next available trial. The two SFS stations in each aviary operated independently from one another, meaning that it was possible for two starlings in an aviary to forage simultaneously on different stations. The operant sessions began at dawn each day (0800) and ended at lights-off (1715). Thus, the maximum number of reinforcements available per day from a single SFS station was 4162 (calculated assuming that a bird always pecked the lit key one second after it illuminated). Continuous foraging continued for 11 consecutive days in each of the two experimental aviaries. 2.8. Outcome Variables 2.8.1. Operant Foraging Behaviour In the current study, a single peck to a lit key always resulted in the food hopper raising for 5 s, meaning that the number of pecks equated to the number of reinforcements earned. The time of each peck was recorded, and the peck was attributed to the bird currently on the perch. Therefore, number of reinforcements earned was available at the individual bird level. Reinforcements earned was expressed as the rate × h−1. 2.8.2. Food Consumption Total crumb consumption in each aviary was estimated daily by calculating the difference in the mass of the SFS food hoppers between the beginning and end of the day and subtracting any crumb collected in a spill tray located beneath each hopper. Therefore, consumption data were only available at the aviary level, but were expressed as g × bird−1·day−1 for ease of comparison between groups of different sizes. 2.8.3. Body Mass In all experiments, body masses were recorded each day between lights-on and lights-off; each mass was assigned to the microchip recorded from the bird currently on the perch. Therefore, body mass was available at the individual bird level. The balances measured masses at a frequency of 6 Hz. A stable mass was recorded for a bird if the balance measured five consecutive masses of >50 g that were within a range of 5 g. These criteria were chosen to eliminate masses from birds that were perching incorrectly (e.g., by placing one foot on the food hopper), but to maximise the number of stable masses recorded from moving birds. Once a stable mass had been recorded, another stable mass could not be recorded until the balance had registered a mass <10 g, indicating that the current bird had left the perch. Balances were checked with a 100-g test mass a minimum of twice daily and calibrated if necessary. In order to control for build-up of guano on the perch over the day, balances were automatically zeroed regularly throughout the day when no bird was present on the perch. Prior to modelling the mass data, any masses greater than 120 g were removed on the grounds that such masses were at least 10 g above the maximum mass ever recorded for a starling in our laboratory and were thus likely to be the result of measurement error. The raw masses showed a clear trend of mass increase over the day; the trajectory was typically non-linear, with mass gain generally being fastest early in the day, slowing down in the middle of the day and either peaking or speeding up again towards dusk. There is substantial random error in mass, due to the imprecision of the balances and movement of the birds whilst on the perch. Furthermore, masses were not always available at all times of every day (example data from one bird are shown in Figure 3). To estimate comparable masses for each bird, we therefore modelled how individual body mass changed as a function of time of day. As long as a minimum of 10 masses were available for a bird on a given day, the masses were fitted with a regression line (on the choice of the model see below). To remove biologically implausible outliers, any masses >10 g from the fitted line were removed and a new cubic polynomial fitted to the remaining data (Figure 3). This latter fit was used to estimate body mass at specific times of the day such as dawn or dusk. To avoid extrapolation beyond the measured data, a dawn or dusk mass was only estimated if there was a mass recording within 1 h of the estimate. Dawn was chosen as the time the lights came on (0800) and dusk as 1600, because birds often stopped foraging considerably before lights-off, reducing the number of masses available in the final hour of the day. 2.9. Inferential Statistics Data were analysed using R version 3.5.1 [31]. Due to the multi-level structure of the data, with individual observations nested within birds and birds nested within aviaries, we used general linear mixed models (GLMMs) for inferential statistics. GLMMs were fitted using restricted maximum likelihood estimation (REML) in the package ‘lme4′ [32] and p-values were calculated using Satterthwaite’s method in the package ‘lmerTest’ [33]. We included a random intercept for bird in all GLMMs. A random intercept for aviary was also initially added to the models, but since in practice the aviary explained little or no variance, this random effect was dropped from the final version of the models reported. A fixed effect of sex was included in all models, because on average, male starlings are skeletally larger and heavier than females, meaning that there is reason to expect the effects of sex on between-subject differences in foraging behaviour and body mass. As expected, the males had longer tarsi than the females, but the difference was marginally non-significant (linear model: βmale ± se = 0.99 ± 0.47; F1,10 = 4.41, p = 0.062). To assess the reliability of the measurements of reinforcements earned and body mass derived from the SFS across the 11 days of the study, we computed intra-class correlation coefficients (ICCs) and their 95% confidence intervals using the R package ‘psych’ [34]. ICCs were based on a two-way random-effects model assuming single measurements and absolute-agreement. 3. Results 3.1. Operant Training The six previously-trained birds all started key pecking on release into their experimental aviary and began their first full day of continuous foraging the following day. The six naïve birds took a total of eight days to learn to peck the illuminated key for food: habituation to the SFS took four days (training days 1–4) and training to key peck a further four days (training days 5–8). In the first session of unconditional auto-shaping (training day 5), four out of six birds pecked on at least one trial and in the second session (training day 6), all six birds pecked on at least three trials. In the third session (training day 7), both aviaries advanced to conditional training and all six birds pecked on at least 41 trials (Figure 4). The next day (training day 8), both aviaries advanced to continuous foraging and, since both groups were pecking well, at the end of the day, the two groups were united in a single experimental aviary and began their first full day of continuous foraging the following day. 3.2. Continuous Foraging The data presented were from 11 consecutive days of continuous foraging, collected while the 12 birds were housed under a closed economy in two groups of six. 3.2.1. Reinforcements Earned Birds earned an average of 261 ± 138 reinforcements bird−1 × day−1 (mean ± sd; see Table 1 for individual descriptive statistics). The ICC for reinforcements bird−1·day−1 was 0.85 (95% CI: 0.73–0.94), indicating moderate to excellent reliability for measurement of this outcome variable. The number of reinforcements earned per day by each bird decreased slightly over the 11 days of continuous foraging (GLMM: β ± se = −3.37 ± 1.56; F1,119 = 4.71, p = 0.032); there was no significant effect of sex on number of reinforcements earned (βmale ± se = 120.18 ± 74.15; F1,10 = 2.63, p = 0.136). Within each day, the rate at which birds earned reinforcements declined significantly with increasing hour of the day (GLMM: β ± se = −1.50 ± 0.14; F1,1195 = 119.28, p < 0.001; Figure 5). 3.2.2. Food Consumption The dataset comprised 10 and 11 daily measurements of total food consumed from aviaries 114 and 116, respectively; day 11 from aviary 114 was missing due to experimenter error. The birds consumed 24.9 ± 1.0 g bird−1 × day−1 (mean ± se) of crumb. A GLMM with a random effect of aviary showed a non-significant increase in daily consumption over the 11 days of continuous foraging (GLMM: β ± se = 0.34 ± 0.17; F1,18.08 = 4.04, p = 0.060). To explore whether variation in operant foraging behaviour predicted variation in food consumption, we asked whether the daily total number of reinforcements earned by all the birds in each aviary predicted the daily total crumb consumption in that aviary. A GLMM with a random effect of aviary showed that food consumption was significantly positively predicted by the number of reinforcements earned (GLMM: β ± se = 0.006 ± 0.002; F1,18.03 = 5.77, p = 0.027). 3.2.3. Body Mass The initial dataset comprised 8857 stable mass measurements, but 16 of these (0.18%) were excluded due to being greater than 120 g (see Methods), yielding 8841 masses for modelling. To establish the highest degree of polynomial necessary to model these data, we compared the fit of linear, quadratic, and cubic polynomials to the mass data from each bird day (data from the ninth day for bird P75 were not fitted due to there being too few measurements—see Methods). The best fitting model on each day was defined as the model with the lowest Akaike’s information criterion (AIC). Overall, the data from 27% of days were fitted best with a linear model, 30% with a quadratic model, and 43% with a cubic model (Figure 6). With the exception of one bird (P79), the majority of days for individual birds fitted best with either a quadratic or cubic polynomial. Therefore, there is strong support for daily mass gain being nonlinear, and specifically for the use of a cubic model capable of capturing two points of inflection. Based on these findings, we elected to use cubic polynomials to estimate how mass changed with time of day. A further 162 masses (1.83%) were excluded as outliers during the fitting process (see Methods), meaning that the results described below were based on a total of 8679 stable masses, yielding an average of 66 ± 3 masses (mean ± sd) bird−1·× day−1. The 131 cubic polynomial fits had an R-squared value of 0.54 ± 0.20 (mean ± sd). Therefore, the cubic polynomials on average explained over half of the variation in the mass measurements. We used the fitted cubic polynomials to estimate body mass at each hour of each day between 0800 and 1600. Selected descriptive statistics for body mass derived from this approach are shown in Table 1. The ICCs for reliability of body mass were as follows: dawn mass 0.90 (95% CI: 0.79–0.97), noon mass 0.91 (95% CI: 0.82–0.97), and dusk mass 0.89 (95% CI: 0.79–0.97). Thus, estimates of body mass had good to excellent reliability, with noon mass being the most reliable. Average noon mass was 84.1 ± 5.7 g (mean ± sd). There was no effect of day of study on noon mass (GLMM: β ± se = 0.067 ± 0.053; F1,118 = 1.59, p = 0.2100). Although, as expected for starlings, the males were heavier than females (86.9 ± 6.8 versus 81.4 ± 3.0 respectively), but the difference was not significant (βmale ± se = 5.44 ± 3.02; F1,10 = 3.24, p = 0.100). Mean dawn mass was highly positively correlated with mean dusk mass (Pearson correlation: r10 = 0.98, p < 0.0001). All birds gained mass between dawn and dusk each day and the gain was 6.0 ± 1.2 g (mean ± sd). Nightly mass loss was 6.1 ± 1.2 g (mean ± sd). The ICCs for the reliability of daily mass gain and nightly loss were 0.16 (95% CI: 0.05–0.44) and 0.16 (95% CI: 0–0.45), respectively. Thus, estimates of gain and loss had poor reliability. However, mean nightly mass loss was highly positively correlated with mean daily mass gain (Pearson correlation: r10 = 0.99, p < 0.0001). Figure 7 shows how predicted mass changed with time of day for each of the 12 birds. 3.2.4. Foraging Effort and Body Mass To explore the association between foraging effort and body mass, we asked whether dusk body mass was predicted by the number of reinforcements earned during the day (equal to the number of key pecks). Dusk mass was significantly positively predicted by the number of reinforcements earned since dawn (GLMM: β ± se = 0.0086 ± 0.0030; F1,125 = 8.11, p = 0.005; Figure 8a); there was no significant effect of sex (βmale ± se = 3.83 ± 3.34; F1,10 = 1.32, p = 0.278). Although we might predict that birds that are heavier overall are heavier because they eat more, this does not follow from the above result, because the mixed model combines within-subject effects due to plastic phenotypic responses, with between-subject effects due to individual differences. We therefore used the method described by van de Pol and Wright [35] to separate within- from between-subject effects of foraging effort on body mass. We used within-subject centring (i.e., subtracting the mean daily reinforcements earned by each subject from each daily observation) to derive a predictor variable that expresses the within-subject component of variation in reinforcements earned. We also derived a predictor variable that expresses the between-subject component of variation in reinforcements earned by replacing all daily observations for a given subject with the mean daily number of reinforcements earned by that subject. A second GLMM with these two new predictor variables (in place of reinforcements earned) showed that while the within-subject effect of reinforcements earned on dusk mass was significant (β ± se = 0.0098 ± 0.0031; F1,117 = 9.96, p = 0.002), the between-subject effect was not (β ± se = −0.0155 ± 0.0125; F1,9 = 1.54, p = 0.247). Furthermore, the parameter estimates for these effects were in opposite directions, with a positive association between reinforcements earned and body mass within subjects, but a negative association, albeit non-significant, between subjects (Figure 8b,c). In this model, males were marginally non-significantly heavier than females (βmale ± se = 6.73 ± 3.30; F1,9 = 4.16, p = 0.072). To test whether the difference in the parameter estimates for the within- and between-subject effects of reinforcements earned was significant, we fitted a third GLMM with both the original predictor, which combines within- and between-subject effects, and our new predictor, which expresses only between-subject variation. The between-subject effect in this model represents the difference between the between- and within-subject effects in the second model [35]. This GLMM showed a significant effect of the original combined predictor (β ± se = 0.0098 ± 0.0031; F1,117 = 9.96, p = 0.002), a marginally non-significant effect of the predictor expressing between-subject variation (which in this model, tests for a difference between within- and between-subject effects; β ± se = −0.0253 ± 0.0129; F1,10.1 = 3.84, p = 0.078) and a marginally non-significant effect of sex (βmale ± se = 6.73 ± 3.30; F1,9 = 4.16, p = 0.072). Thus, there is some evidence for different within- and between-subject effects of the number of reinforcements earned. Whereas earning more reinforcements during the day resulted in higher dusk mass within birds, between-bird differences in dusk mass were not explained by between-bird differences in the number of reinforcements earned. 4. Discussion We described a novel RFID-based technology, a social foraging system (SFS) that permits individual operant foraging behaviour and body masses of group-housed starlings to be measured from dawn until dusk, seven days a week. Our aim was to develop a refined method for studying body mass regulation in small birds that simultaneously delivered improved ecological validity, data quality, and animal welfare. We demonstrated that naïve starlings rapidly learn operant key pecking through an auto-shaping procedure while housed in groups of three birds. Once trained to forage from the SFS, starlings housed in groups of six maintained stable body masses while foraging on a ratio schedule under closed economy for a period of 11 consecutive days. Birds gained 6.0 ± 1.2 g (mean ± sd) between dawn and dusk each day and lost an equal amount overnight. Within each day, the rate of individual foraging behaviour (key pecking) decreased between dawn and dusk and birds gained mass non-linearly, with the rate of mass gain decelerating as the day progressed. There were stable individual differences in mean body mass over the 11 days of data collection. Within-bird variation in daily foraging rate positively predicted within-bird variation in dusk mass. However, between-bird variation in mean foraging rate was uncorrelated with between-bird variation in mean mass. Below, we discuss the scientific and animal welfare implications of these findings. 4.1. Diurnal Mass Gain Trajectories Using data from the SFS, we were able to study how both foraging effort (key pecks) and body mass changed within the day and between days in individual birds. Measures of foraging behaviour and body mass were reliable across days within birds, but differed between birds, suggesting that the measurement error was sufficiently small to allow us to detect stable individual differences in foraging behaviour and body mass. In contrast, estimates of daily mass gain and overnight mass loss were relatively unreliable, presumably because gain and loss were computed from the difference between two measurements of mass, doubling the measurement error. The birds foraged most intensively first thing in the morning and their rate of foraging declined as the day progressed. The majority of diurnal mass gain trajectories were fitted best with a cubic polynomial function. In the majority of birds, body mass increased most rapidly early in the day prior to levelling off later in the day. This diurnal trajectory of mass gain is predicted by optimality models when birds are constrained by only starvation risk and mass-dependent costs (i.e., predation risk) are absent or low [7]. In contrast, the most common diurnal mass gain trajectory observed in wild birds is a double-exponential function with a second period of rapid mass increase prior to dusk [36]. This latter trajectory is predicted when birds are constrained simultaneously by both starvation risk, which favours mass gain early in the day, and mass-dependent risk, which favours delaying mass gain until later in the day [7]. Interestingly, this double-exponential trajectory was also observed in caged coal tits (Periparus ater) caught frequently for manual weighing [37]. Over the 11 days of this study, only one starling in our (P78) showed any evidence of the double-exponential pattern in its mean mass gain trajectory, with a slight increase in the rate of mass gain prior to dusk (see Figure 7). Therefore, one interpretation of the diurnal mass trajectories that we observed was that our starlings perceived their predation risk to be low and regulated their body masses accordingly, by accumulating the majority of their daily mass gain early in the day. This interpretation could be tested by exploring whether experimentally increasing perceived predation risk (e.g., by unpredictably exposing the birds to a stuffed sparrow hawk) causes the predicted change in diurnal mass trajectories with a shift toward delaying some mass gain until later in the day. 4.2. Limitations on Measurement of Food Consumption Due to constraints of our methodology, data on food consumption was only available at the aviary level and we were unable to directly measure how much food each bird consumed each day. However, we were able to demonstrate that the daily total number of reinforcements earned by all the birds in each aviary predicted the daily total food consumption in that aviary, suggesting that the number of reinforcements earned (which is available for each individual) could be used as a proxy for individual food consumption. Obtaining accurate individual data for food consumption is currently only possible in individually housed birds. Due to space constraints in most laboratories, this usually implies housing birds in smaller cages that restrict natural behaviour (and compromise animal welfare) (e.g., [37]). Therefore, there is a trade-off between the level at which some variables can be measured and the ecological validity of experiments, which has implications for the types of effects that can be detected. Given the established importance of the social and physical environment for body mass (e.g., [20,23]), we argue that a move toward more naturalistic experimental paradigms in which opportunities for social interaction and physical activity are present is critical, even if this comes at the cost of limiting individual-level measurement of some variables. 4.3. Food Consumption and Body Mass Within individual birds, we observed a positive association between the foraging effort a bird made in a day (i.e., the total key pecks and hence total reinforcements earned) and its dusk mass that day. If we assume that individual differences in foraging effort relate to individual differences in food consumption, then we have evidence that within-individual variation in dusk mass is explained by the amount of food a bird consumed that day. Interestingly however, the larger between-bird differences in mean body mass that we observed were not predicted by between-bird differences in mean daily foraging effort (Figure 8c), suggesting that variation in mean body mass is not explained by the mean amount of food consumed. In animals for which body mass is stable (as was the case in our birds), daily energy intake must equal daily energy expenditure. It follows that one possible interpretation of the lack of association between mean foraging effort and mean body mass is that there are large between-bird differences in energy expenditure that are unrelated to mean body mass. In humans, a recent analysis of variation in daily energy expenditure reported large individual differences in energy expenditure, even after controlling for effects of fat free mass, sex, and age, suggestive of large unexplained individual differences in resting metabolic rate [38]. If similar variation is proven to exist in wild-caught starlings, this species could provide a good model for research into the causes of individual variation in resting metabolic rate. An alternative explanation for the lack of association between mean foraging effort and mean body mass is that there are large individual differences in the rate at which our birds consumed food from the hopper during the 5-s reinforcements. Further work is required to establish whether there is between-bird variation in the amount of food consumed per reinforcement earned (this could easily be measured by housing birds individually for a few days). However, even without further data, the magnitude of the individual variation seen in the number of reinforcements earned per day (see Table 1) suggests that it is likely that the birds consumed food at very different rates. 4.4. Animal Welfare Benefits The SFS delivers several refinements over the individual caging typically used for operant studies in small birds [26,37,39]. The real-time data on foraging behaviour and body weight provided by the SFS allowed us to closely monitor the welfare of the starlings. Any changes in behaviour or body weight indicative of a potential welfare problem were rapidly detected and investigated. The husbandry of the birds in the SFS is also likely to have delivered welfare improvements over previous methods. The birds were group housed throughout this study, which is likely to be important for a social species such as the starling. Starlings prefer to forage in the presence of conspecifics despite their rate of food intake being slower [40]. It is likely that social species such as starlings avoid lone foraging when possible due to the increased predation risk it presents [41]. Our birds were housed in an aviary that permitted a greater range of behaviour than is possible in smaller individual cages. In addition to allowing the birds enough space to fly properly, the size of the room allowed the birds to retreat to a high perch when humans entered the aviary. Caged wild-caught starlings move to the rear of their cages when a human enters the room, indicating fear of humans [42]. It thus seems likely that the ability to escape to a high perch would reduce the stress caused by human presence. Finally, the mass data provided by the SFS allowed us to minimise the frequency with which we had to catch the birds for manual weighing. Catching and manual weighing was only required by our protocol at the beginning and end of the study and for welfare monitoring in the early stages of habituation to the SFS. In contrast, to collect comparable data on daily mass gain trajectories from coal tits, birds were caught from their cages and restrained in a box for weighing 10 times per day for a period of at least 24 consecutive days [37]. Given that catching and restraint is an established procedure for inducing acute stress in small birds [43,44,45], this latter procedure is likely to have induced chronic stress, with potential effects on body mass. As noted above, the tits in the latter study showed diurnal mass gain trajectories predicted of birds subject to mass-dependent costs, whereas our starlings did not. This difference leads to the interesting possibility that diurnal mass gain trajectories in small birds could be used as a novel welfare indicator to measure perceived predation risk. 5. Conclusions The social foraging system (SFS) for European starlings simultaneously delivers improved ecological validity, data quality, and animal welfare compared with conventional methods in which birds are individually caged and caught frequently for manual weighing. The system delivers reliable estimates of individual foraging behaviour and body mass both within- and between-days. The SFS will facilitate novel research into the biology and psychology of body mass regulation including understanding the impacts of limited and unpredictable access to food (e.g., [2]). Acknowledgments Campden Instruments built the SFS and provided ongoing technical support throughout the project. Jonathon Dunn wrote the original software for the SFS and was involved in much of the original validation of the system not described in the current paper. Clare Andrews and Daniel Nettle assisted in running the study. Supplementary Materials The following are available online at https://www.mdpi.com/article/10.3390/ani12091159/s1, Document: Welfare Monitoring in European Starlings; Figures S1–S11: Supplementary Figures. Click here for additional data file. Author Contributions Conceptualization, M.B.; Methodology, M.B. and R.N.; Formal analysis, M.B.; Data curation, R.N.; Writing—original draft preparation, M.B.; Writing—review and editing, R.N.; Visualization, M.B.; Supervision, M.B.; Project administration, M.B. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted with approval of the Animal Welfare and Ethical Review Body at Newcastle University under a UK Home Office licence to M.B. (PPL P038AB1D3 granted 11 November 2019). Data Availability Statement Data and the analysis script are available at: https://doi.org/10.5281/zenodo.6368487, published on 18 March 2022. Conflicts of Interest The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. Figure 1 Diagram of a single social foraging system station. The wooden perch was 3.7 cm long and was mounted 22.0 cm above the base of the SFS and 8.0 cm from the hopper entrance. The pecking key was a 3.5-cm square of clear Perspex, hinged from the top and illuminated from behind with an array of LEDs. The food hopper was accessed via a 4 cm-wide aperture in the front panel of the SFS located to the left of the pecking key. Figure 2 Scheme showing the study design and timeline for the 12 birds used in this study. During operant training, there was one SFS station per aviary and during continuous foraging, there were two stations per aviary. Thus, the ratio of birds per SFS station was maintained at 3:1 throughout. The entire study lasted 22 days. Figure 3 Example of raw mass data from one bird. Panels 1–11 show scatterplots of individual stable mass measurements (g) by time of day (hours) for 11 consecutive days of continuous foraging under closed economy. Red points indicate masses that were excluded from the final model due to falling more than 10 g from the initial cubic fits (see Methods for details). The black line shows the cubic fit to the remaining grey points that was used to derive the estimated masses for specific times of day. Data from bird P77 are shown; equivalent graphs for the other 11 birds are shown in Supplementary Figures S1–S11. Figure 4 Group-housed experimentally-naïve birds rapidly acquired operant key pecking over three days of training. The six birds were housed in two single-sex groups of three during training (male group: P74, P79, and P98; female group: P75, P91, and P99). Sessions 1–3 took place on training days 5–7, following habituation to the SFS. Line graphs show the number of trials on which each bird pecked the lit key on the SFS by session. In sessions 1 and 2 (auto-shaping; 90 trials each), access to food was unconditional on key pecking, whereas in session 3 (270 trials), access to food was conditional on a bird pecking the lit key. Figure 5 Rate of earning reinforcements declines over the day. Panels show scatterplots of the average reinforcement rate (mean ± 95% CI) by time of day for each of the 12 birds. Each row of panels corresponds to one experimental aviary. Data from male birds are shown in blue and females in pink. Figure 6 Best-fitting polynomial models describing the association between time of day and body mass. Stacked bar chart showing the proportion of days for each bird that were fitted best by a linear, quadratic, or cubic polynomial model. The ninth day of data for bird P75 was not fitted due to insufficient data. Figure 7 Body mass increased non-linearly over the course of the day, slowing or peaking toward dusk. Panels show the mean predicted mass (±95% CI) by time of day for each of the 12 birds. Each row of panels corresponds to one experimental aviary. Data from male birds are shown in blue and females in pink. Figure 8 The positive effect of foraging effort (daily reinforcements earned) on dusk body mass is driven by a within-subjects effect. (a) Scatterplot of dusk mass by the total number of reinforcements·× day−1 with each point representing the data from one bird day. Lines show the fitted values from a standard mixed model that combines within- and between subject effects. (b) Scatterplot of dusk mass by within-subject centred reinforcements·× day−1 with each point representing the data from one bird on one day. The line shows the significant positive within-subject effect. (c) Scatterplot of mean dusk mass by mean reinforcements × day−1 with points representing birds. The line shows the absence of a positive between-subject effect. animals-12-01159-t001_Table 1 Table 1 Details of birds and descriptive statistics (mean ± sd) for reinforcements earned and body mass data †. Aviary Bird ID Sex Tarsus Length (mm) Reinforcements × Day−1 Masses·Day−1 Dawn × Mass Noon × Mass Dusk × Mass Daily × Gain Nightly × Loss 114 P77 F 28.8 243 ± 47 71 ± 20 76.3 ± 1.6 80.1 ± 0.9 81.4 ± 1.4 5.1 ± 1.8 5.0 ± 1.7 114 P92 F 30.5 145 ± 44 44 ± 17 75.2 ± 2.2 79.6 ± 1.4 81.5 ± 2.3 6.3 ± 3.0 6.4 ± 3.9 114 P95 F 29.0 358 ± 65 127 ± 40 72.1 ± 1.8 77.1 ± 2.0 77.8 ± 2.7 5.6 ± 2.9 6.0 ± 1.9 114 P78 M 29.4 348 ± 68 41 ± 10 71.1 ± 1.1 75.9 ± 1.5 76.1 ± 1.7 5.0 ± 1.9 4.9 ± 2.3 114 P80 M 29.9 349 ± 56 69 ± 14 91.4 ± 1.1 94.4 ± 1.3 95.0 ± 2.1 3.6 ± 2.2 3.8 ± 2.0 114 P93 M 30.5 129 ± 23 25 ± 11 85.3 ± 5.2 92.6 ± 1.4 92.1 ± 2.6 6.9 ± 5.8 7.1 ± 6.7 116 P75 F 29.6 140 ± 33 31 ± 11 79.4 ± 1.5 85.2 ± 2.7 85.7 ± 2.3 6.5 ± 2.7 6.3 ± 3.1 116 P91 F 30.5 168 ± 51 81 ± 30 77.9 ± 2.6 83.5 ± 2.3 85.3 ± 2.1 7.4 ± 3.1 7.3 ± 2.9 116 P99 F 28.4 150 ± 38 63 ± 27 76.8 ± 1.6 83.0 ± 1.9 85.1 ± 1.0 8.3 ± 1.5 8.3 ± 1.6 116 P74 M 30.6 568 ± 103 107 ± 38 78.4 ± 2.3 83.1 ± 2.1 84.9 ± 1.5 6.5 ± 2.5 6.4 ± 2.6 116 P79 M 30.9 160 ± 26 42 ± 15 82.0 ± 2.6 86.0 ± 2.3 87.2 ± 2.0 5.2 ± 3.6 5.2 ± 3.0 116 P98 M 31.5 372 ± 81 91 ± 38 84.5 ± 2.6 89.3 ± 2.6 90.5 ± 1.7 6.0 ± 1.5 6.2 ± 2.7 † Mass statistics are based on fits from 11 days of continuous foraging for all birds except P75, which had 10 days due to insufficient data for fitting on one day. ‘Masses × day−1′ is the number of stable masses remaining after all exclusions (see Methods for details). ‘Dawn mass’ is the fitted mass at 0800 h, ‘Noon mass’ is the fitted mass at 1200 h, and ‘Dusk mass’ is the fitted mass at 1600 h. ‘Daily gain’ is the difference between dawn mass and dusk mass. ‘Nightly loss’ is the difference between the dusk mass the previous day and dawn mass. All masses are in grams. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Halsey L.G. Keeping Slim When Food Is Abundant: What Energy Mechanisms Could Be at Play? Trends Ecol. Evol. 2018 33 745 753 10.1016/j.tree.2018.08.004 30241777 2. Bateson M. Andrews C. Dunn J. Egger C.B.C.M. Gray F. Mchugh M. Nettle D. Food Insecurity Increases Energetic Efficiency Not Food Consumption: An Exploratory Study in European Starlings PeerJ 2021 9 e11541 10.7717/peerj.11541 34123601 3. Lima S.L. Predation Risk and Unpredictable Feeding Conditions: Determinants of Body Mass in Birds Ecology 1986 67 377 385 10.2307/1938580 4. Bednekoff P.A. Houston A.I. Optimizing Fat Reserves over the Entire Winter: A Dynamic Model Oikos 1994 71 408 415 10.2307/3545828 5. Higginson A.D. McNamara J.M. Houston A.I. Fatness and Fitness: Exposing the Logic of Evolutionary Explanations for Obesity Proc. R. Soc. B Biol. Sci. 2016 283 20152443 10.1098/rspb.2015.2443 26740612 6. McNamara J.M. Houston A.I. The Value of Fat Reserves and the Tradeoff between Starvation and Predation Acta Biotheor. 1990 38 37 61 10.1007/BF00047272 2109917 7. Polo V. Bautista L.M. Daily Routines of Body Mass Gain in Birds: 1. An Exponential Model Anim. Behav. 2006 72 503 516 10.1016/j.anbehav.2005.09.024 8. Witter M.S. Swaddle J.P. Mass Regulation Response to Change in Food Availability Depends on Initial Body Mass Funct. Ecol. 1997 11 11 15 10.1046/j.1365-2435.1997.00041.x 9. Witter M.S. Swaddle J.P. Cuthill I.C. Periodic Food Availability and Strategic Regulation of Body Mass in the European Starling, Sturnus Vulgaris Funct. Ecol. 1995 9 568 574 10.2307/2390146 10. Cuthill I.C. Maddocks S.A. Weall C.V. Jones E.K.M. Body Mass Regulation in Response to Changes in Feeding Predictability and Overnight Energy Expenditure Behav. Ecol. 2000 11 189 195 10.1093/beheco/11.2.189 11. Andrews C. Zuidersma E. Verhulst S. Nettle D. Bateson M. Exposure to Food Insecurity Increases Energy Storage and Reduces Somatic Maintenance in European Starlings (Sturnus vulgaris) R. Soc. Open Sci. 2021 8 211099 10.1098/rsos.211099 34540262 12. Folwarczny M. Christensen J.D. Li N.P. Sigurdsson V. Otterbring T. Crisis Communication, Anticipated Food Scarcity, and Food Preferences: Preregistered Evidence of the Insurance Hypothesis Food Qual. Prefer. 2021 91 104213 10.1016/j.foodqual.2021.104213 13. Lee A.M. Cardel M.I. Social Status and Adolescent Physical Activity: Expanding the Insurance Hypothesis to Incorporate Energy Expenditure Am. J. Lifestyle Med. 2019 13 156 160 10.1177/1559827618815449 30800021 14. Nettle D. Andrews C. Bateson M. Food Insecurity as a Driver of Obesity in Humans: The Insurance Hypothesis Behav. Brain Sci. 2017 40 e105 10.1017/S0140525X16000947 27464638 15. Nettle D. Bateson M. Food-Insecure Women Eat a Less Diverse Diet in a More Temporally Variable Way: Evidence from the US National Health and Nutrition Examination Survey, 2013–2014 J. Obes. 2019 2019 7174058 10.1155/2019/7174058 31662904 16. Anselme P. Güntürkün O. How Foraging Works: Uncertainty Magnifies Food-Seeking Motivation Behav. Brain Sci. 2019 42 1 59 10.1017/S0140525X18000948 17. Anselme P. Otto T. Güntürkün O. How Unpredictable Access to Food Increases the Body Fat of Small Passerines: A Mechanistic Approach Behav. Process. 2017 144 33 45 10.1016/j.beproc.2017.08.013 18. Brunstrom J.M. Cheon B.K. Do Humans Still Forage in an Obesogenic Environment? Mechanisms and Implications for Weight Maintenance Physiol. Behav. 2018 193 261 267 10.1016/j.physbeh.2018.02.038 29471074 19. Dall S.R.X. Witter M.S. Feeding Interruptions, Diurnal Mass Changes and Daily Routines of Behaviour in the Zebra Finch Anim. Behav. 1998 55 715 725 10.1006/anbe.1997.0749 9514680 20. Miles J.L. Huber K. Thompson N.M. Davison M. Breier B.H. Moderate Daily Exercise Activates Metabolic Flexibility to Prevent Prenatally Induced Obesity Endocrinology 2009 150 179 186 10.1210/en.2008-1035 18772230 21. Miles J.L. Landon J. Davison M. Krageloh C.U. Thompson N.M. Triggs C.M. Breier B.H. Prenatally Undernourished Rats Show Increased Preference for Wheel Running v. Lever Pressing for Food in a Choice Task Br. J. Nutr. 2009 101 902 908 10.1017/S0007114508043353 18761782 22. Oberweger K. Goller F. The Metabolic Cost of Birdsong Production J. Exp. Biol. 2001 204 3379 3388 10.1242/jeb.204.19.3379 11606611 23. Witter M.S. Swaddle J.P. Dominance, Competition, and Energetic Reserves in the European Starling, Sturnus Vulgaris Behav. Ecol. 1995 6 343 348 10.1093/beheco/6.3.343 24. Hursh S.R. Economic Concepts for the Analysis of Behavior J. Exp. Anal. Behav. 1980 34 219 238 10.1901/jeab.1980.34-219 16812188 25. Houston A.I. McNamara J.M. The Value of Food: Effects of Open and Closed Economies Anim. Behav. 1989 37 546 562 10.1016/0003-3472(89)90034-1 26. Feenders G. Bateson M. Hand Rearing Affects Emotional Responses but Not Basic Cognitive Performance in European Starlings Anim. Behav. 2013 86 127 138 10.1016/j.anbehav.2013.05.002 23888084 27. Percie N. Hurst V. Ahluwalia A. Alam S. Avey T. Baker M. Browne W.J. Clark A. Cuthill I.C. Dirnagl U. The ARRIVE Guidelines 2.0: Updated Guidelines for Reporting Animal Research PLoS Biol. 2020 18 e3000410 10.1371/journal.pbio.3000410 32663219 28. Smith E.L. Cuthill I.C. Griffiths R. Greenwood V.J. Goldsmith A.R. Evans J.E. Sexing Starlings Sturnus Vulgaris Using Iris Colour Ringing Migr. 2005 22 193 197 10.1080/03078698.2005.9674332 29. Redfern C.P.F. Clark J.A. Ringers’ Manual British Trust for Ornithology Norfolk, UK 2001 30. Cardinal R.N. Aitken M.R.F. Whisker: A Client-Server High-Performance Multimedia Research Control System Behav. Res. Methods 2010 42 1059 1071 10.3758/BRM.42.4.1059 21139173 31. R Core Team R: A Language and Environment for Statistical Computing R Core Team Vienna, Austria 2018 Available online: https://www.R-project.org/ (accessed on 18 March 2022) 32. Bates D. Maechler M. Bolker B. Walker S. Fitting Linear Mixed-Effects Models Using lme4 J. Stat. Softw. 2015 67 1 48 10.18637/jss.v067.i01 33. Kuznetsova A. Brockhoff P.B. Christensen R.H.B. LmerTest Package: Tests in Linear Mixed Effects Models J. Stat. Softw. 2017 82 1 26 10.18637/jss.v082.i13 34. Revelle W. psych: Procedures for Personality and Psychological Research Northwestern University Evanston, IL, USA 2018 Available online: https://cran.r-project.org/web/packages/psych/index.html (accessed on 18 March 2022) 35. van de Pol M. Wright J. A Simple Method for Distinguishing within- versus between-Subject Effects Using Mixed Models Anim. Behav. 2009 77 753 758 10.1016/j.anbehav.2008.11.006 36. Pravosudov V. Grubb T. Jr. Energy Management in Passerine Birds during the Nonbreeding Season Curr. Ornithol. 1997 14 189 234 37. Polo V. Bautista L.M. Daily Routines of Body Mass Gain in Birds: 2. An Experiment with Reduced Food Availability Anim. Behav. 2006 72 517 522 10.1016/j.anbehav.2005.09.025 38. Pontzer H. Yamada Y. Sagayama H. Ainslie P.N. Andersen L.F. Anderson L.J. Arab L. Baddou I. Bedu-Addo K. Blaak E.E. Daily Energy Expenditure through the Human Life Course Science 2021 373 808 812 10.1126/science.abe5017 34385400 39. Matheson S.M.S.M. Asher L. Bateson M. Larger, Enriched Cages Are Associated with ‘optimistic’ Response Biases in Captive European Starlings (Sturnus Vulgaris) Appl. Anim. Behav. Sci. 2008 109 374 383 10.1016/j.applanim.2007.03.007 40. Vasquez R.A. Kacelnik A. Foraging Rate versus Sociality in the Starling Sturnus Vulgaris Proc. R. Soc. B 2000 267 157 164 10.1098/rspb.2000.0981 10687821 41. Roberts G. Why Individual Vigilance Declines as Group Size Increases Anim. Behav. 1996 51 1077 1086 10.1006/anbe.1996.0109 42. Feenders G. Bateson M. Hand-Rearing Reduces Fear of Humans in European Starlings, Sturnus Vulgaris PLoS ONE 2011 6 e17466 10.1371/journal.pone.0017466 21364770 43. Spencer K.A. Evans N.P. Monaghan P. Postnatal Stress in Birds: A Novel Model of Glucocorticoid Programming of the Hypothalamic-Pituitary-Adrenal Axis Endocrinology 2009 150 1931 1934 10.1210/en.2008-1471 19095740 44. Andrews C. Nettle D. Larriva M. Gillespie R. Reichert S. Brilot B.O.B.O. Bedford T. Monaghan P. Spencer K.A.K.A. Bateson M. A Marker of Biological Age Explains Individual Variation in the Strength of the Adult Stress Response R. Soc. Open Sci. 2017 4 171208 10.1098/rsos.171208 28989794 45. Gott A. Andrews C. Hormigos M.L. Spencer K. Bateson M. Nettle D. Chronological Age, Biological Age, and Individual Variation in the Stress Response in the European Starling: A Follow-up Study PeerJ 2018 6 e5842 10.7717/peerj.5842 30370189

PMC009xxxxxx/PMC9099604.txt

==== Front J Clin Med J Clin Med jcm Journal of Clinical Medicine 2077-0383 MDPI 10.3390/jcm11092638 jcm-11-02638 Article Higher Risk of Anxiety and Depression in Women with Adenomyosis as Compared with Those with Uterine Leiomyoma https://orcid.org/0000-0002-3906-4066 Li Ni Yuan Ming https://orcid.org/0000-0003-0141-4216 Li Qiuju Ji Miaomiao https://orcid.org/0000-0002-4172-6794 Jiao Xue https://orcid.org/0000-0003-4023-084X Wang Guoyun * Guo Sun-Wei Academic Editor Sheiner Eyal Academic Editor Department of Obstetrics and Gynecology, Qilu Hospital of Shandong University, Jinan 250012, China; linismile@126.com (N.L.); sddxqlyyym@163.com (M.Y.); qiuju901220@163.com (Q.L.); jmmdoc0506@163.com (M.J.); jiaoxue987@163.com (X.J.) * Correspondence: wangguoy@sdu.edu.cn; Tel.: +86-18560081729 07 5 2022 5 2022 11 9 263825 3 2022 05 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). The main symptoms of adenomyosis may adversely affect physical and mental health and quality of life (QOL). However, studies are few on this topic. This study evaluated the effect of adenomyosis on anxiety, depression, and QOL. Participants with adenomyosis (n = 90) or leiomyoma (n = 59) completed questionnaires, including the visual analog scale (VAS) for pain, Hospital Anxiety and Depression Scale (HADS), and Short Form (SF)-36. HADS anxiety and depression positive rates, physical (PCS) and mental (MCS) component summary scores, and eight subscale scores of SF-36 were compared between the two groups. Among patients with adenomyosis, the incidence of anxiety symptoms was 28.9% (control group, 10.2%; OR = 3.589, 95% CI: 1.375–9.367), with 10% of patients showing moderate-to-severe symptoms. The incidence of depressive symptoms was 14.4% (control group, 3.4%; OR = 4.812, 95% CI: 1.044–22.168). The case group had significantly lower PCS and MCS scores than the control group. In patients with adenomyosis, being employed (OR = 6.393, 95% CI: 1.153–35.440) and perianal pain (OR = 25.419, 95% CI: 2.504–258.024) were risk factors for anxiety, and perianal pain (OR = 15.208, 95% CI: 3.050–75.836) was a risk factor for depression. Compared with leiomyoma, adenomyosis is associated with a higher risk of anxiety and depression, with a poorer QOL. adenomyosis anxiety depression quality of life the National Natural Science Foundation of China82071621 81901458 the Major Basic Research of Natural Science Foundation of ShandongZR2021ZD34 This research was funded by the National Natural Science Foundation of China (grant numbers 82071621 and 81901458) and the Major Basic Research of Natural Science Foundation of Shandong (grant number ZR2021ZD34). ==== Body pmc1. Introduction Adenomyosis is characterized by uterine enlargement, caused by ectopic endometrial tissue, comprising both glands and stroma, growing deep within the myometrium [1]. It is a common gynecological disease, but its mechanism of action is unclear. Adenomyosis mainly occurs in premenopausal women aged 40–50 years, with a prevalence of 5–70% [2]. In the past, epidemiological data were mainly from patients with hysterectomy. The lack of population-based epidemiological data and inconsistent diagnostic criteria have led to marked changes in disease prevalence. In recent years, with the improvement in magnetic resonance imaging (MRI), two- and three-dimensional transvaginal sonography, and other technologies, the detection of adenomyosis has gradually increased, and it may now be detected at a younger age [3,4,5]. In a United States population-based study (2006–2015), the overall incidence of adenomyosis was 1.03% [6]. Several ultrasonographic criteria have been utilized for the diagnosis of adenomyosis, including uterine enlargement, asymmetry of the anterior and posterior uterine wall thickness, presence of heterogeneous myometrial areas, and poor definition and thickening of the junctional zone (JZ). In an MRI, typical adenomyosis appears as an ill-demarcated low-signal intensity area on T2-weighted images, and thickening of the JZ of the uterus (≥12 mm) [7]. Gureje et al., found a strong and symmetrical relationship between persistent pain and psychological disorders [8]. Iron-deficiency anemia is associated with higher levels of psychological distress [9,10]. An infertility diagnosis is a risk factor for future depression in those undergoing fertility treatments [11,12]. The main symptoms of adenomyosis are progressive dysmenorrhea, chronic pelvic pain, menorrhagia, and infertility, which affect a majority of women of childbearing age [7,13]. Hysterectomy, which is currently the only radical curative method for adenomyosis, is unsuitable for patients who have fertility requirements or wish to preserve their uterus. However, other treatments have shortcomings, such as recurrence and adverse effects. For patients with adenomyosis who undergo conservative treatment, a series of persistent or recurring symptoms, such as dysmenorrhea and menorrhagia, usually have continuous adverse effects on their physical and mental health and quality of life. Compared to patients with uterine myoma, patients with adenomyosis more frequently have a history of depression (up to 57.1%), and their use of antidepressants is also higher [14]. Compared with pain patients without depression or anxiety, pain patients with anxiety or depression, such as those with primary pain, cancer pain, neuropathic pain, and visceral pain, have a worse response to pain treatment, lower satisfaction, more severe and longer-lasting pain symptoms, and reduced pain tolerance [15,16,17]. Drugs and surgery are currently inevitable treatment modalities, but the pain’s etiology cannot be substantially improved in the short term. Psychological evaluation needs to be expanded to improve treatment effects. Investigators have begun to study the impact of adenomyosis on the quality of life of patients and have used indexes to evaluate the effect of treatment [18,19,20]. However, only a few related studies have been conducted in this field. Therefore, this study aimed to evaluate the impact of adenomyosis on anxiety, depression, and quality of life through investigation, to remind gynecologists to pay attention to the psychological state of patients with adenomyosis, provide appropriate treatment, and help in the long-term management of adenomyosis. 2. Materials and Methods The study included patients who were hospitalized at the Department of Gynecology, Qilu Hospital of Shandong University (Jinan, China), from 8 June 2019 to 7 January 2020. The case group included 90 patients with adenomyosis diagnosed by histopathology. The inclusion criteria for the case group were as follows: (1) admission to the gynecological ward; (2) main diagnosis of adenomyosis, and (3) provision of informed consent to participate in the study. The exclusion criteria were as follows: (1) a postoperative histopathological diagnosis excluding adenomyosis or a main diagnosis of non-adenomyosis; (2) co-occurrence of malignant tumor; (3) other serious internal and external diseases and serious cognitive problems that could prevent cooperation with the investigation, and (4) poor compliance and refusal to cooperate with investigators. The control group included 59 patients with leiomyoma. The inclusion criteria for the control group were as follows: (1) main diagnosis of leiomyoma; (2) preparation for surgical treatment in the gynecological ward at the same time as the case group, and (3) provision of informed consent to participate in this study. The exclusion criteria were: (1) postoperative histopathological findings complicated with adenomyosis or endometriosis; (2) co-occurrence of malignant tumor; (3) presence of other serious internal and external diseases and serious cognitive problems that could prevent cooperation with the investigation, and (4) poor compliance and refusal to cooperate with investigators. A total of 149 participants were included in the analysis (Figure 1). 2.1. Assessments and Questionnaires All patients completed questionnaires and underwent assessments after admission and before surgical treatment. 2.1.1. Assessment of Anxiety and Depression Anxiety and depression levels were evaluated using the Hospital Anxiety and Depression Scale (HADS) [21]. This scale is commonly used to screen for anxiety and depression symptoms. This scale includes two subscales, Anxiety (HADS-A) and Depression (HADS-D). The higher the subscale score, the higher the degree of anxiety or depression. The highest possible total HADS-A score is 21. If the score is ≥8, the patient is considered to have anxiety; 0–7 points indicate a normal state, 8–10 points indicate mild anxiety, 11–15 points indicate moderate anxiety, and ≥16 points indicate severe anxiety. The HADS-D scale follows a similar pattern with regard to depression [22]. The Cronbach’s alpha coefficient of the scale in this study was 0.886. 2.1.2. Assessment of Quality of Life Quality of life was measured using the 36-item Short-Form Health Survey (SF-36) [23]. The SF-36 measures eight health concepts: (1) physiological function (PF); (2) role-physical (RP); (3) bodily pain (BP); (4) general health (GH); (5) vitality (VT); (6) social functioning (SF); (7) role-emotional (RE), and (8) mental health (MH). Scores also contribute to two composite scores: physical component summary (PCS) and mental component summary (MCS). The Cronbach’s alpha coefficient of the scale in this study was 0.916. 2.1.3. Pain Assessment The visual analog scale (VAS) was used to evaluate the degree of pain during the patient’s last menstruation. We divided VAS into three grades: 0 to 3, no or mild pain; 4 to 6, moderate pain, and 7 to 10, severe pain. 2.1.4. Demographic and Clinical Variables We used a self-designed questionnaire to obtain information regarding demographic and clinical characteristics, such as age, height, weight, educational level, occupation, smoking history, menstruation, marriage, and childbearing history, past history, operation history, and relevant information on adenomyosis, including clinical symptoms, course of disease, and treatment. 2.2. Outcome Measures The main outcome measures in this study were the comparison of the HADS-A and HADS-D positive rates of the two groups and the PCS and MCS scores in the SF-36 questionnaire. Secondary indicators included the relationship and factors influencing anxiety, depression, and quality of life. 2.3. Statistical Analysis Statistical analysis of the collected data was performed using SPSS version 25 for Mac OS (IBM, Armonk, NY, USA). None of the continuous variables were normally distributed; therefore, they are presented as median (interquartile range). Data analysis of continuous variables and ordinal categorical variables was performed using the Wilcoxon rank-sum test. Data analysis of non-ordinal categorical variables was performed using the Chi-square test. Logistic regression analysis was used to identify the risk factors. Spearman’s correlation test was used for the correlation analysis. Statistical significance was set at p < 0.05. 3. Results 3.1. Demographic and Clinical Characteristics In the case group, the median age was 44 years, and 57.8% of the patients were aged 40–49 years. The body mass index (BMI) was 24.4 (22.2–27.3) kg/m2, and 56.2% of patients were overweight (36%) or obese (20.2%). In the case group, 84.4% had a history of uterine surgery, 43.3% had a history of uterine cavity surgery, and 22.2% had a history of ovarian and fallopian tube surgery. Apart from a history of surgery, there were no significant differences between the two groups in terms of age, BMI, educational background, employment, smoking history, age at menarche, or parity (Table 1). The VAS score of patients with adenomyosis was 9 (6–10) (controls: 0 (0–5.2), p < 0.001)), and 81.1% of the case group had moderate-to-severe pain. When they experienced pain, 65.6% of patients required analgesics to relieve their symptoms. Among the case group, 32.8% had hypermenorrhea. The hemoglobin level in the case group was 110 (92–123) g/L, and 21.8% had moderate-to-severe anemia. There were no significant differences in infertility, hypermenorrhea, or anemia between the two groups (Table 2). 3.2. Anxiety and Depression Nine of the patients with adenomyosis were diagnosed with anxiety or depression in the past, and four of these patients received psychotherapy. No significant differences were found between the two groups regarding the history of anxiety, depression, or treatment (Table 3). The HADS-A scores revealed that 28.9% (n = 26) of the case group presented with anxiety, which was moderate to severe in 10% (n = 9). In the control group, the HADS-A scores revealed that 10.2% (n = 6) of women presented with anxiety, which was moderate to severe in 5.1% (n = 3). There was a significant difference in the rate of anxiety symptoms between the two groups (X2 = 7.405, p = 0.007; OR = 3.589, 95% CI: 1.375–9.367), and there were more people with moderate-to-severe anxiety in the case group (Figure 2). In the case group, the HADS-D scores revealed that 14.4% (n = 13) had depressive symptoms, which were moderate to severe in 6.7% (n = 6). The HADS-D scores revealed that 3.4% (n = 2) of the control group presented with depression, which was moderate to severe in 1.7% (n = 1). There was a significant difference in the rate of depression symptoms between the two groups (X2 = 4.810, p = 0.028; OR = 4.812, 95% CI: 1.044–22.168), and there were more people with moderate-to-severe anxiety in the case group (Figure 2). Univariate logistic regression analysis was used to identify risk factors for anxiety in patients with adenomyosis. Variables with p < 0.1 (employment, dyspareunia, and perianal pain) were introduced into the multivariate logistic regression analysis to determine independent risk factors for anxiety in patients with adenomyosis. Being employed (OR = 6.393, 95% CI: 1.153–35.440) and perianal pain (OR = 25.419, 95% CI: 2.504–258.024) were identified as risk factors for anxiety in patients with adenomyosis. Factors identified as significantly different between the case and control groups underwent univariate binary logistic regression to analyze the risk factors for depression in patients with adenomyosis. Only the p-value for perianal pain was less than 0.1; therefore, multivariate logistic regression analysis was not performed. Perianal pain (OR = 15.208, 95% CI: 3.050–75.836) was identified as a risk factor for depression in patients with adenomyosis. 3.3. Quality of Life Women with adenomyosis had lower scores than controls for both PCS (case: 68.125 (49–78.5) vs. controls: 86 (68.5–91), p < 0.001) and MCS (case: 71.1 (58.95–84.5) vs. control: 80.8 (73.8–87.7), p = 0.002) scores. Except for PF and MH, participants with adenomyosis had lower scores in the other SF-36 health subscales (Table 4 and Figure 3A). 3.4. Relationship between Anxiety, Depression, and Quality of Life in Patients with Adenomyosis There was a significant correlation between anxiety and depression (r = 0.568, p < 0.001) in the case group. In this study, 13.3% (n = 12) of patients with adenomyosis had both anxiety and depression symptoms. Spearman’s correlation revealed that PCS was negatively correlated with HADS-A scores (rs = −0.454, p < 0.001) and HADS-D scores (rs = −0.439, p < 0.001), and MCS was also negatively correlated with HADS-A scores (rs = −0.653, p < 0.001) and HADS-D scores (rs = −0.676, p < 0.001) (Table 5). There was a significant correlation between quality of life, anxiety, and depression. Patients with symptoms of anxiety and depression tended to have a lower quality of life (Figure 3B). Similarly, patients with a low quality of life were more likely to suffer from anxiety and depression. 4. Discussion In this study, women with adenomyosis had higher scores in the anxiety and depression subscales of the HADS, as well as lower scores in all domains and in the PCS and MCS scores of the SF-36 questionnaire, compared to women with leiomyoma. This result is consistent with the study of Alcalde et al., on outpatients with adenomyosis [18]. This phenomenon also exists in patients with other chronic diseases, such as diabetes, cardiovascular disease, and rheumatoid arthritis [24]. However, a previous study in the United States showed that 57.1% of patients with adenomyosis had a history of depression [14], which is notably higher than in this study. This may be because the prevalence of depression in the US population is higher than that in the Chinese population (19.2% vs. 6.5%) [25], and the Chinese have insufficient awareness of mental illness and are unwilling to admit to having mental problems [26]. The endocannabinoid system (ECS) is a neuromodulatory system that can coordinate appropriate behavioral responses, which are essential for long-term survival and the health of the body. Disorders in ECS signal transduction can lead to negative emotional states, such as anxiety and depression [27,28,29]. In women with adenomyosis, cannabinoid receptor 1 expression is downregulated [30]. Therefore, it is speculated that a disorder of the ECS may be the main reason for the increased prevalence of anxiety in patients with adenomyosis. Xu et al., suggested that sympathetic-nerve-derived neurotransmitters, such as noradrenaline, may promote the development of adenomyosis through activation of their respective receptors on adenomyotic lesions [31]. Activation of peripheral presynaptic CB(1) receptors inhibits noradrenaline release from sympathetic nerve terminals [32]. The activation of cannabinoid receptors can reduce anxiety, regulate mood, inhibit the development of adenomyosis, and may be a potential target for the treatment of adenomyosis. Current evidence indicates that women with endometriosis have an increased prevalence of psychological disorders that correlate more with pain itself than with endometriosis per se [33,34,35]. In the present study, quality of life and psychological well-being were not found to be related to the severity of pain. However, Alcalde et al., found that when it is associated with symptoms, the quality of life is further diminished [18]. It has not yet been elucidated whether depression and anxiety determine an increased perception of pain or whether pain causes psychopathological symptoms. However, anxiety and depression could increase pain perception, both emotionally and cognitively, leading to less tolerance to pain and greater sensitivity to physical sensations in general [36]. In our research, it was found that perianal pain is a risk factor for anxiety and depression in patients with adenomyosis. Patients with adenomyosis with perianal pain were more likely to have symptoms of anxiety and depression. Compared with unemployed people, employed patients with adenomyosis were more likely to have symptoms of anxiety. This is because adenomyosis has a clinically relevant impact on work productivity, with higher rates of absenteeism, overall loss of work productivity, and impairment of daily activities [18]. In this study, most patients with anxiety and depression that were screened with the questionnaire did not realize they had anxiety and depression, and most Chinese people are unwilling to see a psychologist [26], which explains why the number of patients with adenomyosis who were previously diagnosed with anxiety or depression was low in the study. Studies have shown that the physical symptoms of chronic disease are significantly higher in patients with anxiety or depression than in those without anxiety and depression and that physical symptoms can be significantly improved after treatment for anxiety and depression [24,37,38]. Anxiety symptoms affect the outcome of pain treatment. Even if analgesics are used, patients with anxiety symptoms are still less satisfied with pain treatments [17]. Research shows that improving the mental health status of patients with chronic diseases can improve their quality of life and improve the treatment effect of diseases [37,38]. Zhao et al., suggests that progressive muscle relaxation training is effective in improving anxiety, depression and QOL in endometriosis patients under GnRH agonist therapy [39]. Considering the high incidence of anxiety and depression in patients with adenomyosis and that both conditions reduce the quality of life of patients, thus, influencing the effect of treatments, attention should be paid to the mental problems of patients, especially those with perianal pain and employed during clinical diagnosis and treatment. Patients with adenomyosis can be screened through a simple scale to identify the symptoms of anxiety and depression early and provide patients with personalized treatment and necessary psychotherapy to improve the treatment and long-term management of adenomyosis. This study found that the quality of life of patients with adenomyosis was poor, both in PCS and MCS, which was similar to previous studies [18,40]. Alcalde et al., compared 89 patients with adenomyosis with 203 normal women and found that the eight dimensions of the SF-36 scale in the observation group were significantly lower than those in the control group. However, in this study, there was no significant difference in PF or MH compared with the control group, which may be related to the different criteria for selecting patients. In this study, the control group comprised patients with leiomyoma, which will also affect the quality of life of the patients [41]. In this study, SF-36 was used to measure the quality of life of the patients because it has universal applicability and met the requirements of this study for comparison with patients with other diseases. However, this scale cannot assess the impact of adenomyosis on the quality of life of patients in terms of sexual life, pregnancy, and response to treatment, which are commonly assessed in other universal scales. The Endometriosis Health Profile (EHP-30), a proprietary scale for endometriosis, provides a comprehensive assessment on the quality of life of patients in daily life, work, sexual relationships, education, social interaction, and psychology [42]. This scale has been used to assess the therapeutic effects of adenomyosis [43]. However, the validity and reliability of this scale in patients with adenomyosis require further research. This was a cross-sectional study, without longitudinal, dynamic observation, and it could only show the patient’s anxiety, depression, and quality of life at the time of evaluation. To obtain a histopathological diagnosis, the patients selected in this study were only hospitalized patients; however, compared with outpatients, the anxiety, depression, and quality of life of inpatients may have been more severe. Moreover, the sample size of this study was small. To determine the prevalence of anxiety and depression in patients with adenomyosis and the influencing factors, large-sample, multicenter cohort studies are warranted. Further research is needed to improve the mental health of these patients. 5. Conclusions In our study, compared to those with leiomyoma, patients with adenomyosis had a higher risk of anxiety and depression, and the quality of life of these patients was poor. Being employed and perianal pain were risk factors for anxiety in patients with adenomyosis, while perianal pain was the risk factor for depression in patients with adenomyosis. Based on these results, patients with adenomyosis can be screened through a simple scale to identify the symptoms of anxiety and depression early and be provided with individualized treatment and psychotherapy to improve treatment outcomes and the long-term management of adenomyosis. Acknowledgments We would like to thank Xiaorong Yang (Clinical Research Center of Shandong University, Clinical Epidemiology Unit, Qilu Hospital of Shandong University, Jinan, Shandong, China) for providing statistical help. Author Contributions Conceptualization, G.W., N.L., M.Y. and Q.L.; methodology, G.W., N.L., M.Y. and Q.L.; formal analysis, N.L.; investigation, N.L., M.Y., Q.L., M.J. and X.J.; resources, G.W.; data curation, N.L.; writing—original draft preparation, N.L.; writing—review and editing, G.W., N.L., M.Y., Q.L., M.J. and X.J.; visualization, N.L.; supervision, G.W. and N.L.; project administration, G.W., N.L., M.Y. and Q.L.; funding acquisition, G.W. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the guidelines of the Declaration of Helsinki and approved by the Ethics Committee (approval number: 2019(112)). Informed Consent Statement Informed consent was obtained from all the subjects involved in the study. Data Availability Statement The data presented in this study are available upon reasonable request from the corresponding author. Conflicts of Interest The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. Figure 1 Data collection flowchart. An incomplete questionnaire indicates a patient did not complete the required questionnaire, such as HADS or SF-36. Figure 2 (A): Distribution of anxiety and depression. * p < 0.05, ** p < 0.01. (B): Severity of anxiety and depression in patients with adenomyosis. Figure 3 SF-36 ** p < 0.01, *** p < 0.001. (A) SF-36 score of patients with adenomyosis and leiomyoma. (B) SF-36 score of anxiety and depression in patients with adenomyosis. jcm-11-02638-t001_Table 1 Table 1 Demographic characteristics. M(P25–75) a/N(%) Cases (n = 90) Controls (n = 103) X2/Z b p Age (years) 44 (40–48) 44 (36–47) −1.021 0.307 BMI (kg/m2) 24.4 (22.2–27.3) 23.3 (21.4–26.5) −1.547 0.122 Educational background 1.857 0.395 Primary Education 13 (14.4%) 6 (10.2%) Secondary Education 63 (70%) 39 (66.1%) Higher Education 14 (15.6%) 14 (23.7%) Being employed 0.241 0.623 No 23 (25.6%) 13 (22%) Yes 67 (74.4%) 46 (78%) Smoking 1.382 0.240 No 86 (97.7%) 60 (100%) Yes 2 (2.3%) 0 (0%) Age at menarche (Years) 14 (13–15) 14 (12–15) −0.967 0.334 Prior surgery History of uterine cavity operation 76 (84.4%) 42 (71.2%) 3.802 0.051 Prior uterine surgery 39 (43.3%) 15 (25.4%) 4.947 0.026 c Adnex surgery history 20 (22.2%) 4 (6.8%) 6.289 0.012 Pelvic and Abdominal Surgery 6 (6.7%) 7 (11.9%) 1.209 0.272 Tubal sterilization 11 (12.2%) 5 (8.5%) 0.522 0.470 Parity 1 (1~2) 1 (1~2) −0.662 0.508 With endometriosis 15 (16.7%) - - - a: M(P25–75) means median with interquartile range. b: It means X2 or Z value. c: Bold values indicate p < 0.05. jcm-11-02638-t002_Table 2 Table 2 Clinical symptoms. M(P25–75)/N (%) Cases Controls X2/Z p Pain VAS 9 (6–10) 0 (0–5.2) −7.666 <0.001 Severity 46.026 <0.001 No or mild pain 17 (18.9%) 41 (69.5%) Moderate pain 10 (11.1%) 9 (47.4%) Severe pain 63 (70%) 9 (15.3%) Types of Pain Dysmenorrhea 78 (86.7%) 19 (32.8%) 45.383 <0.001 Chronic pelvic pain 3 (3.3%) 1 (1.7%) 0.347 0.556 Lumbago 34 (38.2%) 6 (10.2%) 14.137 <0.001 Dyspareunia 6 (6.7%) 0 (0%) 4.146 0.042 Perianal pain 8 (9%) 0 (0%) 5.606 0.018 Analgesics 59 (65.6%) 4 (7.0%) 48.830 <0.001 Hypermenorrhea 51 (57.3%) 24 (41.4%) 3.563 0.059 Hemoglobin (g/L) 110 (92–123) 116 (101–128) −1.719 0.086 Mild anemia 24 (27.6%) 12 (20.3%) Moderate anemia 18 (20.7%) 10 (16.9%) Severe anemia 1 (1.1%) 1 (1.7%) Infertility 15 (18.3%) 7 (13.2%) 0.610 0.435 Bold values indicate p < 0.05. jcm-11-02638-t003_Table 3 Table 3 Anxiety and depression in patients with and without adenomyosis. M(P25–75)/N (%) Cases Controls X2/Z p HADS-A a 5 (2–8) 3 (1–6) −2.731 0.006 b HADS-D c 3 (1–6) 1 (0–3) −2.897 0.004 History of anxiety and depression 9 (10%) 5 (8.5%) 0.097 0.755 Anxiety and depression treatment history 4 (4.4%) 1 (1.7%) 0.831 0.362 a: the score of HADS-anxiety. b: Bold values indicate p < 0.05. c: the score of HADS-depression. jcm-11-02638-t004_Table 4 Table 4 SF-36 in patients with and without adenomyosis. Cases Controls Mean M(P25–75) Mean M(P25–75) Z p PCS 64.0 68.125 (49~78.5) 80.2 86 (68.5–91) −4.980 <0.001 PF 89.2 95 (85–100) 93.7 95 (90–100) −1.710 0.087 RP 61.4 75 (0–100) 81.1 100 (75~100) −3.058 0.002 BP 50.2 51 (22–74) 76.9 80 (62~100) −4.960 <0.001 GH 55.1 56 (40–72) 68.9 72 (50–87) −3.469 0.001 MCS 68.1 71.1 (58.9–84.5) 78.0 80.8 (73.8–87.7) −3.068 0.002 VT 67.3 67.5 (55–85) 76.3 80 (65–90) −2.407 0.016 SF 65.3 77.8 (55.6–77.8) 71.8 77.8 (66.7–77.8) −2.290 0.022 RE 64.8 83.3 (33.3–100) 84.7 100 (66.7–100) −3.086 0.002 MH 75.0 80 (60–92) 79.3 84 (72–88) −1.048 0.295 Bold values indicate p < 0.05. PCS: physical component summary. MCS: mental component summary. PF: physiological function. RP: role-physical. BP: bodily pain. GH: general health. VT: vitality. SF: social functioning. RE: role-emotional. MH: mental health. jcm-11-02638-t005_Table 5 Table 5 Relationship between anxiety, depression, and quality of life in patients with adenomyosis. PCS MCS rs p rs p HADS-A −0.455 <0.001 −0.654 <0.001 HADS-D −0.439 <0.001 −0.676 <0.001 PCS: physical component summary. MCS: mental component summary. HADS-A: the Hospital Anxiety and Depression Scale-anxiety scores. HADS-D: the Hospital Anxiety and Depression Scale-depression scores. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Bulun S.E. Yildiz S. Adli M. Wei J.-J. Adenomyosis pathogenesis: Insights from next-generation sequencing Hum. Reprod. Update 2021 27 1086 1097 10.1093/humupd/dmab017 34131719 2. Vercellini P. Viganò P. Somigliana E. Daguati R. Abbiati A. Fedele L. Adenomyosis: Epidemiological factors Best Pract. Res. Clin. Obstet. Gynaecol. 2006 20 465 477 10.1016/j.bpobgyn.2006.01.017 16563868 3. Tellum T. Matic G.V. Dormagen J.B. Nygaard S. Viktil E. Qvigstad E. Lieng M. Diagnosing adenomyosis with MRI: A prospective study revisiting the junctional zone thickness cutoff of 12 mm as a diagnostic marker Eur. Radiol. 2019 29 6971 6981 10.1007/s00330-019-06308-3 31264010 4. Tellum T. Nygaard S. Skovholt E.K. Qvigstad E. Lieng M. Development of a clinical prediction model for diagnosing adenomyosis Fertil. Steril. 2018 110 957 964.e3 10.1016/j.fertnstert.2018.06.009 30316443 5. Karamanidis D. Nicolaou P. Chrysafis I. Georgopoulos A. Tarres E. Koutsougeras G. OC01: Transvaginal ultrasonography compared with magnetic resonance imaging for the diagnosis of adenomyosis Ultrasound Obs. Gynecol. 2018 52 555 10.1002/uog.19201 30284361 6. Yu O. Schulze-Rath R. Grafton J. Hansen K. Scholes D. Reed S.D. Adenomyosis incidence, prevalence and treatment: United States population-based study 2006–2015 Am. J. Obs. Gynecol. 2020 223 e91 e94 10.1016/j.ajog.2020.01.016 31954156 7. Chapron C. Vannuccini S. Santulli P. Abrão M.S. Carmona F. Fraser I.S. Gordts S. Guo S.W. Just P.A. Noël J.C. Diagnosing adenomyosis: An integrated clinical and imaging approach Hum. Reprod. Update 2020 26 392 411 10.1093/humupd/dmz049 32097456 8. Gureje O. Simon G.E. Von Korff M. A cross-national study of the course of persistent pain in primary care Pain 2001 92 195 200 10.1016/S0304-3959(00)00483-8 11323140 9. Hidese S. Saito K. Asano S. Kunugi H. Association between iron-deficiency anemia and depression: A web-based Japanese investigation Psychiatry Clin. Neurosci. 2018 72 513 521 10.1111/pcn.12656 29603506 10. Vulser H. Wiernik E. Hoertel N. Thomas F. Pannier B. Czernichow S. Hanon O. Simon T. Simon J.M. Danchin N. Association between depression and anemia in otherwise healthy adults Acta Psychiatr. Scand. 2016 134 150 160 10.1111/acps.12595 27238642 11. Volgsten H. Skoog Svanberg A. Ekselius L. Lundkvist O. Sundström Poromaa I. Prevalence of psychiatric disorders in infertile women and men undergoing in vitro fertilization treatment Hum. Reprod. 2008 23 2056 2063 10.1093/humrep/den154 18583334 12. Peterson B.D. Sejbaek C.S. Pirritano M. Schmidt L. Are severe depressive symptoms associated with infertility-related distress in individuals and their partners? Hum. Reprod. 2014 29 76 82 10.1093/humrep/det412 24256990 13. Chen Q. Li Y.W. Wang S. Fan Q.B. Shi H.H. Leng J.H. Sun D.W. Lang J.H. Zhu L. Clinical Manifestations Of Adenomyosis Patients With Or Without Pain Symptoms J. Pain Res. 2019 12 3127 3133 10.2147/JPR.S212117 31814754 14. Taran F.A. Weaver A.L. Coddington C.C. Stewart E.A. Understanding adenomyosis: A case control study Fertil. Steril. 2010 94 1223 1228 10.1016/j.fertnstert.2009.06.049 19643403 15. Dworkin R.H. Gitlin M.J. Clinical aspects of depression in chronic pain patients Clin. J. Pain 1991 7 79 94 10.1097/00002508-199106000-00004 1809423 16. Michaelides A. Zis P. Depression, anxiety and acute pain: Links and management challenges Postgrad. Med. 2019 131 438 444 10.1080/00325481.2019.1663705 31482756 17. Craven P. Cinar O. Madsen T. Patient anxiety may influence the efficacy of ED pain management Am. J. Emerg. Med. 2013 31 313 318 10.1016/j.ajem.2012.08.009 22981626 18. Alcalde A.M. Martínez-Zamora M. Gracia M. Ros C. Rius M. Nicolás I. Carmona F. Impact of Adenomyosis on Women’s Psychological Health and Work Productivity: A Comparative Cross-Sectional Study J. Womens Health 2021 30 1653 1659 10.1089/jwh.2020.8789 19. Nelsen L.M. Lenderking W.R. Pokrzywinski R. Balantac Z. Black L. Pokras S. Enslin M.B. Cooper M. Lukes A.S. Experience of Symptoms and Disease Impact in Patients with Adenomyosis Patient 2018 11 319 328 10.1007/s40271-017-0284-2 29197944 20. Lee J.-S. Hong G.-Y. Lee K.-H. Song J.-H. Kim T.-E. Safety and Efficacy of Ultrasound-Guided High-Intensity Focused Ultrasound Treatment for Uterine Fibroids and Adenomyosis Ultrasound Med. Biol. 2019 45 3214 3221 10.1016/j.ultrasmedbio.2019.08.022 31563479 21. Zigmond A.S. Snaith R.P. The hospital anxiety and depression scale Acta Psychiatr. Scand. 1983 67 361 370 10.1111/j.1600-0447.1983.tb09716.x 6880820 22. Smarr K.L. Keefer A.L. Measures of depression and depressive symptoms: Beck Depression Inventory-II (BDI-II), Center for Epidemiologic Studies Depression Scale (CES-D), Geriatric Depression Scale (GDS), Hospital Anxiety and Depression Scale (HADS), and Patient Health Questionnaire-9 (PHQ-9) Arthritis Care Res. 2011 63 (Suppl. 11) S454 S466 10.1002/acr.20556 23. Ware J.E. Sherbourne C.D. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection Med. Care 1992 30 473 483 10.1097/00005650-199206000-00002 1593914 24. Katon W. Lin E.H.B. Kroenke K. The association of depression and anxiety with medical symptom burden in patients with chronic medical illness Gen. Hosp. Psychiatry 2007 29 147 155 10.1016/j.genhosppsych.2006.11.005 17336664 25. Kessler R.C. Bromet E.J. The epidemiology of depression across cultures Annu. Rev. Public Health 2013 34 119 138 10.1146/annurev-publhealth-031912-114409 23514317 26. Yin H. Wardenaar K.J. Xu G. Tian H. Schoevers R.A. Mental health stigma and mental health knowledge in Chinese population: A cross-sectional study BMC Psychiatry 2020 20 323 10.1186/s12888-020-02705-x 32571270 27. Yin A.-Q. Wang F. Zhang X. Integrating endocannabinoid signaling in the regulation of anxiety and depression Acta Pharm. Sin. 2019 40 336 341 10.1038/s41401-018-0051-5 28. Lutz B. Marsicano G. Maldonado R. Hillard C.J. The endocannabinoid system in guarding against fear, anxiety and stress Nat. Rev. Neurosci. 2015 16 705 718 10.1038/nrn4036 26585799 29. Huang W.-J. Chen W.-W. Zhang X. Endocannabinoid system: Role in depression, reward and pain control (Review) Mol. Med. Rep. 2016 14 2899 2903 10.3892/mmr.2016.5585 27484193 30. Shen X. Duan H. Wang S. Gan L. Xu Q. Li J.-J. Decreased Expression of Cannabinoid Receptors in the Eutopic and Ectopic Endometrium of Patients with Adenomyosis BioMed Res. Int. 2019 2019 5468954 10.1155/2019/5468954 30800671 31. Xu X. Cai X. Liu X. Guo S.-W. Possible involvement of neuropeptide and neurotransmitter receptors in Adenomyosis Reprod. Biol. Endocrinol. 2021 19 25 10.1186/s12958-021-00711-6 33602248 32. Pfitzer T. Niederhoffer N. Szabo B. Search for an endogenous cannabinoid-mediated effect in the sympathetic nervous system Naunyn Schmiedebergs Arch. Pharm. 2005 371 9 17 10.1007/s00210-004-1003-9 15660243 33. Casalechi M. Vieira-Lopes M. Quessada M.P. Arão T.C. Reis F.M. Endometriosis and related pelvic pain: Association with stress, anxiety and depressive symptoms Minerva. Obs. Gynecol. 2021 73 283 289 10.23736/S2724-606X.21.04704-3 34008383 34. Lorençatto C. Petta C.A. Navarro M.J. Bahamondes L. Matos A. Depression in women with endometriosis with and without chronic pelvic pain Acta Obs. Gynecol. Scand. 2006 85 88 92 10.1080/00016340500456118 35. Roth R.S. Punch M. Bachman J.E. Psychological factors in chronic pelvic pain due to endometriosis: A comparative study Gynecol. Obs. Investig. 2011 72 15 19 10.1159/000321392 36. Laganà A.S. La Rosa V.L. Rapisarda A.M.C. Valenti G. Sapia F. Chiofalo B. Rossetti D. Ban Frangež H. Vrtačnik Bokal E. Vitale S.G. Anxiety and depression in patients with endometriosis: Impact and management challenges Int. J. Womens Health 2017 9 323 330 10.2147/IJWH.S119729 28553145 37. Richards S.H. Anderson L. Jenkinson C.E. Whalley B. Rees K. Davies P. Bennett P. Liu Z. West R. Thompson D.R. Psychological interventions for coronary heart disease Cochrane Database Syst. Rev. 2017 4 CD002902 10.1002/14651858.CD002902.pub4 28452408 38. Connolly M.J. Yohannes A.M. The impact of depression in older patients with chronic obstructive pulmonary disease and asthma Maturitas 2016 92 9 14 10.1016/j.maturitas.2016.07.005 27621232 39. Zhao L. Wu H. Zhou X. Wang Q. Zhu W. Chen J. Effects of progressive muscular relaxation training on anxiety, depression and quality of life of endometriosis patients under gonadotrophin-releasing hormone agonist therapy Eur. J. Obs. Gynecol. Reprod. Biol. 2012 162 211 215 10.1016/j.ejogrb.2012.02.029 40. Ozdegirmenci O. Kayikcioglu F. Akgul M.A. Kaplan M. Karcaaltincaba M. Haberal A. Akyol M. Comparison of levonorgestrel intrauterine system versus hysterectomy on efficacy and quality of life in patients with adenomyosis Fertil. Steril. 2011 95 497 502 10.1016/j.fertnstert.2010.10.009 21074150 41. Williams V.S. Jones G. Mauskopf J. Spalding J. DuChane J. Uterine fibroids: A review of health-related quality of life assessment J. Womens Health 2006 15 818 829 10.1089/jwh.2006.15.818 16999637 42. Culley L. Law C. Hudson N. Denny E. Mitchell H. Baumgarten M. Raine-Fenning N. The social and psychological impact of endometriosis on women’s lives: A critical narrative review Hum. Reprod. Update 2013 19 625 639 10.1093/humupd/dmt027 23884896 43. Andersson J.K. Khan Z. Weaver A.L. Vaughan L.E. Gemzell-Danielsson K. Stewart E.A. Vaginal bromocriptine improves pain, menstrual bleeding and quality of life in women with adenomyosis: A pilot study Acta Obs. Gynecol. Scand. 2019 98 1341 1350 10.1111/aogs.13632 31025313

PMC009xxxxxx/PMC9099605.txt

==== Front Int J Environ Res Public Health Int J Environ Res Public Health ijerph International Journal of Environmental Research and Public Health 1661-7827 1660-4601 MDPI 10.3390/ijerph19095099 ijerph-19-05099 Review Machine Learning, Deep Learning, and Mathematical Models to Analyze Forecasting and Epidemiology of COVID-19: A Systematic Literature Review https://orcid.org/0000-0003-2924-2030 Saleem Farrukh 1* https://orcid.org/0000-0001-9259-4536 AL-Ghamdi Abdullah Saad AL-Malaise 1 https://orcid.org/0000-0001-9919-8368 Alassafi Madini O. 2 AlGhamdi Saad Abdulla 3 Mollalo Abolfazl Academic Editor Zhou Shang-Ming Academic Editor 1 Department of Information System, Faculty of Computing and Information Technology, King Abdulaziz University, Jeddah 21589, Saudi Arabia; aalmalaise@kau.edu.sa 2 Department of Information Technology, Faculty of Computing and Information Technology, King Abdulaziz University, Jeddah 21589, Saudi Arabia; malasafi@kau.edu.sa 3 Ministry of Health, King Abdulaziz Hospital, Jeddah 22421, Saudi Arabia; s.a.malaise@gmail.com * Correspondence: fsaleem@kau.edu.sa 22 4 2022 5 2022 19 9 509925 2 2022 20 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). COVID-19 is a disease caused by SARS-CoV-2 and has been declared a worldwide pandemic by the World Health Organization due to its rapid spread. Since the first case was identified in Wuhan, China, the battle against this deadly disease started and has disrupted almost every field of life. Medical staff and laboratories are leading from the front, but researchers from various fields and governmental agencies have also proposed healthy ideas to protect each other. In this article, a Systematic Literature Review (SLR) is presented to highlight the latest developments in analyzing the COVID-19 data using machine learning and deep learning algorithms. The number of studies related to Machine Learning (ML), Deep Learning (DL), and mathematical models discussed in this research has shown a significant impact on forecasting and the spread of COVID-19. The results and discussion presented in this study are based on the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. Out of 218 articles selected at the first stage, 57 met the criteria and were included in the review process. The findings are therefore associated with those 57 studies, which recorded that CNN (DL) and SVM (ML) are the most used algorithms for forecasting, classification, and automatic detection. The importance of the compartmental models discussed is that the models are useful for measuring the epidemiological features of COVID-19. Current findings suggest that it will take around 1.7 to 140 days for the epidemic to double in size based on the selected studies. The 12 estimates for the basic reproduction range from 0 to 7.1. The main purpose of this research is to illustrate the use of ML, DL, and mathematical models that can be helpful for the researchers to generate valuable solutions for higher authorities and the healthcare industry to reduce the impact of this epidemic. epidemiology of COVID-19 basic reproduction rate machine learning deep learning ==== Body pmc1. Introduction The outbreak of a deadly disease called coronavirus (COVID-19) has had a significant global impact. As such, the World Health Organization (WHO) has declared it a pandemic [1]. It has affected all spheres of life; moreover, people from poor nations to developed nations are trapped indoors by the pandemic. In this situation, information and communication technologies (ICT) play an important part in connecting communities, implementing the policies, and guiding the communities by analyzing the large datasets generated from COVID-19. Within a few months after the first COVID-19 case was discovered in Wuhan, China, several researchers published articles, discussing this virus and its impact on society [2,3,4,5]. Moreover, the use of computing technologies has generated substantial support to deal with the virus. Current technological developments such as smart applications [6], Artificial Intelligence (AI) [7], Machine Learning (ML) [8], Deep Learning (DL) [9], and big data analytics [10] have led to numerous solutions, epidemiology analysis, and other clinical findings from the collected data sets. These computing technologies are also assisting healthcare and governmental agencies in controlling the spread of the virus, creating social distancing awareness, and predicting potential growth, positive cases, and mortality rates. To understand the current situation, this study mainly focused on reviewing the published papers related to ML and DL techniques. In addition, we integrated some other factors such as epidemiology, reproduction number, and virus doubling time factors in this study, which make it a different SLR than presented previously [11]. Researchers are trying to make good use of the datasets related to COVID-19 patients such as patients’ demographic data, clinical information, chest X-rays (CXR), and Computed Tomography (CT) images. For example, ML techniques assisted in preparing a learning system, and predicting the future concerns about COVID-19, using a training data set to acquire knowledge from the collected dataset [12]. It is also helpful to estimate the future trend and potential infection rate [13]. On the other side, DL implementation is providing more support by predicting the clinical findings using CXR and CT scan images [14,15]. For instance, analyzing medical images can provide irregularities in those images by highlighting different spots and predicting infected and normal patients [16]. Therefore, these computing strategies are assisting medical and governmental agencies to generate multiple findings using COVID-19 dataset, for example, severity detection, virus spreading and control, creating policies and guidelines for the communities, helping in medicine and vaccine development. Previously, computing scholars proposed productive health solutions to deal with different diseases and treatments [17,18,19,20,21,22]. Similarly, integration of the computing and health industries led to ideas for controlling the spread of the virus, suggestions for future virus containment, and pattern identification from real-world data. In addition, the COVID-19 pandemic has also opened many challenges that have ultimately triggered further development and integration of medical and technology fields. Whereas, ML and DL techniques helped to overcome those challenges by providing various solutions to assist the medical industry and higher authorities. This research provides a systematic literature review and analysis of ML, DL, and mathematical models for different purposes such as predicting future cases, analyzing previous infected cases, estimating basic reproduction numbers and virus doubling time. This research discussed the number of developments and solutions provided by multiple scholars around the world. Furthermore, we discussed a number of common datasets, statistical models, and techniques to understand different factors such as infection growth rates, reproduction rates, and doubling time. The main motivation for this paper is to present a comprehensive review for the research and medical community on the current development and future challenges of ML and DL approaches for COVID-19. The summary of ML and DL techniques for prediction, detection, and treatment of COVID-19 are some of the major findings of this study. Overall, this study reviewed selected studies and contributed in the following ways:The main research categories can be identified in this area of study; Review of machine learning and deep learning techniques for understanding previous data and predicting future cases; Review of different mathematical models for time series analysis and estimating epidemiological factors; Identification of validation strategies and evaluation metrics have been used for model performance. Accordingly, the paper is organized as follows: Section 2 discusses the methodology and search strategy applied in this study. Comprehensive analysis of ML, DL, and mathematical models applied on COVID-19 dataset is presented in Section 3. Finally, Section 4 concludes this study by highlighting future work. 2. Methodology and Search Strategy This research is mainly focused on SLR methodology. SLR is a systematic approach to organize, present, and synthesize previously published papers that can help readers to understand the current situation and potential developments in a specific field of research. Therefore, this research identified published papers that describe the COVID-19 epidemiology, use of ML and DL approaches for prediction and identification, basic reproduction rate, and virus doubling time in different regions. The subsequent sections are further describing the step-wise approach used in this article. 2.1. Protocol and Registration The systematic approach used in this study is based on the PRISMA guidelines [23]. The paper title and abstract are written as per the pre-defined guidelines. The review objectives in the introduction section were defined accordingly. The main inclusion and exclusion criteria are also discussed in Section 2.2, whereas the representation of the SLR used in this study is depicted in Figure 1. 2.2. Search Strategy We performed the searching process using different digital libraries, such as: (i) Web of Science; (ii) Scopus; (iii) Google Scholar; and (iv) Medline, up to the beginning of April 2022. This process was mainly applied under the supervision of one researcher and one clinician. Both researchers performed this task together to perform the initial screening process from computing and medical perspectives. At the first step, the following keywords were used: “COVID-19”, “novel coronavirus”, epidemiological features”, “ML or DL model prediction for COVID-19”. An enormous number of articles are available on these databases due to the large interest of researchers in this area of study. Therefore, papers were selected on the bases of explained inclusion and exclusion criteria. In the next step, the papers refined by excluding out of the scope topics, for example, social network analysis, virtual education, or work from home focused papers. 2.3. Inclusion and Exclusion Criteria We included the number of studies using specific inclusion criteria. As this research area has recorded an enormous list of publications, therefore, the inclusion criteria are important to be defined, and are also mentioned in the PRISMA guidelines document. The inclusion criteria were applied as follows: (1) the selected studies should be published in English; (2) the article must have applied and measured any of the epidemiological factors (i.e., size of estimation, epidemic doubling time, basic reproduction number, demographic features, clinical characteristics); and (3) the implementation of a ML or DL approach to identify, analyze previous cases, and predict future rate of infection and recovery. In addition, some articles were excluded due to several reasons as follows: (1) duplicate entities; (2) title, keywords, and abstract screening; (3) non-peer reviewed articles; and (4) opinion or conceptual framework focused articles. 2.4. Identified Research Questions As per the above discussion, this SLR will answer the following research questions:What are the main research categories that can be identified in this area of study? Which machine learning and deep learning techniques were proposed for predicting the future COVID-19 cases? Which mathematical models were used for time series analysis and for calculating different epidemiological factors? What validation strategies and evaluation metrics were used for measuring the model performance? 2.5. Quality Assessment Finally, the quality check process was applied by two researchers to assess the quality of the contents presented in selected studies. The main purpose of this step was to measure the quality of papers and their impact on this SLR. We used eight quality evaluation questions [24] to evaluate each article as follows: (i) objective relevance; (ii) usefulness; (iii) experimental procedure; (iv) model validation and efficiency; (v) dataset importance; (vi) availability of research limitation; (vii) discussion on future aspects; and (viii) presentation of model evaluation metrics. 3. Results and Discussion After reviewing and analyzing the selected case studies, this section describes the major findings and discussion, as presented in different sub-sections. 3.1. Characteristics of Selected Articles The first section elaborates on the major characteristics of reviewed articles. After going through the long procedure, we short-listed 57 studies out of 218 (first search) based on their relevance to the main objectives of this study. Prior to answering the main research questions, the following are some highlights of selected articles. 3.1.1. Journal-Wise Categorization Given the large number of publications in this area of research, the selection process was not basically dependent on journal venue, rather it was based on the inclusion criteria. Therefore, the researchers’ main focus was to include articles on the bases of defined rules without considering the journal venue. However, all searching databases are well-known for academic and applied research publications. Figure 2 illustrates the selected paper’s publishing venues. Most of the selected papers were published in Elsevier (20), which is one of the prominent venues for publishing quality papers. Furthermore, 10 selected articles belong to MDPI, which is one of the largest publishing venues in academic research. In the other category, we put remaining journals such as Frontiers, Wiley, IEEE, and others. 3.1.2. Country-Wise Statistics We usually selected papers that proposed, implemented, and validated the prediction model using ML, DL, mathematical, or regression techniques and applied the model to the real datasets. The population of the selected case studies belonged to 19 different countries, where the COVID-19 dataset had in particular been collected and applied for different purposes, as depicted in Figure 3. Mainly, most of the studies were associated with the population of China (22%), which has been the focal point of this disease. The researchers from that region have published a number of articles related to predicting techniques [25], estimation of disease-related factors [26], and impact of prevention strategies [27]. The number of studies selected from the United States of America (USA) and the Indian regions constituted 15% and 6%, respectively. In addition, we put some studies under the public dataset category. This category represents the used dataset that either belongs to multiple regions or has been collected from an online portal (i.e., Kaggle, GitHub, and others). A large number of countries and real-world data provided a suitable ground to review the current scenario and future aspects in this area of research. 3.2. Research Domain Most of the selected studies applied prediction strategies using different kinds of models. In brief, we avoid putting most of them under the prediction category and presented them in five categories based on the main research questions mentioned in those articles. Table 1 represents the five domains classification of selected articles as follows: (i) Automated Detection; (ii) Estimation of Disease Related Factors; (iii) Impact of Quarantine and Traveling; (iv) Reporting on COVID-19 Numbers; and (v) Virus Reproduction and Doubling Time. For instance, the “Automatic Detection” category combines different prediction models implemented for automating the process of diagnosing and treatment [28]. In addition, the number of studies that belongs to this category are helpful for automatic feature extraction and improving the learning process. For the most part, those articles used CT and CXR images that played a vital role in the early diagnosis and treatment of COVID-19 disease [29]. Furthermore, the category “Estimation of Disease-Related Factors” comprises multiple studies that demonstrated other factors and their correlation with COVID-19 disease. For example, a study defined the prevalence of depression and anxiety and its associated risk factors in the patients already infected by COVID-19 [30]. High temperature & humidity [31], and geo-location [26], are some other external factors used in the selected studies to measure their impact on COVID-19 spread or control. This classification table is useful for the researchers to find a group of research papers associated with the mentioned domain. ijerph-19-05099-t001_Table 1 Table 1 Classification of Selected Research Articles. Research Domain Classification Authors Automatic Detection [15,28,29,32,33,34,35,36,37,38,39,40,41,42,43,44] Estimation of Disease-Related Factors [25,26,30,31,45,46,47,48,49] Impact of Quarantine and Traveling [27,50,51,52,53,54] Reporting on COVID-19 Numbers [55,56,57,58,59,60,61,62,63,64,65,66,67,68,69] Virus Reproduction and Doubling Time [64,65,66,70,71,72,73,74,75,76,77,78] Predicting COVID-19 is handled in different ways and perspectives, from its detection to prevention there are so many areas where researchers have proposed computing solutions. The categories shown in Figure 4 portray the percentage of selected articles in different domains. “Virus Reproduction and Doubling Time” is the third largest category in this SLR and comprises 20% of the 57 articles. These articles reported epidemic doubling time and basic reproduction rate using previous data [73]. Overall, these estimates were useful for governmental authorities to prepare a number of guidelines for breaking the chain of COVID-19 infection. 3.3. Types of Modeling Applied for Modeling COVID-19 Cases The number of research domains discussed above has applied ML, DL, mathematical, or regression models. For the medical image classification task, DL techniques are considered feasible and suitable for automatic feature extraction and finding out the hidden patterns from those images. On the other side, a large number of ML algorithms are applied for the classification, identification, and analyze of COVID-19 cases. Figure 5 represents that 28% of the selected papers applied ML techniques, whereas 36% implemented DL, or other mathematical models, respectively. The mapping of each article with modeling techniques is shown in Table 2. It can be evident from this table that all kinds of models are almost equally important and proposed several solutions while dealing with COVID-19. It summarizes that 21 out of the 57 selected articles used DL approaches, 16 out of the 57 employed ML, and a final 21 articles used other regression or mathematical models. Whilst the regression model is one of the ML techniques, we put regression models in the “Others” category, due to their dynamics, variety, and association with mathematical and statistical approaches. A detailed review of each type of modeling is presented in the subsequent sections. 3.3.1. Machine Learning Models Of the selected studies, 28% of the studies implemented ML techniques to propose learning procedures or to develop prediction models. As shown in Figure 6, over 23% of the articles employed support vector machines (SVM), whereas 17% Decision Trees (DT), 15% Boosting, 12% Naïve Bayes (NB) and Random Forest (RF), 9% Artificial Neural Net (ANN) and K-Nearest Neighbor (KNN), and MLP implemented recorded the lowest %, at 3%. Previous studies highlighted the importance of the ML algorithm for multi-purpose solution building, which was further justified through measured accuracy of the models. For instance, research was applied to the multi-region datasets for (i) predicting the spread of virus in different regions; (ii) virus transmission rate; (iii) ending point; (iv) weather conditions and their association with the virus [55]. Early assessment and identification of COVID-19 is helpful for effective treatment and it can also reduce the healthcare cost. A study used multi-ML models for predicting the infection status in different states of India [67]. Overall, 5004 patients were recorded with a cross-validation approach used for model implementation. For this, the ensemble model proposed using different classifiers such as SVM, DT, and NB. The model outperformed (accuracy: 0.94) as compared to other studies 0.85 [79] and 0.91 [80]. The use of ML approaches for COVID-19 disease recorded several frameworks. One study analyzed the multiple symptoms to identify risk factors for clinical evaluation of COVID-19 patients [49]. The study used 166 patients of different age groups including demographic features, disease history, and other test information. The study applied a multi-model (ANN, SVM, and Boosting) approach, in which ANN outperformed other classifiers with 96% accuracy. Moreover, it is also useful for real-time forecasting purposes, as discussed in a study applied to the time series data collected from Johns Hopkins [56]. The model provided predictions for the next 3 weeks and the results were suitable for the higher authorities to plan resources and prepare policy accordingly. In the same way, another study proposed a model using SVM and DT that forecasted the next six months in Algeria [57]. Scholars suggested ideas to support the government by predicting numbers on potential virus growth using different variables. In the study, factors such as weather, temperature, pollution, gross domestic product, and population density were used to develop a prediction model [25]. The collected dataset was associated with the different states of the USA. SVM, DT, and regression-based models were applied in this study to forecast the spread of the virus. SVM performance showed 95% more variation than other models. The study further suggested that population density can be a critical factor to analyze the size of the spread. The author explored a good factor, but comparing this factor in high and low population regions can provide better results. In addition, the impact of quarantine was measured using data collected from three countries (Italy, South Korea, and the USA) [51]. The study recommended that strict government policies for isolation played a significant role in halting the virus’ spread. The review process in this study identified several facts about ML techniques. According to the studies selected in this paper, the most useful model is SVM, which has been used in 23% of articles. DT (17%) and Boosting (15%) stand in second and third place. Based on the review performed on selected case studies, the ML approach is useful to predict future growth [55], severity detection [47], analyzing CT radiomic features [63], CT images’ classification [37], measuring the impact of social restrictions on virus spread [27], the importance of travel restrictions in reducing virus spread [52], measuring depression and anxiety in COVID-19 infected people [30], and using population density as the main factor for prediction [25]. The model evaluation has shown extraordinary performance in different studies, such as for severity detection (Classifier: SVM, Accuracy: 81%, China) [47], CT images classification accuracy (Classifier: SVM, Accuracy: 99.68%, China) [37] (Classifier SVM, Accuracy: 92.1%, Multi-region) [55], and spatial visualization (Boosting, R2: 0.72, China) [26]. 3.3.2. Deep Learning Models Another major development presented in this SLR study is to review published papers that performed DL techniques to automate the COVID-19 detection process and predict a number of cases. Fast diagnostic methods and deep analysis can help and control COVID-19 spread and that is strongly supported by DL methods. In this SLR, based on the review performed on the selected cases, Figure 7 elaborates on the DL models and the number of times they are used in selected studies. The figure explains the usefulness of the Convolutional Neural Network (CNN) model as it has been used in 10 different articles from the selected studies. Although LSTM is the modified version of Recurrent Neural Network (RNN), to be more specific, we kept them separated and used the same name as mentioned in the studies. Altogether LSTM and RNN were used in nine different articles. The use of a CNN-based deep neural system for medical image classification has been known for its better feature extractions’ capabilities [15,29]. A research team proposed and used 10 different types of CNN-based models to classify the images into infected and non-infected groups [32]. For this, 1020 CT images, and 108 patients’ records were used for the model implementation and validation process. ResNet-101 and Xception showed the best performance with accuracy measured as 99.51% and 99.02%, respectively, although high accuracy could be tested by adding more images from different classes. In addition, research applied the CNN technique to distinguish the infected and non-infected person using their CXR images. For better accuracy and automatic feature detection, transfer learning with CNN approach applied which helped to achieve accuracy, sensitivity, and specificity as 96.78%, 98.66%, and 96.46%, respectively [28]. As per the recommendations collected from different studies, DL approaches could be helpful in several situations. Commonly, different studies used CNN methods to classify CT and CXR images (Classes: COVID-19 infected, viral pneumonia patients, normal patients) [34,36,44], whereas model accuracy recorded more than 90%. In addition, these strategies most of the time used a split validation approach. Another study proposed CNN-based architecture (STM-RENet) to analyze and identify radiographic patterns and textural variations in CXR images of COVID-19 infected people [39]. The proposed model achieved an accuracy of 96.53%, which can be adapted for detecting COVID-19 infected patients. COVID-Net, a CNN-based network system for automation in clinical decisions [35], detection of COVID-19 using SVM classifiers [42], and predicting severe and critical cases based on clinical data of patients using SVM classifiers [33] are some other valuable researches that can provide potential feedback to the medical and higher authorities. The idea of providing a more robust forecast is presented in a research paper with the help of the LSTM framework and mathematical epidemic model [64]. The paper proposed a model that can predict the number of cases on daily bases for the next 15 days with reasonable interpretation. Similarly, another integration was presented using LSTM and Auto-Regressive Integrated Moving Average (ARIMA) techniques, that can forecast for the next 60 days [65]. LSTM has been applied in another study that used time series analysis, evaluated the model, and forecast the number of cases for the next 15 days, applied to the Moscow dataset [59]. The implementation of DL models assisted positively in this epidemic situation to encounter the issues related to automatic infection detection using CT or CXR [43], finding out hidden features [48], forecasting for the next few days [68], and correlating external factors with COVID-19-like social restrictions [27], or spatiotemporal data [50]. According to the selected studies, the range of forecasting provided was from 15 to 60 days. The most common evaluation metrics used were RMSE and MAPE. In addition, for classification tasks the common evaluation metrics used were sensitivity, specificity, and accuracy, which most of the time measured more than 90% [38,41]. 3.3.3. Others (Regression and Mathematical Models) This category combines different mathematical, statistical, regression, and compartmental models that provided a number of solutions in this epidemic situation. These compartmental models use groups of populations and employ mathematical equations using different disease-related factors [24]. These models are also helpful for early prediction, growth rate, number of deaths, and recoveries, which ultimately can provide assistance to higher authorities in controlling the situation. Figure 8 represents the number of models covered in this category and used in selected case studies. Regression analysis (15) is at the top, which has been proved several times to apply time series analysis and forecast for future infections. In addition, the exponential growth model, the SIR Model (Susceptible, Infectious, Recovered), and its extended version such as SEIR (Susceptible, Exposed, Infectious, Recovered), SIRF (Susceptible, Infectious, Recovered, Fatalities), and SIMLR (Susceptible, Infected, Machine Learning, Recovered) are used in selected cases. SIMLR is an extension of the basic epidemiological SIR model that is integrated with the ML approach, applied to track the changes in policies and guidelines applied by governmental authorities [58]. The main purpose of this model was to forecast one to four weeks in advance in Canada and the United States. The results generated and presented a comparison of MAPE in different states. Using a dataset up to July 6, 2021 (India and Israel) the SIRF model was proposed, which extended the basic SIR model by adding fatalities data and can forecast for the next 100 days [60]. In addition, the third extended version found in the selected studies is SEIR, integrating with the “exposed” parameter. This study proposed a simulation-based approach applied to the past 300 days’ data from China to see the impact of prevention strategies [53]. Multiple regression models were applied in a study to predict the number of positive cases in the next few days [25,30]. The idea was to strengthen government policies in order to reduce the number of infected people [66]. For forecasting purposes, the study collected data (22 January 2020, to 12 July 2021), where the study suggested that if the current number of cases are 5000, it can be doubled in the next 5 days. Similarly, the linear regression method was applied to estimate the basic reproduction rate based on the data (1 March–18 May 2020) collected from different regions of the United States [46]. The main idea of this study was to analyze the impact of face-coverings in different states. The result estimated that the total number of infections at the end of May could reach up to 252,000, which shows the positive impact of face coverings. The regression model was applied in different studies and highlighted multiple factors, such as higher temperature, which would help to reduce the transmission rate in China and the USA [31], while the study conducted in Brazil did not support the same idea [45]. Some other time series forecasting models such as FB Prophet applied in Bangladesh (estimation size: 8 March 2020 to 14 October 2021) [61], India and Israel (estimation size: July 6, 2021) [60], ARIMA in China (estimation size: 22 January 2020 to 7 April 2020) [69] are some useful models that can help their country’s representatives to prepare guidelines and prevention strategies. 3.3.4. Model Validation Strategy In this section, we elaborate on the number of validation strategies applied in selected case studies and their ratio, to understand the most favorable validation method in the current situation. As shown in Figure 9, most of the selected studies employed split validation (77%) strategies. One of the reasons behind split validation could be the availability of a smaller number of datasets. As per the importance and quality of cross-validation strategy discussed in previous studies [81], it could be a critical point for the future researchers to: (i) encourage dataset availability on the public platforms; (ii) assess the difference between both validation strategies. 3.3.5. Quality Evaluation Metrics Used in Selected Studies The evaluation metrics allowed researchers to quantify the work presented in any study. It also allowed the author to present the results in an efficient manner. However, the selection of the evaluation metrics is an important aspect, which is based on the type of model employed in that study. The list of quality metrics used for model evaluation in selected studies is depicted in Figure 10. The important thing to mention here, these numbers are not representing the best or worst evaluation metric, they are just presented to highlight the number of potential metrics that could be used, based on the type of forecasting model. Commonly, after reviewing all papers, we can say that growth rate, doubling time, R0, R2, MAPE, MAE, MSE, and RMSE are evaluation metrics that are useful (but not limited to) for time series, regression, compartmental models, or for other mathematical models. The remaining are possible evaluation metrics when we employed other ML or DL methods. 3.4. Epidemiologic Characteristics and Transmission Factors This section describes epidemiological and transmission factors reviewed from the selected case studies. We present the major findings in two sub-sections: (i) Epidemic Doubling Time; and (ii) Basic Reproduction number as presented in subsequent sections. 3.4.1. Estimated Period and Doubling Time The epidemic’s exponential growth within a short period is reported from all over the world. Different studies proposed solutions to reduce, control, and mitigate the impact of COVID-19. The main purpose of those studies was to provide some useful numbers to the higher authorities for preparing controlling strategies as illustrated in Table 3. Therefore, research conducted in India using the data collected from February 2020 to March 2021, estimated that the epidemic doubled in size every 1.7 to 46.2 days. The minimum and maximum numbers were calculated based on the infected cases in different districts [70]. Using linear regression and SVM approaches, an analysis was conducted on multi-region data, where the mathematical model estimated the size on the basis that if the number of positive cases is 5000, it will double in size every 5 days, whereas 163,840,000 cases would be doubled in 140 days. The equation presented multiple scenarios using different datasets, to make the government aware about the severity level of the epidemic [66]. Using a similar strategy (the exponential growth model) estimated the doubling size in China was every 3.6 days [75], whereas another Chinese study concluded that the doubling size was every 4.2 days [74]. The number of studies presented and reviewed in this study conducted in different regions highlighted multiple factors for the governmental agencies. According to the selected cases, the interval for doubling time occurs between 1.7 to 140 days, based on the number of infected people and estimation size. The recommendations list collected from different articles are compiled and presented in the following table. 3.4.2. Basic Reproduction Number (R0) Basic reproduction number estimation plays a significant role and directly impacts different factors such as procedures, guidelines, travel restrictions, quarantine process, and other related factors. Table 4 represents the R0, identified in selected case studies. Generally, a larger reproduction rate would have a large number of infected people in the future. Mainly, the exponential growth model, SIR, ARIMA, and other mathematical models are used for measuring the rate of reproduction number. In addition, the interval of ranges based on the given studies occurs between 0 to 7.1. In which, 0 is the ideal case discussed in the paper related to some districts in India, which recorded less than 40 isolated cases and no local transmission of infection reported [70]. The highest R0 estimate of 7.1 was measured for the New Jersey, USA, in a study published recently [72], which indicates the virus transmission varies in different states. Another recent study used SIR and applied it to the dataset collected from Spain with ranges for R0 from 0.48 to 5.89. As mentioned in the study, the minimum value is clearly identifying the impact of lockdown as the R0 dropped from 5.89 (before lockdown) to 0.48 (after lockdown) [73]. 4. Conclusions As we are aware, the pandemic has had an impact on the entire world. This research discussed the role of ML and DL techniques that can assist medical and governmental agencies. This SLR reviewed a number of papers to identify ML, DL, and mathematical models that can predict the potential impact, transmission growth rate, and virus identification. The research identifies that understanding epidemiology and forecasting models are important to mitigate the impact of this epidemic situation. As for now, the virus transmission is continuing to spread around the world, and the integration of multiple strategies can help to control the situation. In the future, we need to select the most recent papers, while presenting the work using different SLR tools. We discussed a number of key findings that can be helpful for policymakers and future researchers. This type of study should be conducted in the future to understand, analyze, and collect the recent advancement in this area of research. Author Contributions Conceptualization, F.S. and A.S.A.-M.A.-G.; model, F.S.; literature review, M.O.A., S.A.A.; dataset review, M.O.A., S.A.A. and A.S.A.-M.A.-G.; Search strategy, F.S. and A.S.A.-M.A.-G.; ML and DL analysis, F.S. and A.S.A.-M.A.-G.; writing—original draft preparation, F.S., A.S.A.-M.A.-G. and M.O.A.; writing—review and editing, F.S., A.S.A.-M.A.-G. and M.O.A.; visualization, F.S.; supervision, A.S.A.-M.A.-G.; project administration, M.O.A. All authors have read and agreed to the published version of the manuscript. Funding This Project was funded by the Deanship of Scientific Research (DSR) at King Abdulaziz University, Jeddah, under grant no. GCV19-7-1441. The authors, therefore, acknowledge with thanks DSR for technical and financial support. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. In addition, the funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. Figure 1 Study Selection Workflow based on PRISMA. Figure 2 Selected Studies Publishing Journals. Figure 3 Region of Selected Studies. Figure 4 Research Domain Classification. Figure 5 Types of Modeling in Selected Studies. Figure 6 Ratio of ML Models in Selected Studies. Figure 7 Ratio of DL Models in Selected Studies. Figure 8 Ratio of Mathematical and Regression Models in Selected Studies. Figure 9 Ratio of Validation Strategies in Selected Studies. Figure 10 List of Evaluation Metrics used in the Selected Studies. ijerph-19-05099-t002_Table 2 Table 2 Number of Articles and Types of Modeling in Selected Studies. Types of Modeling Authors Deep Learning Models [15,27,28,29,32,33,34,35,36,38,39,41,42,43,44,48,50,59,64,65,68] Machine Learning Models [25,26,27,30,37,47,49,51,52,54,55,56,57,63,67,72] Others (Regression and Mathematical Models) [25,30,31,45,46,53,58,60,61,62,64,66,69,70,71,73,74,75,76,77,78] ijerph-19-05099-t003_Table 3 Table 3 Epidemic Doubling Time in Selected Studies. Author Country Method Dataset Doubling Time Tool Used Recommendation by Author [70] India Exponential Growth Model February 2020–March 2021 1.7 to 46.2 days (based on districts) Q-GIS software no uniformity across country to analyze and study epidemics in future [74] China Global Epidemic and Mobility Model (GLEAM) By 23 January 2020 4.2 days - travel restrictions [66] Multi-Countries Linear Regression and Support Vector Machine 22 January 2020, to 12 July 2021 Min = if (5000 cases) double in 5 days Max = if (163,840,000 cases) double in 140 days - government and individuals aware about the severity [75] China Exponential Growth Model 1–23 January 2020 3.6 days - prevention measures were effective [76] South Africa Susceptible–Exposed– Infectious–Recovered (SEIR) model By 23 November 2021 3.3 days - immune evasion is more concerning increased transmissibility [78] Argentina Agent-based Model Multiple Scenario 2.0 to 7.14 days social distancing measures ijerph-19-05099-t004_Table 4 Table 4 Epidemic Basic Reproduction Number in Selected Studies. Author Country Dataset Basic Reproduction Number Method Confidence Interval (CI) Tool Used [71] China 1–15 January 2020 2.56 Exponential Growth Model 95% CI - [70] India February 2020–March 2021 0 to > 7 (based on district) Exponential Growth Model - Q-GIS software [72] USA 21 January 2020–21 June 2020 2.3 to 7.1 (based on different states) Bayesian inference 95% CI PyBioNetFit [73] Spain March–April 2020 0.48 to 5.89 (different conditions) SIR (Susceptible-Infected-Recovered) 95% CI - [64] USA 22 January 2020–10 August 2020 2.747 to 3.856 (increase as days increase) Mathematical Epidemic Model (MEM) + DL - MATLAB [65] Morocco 22 January 2020–22 November 2020 0.9 and 1.3 (increase as days increase) Auto-Regressive Integrated Moving Average (ARIMA) and Long short-term memory (LSTM) 95% CI Python [74] China By 23 January 2020 2.57 Global Epidemic and Mobility Model (GLEAM) 90% CI - [75] China 1–23 January 2020 4.2 Exponential Growth Model 95% CI - [77] Malaysia 1 February 2020–8 November 2020 3.96 Susceptible-Exposed-Infectious-Removed (SEIR) Model 95% CI Excel [31] China, USA By 10 February 2020 0.023 (China) 0.020 (USA) Retrospective Regression Analysis 95% CI Python [46] USA 8 March–12 April 3.96 Linear Regression 95% CI - [25] USA By 16 April 2020 3.81 to 4.07 (based on method) SIR (Susceptible-Infected-Recovered) 95% CI - Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. WHO WHO Director-General’s Opening Remarks at the Media Briefing on COVID-19. 11 March 2020 Available online: https://www.who.int/dg/speeches/detail/who-director-general-s-opening-remarks-at-the-media-briefing-on-COVID-19---11-march-2020 (accessed on 15 August 2020) 2. The Novel Coronavirus Pneumonia Emergency Response Epidemiology Team The epidemiological characteristics of an outbreak of 2019 novel coronavirus diseases (COVID-19) in China Zhonghua Liu Xing Bing Xue Za Zhi 2020 41 145 32064853 3. Wang D. Yin Y. Hu C. Liu X. Zhang X. Zhou S. Jian M. Xu H. Prowle J. Hu B. Clinical course and outcome of 107 patients infected with the novel coronavirus, SARS-CoV-2, discharged from two hospitals in Wuhan, China Crit. Care 2020 24 188 10.1186/s13054-020-02895-6 32354360 4. Yang X. Yu Y. Xu J. Shu H. Liu H. Wu Y. Zhang L. Yu Z. Fang M. Yu T. Clinical course and outcomes of critically ill patients with SARS-CoV-2 pneumonia in Wuhan, China: A single-centered, retrospective, observational study Lancet Respir. Med. 2020 8 475 481 10.1016/S2213-2600(20)30079-5 32105632 5. Xu J. Yang X. Yang L. Zou X. Wang Y. Wu Y. Zhou T. Yuan Y. Qi H. Fu S. Clinical course and predictors of 60-day mortality in 239 critically ill patients with COVID-19: A multicenter retrospective study from Wuhan, China Crit. Care 2020 24 394 10.1186/s13054-020-03098-9 32631393 6. Alshaikh K. Maasher S. Bayazed A. Saleem F. Badri S. Fakieh B. Impact of COVID-19 on the Educational Process in Saudi Arabia: A Technology–Organization–Environment Framework Sustainability 2021 13 7103 10.3390/su13137103 7. Li L. Qin L. Xu Z. Yin Y. Wang X. Kong B. Bai J. Lu Y. Fang Z. Song Q. Artificial intelligence distinguishes COVID-19 from community acquired pneumonia on chest CT Radiology 2020 296 E65 E71 10.1148/radiol.2020200905 32191588 8. Debnath S. Barnaby D.P. Coppa K. Makhnevich A. Kim E.J. Chatterjee S. Tóth V. Levy T.J. Paradis M.D. Cohen S.L. Machine learning to assist clinical decision-making during the COVID-19 pandemic Bioelectron. Med. 2020 6 14 10.1186/s42234-020-00050-8 32665967 9. Zeroual A. Harrou F. Dairi A. Sun Y. Deep learning methods for forecasting COVID-19 time-Series data: A Comparative study Chaos Solitons Fractals 2020 140 110121 10.1016/j.chaos.2020.110121 32834633 10. Wang C.J. Ng C.Y. Brook R.H. Response to COVID-19 in Taiwan: Big data analytics, new technology, and proactive testing JAMA 2020 323 1341 1342 10.1001/jama.2020.3151 32125371 11. Alyasseri Z.A.A. Al-Betar M.A. Doush I.A. Awadallah M.A. Abasi A.K. Makhadmeh S.N. Alomari O.A. Abdulkareem K.H. Adam A. Damasevicius R. Review on COVID-19 diagnosis models based on machine learning and deep learning approaches Expert Syst. 2022 39 e12759 10.1111/exsy.12759 12. Ardabili S.F. Mosavi A. Ghamisi P. Ferdinand F. Varkonyi-Koczy A.R. Reuter U. Rabczuk T. Atkinson P.M. COVID-19 outbreak prediction with machine learning Algorithms 2020 13 249 10.3390/a13100249 13. Li Q. Feng W. Quan Y.-H. Trend and forecasting of the COVID-19 outbreak in China J. Infect. 2020 80 469 496 32113991 14. Huang L. Han R. Ai T. Yu P. Kang H. Tao Q. Xia L. Serial quantitative chest ct assessment of COVID-19: Deep-learning approach Radiol. Cardiothorac. Imaging 2020 2 e200075 10.1148/ryct.2020200075 33778562 15. Ismael A.M. Şengür A. Deep learning approaches for COVID-19 detection based on chest X-ray images Expert Syst. Appl. 2020 164 114054 10.1016/j.eswa.2020.114054 33013005 16. Muhammad L.J. Algehyne E.A. Usman S.S. Mohammed I.A. Abdulkadir A. Jibrin M.B. Malgwi Y.M. Deep Learning Models for Predicting COVID-19 Using Chest X-Ray Images Trends and Advancements of Image Processing and Its Applications Springer Berlin/Heidelberg, Germany 2022 127 144 17. Yu K.-H. Beam A.L. Kohane I.S. Artificial intelligence in healthcare Nat. Biomed. Eng. 2018 2 719 731 10.1038/s41551-018-0305-z 31015651 18. Seto E. Leonard K.J. Cafazzo J.A. Barnsley J. Masino C. Ross H.J. Developing healthcare rule-based expert systems: Case study of a heart failure telemonitoring system Int. J. Med. Inform. 2012 81 556 565 22465288 19. Cresswell K. Cunningham-Burley S. Sheikh A. Health care robotics: Qualitative exploration of key challenges and future directions J. Med. Internet Res. 2018 20 e10410 10.2196/10410 29973336 20. Bengtsson E. Computerized cell image processing in healthcare Proceedings of the 7th International Workshop on Enterprise networking and Computing in Healthcare Industry, 2005. HEALTHCOM 2005 Busan, Korea 23–25 June 2005 11 17 21. Esteva A. Robicquet A. Ramsundar B. Kuleshov V. DePristo M. Chou K. Cui C. Corrado G. Thrun S. Dean J. A guide to deep learning in healthcare Nat. Med. 2019 25 24 29 10.1038/s41591-018-0316-z 30617335 22. Gerup J. Soerensen C.B. Dieckmann P. Augmented reality and mixed reality for healthcare education beyond surgery: An integrative review Int. J. Med. Educ. 2020 11 1 18 10.5116/ijme.5e01.eb1a 31955150 23. Sarkis-Onofre R. Catalá-López F. Aromataris E. Lockwood C. How to properly use the PRISMA Statement Syst. Rev. 2021 10 117 10.1186/s13643-021-01671-z 33875004 24. Hassan A. Prasad D. Rani S. Alhassan M. Gauging the Impact of Artificial Intelligence and Mathematical Modeling in Response to the COVID-19 Pandemic: A Systematic Review BioMed Res. Int. 2022 2022 7731618 10.1155/2022/7731618 35309167 25. Gupta A. Gharehgozli A. Developing a Machine Learning Framework to Determine the Spread of COVID-19. SSRN 2020 Available online: https://ssrn.com/abstract=3635211 (accessed on 20 March 2022) 26. Wu C. Zhou M. Liu P. Yang M. Analyzing COVID-19 using multisource data: An integrated approach of visualization, spatial regression, and machine learning GeoHealth 2021 5 e2021GH000439 10.1029/2021GH000439 34377880 27. Naemi M. Naemi A. Ekbatani R.Z. Ebrahimi A. Schmidt T. Wiil U.K. Modeling and Evaluating the Impact of Social Restrictions on the Spread of COVID-19 Using Machine Learning Smart and Sustainable Technology for Resilient Cities and Communities Springer Berlin/Heidelberg, Germany 2022 107 118 28. Apostolopoulos I.D. Mpesiana T.A. Covid-19: Automatic detection from x-ray images utilizing transfer learning with convolutional neural networks Phys. Eng. Sci. Med. 2020 43 635 640 10.1007/s13246-020-00865-4 32524445 29. Ozturk T. Talo M. Yildirim E.A. Baloglu U.B. Yildirim O. Acharya U.R. Automated detection of COVID-19 cases using deep neural networks with X-ray images Comput. Biol. Med. 2020 121 103792 10.1016/j.compbiomed.2020.103792 32568675 30. Nie X.-D. Wang Q. Wang M.-N. Zhao S. Liu L. Zhu Y.-L. Chen H. Anxiety and depression and its correlates in patients with coronavirus disease 2019 in Wuhan Int. J. Psychiatry Clin. Pract. 2021 25 109 114 10.1080/13651501.2020.1791345 32662692 31. Wang J. Tang K. Feng K. Lv W. High temperature and high humidity reduce the transmission of COVID-19 BMJ Open 2020 Available online: https://www.scienceopen.com/document_file/ff8b579c-26ff-4c3c-9aa5-a141bc8e35f6/PubMedCentral/ff8b579c-26ff-4c3c-9aa5-a141bc8e35f6.pdf (accessed on 15 January 2022) 10.2139/ssrn.3551767 32. Ardakani A.A. Kanafi A.R. Acharya U.R. Khadem N. Mohammadi A. Application of deep learning technique to manage COVID-19 in routine clinical practice using CT images: Results of 10 convolutional neural networks Comput. Biol. Med. 2020 121 103795 10.1016/j.compbiomed.2020.103795 32568676 33. Sun L. Liu G. Song F. Shi N. Liu F. Li S. Li P. Zhang W. Jiang X. Zhang Y. Combination of four clinical indicators predicts the severe/critical symptom of patients infected COVID-19 J. Clin. Virol. 2020 128 104431 10.1016/j.jcv.2020.104431 32442756 34. Song Y. Zheng S. Li L. Zhang X. Zhang X. Huang Z. Chen J. Wang R. Zhao H. Chong Y. Deep Learning Enables Accurate Diagnosis of Novel Coronavirus (COVID-19) with CT Images IEEE/ACM Trans. Comput. Biol. Bioinform. 2021 18 2775 2780 10.1109/TCBB.2021.3065361 33705321 35. Wang L. Lin Z.Q. Wong A. COVID-Net: A tailored deep convolutional neural network design for detection of COVID-19 cases from chest X-ray images Sci. Rep. 2020 10 19549 10.1038/s41598-020-76550-z 33177550 36. Xu X. Jiang X. Ma C. Du P. Li X. Lv S. Yu L. Ni Q. Chen Y. Su J. A deep learning system to screen novel coronavirus disease 2019 pneumonia Engineering 2020 6 1122 1129 10.1016/j.eng.2020.04.010 32837749 37. Barstugan M. Ozkaya U. Ozturk S. Coronavirus (COVID-19) classification using ct images by machine learning methods arXiv 2020 2003.09424 38. Al Rahhal M.M. Bazi Y. Jomaa R.M. AlShibli A. Alajlan N. Mekhalfi M.L. Melgani F. COVID-19 Detection in CT/X-ray Imagery Using Vision Transformers J. Pers. Med. 2022 12 310 10.3390/jpm12020310 35207797 39. Khan S.H. Sohail A. Khan A. Lee Y.-S. COVID-19 detection in chest X-ray images using a new channel boosted CNN Diagnostics 2022 12 267 10.3390/diagnostics12020267 35204358 40. Sarki R. Ahmed K. Wang H. Zhang Y. Wang K. Automated Detection of COVID-19 through Convolutional Neural Network using Chest x-ray images PLoS ONE 2022 17 e0262052 10.1371/journal.pone.0262052 35061767 41. Mousavi Z. Shahini N. Sheykhivand S. Mojtahedi S. Arshadi A. COVID-19 detection using chest X-ray images based on a developed deep neural network SLAS Technol. 2022 27 63 75 10.1016/j.slast.2021.10.011 35058196 42. Sethy P.K. Behera S.K. Ratha P.K. Biswas P. Detection of Coronavirus Disease (COVID-19) Based on Deep Features and Support Vector Machine 2020 Available online: https://pdfs.semanticscholar.org/9da0/35f1d7372cfe52167ff301bc12d5f415caf1.pdf (accessed on 10 January 2022) 43. Nour M. Cömert Z. Polat K. A novel medical diagnosis model for COVID-19 infection detection based on deep features and Bayesian optimization Appl. Soft Comput. 2020 97 106580 10.1016/j.asoc.2020.106580 32837453 44. Aggarwal S. Gupta S. Alhudhaif A. Koundal D. Gupta R. Polat K. Automated COVID-19 detection in chest X-ray images using fine-tuned deep learning architectures Expert Syst. 2022 39 e12749 10.1111/exsy.12749 45. Prata D.N. Rodrigues W. Bermejo P.H. Temperature significantly changes COVID-19 transmission in (sub) tropical cities of Brazil Sci. Total Environ. 2020 729 138862 10.1016/j.scitotenv.2020.138862 32361443 46. Li Y. Zhang R. Zhao J. Molina M.J. Understanding transmission and intervention for the COVID-19 pandemic in the United States Sci. Total Environ. 2020 748 141560 10.1016/j.scitotenv.2020.141560 32798884 47. Yao H. Zhang N. Zhang R. Duan M. Xie T. Pan J. Peng E. Huang J. Zhang Y. Xu X. Severity detection for the coronavirus disease 2019 (COVID-19) patients using a machine learning model based on the blood and urine tests Front. Cell Dev. Biol. 2020 8 683 10.3389/fcell.2020.00683 32850809 48. Loey M. Manogaran G. Taha M.H.N. Khalifa N.E.M. A hybrid deep transfer learning model with machine learning methods for face mask detection in the era of the COVID-19 pandemic Measurement 2021 167 108288 10.1016/j.measurement.2020.108288 32834324 49. Kocadagli O. Baygul A. Gokmen N. Incir S. Aktan C. Clinical prognosis evaluation of COVID-19 patients: An interpretable hybrid machine learning approach Curr. Res. Transl. Med. 2022 70 103319 10.1016/j.retram.2021.103319 34768217 50. Wang L. Xu T. Stoecker T. Stoecker H. Jiang Y. Zhou K. Machine learning spatio-temporal epidemiological model to evaluate Germany-county-level COVID-19 risk Mach. Learn. Sci. Technol. 2021 2 35031 10.1088/2632-2153/ac0314 51. Dandekar R. Barbastathis G. Quantifying the effect of quarantine control in COVID-19 infectious spread using machine learning medRxiv 2020 1 13 10.1101/2020.04.03.20052084 52. Ogundokun R.O. Lukman A.F. Kibria G.B.M. Awotunde J.B. Aladeitan B.B. Predictive modelling of COVID-19 confirmed cases in Nigeria Infect. Dis. Model. 2020 5 543 548 10.1016/j.idm.2020.08.003 32835145 53. Zou Y. Yang W. Lai J. Hou J. Lin W. Vaccination and Quarantine Effect on COVID-19 Transmission Dynamics Incorporating Chinese-Spring-Festival Travel Rush: Modeling and Simulations Bull. Math. Biol. 2022 84 30 10.1007/s11538-021-00958-5 35006388 54. Andariesta D.T. Wasesa M. Machine learning models for predicting international tourist arrivals in Indonesia during the COVID-19 pandemic: A multisource Internet data approach J. Tour. Futures 2022 1 17 10.1108/JTF-10-2021-0239 55. Yadav M. Perumal M. Srinivas M. Analysis on novel coronavirus (COVID-19) using machine learning methods Chaos Solitons Fractals 2020 139 110050 10.1016/j.chaos.2020.110050 32834604 56. Singh V. Poonia R.C. Kumar S. Dass P. Agarwal P. Bhatnagar V. Raja L. Prediction of COVID-19 corona virus pandemic based on time series data using Support Vector Machine J. Discrete Math. Sci. Cryptogr. 2020 23 1583 1597 10.1080/09720529.2020.1784535 57. Lounis M. Khan F.M. Predicting COVID-19 cases, deaths and recoveries using machine learning methods Eng. Appl. Sci. Lett. 2021 4 43 49 10.30538/psrp-easl2021.0079 58. Vega R. Flores L. Greiner R. SIMLR: Machine Learning inside the SIR model for COVID-19 Forecasting Forecasting 2022 4 72 94 10.3390/forecast4010005 59. Pavlyutin M. Samoyavcheva M. Kochkarov R. Pleshakova E. Korchagin S. Gataullin T. Nikitin P. Hidirova M. COVID-19 Spread Forecasting, Mathematical Methods vs. Machine Learning, Moscow Case Mathematics 2022 10 195 10.3390/math10020195 60. Shwetha S. Sunagar P. Rajarajeswari S. Kanavalli A. Ensemble Model to Forecast the End of the COVID-19 Pandemic Proceedings of the Third International Conference on Communication, Computing and Electronics Systems Coimbatore, India 28–29 October 2021 Springer Singapore 2022 815 829 61. Babu M.A. Ahmmed M.M. Ferdousi A. Mostafizur Rahman M. Saiduzzaman M. Bhatnagar V. Raja L. Poonia R.C. The mathematical and machine learning models to forecast the COVID-19 outbreaks in Bangladesh J. Interdiscip. Math. 2022 1 20 10.1080/09720502.2021.2015095 62. Krivorotko O. Sosnovskaia M. Vashchenko I. Kerr C. Lesnic D. Agent-based modeling of COVID-19 outbreaks for New York state and UK: Parameter identification algorithm Infect. Dis. Model. 2022 7 30 44 10.1016/j.idm.2021.11.004 34869960 63. Shiri I. Salimi Y. Pakbin M. Hajianfar G. Avval A.H. Sanaat A. Mostafaei S. Akhavanallaf A. Saberi A. Mansouri Z. COVID-19 prognostic modeling using CT radiomic features and machine learning algorithms: Analysis of a multi-institutional dataset of 14,339 patients Comput. Biol. Med. 2022 145 105467 10.1016/j.compbiomed.2022.105467 35378436 64. Masum M. Masud M.A. Adnan M.I. Shahriar H. Kim S. Comparative study of a mathematical epidemic model, statistical modeling, and deep learning for COVID-19 forecasting and management Socio-Econ. Plan. Sci. 2022 80 101249 10.1016/j.seps.2022.101249 65. Rguibi M.A. Moussa N. Madani A. Aaroud A. Zine-Dine K. Forecasting COVID-19 Transmission with ARIMA and LSTM Techniques in Morocco SN Comput. Sci. 2022 3 133 10.1007/s42979-022-01019-x 35043096 66. Khan M.A. Khan R. Algarni F. Kumar I. Choudhary A. Srivastava A. Zhao 2020 Ain Shams Eng. J. 2022 13 101574 10.1016/j.asej.2021.08.016 67. Guleria P. Ahmed S. Alhumam A. Srinivasu P.N. Empirical Study on Classifiers for Earlier Prediction of COVID-19 Infection Cure and Death Rate in the Indian States Healthcare 2022 10 85 10.3390/healthcare10010085 35052249 68. Ayris D. Imtiaz M. Horbury K. Williams B. Blackney M. See C.S.H. Shah S.A.A. Novel Deep Learning Approach to Model and Predict the spread of COVID-19 Intell. Syst. Appl. 2022 14 200068 10.1016/j.iswa.2022.200068 69. Chyon F.A. Suman M.N.H. Fahim M.R.I. Ahmmed M.S. Time series analysis and predicting COVID-19 affected patients by ARIMA model using machine learning J. Virol. Methods 2022 301 114433 10.1016/j.jviromet.2021.114433 34919977 70. Shil P. Atre N.M. Patil A.A. Tandale B.V. Abraham P. District-wise estimation of Basic reproduction number (R0) for COVID-19 in India in the initial phase Spat. Inf. Res. 2022 30 37 45 10.1007/s41324-021-00412-7 71. Zhao S. Musa S.S. Lin Q. Ran J. Yang G. Wang W. Lou Y. Yang L. Gao D. He D. Estimating the unreported number of novel coronavirus (2019-nCoV) cases in China in the first half of January 2020: A data-driven modelling analysis of the early outbreak J. Clin. Med. 2020 9 388 10.3390/jcm9020388 32024089 72. Mallela A. Neumann J. Miller E.F. Chen Y. Posner R.G. Lin Y.T. Hlavacek W.S. Bayesian Inference of State-Level COVID-19 Basic Reproduction Numbers across the United States Viruses 2022 14 157 10.3390/v14010157 35062361 73. Hyafil A. Moriña D. Analysis of the impact of lockdown on the reproduction number of the SARS-Cov-2 in Spain Gac. Sanit. 2022 35 453 458 10.1016/j.gaceta.2020.05.003 32571528 74. Chinazzi M. Davis J.T. Ajelli M. Gioannini C. Litvinova M. Merler S. Pastore y Piontti A. Mu K. Rossi L. Sun K. The effect of travel restrictions on the spread of the 2019 novel coronavirus (COVID-19) outbreak Science 2020 368 395 400 10.1126/science.aba9757 32144116 75. Zhao Q. Chen Y. Small D.S. Analysis of the epidemic growth of the early 2019-nCoV outbreak using internationally confirmed cases MedRxiv 2020 1 10 10.1101/2020.02.06.20020941 76. Grabowski F. Kochańczyk M. Lipniacki T. The spread of SARS-CoV-2 variant Omicron with a doubling time of 2.0–3.3 days can be explained by immune evasion Viruses 2022 14 294 10.3390/v14020294 35215887 77. Herng L.C. Singh S. Sundram B.M. Zamri A.S.S.M. Vei T.C. Aris T. Ibrahim H. Abdullah N.H. Dass S.C. Gill B.S. The effects of super spreading events and movement control measures on the COVID-19 pandemic in Malaysia Sci. Rep. 2022 12 2197 10.1038/s41598-022-06341-1 35140319 78. Simoy M.I. Aparicio J.P. Socially structured model for COVID-19 pandemic: Design and evaluation of control measures Comput. Appl. Math. 2021 41 1 23 10.1007/s40314-021-01705-1 79. Agbelusi O. Olayemi O.C. Prediction of mortality rate of COVID-19 patients using machine learning techniques in nigeria Int. J. Comput. Sci. Softw. Eng. 2020 9 30 34 80. An C. Lim H. Kim D.-W. Chang J.H. Choi Y.J. Kim S.W. Machine learning prediction for mortality of patients diagnosed with COVID-19: A nationwide Korean cohort study Sci. Rep. 2020 10 18716 10.1038/s41598-020-75767-2 33127965 81. Yadav S. Shukla S. Analysis of k-fold cross-validation over hold-out validation on colossal datasets for quality classification Proceedings of the 2016 IEEE 6th International Conference on Advanced Computing (IACC) Bhimavaram, India 27–28 February 2016 78 83

PMC009xxxxxx/PMC9099606.txt

==== Front Sensors (Basel) Sensors (Basel) sensors Sensors (Basel, Switzerland) 1424-8220 MDPI 10.3390/s22093188 sensors-22-03188 Article Implementing a Statistical Parametric Speech Synthesis System for a Patient with Laryngeal Cancer https://orcid.org/0000-0001-6540-1671 Szklanny Krzysztof * Lachowicz Jakub Woo Wai Lok Academic Editor Multimedia Department, Polish-Japanese Academy of Information Technology, 02-008 Warsaw, Poland; s12054@pjwstk.edu.pl * Correspondence: kszklanny@pjwstk.edu.pl 21 4 2022 5 2022 22 9 318823 2 2022 13 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Total laryngectomy, i.e., the surgical removal of the larynx, has a profound influence on a patient’s quality of life. The procedure results in a loss of natural voice, which in effect constitutes a significant socio-psychological problem for the patient. The main aim of the study was to develop a statistical parametric speech synthesis system for a patient with laryngeal cancer, on the basis of the patient’s speech samples recorded shortly before the surgery and to check if it was possible to generate speech quality close to that of the original recordings. The recording made use of a representative corpus of the Polish language, consisting of 2150 sentences. The recorded voice proved to indicate dysphonia, which was confirmed by the auditory-perceptual RBH scale (roughness, breathiness, hoarseness) and by acoustical analysis using AVQI (The Acoustic Voice Quality Index). The speech synthesis model was trained using the Merlin repository. Twenty-five experts participated in the MUSHRA listening tests, rating the synthetic voice at 69.4 in terms of the professional voice-over talent recording, on a 0–100 scale, which is a very good result. The authors compared the quality of the synthetic voice to another model of synthetic speech trained with the same corpus, but where a voice-over talent provided the recorded speech samples. The same experts rated the voice at 63.63, which means the patient’s synthetic voice with laryngeal cancer obtained a higher score than that of the talent-voice recordings. As such, the method enabled for the creation of a statistical parametric speech synthesizer for patients awaiting total laryngectomy. As a result, the solution would improve the quality of life as well as better mental wellbeing of the patient. speech synthesis parametrical synthesis deep neural networks laryngeal cancer ==== Body pmc1. Introduction The larynx is the most common localization of malignant head and neck cancers. In Poland, laryngeal cancer accounts for 2.3% of all cancers in men and 0.5% cancers in women [1,2,3,4]. Symptoms of laryngeal cancer include persistent hoarseness, globus sensation, a sore throat, an earache, a cough or weight loss. The risk factors include alcohol consumption, smoking, HPV-16 infection, reflux and exposure to toxic fumes of nickel compounds, sulfuric acid, asbestos or heavy metals [5,6,7]. HPV-16 (human papilloma virus) infection can lead to uncontrolled cell divisions of the cervical epithelium, which can end in cervical cancer [8,9]. In its initial stage, laryngeal cancer may not display clear symptoms, which can lead to a late diagnosis and, consequently, to a more aggressive treatment: surgery and/or chemotherapy and/or radiotherapy [1,6,9]. While early, locally advanced cancer can be treated effectively, for instance by means of microsurgery, but more advanced laryngeal cancer may require a complete removal of the larynx (total laryngectomy) [9]. This will always have a profound impact on the patient’s quality of life, as the loss of natural voice constitutes a significant socio-psychological problem for patients. Regrettably, in many cases, this often leads to a patient’s social isolation and depression [9,10,11,12]. There are three methods of voice restoration following laryngectomy [13]. The first involves the implantation of an artificial larynx. Thanks to the implant, the air can be directed from the lungs to the esophagus in order to create the primary laryngeal tone [14]. In order to be able to speak, the patient has to close off the trach tube opening, which is a major inconvenience. However, patients recover their voice fairly quickly, usually within several days. This kind of speech is known as tracheoesophageal speech (TE) [15,16]. Another method of voice recovery involves the learning of esophageal speech (ES) [17]. It requires the patient to learn to burp out the air returning from the stomach or esophagus. This is far more difficult to learn, while patients often feel uneasy about burping, as it is thought to be rude. Statistically, 40% of all patients manage to master this method, but merely 15% of them actually make use of it [11,18]. The third method involves the use of an electrolarynx [19], a device that generates the fundamental frequency when held against the neck. The generated voice sounds artificial and flat, similar in quality to that of formant synthesis (defined below). Clearly then, there is a need to create augmentative and alternative communication methods, allowing those who cannot produce speech, or have a limited ability to produce speech, to communicate. These include sign language as well as voice output communication aids (VOCAs) [20]. There are several types of speech synthesis used in the VOCA systems, such as formant synthesis, concatenation synthesis, unit selection speech synthesis, and statistical parametric speech synthesis based on the hidden Markov model [21]. The concept of a digital formant speech synthesizer was introduced by Dennis Klatt in 1979 [22]. This kind of synthesis involves using cascade and/or parallel digital filters to model the vocal tract transfer function in the frequency domain. The sound generated in this way has a characteristic tone quality, reproducing the typical formants of speech sounds. Generating intelligible speech requires the reproduction of three formants. Five formants make it possible to generate speech of sufficiently high quality. Each of the formants is modelled with a formant frequency and a resonance band [23]. Models for concatenative speech synthesis, developed since the 1970s, have gained considerable popularity due to their ability to generate high-quality natural-sounding speech. In concatenative synthesis, speech is generated by concatenating acoustic segments, such as phones, diphones, triphones and syllables [24]. Thanks to its sound-to-sound transition characteristic, the diphone is the most common unit which ensures high-quality natural speech. The small size of its database is an advantage of this type of synthesis. The smaller the database, the better, as speech will be generated more quickly, and the hardware requirements will be less demanding [25]. Rather than having a database containing a single occurrence of a given sound unit, unit selection (corpus-based) speech synthesis relies on a special corpus that comprises a number of its occurrences in different contexts, making use of units of varying duration. Owing to this, it is often possible to avoid artificial concatenation points, allowing for more natural-sounding speech [26]. The most important element responsible for the acoustic segment selection is the cost function. It consists of target cost and a concatenation cost (joint cost). The concatenation cost is used to assess the degree to which two units match if they are not in adjacent positions in the acoustic database. The unit selection cost searches out units that will most closely match the linguistic features of the target sentence [27,28]. The HMM-based speech synthesis system (HSS) utilizes the hidden Markov models (HMMs) [28]. In a way, it is similar to concatenation. However, in this case, instead of using segments of natural speech, the synthesis process relies on context-dependent HMMs. These models are concatenated according to the text to be synthesized, and the resultant feature vectors (observations) serve as a basis for the speech synthesis implemented by a particular filter. It should be noted that parameters related to the spectrum (or cepstrum) and the laryngeal tone parameters (f0, voicedness) are modeled separately. What is interesting in the HSS synthesis is that the models are trained on a large acoustic database before being adapted for a particular speaker. Such an approach makes it much easier to create a new synthesizer [29]. In 2016, Deep Mind Technologies published the findings of its study into the WaveNet system [30]. This type of speech synthesis is called parametrical synthesis. According to the authors, the system narrows the gap between the best available speech synthesis and natural speech by over 50%. Like the HSS synthesis, this method is also based on acoustic modeling. What makes it different is the elimination of the Vocoder (Voice Encoder), a coder used for analyzing and synthesizing the human speech. The audio signal is modeled directly by the same model. Because of its high computational complexity at the time of the publication, WaveNet was unable to generate a real-time speech signal, which is why this kind of synthesis is not included in this study. Later, Deep Mind went on to develop an improved model which served to create a TTS system, accessible in a virtual cloud [31]. VOCA devices make use of professional commercial voices, but their high quality is not the most important aspect for patients, who would rather hear their own voice. Unfortunately, the technology currently used in these systems does not allow for the provision of personalized voices [20]. Perhaps the most famous user of such a device was the British astrophysicist Stephen Hawking, who suffered from amyotrophic lateral sclerosis. Hawking used software made by the Speech Plus company. In the initial stages of the disease, he controlled the speech synthesizer with a joystick. Having lost use of his hands, he operated the device with his cheek. Currently, there are several companies that produce custom-made synthetic voices [32]. ModelTalker for example, a US-based company, offers to build personalized synthetic voices for the English language. The prospective user has to record between 400 and 1800 speech samples. The systems that are offered include concatenative, corpus-based and parametrical syntheses. Parametrical synthesis makes use of Deep Neural Networks (DNN). The Polish language is currently unavailable. OKI Electric Industry Co., Ltd. in Japan employs a hybrid speech synthesizer Polluxstar to build a personalized voice that is a combination of statistical and corpus-based speech. It makes use of both acoustic units and Markov models [33]. The Google Cloud Text-to-Speech also offers a Custom Voice feature. Custom Voice allows training of a custom voice model using own studio-quality audio recordings to create a unique voice. In addition, it is possible to synthesize audio using the Cloud Text-to-Speech API. Currently, only American English (en-US), Australian English (en-AU), and American Spanish (es-US) are supported [34]. Amazon Web Services implemented a feature in Amazon Polly called Brand Voice. Amazon Polly is a service that turns text into lifelike speech, allowing one to create applications that talk and build new categories of speech-enabled products. With the Brand Voice feature, it is possible to make Neural Text-to-Speech (NTTS) voice representing your Brand’s persona. Brand Voice allows differentiating your Brand by incorporating a unique vocal identity into your products and services. There is no Polish language neural voice present [35]. Edinburgh-based CereProc is another company that offers to build synthetic voices for individual customers [36]. The technology makes use of corpus-based synthesis, and the voice building involves the adaptation of an acoustic model based on approximately four hours of recorded speech. A female voice (Pola) is available for the Polish language, but it is not possible to adjust the synthesizer to simulate one’s own voice. Acapela is another company producing custom-made synthetic voices. Again, 19 languages are available for voice banking, but Polish still is not offered. Voice Keeper is another company that supports voice banking, but it is available only for English and Hebrew. Similarly, VocalID company also supports voice banking, but only for English [37]. Microsoft Azure offers Custom Neural Voice, a set of online tools for creating voice for brands [38]. In Custom Neural Voice Pro version, 300–2000 utterances are required. Here, the Polish language is available. In their study, Ahmad Khan et al. developed a speech synthesizer based on a patient’s voice recorded just before laryngectomy. The system of statistical speech synthesis was trained on many speakers and adapted to a 6–7min sample of the patient’s speech. Despite its low sound quality, the output resembled natural speech [20]. It is then possible to employ the existing technologies to generate high-quality speech, but it still begs the question of what quality can be obtained for a dysphonic voice. The following study aimed to prepare speech synthesis voice for a patient with changes in the larynx, causing hoarseness, affecting perceptual judgment and the acoustic signal parameters. In addition, we checked if it is possible to generate speech quality close to the original recordings using the MUSHRA listening test. Finally, the obtained synthetic voice was compared to the voice of a professional speaker, and after comparison, the result received a higher quality relative score to the synthetic professional voice. 2. Materials and Methods Back in 2014, the authors were approached by a person seeking help for someone close who had cancer. It turned out that in a few days the sick person was to undergo total laryngectomy, which would result in a loss of natural voice. At the time, it was impossible to predict the course of disease following the surgery. However, the authors were promptly engaged in a project aimed at the design of a speech synthesizer using prosody, close to natural speech. In practical terms, a task like this involves designing a corpus-based synthesizer using unit-selection speech synthesis, or one based on a statistical parametric speech synthesis system. The solution described in this paper guaranteed repeatability as well as versatility, allowing for the implementation of such projects on a larger scale. In both types of synthesis, it was very important to build a sufficiently extensive acoustic data repository to serve as the heart of the system. An acoustic database should include a variety of acoustic units (phones, diphones, syllables) in a number of different contexts and occurrences, and of varying durations. The first stage of building an acoustic database involved creating a balanced text corpus. This required extracting from a large text database a certain number of sentences that would best meet the input criteria, for example, the minimum and maximum number of acoustic units in a sentence. The larger the database, the more likely it was that the selected sentences will meet the set criteria. It was then important to find a balance that would ensure an optimal database size while maintaining the right proportion of acoustic units characteristic of a particular language. The speech corpus was built in a semi-automatic way and then corrected manually. Sentences selected with this method had to be manually verified in order to eliminate any markers, abbreviations and acronyms which were not expanded in the initial preprocessing. The sentences were selected by the greedy algorithm. The operation of this algorithm consists of iterative extraction of a number of sentences from a very large text set. All the sentences were also manually checked to ensure that they did not contain material that would be too hard to pronounce or contains obscene or otherwise loaded material which would introduce an emotional bias to the recordings. More information about balancing corpus is included in these articles [39,40]. The recordings were made in a recording studio during a number of several hours’ long sessions. Each consecutive session was preceded by a hearing of the previously recorded material in order to establish a consistent volume, timbre, manner of speaking, etc. [27,39]. The final stage in the construction of an acoustic database, following the recordings, was the appropriate labeling and segmentation. The segmentation of the database was carried out automatically, using statistical models, or heuristic methods, such as neural networks. Such a database should then be verified for the accuracy of the alignment of the defined boundaries of acoustic units. 2.1. Constructing the Corpus The corpus built for the recordings contained a selection of parliamentary speeches. Initially, it was a 300 MB text file containing 5,778,460 sentences. All the metadata was removed, and all the abbreviations, acronyms and numbers were replaced by full words. Then, the SAMPA phonetic alphabet was used to generate a phonetic transcription. The SAMPA phonetic is a computer-readable phonetic alphabet. A SAMPA transcription is designed to be uniquely parsable. As with the ordinary IPA, a string of SAMPA symbols does not require spaces between successive symbols. Two algorithms of the phonetic transcription were compared: the rule-based system developed for the Festival system, and the automatic method based on decision trees. The use of decision trees proved to be far more effective, ensuring higher accuracy in the phonetic transcription [39]. The balancing of the corpus was implemented by means of a greedy algorithm. This solution best fulfilled the given input criteria such as the number of phonemes, diphones, triphones making up the length of the sentence, or the number of segments in the final corpus. For the purpose of balancing the CorpusCrt program was used, which was written by Alberto Sesma Bailador 1998 at the Polytechnic University of Catalonia and was distributed as freeware [40]. An example input sentence in our initial corpus is in its orthographic and phonetic form represented by (a) orthography, (b) phonemes, (c) diphones, and (d) triphones. z jakim niezrównanym poczuciem humoru opisuje pan swoją marszczącą się wątrobę # z j a k i m n’ e z r u v n a n I m p o tS u ts’ e m x u m o r u o p i s u j e p a n s f o j o~ m a r S tS tS o n ts o~ s’ e~ v o n t r o b e~ # #z zj ja ak ki im mn’ n’e ez zr ru uv vn na an nI Im mp po otS tSu uts’ ts’e em mx xu um mo or ru uo op pi is su uj je ep pa an ns sf fo oj jo~ o~m ma ar rS StS tStS tSo on nts tso~ o~s’ s’e~ e~v vo on nt tr ro ob be~ e~# #zj zja jak aki kim imn’ mn’e n’ez ezr zru ruv uvn vna nan anI nIm Imp mpo potS otSu tSuts’ uts’e ts’em emx mxu xum umo mor oru ruo uop opi pis isu suj uje jep epa pan ans nsf sfo foj ojo~ jo~m o~ma mar arS rStS StStS tStSo tSon onts ntso~ tso~s’ o~s’e~ s’e~v e~vo von ont ntr tro rob obe~ be~# The parliamentary speech corpus was divided into 12 sub-corpora, 20 MB each [20]. The division was made on the grounds of the maximum corpus size that can be accepted by the Corpus CRT program. The following criteria were applied for the selection of the most representative and balanced sentences:Each sentence should contain a minimum of 30 phonemes; Each sentence should contain a maximum of 80 phonemes; The output corpus should contain 2,500 sentences; Each phoneme should occur at least 40 times in the corpus; Each diphone should occur at least 4 times in the corpus; Each triphone should occur at least 3 times in the corpus (this particular criterion can only be met for the most frequently used triphones). These assumptions were made on the basis of [41,42,43]. After the first balancing process, 12 different sub-corpora, each containing 2500 sentences, were created. Each sub-corpus contained approximately 189,000 phonemes. The frequencies of phonemes proved to be very similar in all of the sub-corpora. Figure 1 illustrates the percentage value of frequency distribution in two randomly selected parliamentary sub-corpora. After the second balancing process, the total number of diphones had increased (from 148,479 to 150,814), the number of diphones occurring less than four times had decreased (from 175 to 68), and the number of different diphones had increased (from 1096 to 1196). The total number of triphones had increased (from 145,979 to 148,314), and so had the number of different triphones (from 11,524 to 13,882). The ultimate corpus contains interrogative and imperative sentences and was also supplemented with words of less frequent occurrence. The frequency distribution of particular phonemes is shown in Figure 2. The 15 most common diphones are shown in Figure 3, and the 15 most common triphones are shown in Figure 4. The final stage of the corpus construction involved manual correction, which allowed for the elimination of sentences that were meaningless or difficult to utter. Ultimately, the corpus is made up of 2150 sentences. In its final form, the corpus was used in a doctoral dissertation concerned with the optimization of cost function in corpus-based synthesis for the Polish language [39]. 2.2. Recordings Due to the patient’s condition and the time limitations resulting from the planned surgery, the recordings could not be held in a recording studio. Instead, they were made in the patient’s home. To ensure a better quality, an acoustic booth was used. The recordings were carried out with the help of EDIROL R-09HR, which was placed 60 cm from the mouth. EDIROL R-09HR is a professional, high-resolution recorder with built-in stereo condenser microphone. During the recording, a written text was displayed for the speaker and the person in charge of the recording. The acoustic database was recorded with a 48 kHz sampling frequency and a 16-bit resolution in the WAV format. Each consecutive session was preceded by an examination of the previously recorded material in order to establish a consistent intonation and manner of speaking. The first session had to be repeated as the sentences had been read too quickly. At the second attempt, the recording process was improved as the patient tried to articulate the sentences in a louder voice, and the microphone was placed closer to the speaker, i.e., 50 cm. The entire recording was completed in two 2 h sessions, finishing, a few hours before the patient was transferred to the hospital. The whole of the corpus, consisting of 2150 sentences, was recorded. The synthetic voice was trained on 2000 sentences. 100 sentences were selected as a validation set and were used to determine the best model during after the training was completed. Finally, out of 100 sentences a set of 50 sentences was used to carry out the listening tests (MUSHRA). The corpus containing 2000 sentences has been used in the very first unit selection speech synthesis system programmed by the authors of this paper for non-commercial use. All of the audio files used in this system have been accepted as the acoustic database of the ELRA project (http://catalog.elra.info/product_info.php?cPath=37_39&products_id=1164; http://syntezamowy.pjwstk.edu.pl/korpus.html accessed on 5 April 2022). ELRA is involved in a number of projects at the European and international levels. These projects address various issues related to Language Resources, including production, validation, and standardisation. 2.3. Acoustic and Auditory-Perceptual Assessment of Voice Quality Due to dysphonia in the patient’s voice, the RBH auditory-perceptual scale was used to assess its quality [44]. The RBH auditory-perceptual scale is used in German clinics and is recommended by the Committee on Phoniatrics of the European Laryngological Society [45]. The RBH acronym is used to denote the following features:R—Rauigkeit (roughness) – the degree of voice roughness deviation caused by irregular vocal fold vibrations; B—Behauchtheit (breathiness) – the degree of breathiness deviation caused by glottic insufficiency; H—Heiserkeit (hoarseness) – the degree of hoarseness deviation. Ratings of 0, 1, 2, and 3 are used for all parameters on the RBH scale, with reference to the different degrees of vocal disorder: ‘0’ = normal voice, ‘1’ = a slight degree, ‘2’ = a medium degree, and ‘3’ = a high degree. The perceptual voice assessment was performed by two independent specialists who had completed an RBH training program and had extensive experience in voice signal evaluation. The experts were trained at a university. The training process was divided into three stages; each stage lasted 28 h. After each step, an exam checked the quality of annotation. Upon successfully finishing the training, another learning process was introduced with RBH Learning and Practice mobile application. The experts had been working for three years with an annotation of the speech signal. The assessment showed dynamic voice changes throughout the recordings, with R = 0, B = 1, H = 0 at the beginning of the recordings, and R = 1, B = 1, H = 1 at the end. These ratings indicated dysphonic changes in voice quality, pointing to dynamic changes taking place during the recordings. To better illustrate the changes, an acoustical analysis using AVQI (v. 02.03) was carried out (The Acoustic Voice Quality Index) [45,46]. The Acoustic Voice Quality Index is a relatively new clinical method used to quantify dysphonia severity. The method is calculated on the basis of a signal from a sustained vowel and samples of speech. To determine its value, a weighted combination of 6 parameters is taken into account: shimmer local, shimmer local dB, harmonics-to-noise ratio (HNR), general slope of the spectrum and tilt of the regression line through the spectrum and smoothed cepstral peak prominences (CPPs). The AVQI score obtained for the patient with laryngeal cancer was 5.62, which indicates largely altered voice quality. AVQI values range from 0 to 10. It was assumed that scores ≤ 3 indicate a normal, unchanged voice [45]. The patient’s voice was compared with that of a professional speaker recorded for the corpus-based speech synthesis, both using the same sentences. In order to select the professional speaker, voice samples from 30 voice talents were collected and then assessed by 8 voice analysis experts. Ultimately, the experts chose a female voice. The recordings, which were conducted in the recording studio of the Polish-Japanese Academy of Information Technology, were performed with an Audio-Technica AT2020 microphone with a pop filter, 30 cm from the microphone. The signal was recorded in the AIFF format with a 48 kHz sampling frequency and a 24-bit resolution, using the audio Focusrite Scarlett 2i4 interface. The corpus was recorded during 15 two-hour sessions, with each prompt being recorded as a separate file. After each session, the files were exported in the WAV format with file names corresponding to the prompt numbers in the corpus. The recordings were then checked for distortions and external noises, as well as for mistakes made by the speaker. A total of 480 prompts were re-recorded [27]. The values obtained for the voice were: AVQI = 1.61, and R = 0, B = 0, H = 0 on the perceptual scale. Figure 5 shows a graph with the acoustical analysis using AVQI calculated for the patient. 2.4. Segmentation of Audio File The next step, after recordings, was an automatic segmentation of the corpus. This was carried out by means of a program based on the Kaldi project [47]. Kaldi is an open-source speech recognition toolkit, written in C++. The segmentation was performed using a technique called ‘forced alignment’, which involves matching phone boundaries on the basis of a file containing phonetic transcription. First, the program created an FST graph whose states correspond to the consecutive segmental phonemes of the analyzed phrase. The phonetic transcription for the segmentation was prepared on the basis of an orthographic transcription using a Polish language dictionary with SAMPA transcriptions. Foreign words and proper nouns were transcribed manually. 2.5. Creating Synthetic Voice The authors set out to create a new voice using the Merlin library [48], a toolkit for building statistical parametric speech synthesis by means of Deep Neural Network. This approach must be used in combination with metasystem Festival, responsible for implementing phonetic transcription, linguistic features and the World library as a vocoder [49]. The World library also provides tools for analysis, processing and recording. In Festival, the following features were calculated:Context dependent phones (previous phoneme, next phoneme); Syllable structure (current, previous and next syllable); For each of syllable (stress accent and length of syllable); Position phoneme in syllable; Position phoneme in phrase; Position of stressed syllable in phrase. The first step was to define acoustic parameters based on the recordings. This involved calculating the values of fundamental frequency (f0), voicing levels, mel-generalized cepstral coefficients (MGCC) [28], and band aperiodicity, which expresses the value of the aperiodic energy signal. Each of the parameters were normalized to the mean value of 0, and their variance value equaled 1. All the parameter values for a given frame constitute its vector of acoustic properties. For f0 only values corresponding to the voiced signal frame was used, for non-voiced frames value 0 was used. Additionally, the delta and delta–delta were calculated for the F0 and MGCC parameters. Thus, the F0 for every signal frame is represented by three values. Each of the MGCC parameters is defined by 60 parameters representing the amount of energy for each sub-band. Ultimately, together with the delta and delta–delta, each signal frame is represented by 180 values. Once the sentence to be synthesized has been entered, the acoustic model predicts the acoustic parameter values using the obtained linguistic parameters. The latter were extracted at the phoneme level, while the acoustic parameters were extracted at the frame level. Their numbers differ, which makes model training difficult. In order to resolve the problem, information about the boundaries of phoneme states obtained in the segmentation process was used. Each state was matched with corresponding frames. The vector representing linguistic properties of a given state was copied a required number of times, and an index was added to it. Data prepared in this way contained for each frame its vector of linguistic properties and the corresponding vector of acoustic parameters. This represents, respectively, the input and the desired output required to train an acoustic model. However, the information about the states is not available during the synthesis process. For this reason, a model that will predict their duration on the basis of their linguistic parameters needs to be developed. The acoustic models and phoneme duration models were trained using Python Theano library [50]. The Theano library is integrated with Merlin and contains implemented statistical models based on deep neural networks. In addition, this allows for a very fast computation of mathematical expressions by using specialized GPUs. 3. Results 3.1. Experiments A number of experiments were carried out where voices were built with varying amounts of training data and different acoustic model architectures. In order to compare the models, values of the error function calculated for the verification data were used. The verification data constituted 10% of the training set. The mean squared error was used in the process. The values shown in Figure 6, Figure 7, Figure 8 and Figure 9 are the MSE sum for 180 mel-generalized cepstral coefficients, 3 parameters describing the fundamental frequency (f0) and 3 parameters describing the aperiodic band. The parameters were normalized to the mean value of 0 and the variance of 1. In all experiments, the models were trained for 25 epochs and a model from the best performing epoch was used. 3.1.1. Experiment 1: Building a Voice with 100 Sentences The first model was used to verify the system, so it was trained on a small number of sentences. 100 sentences were randomly selected from the corpus, of which 90 were used to train the models (training data). The remaining 10 sentences were used for verification purposes (verification data). A multilayer perceptron was used for the acoustic modelling. It consisted of an input layer (1), hidden layers (2) and an output layer (3). There were 6 hidden layers, each consisting of 1024 neurons. The hyperbolic tangent was chosen to act as the activation function. An identical neural network was employed for the modelling of phoneme state durations. In both cases, computations were performed without a GPU. They were made on a computer with an 8-core processor Intel Core i7-4790 3.60 GHz, 16 GB RAM. As Theano performs automatic data-parallel computations, all of the processor cores were utilized. The speech generated by the resultant models was comprehensible, though not very natural sounding. However, the experiment helped verify the correct functioning of the system. 3.1.2. Experiment 2: Building a Voice with 2000 Sentences The training data set and the verification set consisted of 2000 and 100 sentences, respectively. Both models were trained with the same neural network architecture as in the first experiment. In both cases, computations were performed with a CPU only. The resultant models made it possible to generate speech that sounded noticeably more natural than the speech generated in experiment 1. Figure 6 and Figure 7 show a graph of the error function during the voice training stage. The problem of overfitting was significantly reduced compared to the model trained with 100 sentences. 3.1.3. Experiment 3: Building a Voice with an Acoustic Model Based on a Recurrent Network This experiment was carried out with the same data set as in experiment 2. What made it different was an altered architecture of the acoustic model neural network (the model of the phoneme state durations remained unchanged). The last two layers of the perceptron were replaced with two LSTM layers [51,52]. The LSTM layer was recurrent, which means that the value predicted for the prior sample was at once the input value for the current sample. Thanks to this property, neural networks containing LSTM layers were used for sequence modelling. Apart from a perceptron with a hyperbolic tangent, a single LSTM block contained three perceptrons with a sigmoid activation function. The first of these was a forget gate, designed to discard any unimportant information from prior elements of the sequence. Next was the input gate, which filtered information in the current element. The third gate was the output gate, which decided which information should be passed to the subsequent elements of the sequence. Each of the LSTM layers consisted of 384 blocks. Computations performed in a single LSTM block were more complex than those in the perceptron. The time needed to train the model with a processor was estimated at 500 h. Therefore, it was decided that a GPU would be used. The GPU processor (Nvidia GTX 760) made it possible to train the model in 31 h and 27 min. Figure 8 shows a graph of the error function values during the model training. The application of LSTM layers practically eliminated the problem of overfitting. 3.1.4. Experiment 3: Building a Voice for 100, 200, 400, 650, 1000 and 1500 Sentences In order to investigate the effect of data volume on voice quality, additional acoustic models were built for 200, 400, 650, 1000 and 1500 sentences, respectively [39]. Figure 9 shows a graph of the error function values for a varying number of sentences. There was a striking difference between 100 and 200 sentences. There was also a noticeable leap between 200 and 400 sentences. A further increase in the number of sentences used did not affect the rate at which the error function values fell. The experiments discussed above led to the construction of 3 synthetic voices: two for 100 and 2000 sentences using an MLP network, and a third voice built on the basis of 2000 recordings using a recurrent network with LSTM layers. The voices built in experiment 4 were designed to examine the impact of different amounts of data on the quality of the models and were excluded from further evaluation. 3.2. MUSHRA The listening tests were conducted using the MUSHRA methodology. In a MUSHRA test, the listener is presented with a professional voice-over talent recording as the reference, (so called proper reference) and samples of generated speech to be evaluated. These generated systems include a so-called anchor. In addition, one of the systems served as a hidden reference. The hidden reference used in our tests is the same a voice-over talent recording as it was used as the proper reference. Such an approach made it possible to verify that the listeners assessed the systems against the reference. An anchor was required to be perceived as inferior in quality to the hidden reference. The tests were carried out by means of webMUSHRA. A total of 25 expert listeners participated in the tests, each of whom assessed 10 sentences in one test. The listeners were instructed to first listen to the reference recording and then assess each system on a 0–100 numerical scale. The results are shown in Table 1. The sentences used in the test came from a specially designed test corpus also named validation corpus and were not used for training or verification purposes. The purpose of creating the corpus was to obtain a set of sentences that would meet specific requirements different from those used to develop the main corpus [53]. It was decided to get a small corpus and, at the same time, the biggest possible coverage of different acoustic units, different from the ones included in the acoustic database. The variety of corpora was supposed to ensure the naturalness and comprehensibility of generated phrases occasionally occurring in the main corpus. The test corpus was prepared in the CorpusCrt application [40]. Sentences were compiled from three different linguistic bases, containing texts from newspapers on various subjects. Before the test corpus was created, it was required to generate the phonetic transcription for phonemes diphones and triphones for the whole database. It was decided to limit the size of the test corpus to 100 short statements (max. 60 phonemes in each sentence). The criteria of the sentence selection referred to their maximum length, the number of occurrences of various acoustic units, and different phoneme configurations. During corpus balancing, it was decided that each phoneme should occur at least 25 times, each diphone and triphone should occur at least once. Because of the small size of the corpus, obtaining all the diphones and triphones was impossible; however, the necessary condition ensured a variety of occurrences of mentioned acoustical units. The results obtained in the tests indicated a very high quality of the synthetic voice of the patient (Table 1). A difference of 0.05 in relative score in favor of the best patient’s synthetic voice 3 LSTM compared to the best professional synthetic voice accounted for by a better adjustment of the acoustic parameters (Table 1). The obtained results indicated that the best synthetic patient voice is more matched to the original patient recordings than the professional synthetic voice concerning the professional recordings. As voice 3 required the use of GPU, systems 2 and 3 were compared to see if their ratings were different. The ratings of both systems had an equal modal distribution and equal variances. P value = 0.651 and t = 0.452 indicated that the ratings of the two systems are not statistically different. This ultimately led to a decision to transfer the system to a virtual speech synthesizer, which would benefit the patient. Due to the computational complexity of the LSTM layer, the model trained with LSTM was too slow to be used on a computer without a GPU. For this reason, it was not placed on the virtual machine that was presented to the non-professional voice. The Analysis of Acoustic Parameters Errors Table 2 shows the values of acoustic parameters errors calculated on the basis of test sentences generated by means of 3 trained systems. The following acoustic parameters were applied: MCD—Mel-Cepstral Distortion; BAPD—band aperiodicity distortion; F0-RMS represents the root mean square of deviations in fundamental frequency values; F0-correlation, the value of Pearson’s correlation coefficient for the fundamental frequency; VUV (voiced-unvoiced error rate) indicates the percentage of incorrect predictions of voicedness [48]. The analysis of the data shown in Table 2 indicated that voice 2, i.e., a voice for the cancer patient, had the best BAPD value as well as having the highest correlation between F0 and the recording. On the other hand, the best values of MCD, F0-RMS and VUV were obtained for the LSTM-trained voice. The professional voice had better MCD and VUV% values while its BAPD value was equal to that of the cancer patient. The parameter values of the fundamental tone deviation (F0-RMS) and the fundamental frequency F0-correlation proved to be worse in the professional voice. In order to enable researchers to repeat or modify the conducted experiments, a GIT repository was created. The repository contains the Merlin Repository with all the modifications and scripts. The Supplementary Materials contain recordings of voice talent, patient voice, and synthetic voice of voice talent and patient. The synthetic voice was made available to the patient in the form of a virtual machine in the VirtualBox13 environment. The text is synthesized with a single command at the terminal level. The synthesizer works fast in Linux. However, transferring it to a virtual machine affects its operating speed. 4. Discussion and Conclusions The study aimed to prepare speech synthesis voice with changes in the larynx, causing hoarseness and affecting perceptual and acoustic signal parameters. The quality for the person with voice changes obtained a relative score of 0.71 for MLP and LSTM, where the relative score is defined as a recording’s mean value divided by synthetic voice mean. Interestingly, a higher voice quality than professional voice was achieved, where the relative score equals 0.66. MUSHRA results of patient MLP voice trained on 2000 sentences obtained 69.40 compared to 63.63 for professional voice. Creating such a voice was possible, but perceptual differences indicated that the patient voice sounded better than the professional voice. According to their study, Repova et al. [21], 61 patients were scheduled for total laryngectomy for T3-T4a laryngeal or hypopharyngeal cancer with uni- or bilateral neck dissection the regional lymph node involvement. A total of 31 patients were assessed as unsuitable for voice recordings due to low voice quality before surgery or unsatisfactory cooperation and compliance. Of the remaining 30 patients, 18 were willing and able to complete voice recordings. Of the 18 patients, 11 patients had a voice prosthesis implanted. Each patient recorded between 210 and 1400 sentences. For most, unit selection (US) or hidden Markov model (HMM) systems were used to perform personalized speech synthesis. However, the quality of speech synthesis was not evaluated. Overall, only 7 patients eventually began using TTS technology in the early postoperative period. However, the frequency and total time of use were significantly better in the first postoperative week than later in the hospital stay, when the device’s effort gradually decreased. Finally, 6 patients are actively using the software. One of these patients was a lecturer. The frequency and total device use time were significantly better in the first postoperative week than later in the hospital stay, when the effort to use the device gradually decreased. The gold standard for voice rehabilitation after total laryngectomy is tracheoesophageal speech with the voice prosthesis placement. The disadvantage of this approach is the necessity of regular replacement of the voice prosthesis due to the device’s lifetime. In their study, Repova et al. [21] obtained results that indicate that voice banking and speech synthesis can be an opportunity to increase the quality of life. Statistical speech synthesis created by recording complete corpus allows the generation of more natural-sounding speech than that obtained by adapting acoustic models to a particular patient, as reported in Ahmad Khan et al. [20]. The authors used the system of statistical speech synthesis was that trained on many speakers and adapted on a 6–7-min’ sample of the patient’s speech. Despite its low sound quality, the output resembled natural speech. The created corpus in this study is representative of the Polish language. It enables high-quality, corpus-based and HSS speech synthesis. The signal segmentation methods developed in the study ensure a high degree of accuracy, as confirmed by the author’s previous studies [27,39,41,54]. This work is innovative for the Polish language. The method developed in the course of the study makes it possible to create a new synthetic voice for the Polish language by means of a statistical parametric speech synthesis system. Despite significant changes in the patient’s voice, reflected in the RBH scale features and the AVQI parameters, the results obtained in the study were very promising, as confirmed by the MUSHRA test. As a result, this method can be employed to develop a synthetic voice for a person awaiting total laryngectomy, allowing them to speak with their own voice, which ensures the patient’s better mental wellbeing. Having been presented with the speech synthesizer, the total laryngectomy patient was clearly moved being able to hear his own voice and expressed full approval of the quality of the synthesis. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/s22093188/s1. The materials contain recordings of voice talent, patient voice, and synthetic voice of voice talent and patient. Click here for additional data file. Author Contributions Conceptualization, K.S., J.L.; methodology, J.L., K.S.; software, J.L.; validation, J.L.; formal analysis, J.L., K.S.; investigation, J.L., K.S.; resources, K.S.; writing—original draft preparation, K.S.; writing—review and editing, K.S.; visualization, K.S., J.L.; supervision K.S. All authors have read and agreed to the published version of the manuscript. Funding This research received no external funding. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement The GIT repository can be accessed at https://github.com/kubapb/merlin, (accessed on 5 April 2022). Conflicts of Interest The authors declare no conflict of interest. Figure 1 A comparison of frequency distribution of phonemes in two random parliamentary sub-corpora. Figure 2 Phoneme frequency distribution in the final version of the corpus. Figure 3 The 15 most common diphones. They account for 14.22% of all diphones in the final corpus. Figure 4 The 15 most common triphones. They account for 4.09% of all triphones in the final corpus. Figure 5 Acoustic assessment of the patient’s voice. Figure 6 Error function values for a voice trained on 100 sentences. Figure 7 Error function values for a voice trained on 2000 sentences. Figure 8 Error function values for a voice trained on 2000 sentences using LSTM layers. Figure 9 Error function values for a varying number of sentences. sensors-22-03188-t001_Table 1 Table 1 MUSHRA test results. Patient’s Voice Patient’s Synthetic Voice 1 MLP Patient’s Synthetic Voice 2 MLP Patient’s Synthetic Voice 3 LSTM Professional Voice Professional Synthetic Voice MLP Relative score * - 0.36 0.71 0.71 - 0.66 Mean value 97.62 35.48 69.40 69.74 96.21 63.63 Median 100 34 71 72 100 66 STD 5.54 21.52 19.42 18.35 8.01 19.41 * Relative score = recording’s mean value/synthetic voice mean value. sensors-22-03188-t002_Table 2 Table 2 Values of acoustic parameter errors calculated for verification data. Voice ID MCD (dB) BAPD (dB) F0-RMS (Hz) F0-Correlation VUV % Voice 1 5.489 0.142 29.787 0.489 11.792 Voice 2 4.779 * 0.133 26.096* 0.635 * 9.059* Voice 3 LSTM 4.731 0.134 25.438 0.629 8.308 Professional voice 4.186 0.133 31.116 0.558 5.689 * Statistically significant in comparison to professional voice, p-value <0.05. All differences between acoustic parameters except the BAPD (dB) parameter for voice 2 and professional voice are statistically significant. Between voice 2 and voice 3, only the VUV % parameter is statistically significant (p-value <0.05). The bold font is used to indicate the best acoustic parameter among all synthetic voices. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Religioni U. Cancer incidence and mortality in Poland Clin. Epidemiol. Glob. Health 2020 8 329 334 10.1016/j.cegh.2019.12.014 2. Chatenoud L. Garavello W. Pagan E. Bertuccio P. Gallus S. La Vecchia C. Negri E. Bosetti C. Laryngeal cancer mortality trends in European countries Int. J. Cancer Res. 2016 138 833 842 10.1002/ijc.29833 26335030 3. Raport on Laryngeal Cancer Available online: http://onkologia.org.pl/nowotwory-zlosliwe-krtani-c32/ (accessed on 15 February 2021) 4. Osowiecka K. Rucińska M. Nawrocki S. Survival of patients with laryngeal cancer treated with irradiation in the years 2003-2006 in the Independent Public Health Care Centre of the Ministry of Interior and the Warmia and Mazury Oncology Centre in Olsztyn J. Oncol. 2014 64 237 245 5. Bobdey S. Jain A. Balasubramanium G. Epidemiological review of laryngeal cancer: An Indian perspective Indian J. Med. Paediatr. Oncol. 2015 36 154 160 10.4103/0971-5851.166721 26855523 6. Obid R. Redlich M. Tomeh C. The treatment of laryngeal cancer Oral Maxillofac. Surg. Clin. 2019 31 1 11 10.1016/j.coms.2018.09.001 30449522 7. De Stefani E. Correa P. Oreggia F. Leiva J. Rivero S. Fernandez G. Deneo-Pellegrini H. Zavala D. Fontham E. Risk factors for laryngeal cancer Cancer 1987 60 3087 3091 10.1002/1097-0142(19871215)60:12<3087::AID-CNCR2820601238>3.0.CO;2-6 3677031 8. Münger K. Baldwin A. Edwards K.M. Hayakawa H. Nguyen C.L. Owens M. Grace M. Huh K. Mechanisms of human papillomavirus-induced oncogenesis J. Virol. 2004 78 11451 11460 10.1128/JVI.78.21.11451-11460.2004 15479788 9. Jones T.M. De M. Foran B. Harrington K. Mortimore S. Laryngeal cancer: United Kingdom national multidisciplinary guidelines J. Laryngol. Otol. 2016 130 S75 S82 10.1017/S0022215116000487 27841116 10. Cox S.R. Theurer J.A. Spaulding S.J. Doyle P.C. The multidimensional impact of total laryngectomy on women J. Commun. Disord. 2015 56 59 75 10.1016/j.jcomdis.2015.06.008 26186255 11. Sharpe G. Camoes Costa V. Doubé W. Sita J. McCarthy C. Carding P. Communication changes with laryngectomy and impact on quality of life: A review Qual. Life Res. 2019 28 863 877 10.1007/s11136-018-2033-y 30417205 12. Schwartz S.R. Yueh B. Maynard C. Daley J. Henderson W. Khuri S.F. Predictors of wound complications after laryngectomy: A study of over 2000 patients Otolaryngol. Head Neck Surg. 2004 131 61 68 10.1016/j.otohns.2003.08.028 15243559 13. Braz D.S.A. Ribas M.M. Dedivitis R.A. Nishimoto I.N. Barros A.P.B. Quality of life and depression in patients undergoing total and partial laryngectomy Clinics 2005 60 135 142 10.1590/S1807-59322005000200010 15880250 14. Kapila M. Deore N. Palav R.S. Kazi R.A. Shah R.P. Jagade M.V. A brief review of voice restoration following total laryngectomy Indian J. Cancer 2011 48 99 21248437 15. Blom E.D. Singer M.I. Hamaker R.C. A prospective study of tracheoesophageal speech Arch. Otorhinolaryngol.-Head Neck Surg. 1986 112 440 447 10.1001/archotol.1986.03780040080017 3081019 16. Debry C. Dupret–Bories A. Vrana N.E. Hemar P. Lavalle P. Schultz P. Laryngeal replacement with an artificial larynx after total laryngectomy: The possibility of restoring larynx functionality in the future Head Neck 2014 36 1669 1673 10.1002/hed.23621 24677757 17. Gates G.A. Hearne III E.M. Predicting esophageal speech Ann. Otol. Rhinol. Laryngol. 1982 91 454 457 10.1177/000348948209100428 7114732 18. Pruszewicz A. Clinical phoniatrics (Foniatria kliniczna) Państ. Zakład Wydawnictw Lekarskich Warszawa, Poland 1992 19. Liu H. Ng M.L. Electrolarynx in voice rehabilitation Auris Nasus Larynx 2007 34 327 332 10.1016/j.anl.2006.11.010 17239553 20. Ahmad Khan Z. Green P. Creer S. Cunningham S. Reconstructing the voice of an individual following laryngectomy Augment. Altern. Comm. 2011 27 61 66 10.3109/07434618.2010.545078 21. Repova B. Zabrodsky M. Plzak J. Kalfert D. Matousek J. Betka J. Text-to-speech synthesis as an alternative communication means after total laryngectomy Biomed Pap. Med. Fac. Univ. Palacky Olomouc. Czech Repub. 2021 165 192 197 10.5507/bp.2020.016 32367081 22. Klatt D.H. Software for a cascade/parallel formant synthesizer J. Acoust. Soc. Am. 1980 67 971 995 10.1121/1.383940 23. Klatt D.H. Review of text-to-speech conversion for English J. Acoust. Soc. Am. 1987 82 737 793 10.1121/1.395275 2958525 24. Khan R.A. Chitode J.S. Concatenative speech synthesis: A Review Int. J. Comput. Appl. 2016 136 1 6 25. Taylor P. Text-to-Speech Synthesis Cambridge University Press New York, NY, USA 2009 26. Kishore S.P. Black A.W. Unit size in unit selection speech synthesis Proceedings of the Eighth European Conference on Speech Communication and Technology Geneva, Switzerland 1–4 September 2003 27. Szklanny K. Koszuta S. Implementation and verification of speech database for unit selection speech synthesis Proceedings of the 2017 Federated Conference on Computer Science and Information Systems (FedCSIS) Prague, Czech Republic 3–6 September 2017 28. Tokuda K. Kobayashi T. Masuko T. Imai S. Mel-generalized cepstral analysis-a unified approach to speech spectral estimation Proceedings of the Third International Conference on Spoken Language Processing Tokyo, Japan 18–22 September 1994 29. Zen H. Nose T. Yamagishi J. Sako S. Masuko T. Black A.W. Tokuda K. The HMM-based speech synthesis system (HTS) version 2.0 SSW 2007 6 294 299 30. Oord A.V.D. Dieleman S. Zen H. Simonyan K. Vinyals O. Graves A. Kalchbrenner N. Senior A. Kavukcuoglu K. Wavenet: A generative model for raw audio arXiv 2006 Available online: 10.48550/arXiv.1609.03499 (accessed on 5 April 2022) 1609.03499 31. Wavenet Website Available online: https://cloud.google.com/text-to-speech/docs/wavenet (accessed on 15 February 2021) 32. ModelTalker Website Available online: www.modeltalker.com (accessed on 15 February 2021) 33. Yamagishi J. Veaux C. King S. Renals S. Speech synthesis technologies for individuals with vocal disabilities: Voice banking and reconstruction Acoust. Sci. Technol. 2012 33 1 5 10.1250/ast.33.1 34. Google Cloud Available online: https://cloud.google.com/text-to-speech/custom-voice/docs (accessed on 5 April 2022) 35. Amazon Polly Available online: https://docs.aws.amazon.com/polly/latest/dg/what-is.html (accessed on 5 April 2022) 36. CereProc Company Website Available online: http://www.cereproc.com/ (accessed on 5 April 2022) 37. Acapela Group Website Available online: https://www.acapela-group.com/ (accessed on 5 April 2022) 38. Train Your Voice Model Available online: https://docs.microsoft.com/en-us/azure/cognitive-services/speech-service/how-to-custom-voice-create-voice (accessed on 5 April 2022) 39. Szklanny K. Optymalizacja funkcji kosztu w korpusowej syntezie mowy polskiej Ph.D. Thesis Polsko-Japońska Wyższa Szkoła Technik Komputerowych Warszawa Warszawa, Poland September 2009 40. Bailador A.S. CorpusCrt Technical Report Polytechnic University of Catalonia (UPC) Barcelona, Spain 1998 41. Oliver D. Szklanny K. Creation and analysis of a Polish speech database for use in unit selection synthesis Proceedings of the Fifth International Conference on Language Resources and Evaluation Genoa, Italy 24–26 May 2006 42. Bozkurt B. Ozturk O. Dutoit T. Text design for TTS speech corpus building using a modified greedy selection Proceedings of the Eighth European Conference on Speech Communication and Technology Geneva, Switzerland 1–4 September 2003 43. Clark R.A. Richmond K. King S. Multisyn: Open-domain unit selection for the Festival speech synthesis system Speech Commun. 2007 49 317 330 10.1016/j.specom.2007.01.014 44. Nawka T. Anders L.C. Wendler J. Die auditive Beurteilung heiserer Stimmen nach dem RBH-System Sprache-Stimme-Gehor. 1994 18 130 300 45. Dejonckere P.H. Bradley P. Clemente P. Cornut G. Crevier-Buchman L. Friedrich G. Van De Heyning P. Remacle M. Woisard V. A basic protocol for functional assessment of voice pathology, especially for investigating the efficacy of (phonosurgical) treatments and evaluating new assessment techniques Eur. Arch. Otorhinolaryngol. 2001 258 77 82 10.1007/s004050000299 11307610 46. Maryn Y. De Bodt M. Barsties B. Roy N. The value of the Acoustic Voice Quality Index as a measure of dysphonia severity in subjects speaking different languages Eur. Arch. Otorhinolaryngol. 2014 271 1609 1619 10.1007/s00405-013-2730-7 24162765 47. Povey D. Ghoshal A. Boulianne G. Burget L. Glembek O. Goel N. Hannemann M. Motlicek P. Qian Y. Schwarz P. The Kaldi speech recognition toolkit Proceedings of the IEEE 2011 workshop on automatic speech recognition and understanding Big Island, HI, USA 11–15 December 2011 48. Wu Z. Watts O. King S. Merlin: An Open Source Neural Network Speech Synthesis System Proceedings of the 9th ISCA Speech Synthesis Workshop Sunnyvale, CA, USA 13–15 September 2016 49. Morise M. Yokomori F. Ozawa K. WORLD: A vocoder-based high-quality speech synthesis system for real-time applications IEICE Trans. Inf. Syst. 2016 99 1877 1884 10.1587/transinf.2015EDP7457 50. Theano Library Website Available online: https://github.com/Theano/Theano (accessed on 15 February 2021) 51. Zen H. Acoustic Modeling in Statistical Parametric Speech Synthesis-from HMM to LSTM-RNN 2015 Available online: https://research.google/pubs/pub43893/ (accessed on 5 April 2022) 52. Olah C. Understanding Lstm Networks 2015 Available online: https://research.google/pubs/pub45500/ (accessed on 5 April 2022) 53. Villaseñor-Pineda L. Montes-y-Gómez M. Pérez-Coutiño M.A. Vaufreydaz D. A corpus balancing method for language model construction Proceedings of the International Conference on Intelligent Text Processing and Computational Linguistics Mexico City, Mexico 16–22 February 2003 54. Szklanny K. Multimodal Speech Synthesis for Polish Language Springer Cham, Switzerland 2014 Volume 242 325 333

PMC009xxxxxx/PMC9099607.txt

==== Front Int J Environ Res Public Health Int J Environ Res Public Health ijerph International Journal of Environmental Research and Public Health 1661-7827 1660-4601 MDPI 10.3390/ijerph19095407 ijerph-19-05407 Article Effects of a Cool-Down after Supramaximal Interval Exercise on Autonomic Modulation https://orcid.org/0000-0001-7357-7685 Parks Jason C. 1* https://orcid.org/0000-0002-1585-9788 Marshall Erica M. 2 https://orcid.org/0000-0002-7409-3030 Humm Stacie M. 3 Erb Emily K. 3 https://orcid.org/0000-0002-4241-0169 Kingsley J. Derek 3 Tchounwou Paul B. Academic Editor 1 Kinesiology, State University of New York at Cortland, Cortland, NY 13045, USA 2 Exercise Science, Florida Southern College, Lakeland, FL 33801, USA; emarshall@flsouthern.edu 3 Exercise Science and Exercise Physiology, Kent State University, Kent, OH 44242, USA; shumm2@kent.edu (S.M.H.); eerb4@kent.edu (E.K.E.); jkingsle@kent.edu (J.D.K.) * Correspondence: jason.parks@cortland.edu 29 4 2022 5 2022 19 9 540709 4 2022 27 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Supramaximal interval exercise alters measures of autonomic modulation, while a cool-down may speed the recovery of vagal modulation. The purpose of this study was to compare the effects of a cool-down (pedaling a cycle ergometer at 50 rpm against a resistance of 45 W) versus passive recovery (no pedaling) after supramaximal interval exercise on autonomic modulation. Sixteen moderately active individuals (Mean ± SD: 23 ± 3 years (men: n = 10; women: n = 6) were assessed for autonomic modulation at Rest, and 15 (R15), 30 (R30), 45 (R45) and 60 (R60) min following supramaximal interval exercise. Linear measures of autonomic modulation included natural log (ln) total power (lnTP), high-frequency power (lnHF), the ratio of low frequency (LF) to HF ln(LF/HF) ratio, root mean square of successive differences between normal heartbeats (lnRMSSD), while non-linear measures included sample entropy (SampEn) and Lempel–Ziv entropy (LZEn). Two-way repeated ANOVAs were used to evaluate the main effects of condition (cool-down, passive recovery) across time (Rest, and R15, R30, R45 and R60). There were significant (p ≤ 0.05) condition by time interactions for SampEn and LZEn, such that they decreased at 15, 30, 45 and 60 min during passive recovery compared to Rest, with the recovery of SampEn and LZEn by 60 and 45 min, respectively, during cool-down. There were significant (p ≤ 0.05) main effects of time for lnTP, lnHF and lnRMSSD, such that lnTP, lnHF and lnRMSSD were attenuated, and lnLF/HF ratio was augmented, at all recovery times compared to Rest. These data demonstrate that a cool-down increases the recovery of nonlinear measures of vagal modulation within 45–60 min after supramaximal interval exercise, compared to passive recovery in moderately active individuals. heart rate variability heart rate complexity active recovery passive recovery lactate This research received no external funding. ==== Body pmc1. Introduction Supramaximal interval exercise is a popular mode of exercise used in healthy individuals [1,2]. Supramaximal interval exercise involves the control of intensity and rest intervals during repeated bouts to increase the volume of supramaximal work while reducing total exercise time, resulting in a time-efficient means to induce positive adaptations in skeletal muscle [3]. Researchers have demonstrated that acute bouts of supramaximal interval exercise may result in transient effects on heart rate (HR) and autonomic modulation measured via linear methods of heart rate variability (HRV) or nonlinear methods using heart rate complexity (HRC). The transient effects on HR and autonomic modulation after supramaximal interval exercise were demonstrated by increases in HR and sympathovagal dominance [4,5], with reductions in linear and nonlinear measures of vagal modulation immediately following [6], and for at least 60 min after the supramaximal exercise [4,5]. This information is important, considering the transient risk of a cardiovascular event may increase after vigorous exertion [7], and reductions in measures of vagal modulation have been observed prior to the onset of atrial fibrillation [8,9]. Taken together, it is clear that supramaximal interval exercise increases HR and sympathovagal dominance with concomitant reductions in vagal modulation, at least temporarily [4], but the data are sparse. The American College of Sports Medicine recommends a cool-down consisting of at least 5–10 min of light-to-moderate aerobic and muscular endurance exercise to facilitate the gradual recovery of HR and blood pressure (BP), and to aid in the clearance of metabolites, such as hydrogen ions and blood lactate after vigorous exercise [10], with no recommendations following supramaximal exercise. Data have demonstrated that a cool-down may increase vagal modulation, thereby attenuating HR after continuous vigorous-intensity exercise [11], but no such data exist regarding supramaximal interval exercise. However, based on the data, the responses to both continuous vigorous-intensity exercise and supramaximal interval exercise, in terms of cardiovascular recovery, appear to be relatively similar [2,4,12,13]. Given that the cardiovascular responses to both intensities appear similar, it is likely that the prescription of a cool-down after supramaximal interval exercise may be beneficial in promoting autonomic recovery. Despite ACSM’s recommendation of a cool-down after vigorous exercise, not one of the existing studies examining the effects of supramaximal interval exercise on the cardiovascular system included a cool-down of at least five min in their study protocol [4,5,13,14], which may have resulted in a misleading or inaccurate interpretation of the data; thus more data on a cool-down following supramaximal exercise is both pertinent and necessary. Understanding how the use of a cool-down influences HR and autonomic modulation after supramaximal exercise may be an important step in improving the exercise prescription. Therefore, the purpose of this study was to determine whether utilizing a cool-down after supramaximal interval exercise has any effect on the recovery of HR and autonomic modulation up to 60 min in moderately active individuals, compared to passive recovery. We hypothesized that HR and sympathovagal dominance would significantly increase, and vagal modulation, measured using both linear and nonlinear assessments of autonomic modulation, would significantly decrease for up to 60 min after acute supramaximal interval exercise, compared to rest, regardless of the recovery condition. We also hypothesized that the use of a cool-down would result in the earlier recovery of HR, as well as linear and nonlinear assessments of autonomic modulation within 60 min after exercise. 2. Materials and Methods 2.1. Participants Eighteen moderately-active individuals participated in this study. However, sixteen young (20–26 years), healthy, moderately active individuals (men = 12, women = 6) completed this study, as two participants were not able to complete the exercise protocol. Participants were recruited through flyers posted on bulletin boards and through social media. Participants were deemed moderately active based on the Lipid Research Clinics Physical Activity Questionnaire, which consists of four questions and uses a four-point scoring method to determine each participant’s physical activity classification [15]. Participants were excluded from the study if they had known vascular or metabolic disease, uncontrolled hypertension (resting brachial BP ≥ 130/80 mmHg), a recent smoking history (<6 months since last tobacco use), obesity (defined as a body mass index ≥ 30 kg/m2), cancer, orthopedic problems, history of blood clots, and/or open wounds. Participants were excluded if they were taking supplements or medications known to affect HR, BP, or vascular function. Inclusionary criteria were self-reported All women were tested during the early follicular phase (days 1–7) of their menstrual cycle. This project was approved by the Institutional Review Board, and data were collected in agreement with the Declaration of Helsinki. All participants were informed of the risks and benefits of the study, and signed an institutionally approved informed consent form prior to the collection of any data. 2.2. Study Design The present study utilized a repeated measures design in which each participant acted as their own control. Participants reported to the laboratory on three different days. The first visit consisted of an orientation and anthropometric assessment, as well as familiarization with the Wingate anaerobic test (WAT), while the second and third visits were used for data collection. On the data collection days, participants arrived at the laboratory having abstained from strenuous exercise, alcohol, and caffeine for a minimum of 24 h, and food for 3 h prior to testing. The order of the second and third visits were counterbalanced and consisted of two 30 s WATs, separated by a two-minute active recovery (AR) followed by either the cool-down or passive recovery condition (see Figure 1). The cool-down consisted of pedaling a cycle ergometer at 50 rpm against a light resistance of 45 W for 5 min and the passive recovery required the participant to sit upright in a chair for the same period of time. There were at least 72 h between exercise sessions, and participants were asked to refrain from strenuous exercise between sessions. The two exercise sessions were performed during the same time of day (±1 h) for each participant. All testing occurred between the hours of 7 am and 12 pm to control for diurnal variation. Prior to data collection, participants rested in a supine position for 10 min while all instrumentation was applied. Once the resting data were collected, participants performed two WATs. After the second WAT, the participant completed the 5 min cool-down or passive recovery condition, followed by 5 min of supine rest prior to recovery data collection. During recovery, data were collected at Rest and 15, 30, 45, and 60 min after exercise (Figure 1). 2.3. Anthropometrics Height and weight were measured using a stadiometer and balance platform scale, respectively. Height was measured to the nearest 0.1 cm, and weight was measured to the nearest 0.1 lb and then converted to kg. 2.4. Autonomic Modulation Continuous 3-lead electrocardiogram (ECG) signals were collected via a modified CM5 configuration (PowerLab; AD Instruments, Colorado Springs, CO, USA) at a rate of 1000 Hz. HR was determined via the ECG recordings. Respiration rate was set at 12 breaths per min during all data collection. The ECG signals were imported into WinCPRS analyzing software Version 1.162 (Absolute Aliens; Turku, Finland) after visual inspection for noise, ectopic beats, as well as artifacts, and no interpolation was performed. Fast Fourier transform was used to generate spectral power. Linear measures of autonomic modulation (HRV) were assessed in both the frequency and time domains. In the frequency domain, total power (TP) represented the total autonomic activity (6). Low-frequency power (LF) consisted of both sympathetic activity and vagal modulation, with a range between 0.04–0.15 Hz, and high-frequency power (HF) was a measure of vagal modulation, with a range between 0.15 and 0.4 Hz (6). The LF/HF ratio was used to represent sympathovagal dominance [16]. In the time domain, the root mean square of successive differences between normal heartbeats (RMSSD) was used as a measure of vagal modulation [17]. The non-linear measures of autonomic modulation (HRC) were sample entropy (SampEn) and Lempel–Ziv entropy (LZEn). These measures were used to examine the complexity of the R-R intervals over a 5 min epoch, after the removal of the linear trend. SampEn is defined as the probability of similar sequences or successive matches over a short period of time, with a range from 0 to 2 [18]. Values nearer to 0 are indicative of a more predictive signal and values approaching 2 are indicative of a more chaotic signal [18]. LZEn involves counting the number of distinct and repeating patterns within a time sequence [19]. Each distinct pattern, or subsequence, is assigned a symbol and then converted to 0 s and 1 s to form a binary substring. When reading the time sequence from left to right, each time a new substring of consecutive digits is encountered, a new symbol is added [20]. The size of the symbol vocabulary and rate at which distinct symbols occur along a sequence determines the complexity of that sequence [20]. LZEn values further from zero indicate an increase in the complexity of the sequence [21]. 2.5. Blood Lactate Blood lactate levels were measured by making a small puncture in the ear lobe using a Unistik 3, single-use safety lancet (Owen Mumford Ltd.; Oxfordshire, UK). A 0.2 µL blood sample was collected and blood lactate concentration was measured with a lactate Plus analyzer (Nova Biomedical; Waltham, MA, USA). The lactate Plus analyzer has been found to be a reliable and accurate alternative to laboratory-based bench top analyzers [22]. Lactate Plus, single use, disposable test strips (Nova Biomedical; Waltham, MA, USA) were used to collect the sample during the study protocol. 2.6. Supramaximal Interval Exercise Prior to completing the supramaximal interval exercise protocol, each participant completed a 5 min warmup on the cycle ergometer pedaling at 50 rpm against a light resistance of 45 W. The supramaximal interval exercise consisted of two, maximal effort, 30 s WATs performed using a mechanically braked cycle ergometer (Monarch Model 894 E; Stockholm, Sweden). The WAT involved pedaling against a constant force [23] and the resistance was set to 7.5% of the participant’s body weight based on previous studies [5,14]. Participants performed 2 min of active recovery (pedaling at 50 rpm against a light resistance of 45 W) between the exercise bouts. After the second and final WAT, participants completed the 5 min cool-down (pedaling at 50 rpm against a light resistance of 45 W) or passive recovery condition (sitting relatively still in a chair). Once these conditions were completed, all participants moved into a supine position for five 5 min prior to the collection of recovery data (see Figure 1). To ensure the safety of the participants, HR and continuous BP were monitored during the exercise bout and the 5 min cool-down or passive recovery conditions (NIBP Nano monitoring system, AD Instruments; Colorado Springs, CO, USA) and sampled at 1000 and 200 Hertz (Hz), respectively. 2.7. Statistical Analysis A 2 × 5 repeated measures analysis of variance (ANOVA) was used to determine the effects of condition (cool-down or passive recovery) across the repeated measure of time (Rest, 15, 30, 45, and 60 min) on HR, autonomic modulation (TP, LF, HF, LF/HF ratio, RMSSD, SampEn and LZEn) and blood lactate after supramaximal interval exercise. Because TP, LF, HF, LF/HF ratio, and RMSSD were not normally distributed, as revealed by the Shapiro–Wilk test, these values underwent log transformation (ln). Paired t-tests were performed for all post hoc comparisons with a Benjamini–Hochberg correction [24] to control for alpha inflation. Partial eta squared (ηp2) was used to assess the effect size of each dependent variable [25]. Percent change was calculated using the formula New number—Original number/Original number × 100, but was not run statistically. Significance was set a priori at p ≤ 0.05. Values are presented as the mean ± standard deviation (SD). All statistical analyses were completed using IBM SPSS (Version 25, IBM, Armonk, NY, USA). The sample size of 16 participants was based on data collected by Kingsley et al. [14]. The G*Power sample size calculator, version 3.1.9.2 (Heinrich-Heine-Universität Düsseldorf, Dusseldorf, Germany) [26], was used to estimate a minimum sample size of eight to achieve a power of 80% based on an effect size (Cohen’s f) of 1.13 and an alpha of 0.05. 3. Results Participant characteristics are presented in Table 1. Measures of HR and autonomic modulation are presented in Table 2. There were no significant (p > 0.05) differences in average power (W/kg) during the WATs, when comparing the cool-down and passive recovery conditions. There were no significant interactions for HR, lnTP, lnLF, lnHF or lnRMSSD. However, there was a significant condition by time interaction for the lnLF/HF ratio (F4,79 = 3.07, p = 0.023, ηp2 = 0.17), such that there was a significant increase from 15 to 30 min after supramaximal interval exercise during the passive recovery condition, with no significant increase during the cool-down condition. There were significant main effects of time for HR (F4,79 = 108.66, p ≤ 0.001, ηp2 = 0.88) and the ln(LF/HF) ratio (F4,79 = 17.68, p ≤ 0.001, ηp2 = 0.54), such that they were augmented at 15, 30, 45 and 60 min after supramaximal interval exercise compared to Rest, during both conditions. There were significant main effects of time for lnTP (F4,79 = 34.94, p ≤ 0.001, ηp2 = 0.79), lnLF (F4,79 = 59.78, p ≤ 0.001, ηp2 = 0.80), lnHF (F4,79 = 70.48, p ≤ 0.001, ηp2 = 0.83) (Table 2), and lnRMSSD (F4,79 = 75.72, p ≤ 0.001, ηp2 = 0.84), such that they were attenuated at 15, 30, 45 and 60 min after supramaximal interval exercise compared to Rest, during both conditions. There were significant condition by time interactions for SampEn (F4,79 = 5.79, p ≤ 0.001, ηp2 = 0.28) and LZEn (F4,79 = 2.66, p = 0.041, ηp2 = 0.15), such that there were significant increases from 30 to 45 min after exercise during the cool-down condition, with no significant increase during the passive recovery condition (Figure 2). There were also significant main effects of time for SampEn (F4,79 = 17.43, p ≤ 0.001, ηp2 = 0.54) and LZEn (F4,79 = 18.01, p ≤ 0.001, ηp2 = 0.55), such that they were attenuated at 15, 30, 45 and 60 min after supramaximal interval exercise during the passive recovery condition, compared to Rest, with the recovery of SampEn and LZEn by 60 and 45 min, respectively, during the cool-down condition (Figure 2). There was a significant condition by time interaction for blood lactate (F4,79 = 15.99, p ≤ 0.001, ηp2 = 0.50), such that it was significantly lower with a cool-down at 15, 30, 45 and 60 min after supramaximal interval exercise, compared to the passive recovery condition. There was a significant main effect of time for blood lactate (F4,79 = 504.81, p ≤ 0.001, ηp2 = 0.97), such that it was augmented at 15, 30, 45 and 60 min after supramaximal interval exercise compared to Rest during both conditions (Table 2). 4. Discussion The present study sought to compare changes in HR and autonomic modulation using a cool-down versus passive recovery for up to 60 min after an acute bout of supramaximal interval exercise in moderately active individuals. The primary findings were that an acute bout of supramaximal interval exercise increases HR, decreases linear measures of vagal modulation, and increases sympathovagal dominance for up to 60 min with both a cool-down and passive recovery. However, non-linear measures of vagal modulation recover within 45–60 min after supramaximal interval exercise with the use of a cool-down, but remains depressed for at least 60 min with passive recovery. Collectively, these findings indicate that an acute bout of supramaximal interval exercise results in a pronounced loss of vagal modulation, with increases in sympathovagal dominance in moderately active individuals, and that the use of a cool-down may facilitate the earlier recovery of non-linear measures of vagal modulation. The present study demonstrated that an acute bout of supramaximal interval exercise resulted in significant increases in HR of 55%, 39%, 27% and 22%, compared to Rest, at 15, 30, 45, and 60 min after exercise, respectively, with a cool-down and passive recovery. These results are in contrast with our hypothesis which stated the HR response would be attenuated within 60 min after exercise using a cool-down, compared to passive recovery. The increases in HR are consistent with Millar et al. (2009) who reported significant increases in HR of 87% and 43%, compared to Rest, at 5–20 and 45–60 min, respectively, during recovery from four repeated WATs with no cool-down after the intervals [4]. The results of the present study are also in agreement with Kingsley et al. (2016) who reported increases in HR of 46% at 10–15 min after three repeated WATs with no cool-down after the intervals [14]. Specifically, Kingsley et al. (2016) had participants return to the supine position within one minute of completing their third and final WAT, compared to five min in the present study [14]. The significant increases in HR during recovery from an acute bout of supramaximal interval exercise demonstrated in these studies indicate a loss of vagal modulation. Data from the present study demonstrate that the use of a cool-down after supramaximal interval exercise does not attenuate HR during recovery compared to passive recovery. Collectively, these data demonstrate that an acute bout of supramaximal interval exercise results in the augmentation of HR for up to 60 min after exercise with or without a cool-down. Our study is the first to examine the effects of a cool-down and supramaximal interval exercise on autonomic modulation in moderately active individuals. Our data demonstrate that supramaximal interval exercise decreases linear measures of vagal modulation, demonstrated by the attenuation of lnHF and lnRMSSD for up to 60 min after acute supramaximal interval exercise, compared to Rest, with a cool-down and passive recovery. These results are in partial agreement with our hypothesis which stated linear measures of vagal modulation would be attenuated for up to 60 min, but would be augmented within 60 min after exercise using a cool-down, compared to passive recovery. There were no significant differences between the conditions. However, the decrease in lnHF of 51%, 41%, 29% and 21% at 15, 30, 45 and 60 min, respectively, compared to Rest, after supramaximal interval exercise in the present study, are in agreement with Millar et al. (2009) who reported decreases in lnHF of 95% and 51%, compared to Rest, at 5–20 and 45–60 min, respectively, after four repeated WATs [4]. The difference in the magnitude of change in lnHF observed between these studies support the findings of Millar et al. (2009) who reported a dose–response relationship between the total volume of supramaximal work and vagal modulation [4]. The study by Millar et al. (2009) also included data from 5–10 min after exercise, which may include a time period of greater vagal withdrawal, compared to 15 min after exercise in the present study. Data from the present study revealed no differences in the lnRMSSD between conditions, but demonstrated a significant decrease of 49%, 38%, 27% and 22%, compared to Rest, at 15, 30, 45 and 60 min, respectively, compared to Rest. To our knowledge, the present study is the first to report lnRMSSD data after acute supramaximal interval exercise. Perkins et al. (2015) reported a significant decrease in RMSSD immediately after supramaximal interval exercise [6]. However, these data were not reported as log transformed and cannot be directly compared with data from the present study. The significant decreases in linear measures of vagal modulation reported in the present study reflect an increase in vagal withdrawal. Taken together, the data from the present study demonstrate that acute supramaximal interval exercise, with and without a cool-down, decreases linear measures of vagal modulation for up to 60 min. Our data demonstrated significant increases in the ln(LF/HF) ratio at 15, 30, 45 and 60 min, respectively, after an acute bout of supramaximal interval exercise, compared to Rest, with or without a cool-down. The findings of the present study support the work of Millar et al. (2009) who also reported significant increases in the ln(LF/HF) ratio compared to Rest for up to 60 min after four repeated WATs [4]. Decreases in lnLF and lnHF after exercise were reported in the present study and by Millar et al. (2009), supporting the concept that the LF/HF ratio is a measure of sympathovagal dominance as opposed to balance [16]. In both studies, the magnitude of vagal withdrawal was greater than the decrease in lnLF, mediating the increase in the LF/HF ratio. Collectively, these data suggest that an acute bout of supramaximal interval exercise augments sympathovagal dominance for up to 60 min after exercise, regardless of whether a cool-down is utilized. Data from the present study demonstrate that supramaximal interval exercise, with passive recovery, decreases nonlinear measures of vagal modulation, as demonstrated by the attenuation of SampEn and LZEn for up to 60 min after acute supramaximal interval exercise compared to Rest. However, the use of a cool-down after exercise resulted in the recovery of SampEn and LZEn within 60 and 45 min, respectively, which supported our hypothesis. SampEn and LZEn are measures of the complexity of vagal modulation. Data from the present study demonstrate decreases in SampEn with a passive recovery of 23%, 24%, 25% and 20% at 15, 30, 45 and 60 min after exercise, respectively. The attenuation of SampEn when utilizing passive recovery in the present study supports the findings of Millar et al. (2009), who demonstrated a significant decrease in SampEn of 52% and 40% at 5–20 and 45–60 min, respectively, compared to Rest, after four repeated WATs, in recreationally active men [4]. Data from the present study demonstrated that when utilizing a cool-down, there were significant decreases in SampEn of 29%, 30% and 14% at 15, 30 and 45 min, respectively, with recovery at 60 min after exercise. To our knowledge, the present study is the first to report changes in LZEn after supramaximal interval exercise. Data demonstrated significant decreases in LZEn of 30%, 28%, 24% and 19% at 15, 30, 45 and 60 min, respectively, when utilizing passive recovery. However, when utilizing a cool-down, LZEn decreased by 30% at 15 and 30 min, but recovered within 45 min after exercise. The early recovery of vagal complexity with the use of a cool-down in the present study may reflect the decreased metabolic demand, as demonstrated by significantly lower blood lactate levels during and after exercise compared to passive recovery. Taken together, these data suggest that supramaximal interval exercise attenuates the complexity of vagal modulation for up to 60 min with passive recovery, but utilizing a cool-down results in the recovery of vagal modulation within 45–60 min. This is important because transient decreases in vagal modulation after vigorous exertion may result in increased cardiovascular risk [7,8]. As expected, there were significant increases in blood lactate levels for up to 60 min after exercise, compared to Rest under both conditions, with significantly lower levels when a cool-down was utilized. The increases in blood lactate in the present study are consistent with the findings of Stuckey et al. (2012) who demonstrated a significant increase in blood lactate concentration in recreationally active men for up to 60 min after four repeated WATs [5]. The performance of a cool-down was demonstrated to accelerate lactate clearance after exercise [27]. It is well known that the enhanced buffering capacity during light-to-moderate exercise resulting from cardiovascular activation improves lactate clearance. There were some limitations to the present study. The results of this study may be limited to moderately active individuals. The prescription of supramaximal interval exercise may limit the ability to generalize the results in comparison to modalities performed at lower intensities. Supramaximal interval exercise may also be performed at an intensity that is too high for some individuals with cardiovascular or metabolic disease. In addition, although inclusionary criteria for participation in this study were met using validated questionnaires, responses were self-reported. 5. Conclusions These data demonstrate that the use of a cool-down after an acute bout of supramaximal interval exercise results in the early recovery of nonlinear measures of vagal modulation, indicating an increase in vagal modulation within 45–60 min after exercise when compared to passive recovery. These data also demonstrate that an acute bout of supramaximal interval exercise results in significant increases in HR and sympathovagal dominance, with concomitant decreases in linear measures of vagal modulation for at least 60 min after exercise compared to Rest, with or without a cool-down. Future research should increase the duration of the cool-down protocol after vigorous exercise to determine whether a longer cool-down period facilitates the recovery of linear measures of vagal modulation and accelerates the recovery of nonlinear measures of vagal modulation. Acknowledgments The authors would like to thank the participants who volunteered for this investigation. Author Contributions Conceptualization, J.C.P. and J.D.K.; data curation, J.C.P., E.M.M., S.M.H., E.K.E. and J.D.K.; formal analysis, J.C.P., E.M.M., S.M.H., E.K.E. and J.D.K.; project administration, J.D.K.; supervision, J.D.K.; writing—original draft, J.C.P. and J.D.K.; writing—review and editing, J.C.P., E.M.M., S.M.H., E.K.E. and J.D.K. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement This study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of Kent State University (protocol 19–167 date of approval 24 March 2019). Informed Consent Statement Written informed consent was obtained from all subjects involved in the study. Data Availability Statement Data presented in this study are available upon request from the corresponding author. The data are not publicly available due to privacy/ethical concerns. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Experimental trial timeline. WAT = Wingate anaerobic test. AR = active recovery. Figure 2 Sample entropy and Lempel–Ziv entropy at Rest and 15, 30, 45 and 60 min during recovery in moderately active individuals. Values are mean ± SD. * Significantly different from Rest (p ≤ 0.05), † Significantly different from 30 min (p ≤ 0.05). ijerph-19-05407-t001_Table 1 Table 1 Participant characteristics (n = 16). Indicators Age (y) 23 ± 3 Height (m) 1.76 ± 0.10 Weight (kg) 74.6 ± 13.7 BMI (kg·m2) 24.0 ± 2.7 Data presented are mean ± SD. BMI = body mass index. ijerph-19-05407-t002_Table 2 Table 2 Heart rate, autonomic variables, and blood lactate at Rest and during recovery from supramaximal interval exercise in moderately active individuals (n = 16). Cool Down Passive Recovery Rest 15 min 30 min 45 min 60 min Rest 15 min 30 min 45 min 60 min Heart rate (bpm) 58 ± 8 91 ± 9 * 81 ± 9 * 73 ± 9 * 71 ± 8 * 59 ± 8 90 ± 10 * 82 ± 10 * 75 ± 9 * 72 ± 10 * Total power (ln ms2) 8.5 ± 0.9 5.8 ± 0.9 * 6.9 ± 0.9 * 7.5 ± 1.0 * 7.6 ± 1.0 * 8.8 ± 0.9 5.9 ± 1.1 * 6.9 ± 1.2 * 7.7 ± 1.1 * 7.8 ± 0.9 * LF (ln ms2) 6.8 ± 0.6 4.2 ± 0.7 * 5.4 ± 0.9 * 6.1 ± 1.2 * 6.2 ± 0.8 * 7.3 ± 0.9 4.1 ± 0.9 * 5.6 ± 1.2 * 6.1 ± 0.7 * 6.2 ± 0.6 * HF (ln ms2) 8.0 ± 1.0 3.8 ± 1.3 * 4.8 ± 1.8 * 5.8 ± 1.6 * 6.4 ± 1.4 * 8.0 ± 1.1 4.1 ± 1.4 * 4.7 ± 1.6 * 5.7 ± 1.6 * 6.1 ± 1.5 * LF/HF ratio (ln) 3.5 ± 0.6 5.1 ± 0.9 * 5.2 ± 1.2 *‡ 4.9 ± 1.1 * 4.4 ± 1.0 * 3.9 ± 0.7 4.7 ± 1.0 * 5.5 ± 1.1 * 5.1 ± 1.1 * 4.7 ± 1.1 * RMSSD (ln ms) 4.5 ± 0.5 2.3 ± 0.7 * 2.9 ± 0.8 * 3.4 ± 0.8 * 3.6 ± 0.7 * 4.5 ± 0.6 2.3 ± 0.9 * 2.8 ± 0.8 * 3.3 ± 0.8 * 3.6 ± 0.8 * Blood lactate (mmol) 0.6 ± 0.2 11.1 ± 1.8 *# 6.7 ± 1.7 *# 4.2 ± 1.1 *# 2.8 ± 0.8 *# 0.7 ± 0.3 13.0 ± 2.0 * 8.3 ± 2.2 * 5.2 ± 1.4 * 3.3 ± 1.0 * Data presented are mean ± SD. * Significantly different from Rest (p ≤ 0.05), ‡ Significantly different from 15 min (p ≤ 0.05), # Significantly different from passive recovery (p ≤ 0.05). HF = high-frequency power; LF = low-frequency power; RMSSD = root mean square of successive differences between normal heart beats. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Rakobowchuk M. Tanguay S. Burgomaster K.A. Howarth K.R. Gibala M.J. MacDonald M.J. Sprint interval and traditional endurance training induce similar improvements in peripheral arterial stiffness and flow-mediated dilation in healthy humans Am. J. Physiol. Regul. Integr. Comp. Physiol. 2008 295 R236 R242 10.1152/ajpregu.00069.2008 18434437 2. Sugawara J. Komine H. Miyazawa T. Imai T. Ogoh S. Influence of single bout of aerobic exercise on aortic pulse pressure Eur. J. Appl. Physiol. 2015 115 739 746 10.1007/s00421-014-3061-0 25428726 3. Rodas G. Ventura J.L. Cadefau J.A. Cussó R. Parra J. A short training programme for the rapid improvement of both aerobic and anaerobic metabolism Eur. J. Appl. Physiol. 2000 82 480 486 10.1007/s004210000223 10985604 4. Millar P.J. Rakobowchuk M. McCartney N. MacDonald M.J. Heart rate variability and nonlinear analysis of heart rate dynamics following single and multiple Wingate bouts Appl. Physiol. Nutr. Metab. 2009 34 875 883 10.1139/H09-086 19935849 5. Stuckey M.I. Tordi N. Mourot L. Gurr L.J. Rakobowchuk M. Millar P.J. Toth R. MacDonald M.J. Kamath M.V. Autonomic recovery following sprint interval exercise Scand. J. Med. Sci. Sports 2012 22 756 763 10.1111/j.1600-0838.2011.01320.x 21535187 6. Perkins S.E. Jelinek H.F. Al-Aubaidy H.A. de Jong B. Immediate and long term effects of endurance and high intensity interval exercise on linear and nonlinear heart rate variability J. Sci. Med. Sport 2017 20 312 316 10.1016/j.jsams.2016.08.009 27568073 7. Albert C.M. Mittleman M.A. Chae C.U. Lee I.M. Hennekens C.H. Manson J.E. Triggering of sudden death from cardiac causes by vigorous exertion N. Engl. J. Med. 2000 343 1355 1361 10.1056/NEJM200011093431902 11070099 8. Tuzcu V. Nas S. Borklu T. Ugur A. Decrease in the heart rate complexity prior to the onset of atrial fibrillation Europace 2006 8 398 402 10.1093/europace/eul031 16687424 9. Vikman S. Mäkikallio T.H. Yli-Mäyry S. Pikkujämsä S. Koivisto A.-M. Reinikainen P. Airaksinen K.J. Huikuri H.V. Altered complexity and correlation properties of RR interval dynamics before the spontaneous onset of paroxysmal atrial fibrillation Circulation 1999 100 2079 2084 10.1161/01.CIR.100.20.2079 10562264 10. Riebe D. Ehrman J. Liguori G. Magal M. American College of Sports Medicine’s Guidelines for Exercise Testing and Prescription 10th ed. Wolters Kluwer Philadelphia, PA, USA 2017 11. Takahashi T. Okada A. Hayano J. Tamura T. Influence of cool-down exercise on autonomic control of heart rate during recovery from dynamic exercise Front. Med. Biol. Eng. 2002 11 249 259 10.1163/156855701321138914 12735426 12. Forjaz C.L. Cardoso C.G. Jr. Rezk C.C. Santaella D.F. Tinucci T. Postexercise hypotension and hemodynamics: The role of exercise intensity J. Sports Med. Phys. Fit. 2004 44 54 62 13. Rakobowchuk M. Stuckey M.I. Millar P.J. Gurr L. Macdonald M.J. Effect of acute sprint interval exercise on central and peripheral artery distensibility in young healthy males Eur. J. Appl. Physiol. 2009 105 787 795 10.1007/s00421-008-0964-7 19125283 14. Kingsley J.D. Tai Y.L. Vaughan J.A. Mayo X. High-intensity interval cycling exercise on wave reflection and pulse wave velocity J. Strength Cond. Res. 2016 31 1313 1320 10.1519/JSC.0000000000001598 15. Ainsworth B.E. Jacobs D.R. Jr. Leon A.S. Validity and reliability of self-reported physical activity status: The Lipid Research Clinics questionnaire Med. Sci. Sports Exerc. 1993 25 92 98 10.1249/00005768-199301000-00013 8423761 16. Billman G.E. The LF/HF ratio does not accurately measure cardiac sympatho-vagal balance Front. Physiol. 2013 4 26 10.3389/fphys.2013.00026 23431279 17. Malik M. Heart rate variability: Standards of measurement, physiological interpretation, and clinical use: Task force of the European society of cardiology and the north American society for pacing and electrophysiology Ann. Noninvasive Electrocardiol. 1996 1 151 181 10.1111/j.1542-474X.1996.tb00275.x 18. Richman J.S. Moorman J.R. Physiological time-series analysis using approximate entropy and sample entropy Am. J. Physiol. Heart Circ. Physiol. 2000 278 H2039 H2049 10.1152/ajpheart.2000.278.6.H2039 10843903 19. Zhang X.-S. Roy R.J. Jensen E.W. EEG complexity as a measure of depth of anesthesia for patients IEEE Trans. Biomed. Eng. 2001 48 1424 1433 10.1109/10.966601 11759923 20. Lempel A. Ziv J. On the complexity of finite sequences IEEE Trans. Biomed. Eng. 1976 22 75 81 10.1109/TIT.1976.1055501 21. Kanaley J. Goulopoulou S. Franklin R. Baynard T. Holmstrup M. Carhart Jr R. Weinstock R. Fernhall B. Plasticity of heart rate signalling and complexity with exercise training in obese individuals with and without type 2 diabetes Int. J. Obes. 2009 33 1198 10.1038/ijo.2009.145 22. Tanner R.K. Fuller K.L. Ross M.L. Evaluation of three portable blood lactate analysers: Lactate Pro, Lactate Scout and Lactate Plus Eur. J. Appl. Physiol. 2010 109 551 559 10.1007/s00421-010-1379-9 20145946 23. Bar-Or O. The Wingate anaerobic test. An update on methodology, reliability and validity Sports Med. 1987 4 381 394 10.2165/00007256-198704060-00001 3324256 24. Benjamini Y. Hochberg Y. Controlling the false discovery rate: A practical and powerful approach to multiple testing J. R. Stat. Soc. Ser. B 1995 57 289 300 10.1111/j.2517-6161.1995.tb02031.x 25. Keppel G. Design and Analysis: A Researcher’s Handbook 4th ed. Prentice-Hall, Inc. Upper Saddle River, NJ, USA 1991 26. Faul F. Erdfelder E. Lang A.G. Buchner A. G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences Behav. Res. Methods 2007 39 175 191 10.3758/BF03193146 17695343 27. Martin N.A. Zoeller R.F. Robertson R.J. Lephart S.M. The comparative effects of sports massage, active recovery, and rest in promoting blood lactate clearance after supramaximal leg exercise J. Athl. Train. 1998 33 30 35 16558481

PMC009xxxxxx/PMC9099608.txt

==== Front Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells11091497 cells-11-01497 Article Effects of Epigenetic Modification of PGC-1α by a Chemical Chaperon on Mitochondria Biogenesis and Visual Function in Retinitis Pigmentosa https://orcid.org/0000-0003-4797-5705 Ozawa Yoko 1234* Toda Eriko 34 https://orcid.org/0000-0002-6253-8047 Homma Kohei 34 Osada Hideto 34 Nagai Norihiro 1234 https://orcid.org/0000-0002-8874-7111 Tsubota Kazuo 4 https://orcid.org/0000-0001-7482-5935 Okano Hideyuki 5 Chartier-Harlin Marie-Christine Academic Editor Kalyuzhny Alexander E. Academic Editor 1 Department of Ophthalmology, St. Luke’s International Hospital, 9-1 Akashi-cho, Chuo-ku, Tokyo 104-8560, Japan; nagai@a5.keio.jp 2 Laboratory of Retinal Cell Biology, St. Luke’s International University, 9-1 Akashi-cho, Chuo-ku, Tokyo 104-8560, Japan 3 Laboratory of Retinal Cell Biology, Department of Ophthalmology, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan; riko.4267@gmail.com (E.T.); hommak@keio.jp (K.H.); 49hidet00sada@gmail.com (H.O.) 4 Department of Ophthalmology, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan; tsubota@tsubota-lab.com 5 Department of Physiology, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan; hidokano@keio.jp * Correspondence: ozawa@a5.keio.jp or ozaway@luke.ac.jp; Tel.: +81-33541-5151 29 4 2022 5 2022 11 9 149714 2 2022 27 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Retinitis pigmentosa (RP) is a hereditary blinding disease characterized by gradual photoreceptor death, which lacks a definitive treatment. Here, we demonstrated the effect of 4-phenylbutyric acid (PBA), a chemical chaperon that can suppress endoplasmic reticulum (ER) stress, in P23H mutant rhodopsin knock-in RP models. In the RP models, constant PBA treatment led to the retention of a greater number of photoreceptors, preserving the inner segment (IS), a mitochondrial- and ER-rich part of the photoreceptors. Electroretinography showed that PBA treatment preserved photoreceptor function. At the early point, ER-associated degradation markers, xbp1s, vcp, and derl1, mitochondrial kinetic-related markers, fis1, lc3, and mfn1 and mfn2, as well as key mitochondrial regulators, pgc-1α and tfam, were upregulated in the retina of the models treated with PBA. In vitro analyses showed that PBA upregulated pgc-1α and tfam transcription, leading to an increase in the mitochondrial membrane potential, cytochrome c oxidase activity, and ATP levels. Histone acetylation of the PGC-1α promoter was increased by PBA, indicating that PBA affected the mitochondrial condition through epigenetic changes. Our findings constituted proof of concept for the treatment of ER stress-related RP using PBA and revealed PBA’s neuroprotective effects, paving the way for its future clinical application. retinitis pigmentosa rhodopsin ER stress PGC-1α mitochondria the Japan Society of the Promotion of Science21K09683 This research was funded in part by a Grants-in-Aid for Scientific Research by the Japan Society of the Promotion of Science to Y.O. (21K09683). ==== Body pmc1. Introduction Recent progress in medical science has led to the development of new therapeutic approaches in various fields; this is also applicable for blinding diseases such as age-related macular degeneration and diabetic retinopathy [1]. However, there are currently no definitive treatments for retinitis pigmentosa (RP), a hereditary blinding disease. RP is a leading cause of blindness with a prevalence of 1:4000 [2,3], affecting more than 1.5 million persons worldwide [3]. While the associated gene mutations are congenital, visual disability progresses with age in patients with RP, and the aging society will lead to increasing numbers of affected patients, rendering RP an urgent social issue. Rhodopsin is an evolutionarily conserved, seven-transmembrane protein, specifically produced in the rod photoreceptor cells, and acts as a visual pigment; rhodopsin mutations cause RP [4]. Among them, the P23H rhodopsin, which has a substitution of proline to histidine at position 23, is the most common cause of human autosomal dominant RP in the USA [5]. The P23H mutant rhodopsin protein has been classified as a Class II mutant causing protein misfolding, leading to disordered protein trafficking in the endoplasmic reticulum (ER) and ER stress [6,7]. More recently, autosomal dominant RPs caused by rhodopsin mutations are classified into four clusters, and P23H mutation is involved in cluster two [8]. Under ER-stress conditions, the unfolded protein response is triggered to promote adaptive alteration requiring energy expenditure. The signaling induces folding enzymes and chaperones, i.e., components of the ER-associated degradation (ERAD) machinery, to clear unfolded proteins and restore ER homeostasis. However, when ER stress exceeds the capacity of the adaptive actions, cellular function deteriorates, often leading to cell death, inducing C/EBP homologous protein (CHOP). Thus, several therapeutic approaches have been proposed for treating ER stress-related RP. Using induced pluripotent stem cells derived from a patient with RP with a rhodopsin mutation, we previously reported that rapamycin and other ER-stress inhibitors could attenuate rod photoreceptor death [7], although the in vivo effect was obscure. Others have reported that gene administration of binding immunoglobulin protein, BiP, an ER-localized chaperone that acts as a sensor of ER stress, improved photoreceptor survival and visual function in P23H mutant animals [9]. However, gene therapy cannot cover the whole area of the retina and could be invasive. Here, we focused on 4-phenylbutyric acid (PBA), a chemical chaperon that can suppress ER stress [10,11]. PBA is a Food and Drug Administration-approved drug used to treat urea-cycle disorders, as it also acts as an ammonia scavenger [11,12]. The mechanisms of PBA action on ER-stress regulation are reported to involve its action as a chaperon, which increases the solubility of the abnormal protein [11,13,14], and/or as a transcriptional regulator of molecules involved in the ER quality control system [14]. Alternatively, its action as a histone deacetylase (HDAC) inhibitor may stabilize the transcriptional activity of spliced XBP1 [15], a key regulator of ER stress and ERAD, leading to recovery from stress [16,17]. In fact, our previous report showed that ER-stress markers were suppressed by PBA, resulting in photoreceptor protection in the retina of an acute photo-stress model [18]. However, PBA could have various roles as a chaperon [11], and the current study was designed to show the probable roles and effects of PBA in RP treatment. In this study, we treated human P23H knock-in mice exhibiting progressive rod degeneration with daily PBA administration to show PBA’s neuroprotective effects and probable mechanisms and propel the future clinical application of PBA for retinal neuroprotection. This minimally invasive approach that uses drugs to slow degeneration and visual disability could be applicable to a large number of patients with RP in the future. 2. Materials and Methods 2.1. Animals The P23H rhodopsin knock-in mice backcrossed to C57B6J background (provided by Dr. Palczewski, Case-Western University, Cleveland, OH, USA) [19,20] were maintained under a 12-h light/dark cycle (lights on from 8 a.m. to 8 p.m.) with free access to food and water, in an air-conditioned room (22 °C) at the animal facility of Keio University School of Medicine. Male animals were used for the study and intraperitoneally treated either with sodium phenylbutyric acid (PBA, SML0309, Sigma-Aldrich, St. Louis, MO, USA) at 10 mg/kg body weight (BW) or with phosphate-buffered saline (PBS) as a vehicle, five times a week from 2 weeks of age. All animal experiments were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and the guidelines of the Animal Care Committee of Keio University (Approval number; 09203). 2.2. Histological Analyses Mouse eyes were enucleated and fixed in 4% paraformaldehyde. Paraffin (Sakura Finetek, Tokyo, Japan) sections (8 μm), including the optic nerve head to the most peripheral region of the superior and inferior retina, were deparaffinized, followed by staining with hematoxylin and eosin. Sections were examined under a microscope equipped with a digital camera (Olympus Co., Tokyo, Japan). For immunohistochemical analyses, sections were incubated with anti-Tom20 (Santa Cruz Biotechnology, Santa Cruz, CA, USA; #sc-11415) at 1:500 and anti-rhodopsin (Thermo Fisher Scientific, Waltham, MA, USA; #MA1-722) at 1:10,000 antibodies; signals were obtained using Alexa 488-conjugated goat anti-mouse and Alexa 555-conjugated goat anti-rabbit IgGs, respectively. Nuclei were stained with Cellstain-4’,6-diamidino-2-phenylindole solution (Dojindo Molecular Technologies, Kumamoto, Japan, 2 μg/mL). Fluorescent images were obtained using a confocal microscope (TCS-SP5; Leica, Tokyo, Japan). The number of cells in the outer nuclear layer (ONL), the photoreceptor layer, and quantification of inner segment (IS) areas in a 50 μm length of the retina was evaluated at each point using ImageJ software (National Institutes of Health, Bethesda, MD, USA; available at http://rsb.info.nih.gov/ij/index.html (accessed 27 September 2021) and averaged as described previously [21,22,23,24]; data at 200 μm distance from the optic nerve in the superior retina were shown. 2.3. Real-Time Reverse Transcription-Polymerase Chain Reaction (RT-PCR) Total RNA in the neural retina was isolated using TRIzol reagent (Life Technologies, Carlsbad, CA, USA) and reverse transcribed by the SuperScript VILO master mix (Life Technologies). Real-time PCR was performed using the StepOnePlus™ PCR system (Applied Biosystems, Foster City, CA, USA), and gene expression levels were quantified by the 2−ΔΔCT method and normalized to the expression of Gapdh. Primers are listed in Supplementary Table S1. 2.4. Electroretinography (ERG) Recordings Mice were dark-adapted for at least 12 h before conducting the ERGs modifying the methods previously described [21,25,26,27]. Briefly, mice were anesthetized with intraperitoneal combined anesthetics [midazolam 4 mg/kg of BW (Sandoz Japan, Tokyo, Japan), medetomidine 0.75 mg/kg BW (Nippon Zenyaku Kogyo Co., Ltd., Fukushima, Japan), butorphanol tartrate 5 mg/kg BW (Meiji Seika Pharma Co., Ltd., Tokyo, Japan)] and maintained on a heating pad under dim-red illumination throughout the experiment. Mouse pupils were dilated by a single eye drop of a mixture of tropicamide and phenylephrine (0.5% each; Mydrin-P®; Santen, Osaka, Japan). The respective ground and reference electrodes were placed on the tail and in the mouth, and the active gold wire electrodes were on the cornea. Recordings were acquired using a PowerLab System 2/25 (AD Instruments, Bella Vista, New South Wales, Australia). Each response was differentially amplified and filtered through a digital bandpass filter ranging from 0.3 to 1000 Hz. Each stimulus was delivered using a commercial stimulator (Ganzfeld System SG-2002; LKC Technologies, Inc., Gaithersburg, MD, USA). Full-field scotopic ERGs were recorded in response to a flash stimulus at intensities ranging from −2.1 to 2.9 log cd s/m2. Photopic ERGs were recorded after 10 min of light adaptation at flash stimuli ranging from 0.4 to 1.4 log cd s/m2 with a background of 30 cd s/m2. The a-wave amplitude was measured from the baseline to the trough, and the b-wave amplitude was from the trough of the a-wave to the peak of the b-wave. The implicit times of the a and b waves were measured from the onset of the stimulus to the peak of each wave. The automatically indicated peak points by the system were confirmed by the examiner. 2.5. Cell Culture HEK293 cells (EC85120602−G0, European Collection of Authenticated Cell Cultures, Public Health England) were maintained in Dulbecco’s modified Eagle’s medium (#08459-35; Nacalai tesque, Kyoto, Japan) supplemented with 10% fetal bovine serum (Life Technologies), 100 unit/mL penicillin, and 100 μg/mL streptomycin (Nacalai tesque) at 37 °C in a humidified atmosphere of 5% CO2. For passaging, cells were dissociated with 0.25% trypsin/EDTA (Thermo Fisher Scientific, Waltham, MA, USA). 2.6. Mitochondrial Membrane Potential Measurement HEK293 cells were plated on glass-based dishes (Non-coat 35 mm glass-bottom dish, D11530H, Matsunami Glass, Osaka, Japan) a day before the experiment. Cells were treated with or without 2.5 mM PBA in the culture medium for 24 h. On the day of the experiment, cells were stained with tetramethylrhodamine ethyl ester (TMRE, Thermo Fisher Scientific, Waltham, MA, USA; #T669) at 10 μM. Fluorescent images were obtained every 2 s using a fluorescent microscope (BZ-X710, KEYENCE, Osaka, Japan) with a TRITC filter (OP-87764, KEYENCE). During the imaging, (2-fluorophenyl) 6-[(2-fluorophenyl) amino] (1,2,5-oxadiazolo[3,4-e]pyrazin-5-yl) amine (BAM15, SML1760, Sigma-Aldrich, St. Louis, MO, USA) was added to the recording dish at 10 μM. 2.7. Cytochrome c Oxidase (CcO) Activity Measurement HEK293 cells were treated with or without 2.5 mM PBA in the culture medium for 24 h. Then, the treated cells were incubated in the buffer in the kit before measuring the CcO activity using the Complex IV Rodent Enzyme Activity Microplate Assay Kit (Abcam, Cambridge, UK) according to the manufacturer’s instructions and our previous work [21]. The luminescent signals were measured using the Cytation 5 system (BioTek, Winooski, VT, USA). 2.8. ATP Measurement HEK293 cells were treated with or without 2.5 mM PBA in the culture medium for 24 h. After treatment, cells S2 were recovered from the culture dish with TE-saturated phenol (NIPPON GENE Co., Ltd., Tokyo, Japan). The ATP levels in the samples were measured using the ATP Bioluminescence Assay Kit CLSII (Sigma-Aldrich, St. Louis, MO, USA) according to the manufacturer’s instructions and our previous work [21]. The luminescent signals were measured using the Cytation 5 system (BioTek, Winooski, VT, USA). 2.9. Chromatin Immunoprecipitation Quantitative Real-Time PCR (ChIP-qPCR) HEK293 cells were treated with or without 2.5 mM PBA in the culture medium for 24 h. After the treatment, cells were fixed with 16% formaldehyde (28906, Thermo Fisher Scientific, Waltham, MA, USA) for 15 min at room temperature. Fixed and fractionated nuclei from HEK293 cells were resuspended in ChIP buffer from ChIP-IT High Sensitivity Kit (Active Motif cat #53040) and sonicated in microtube AFA Fiber Pre-slit Snap-Cap set on Covaris S2 instrument (Covaris, Woburn, MA, USA) at 5% duty cycle, the intensity of 4, and 200 cycles per burst for 16 min. For the immunoprecipitation, ChIP-IT High Sensitivity Kit (#53040, Active Motif, Inc., Carlsbad, CA, USA) and histone H3K27ac antibody (#39133, Active Motif) was used according to the manufacturer’s instructions. After the purification of DNA, the precipitated fraction of DNA was quantified by qPCR by using Fast SYBR Green Master Mix (#4385614, Thermo Fisher Scientific, Waltham, MA, USA) and StepOnePlus™ PCR system (Applied Biosystems). The fold-change of % inputs at the probes to that at GAPDH promoter in each sample was quantified by Ct values of the probes. Primers of the probes are listed in Supplementary Table S2. 2.10. Statistical Analysis All results are expressed as the mean ± standard deviation. The values were processed for statistical analyses using one-way analysis of variance with Tukey’s post hoc tests for comparisons among three groups or two-tailed Student’s t-tests for comparisons between two groups using SPSS Statistics 24 (IBM, Armonk, Hamlet, NY, USA). Differences were considered statistically significant at p < 0.05. 3. Results 3.1. PBA Promoted Photoreceptor Survival in P23H Knock-In Heterozygotes (P23H RP Models) A previous report showed that the thickness of the photoreceptor layer does not change in P23H knock-in heterozygous retinitis pigmentosa models (P23H RP models), compared with wild-type, until postnatal day 12 and that degeneration gradually progresses thereafter [19]. Thus, continuous administration of either PBA or the control vehicle was started at the age of 2 weeks in the current study (Figure 1A). Consequently, the remaining photoreceptors, as assessed by the number of photoreceptor nuclei, were significantly greater at the age of 10 weeks in the retinas of P23H RP models treated with PBA than in those treated with vehicle; at that time, photoreceptor loss was significant in the P23H RP models (Figure 1B–D). The remaining area of the photoreceptor IS, where mitochondria, labeled with Tom20 and ER, are densely distributed, was also greater in the PBA-treated mice (Figure 1B,C,E). These results indicated that PBA promoted photoreceptor survival and suppressed IS degeneration in the P23H RP models. The effects of PBA were significantly observed in the superior part of the retina (Figure S1A,B). 3.2. PBA Preserved Visual Function in the P23H RP Models Impairment of rod photoreceptor function measured with scotopic ERG is already found at the age of 6 weeks, and cone function measured with photopic ERG is found at 10 weeks, both followed by a gradual progress in the P23H RP models as previously reported [20]. Scotopic (Figure 2A–E) and photopic (Figure 2F–H) ERGs were measured to examine whether histological rescue by PBA treatment would contribute to the retention of visual function. In the scotopic ERG, a-wave amplitudes reflecting rod photoreceptor function were greater, and the respective b-wave amplitudes and implicit times reflecting subsequent retinal neural function to photoreceptors were greater and shorter in the PBA-treated P23H RP models than in vehicle-treated models at the age of 10 weeks, indicating that the function of the rod system was partly retained by continuous PBA treatment. Respective greater amplitudes of a and b waves after PBA treatment had most likely reflected the number of survived rod photoreceptors, and shorter implicit time of b-wave may have resulted from the better synaptic function of the remaining rod system. In b-wave photopic ERG, which mainly shows cone system function, the implicit time was shorter after treatment with PBA, indicating that the cone system was also partly protected by PBA treatment in the P23H RP models. 3.3. The Protective Effect of PBA Was Detected in the Photoreceptors of P23H RP Models before Photoreceptor Loss At 4 weeks of age, when the PBA treatment had been continued for 2 weeks, there was no difference in the number of remaining photoreceptor cells between the PBA- and vehicle-treated groups in the P23H RP models (Figure 3A and Figure S2). However, the IS area, evaluated with Tom20 staining in IS, exhibited a difference between the groups with or without PBA treatment (Figure 3A,A’,B). The data suggested that mitochondrial damage may have been attenuated by PBA in the models. At that time, the mRNA levels of photoreceptor markers, rhodopsin (Figure 3C), and crx, a transcription factor upstream of rhodopsin (Figure 3D), were at significantly higher levels in the retina of PBA-treated P23H RP models, indicating that PBA treatment retained visual pigment expression in the rod photoreceptor cells compared with vehicle treatment. 3.4. PBA Induced ERAD and Mitochondrial Markers in the Retina of P23H RP Models While P23H mutated rhodopsin causes ER stress [19,28,29,30,31,32,33], the ERAD system that disposes of abnormal proteins [16,17] is also induced in the models [19]. ERAD is activated by IRE1-related conversion of XBP1 to XBP1s, leading to induction of VCP (also known as Cdc48 or p97) interacting with Derlin 1 to transport the specific misfolded proteins from the ER to the cytosol utilizing its adenosine triphosphate (ATP) ase activity [34]. The misfolded protein delivered to the cytosol is processed to be degraded through the ubiquitin proteasome system (UPS) [16,17]. We found that PBA treatment upregulated xbp1s (Figure 4A), vcp (Figure 4B), and derlin 1 (Figure 4C) in the retina of P23H RP models at the age of 4 weeks. These results suggested that PBA promoted ERAD induction, which could eliminate pathological P23H rhodopsin. ER stress also affects mitochondrial quality control by regulating mitochondrial fission and fusion by VCP [34,35]; VCP induces fission to eliminate abnormal mitochondria through autophagy [35,36], whereas it induces degradation of fusion-related mitofusin proteins through UPS [34,35] to preserve cellular homeostasis [37]. The system is reported to be involved in neural plasticity and survival [38]. We found that PBA treatment increased the mRNA levels of the mitochondrial fission marker, fis1 (Figure 4D), and an autophagy marker, lc3b (Figure 4E), in the retina of P23H RP models. The mRNA levels of fusion markers, mfn1 (Figure 4F) and mfn2 (Figure 4G), were also affected by PBA. Mitochondrial biogenesis, which replaces damaged mitochondria with new and healthy ones [39], is affected during mitochondrial quality control [35]. We found that pgc-1α (Figure 4H), which regulates mitochondrial biogenesis [40], and tfam (Figure 4I), which represents mitochondrial DNA levels and is a transcription factor for inducing mitochondrial DNA encoding respiratory molecules [41], were both upregulated in the retina of P23H RP models treated with PBA, suggesting that mitochondrial biogenesis was activated by PBA treatment. 3.5. PBA Activated Oxidative Phosphorylation (OXPHOS) As shown above, the mitochondrial-concentrated part of the photoreceptor, IS, was preserved by PBA, and the mitochondrial biogenesis marker was upregulated by PBA in vivo. Therefore, we further analyzed the potential effects of PBA on mitochondria using a HEK293 cell line, where PBA upregulated mRNA of pgc-1α (Figure 5A) and tfam (Figure 5B) in a dose-dependent manner. The mitochondrial membrane potential, which is indispensable for ATP synthesis and assessed using the subtraction method of the potential before and after administration of a protonophore uncoupler, BAM15, was increased in the cells treated by PBA (Figure 5C,D). Moreover, the activity of complex IV in the electron transport chain, also named cytochrome c oxidase (CcO), was increased by PBA (Figure 5E), and the ATP levels were increased by PBA (Figure 5F). These results indicated that PBA could promote mitochondrial function to increase ATP levels, which is required to eliminate abnormal unfolded proteins and for cytoprotection [21,42,43]. 3.6. PBA Increased Histone Acetylation of the PGC-1α Promoter We performed ChIP-qPCR assays at promoters A and B of the PGC-1α gene [44] using HEK293 cells to explore the mechanism of pgc-1α induction by PBA. We found that the acetylation of histone (H3K27ac) was increased by PBA at promoter A as shown by probes one, two, and three (Figure 6A,B), indicating that PBA can promote pgc-1α transcription through epigenetic modification, whereas promoter B of PGC-1α may not be involved in PBA-induced upregulation of pgc-1α (Figure 6B). 4. Discussion Photoreceptor loss in the retina of P23H RP models was attenuated, and visual function was retained by constant administration of PBA. In the retina of P23H RP models, biomarkers for ERAD and mitochondrial quality control as well as mitochondrial biogenesis were upregulated. In vitro experiments showed that PBA induced pgc-1α and tfam mRNA and increased the mitochondrial membrane potential, CcO activity, and ATP levels. Histone acetylation of PGC-1α promoter A was increased by PBA. The P23H mutant rhodopsin protein forms aggregates, causing ER stress [19,28,29,30,31,32,33], and is not delivered to OS but remains in IS, where ERs exist [45]. However, a previous report showed that metformin administration, which reduces aggregates of mutant rhodopsin and allows delivery of mutant rhodopsin to the OS, did not suppress but exacerbated photoreceptor death in P23H mutant models [46]. Moreover, CHOP, which is a death signal that is induced when the cells cannot compensate for the ER stress, was not involved in photoreceptor death in the P23H RP models; a CHOP knock-out background does not improve photoreceptor survival in the models [20,47]. In addition, when treated with a PERK inhibitor to suppress one of the ER-stress pathways, the visual function was exacerbated in a P23H transgenic rat model [48], and aggregates of mutant rhodopsin were increased in cells overexpressing P23H mutant rhodopsin [48]. This could be related to the fact that PERK can activate Nrf2, an antioxidative and protective transcription factor, in P23H mutants [49]. These results indicate that simple suppression of ER-stress signals is not an effective treatment in P23H RP models. In contrast, we succeeded in demonstrating the neuroprotective effect of PBA, a chemical chaperon that is well-known as an ER-stress inhibitor [10,11] in the P23H RP models. In fact, PBA increased the ER stress-related molecule, xbp1s, which most likely increased the expression of its downstream molecules related to ERAD, such as vcp and derlin1. Thus, PBA may have promoted adaptive alteration by activating the ERAD pathway. XBP1s promote VCP-dependent translocation of unfolded proteins in an ATP-dependent manner to finally process the proteins to be degraded through UPS, which also consumes ATP [16,17]. A previous report has shown the reduction in the ratio of the insoluble protein and ER stress markers by PBA treatment [13], which could have been, at least in part, through activating ERAD, in addition to correcting the folding of misfolded and unfolded proteins [13]. Moreover, XBP1s-induced VCP also acts on mitochondrial quality control; it promotes fission and suppresses the fusion of mitochondria [16,17]. Upregulation of both fis1 and lc3b in retinas treated with PBA in P23H RP models may have increased mitochondrial fission followed by mitophagy to eliminate unhealthy mitochondria through degradation [37]. This may have been advantageous to effectively respond to the increased energy demand required for ERAD. Moreover, activation of autophagy supports suppressing ER stress [50], which could have decreased energy demand. Meanwhile, autophagy itself may not be supportive of photoreceptor protection, according to previous reports of Xenopus P23H rhodopsin mutants [51]. Increased expression of the fusion marker, mfn1, and mfn2 mRNA was most likely due to negative feedback related to the degradation of the fusion proteins by VCP [16,17]. pgc-1α and tfam, which represent mitochondrial function, and OXPHOS, for energy production [52] were also upregulated by PBA treatment. This may have resulted in preserving the IS area of the photoreceptors where mitochondria are distributed by PBA treatment. Previous reports have shown that metabolic failure was observed in the process of photoreceptor degeneration in rhodopsin-mutant Drosophila [53], and supplying sufficient energy contributes to photoreceptor survival in the retina under stress conditions [21,42,43,54]. Activation of adenosine 5′-monophosphate-activated protein kinase (AMPK) leading to energy supply by increasing ATP [21] or suppressing ATP degeneration [54,55] contributes to photoreceptor survival in mouse models. PBA may have had a role in mitochondrial quality control and effectively promoted the energy supply that was required for ERAD activation by inducing pgc-1α, a master regulator of mitochondrial biogenesis [56,57]. We demonstrated that PBA could increase promoter activation of the PGC-1α gene by histone acetylation of promoter A, which most likely upregulated OXPHOS and ATP production. A previous report showed that the MEF2-binding site of the Pgc-1α promoter could be deacetylated by HDAC5, one of the class IIa HDACs [58], leading to repression of pgc-1α transcriptional induction [59]. PBA is reported to be an inhibitor of class I and IIa HDAC [60], and our results showed that it increased the frequency of open chromatin at promoter A, where the MEF2-binding site is included. To the best of our knowledge, the current study firstly showed the epigenetic action of PBA on PGC-1α. Inhibition of excessively activated HDACs I/IIa in the degenerating retina of rd1 RP model mice, which have a loss-of-function mutation in the gene encoding for the β-subunit of rod photoreceptor cGMP phosphodiesterase-6, was reported to suppress photoreceptor death [61]; whether the effect involved PGC-1α upregulation would be an interesting point as a future study. Because the P23H mutant rhodopsin protein is expressed in the rod photoreceptor cells, rod degeneration is the dominant change; however, the cone photoreceptors are also affected [20] and were also protected by PBA in the current study. The cone photoreceptors are protected by rod-derived cone viable factor (RdCVF) [62] through regulation of glucose metabolism [63] and oxidative stress [64], and administration of RdCVF was reported to rescue the cone photoreceptors in P23H rhodopsin-mutant rats [65]. PBA would have protected the cone photoreceptors by protecting the rod photoreceptors, and a direct pathway might have also been involved. Because ER stress-mediated RP is also observed in the other mutations [66], whether PBA can be applicable to mutations other than P23H rhodopsin should also be studied in the future. Apart from RP, ER-stress-mediated pathogenesis is also reported in other fields, e.g., as a diabetic complication [67]. In addition, it will be interesting to study whether drug therapy with PBA can increase the therapeutic effect when used in combination with other treatments. In summary, constant PBA administration attenuated photoreceptor death and rod and cone photoreceptor-related visual-function impairment in P23H RP models. The neuroprotective effect of PBA may have involved ERAD activation with the ATP supply required to operate the ERAD system and self-defense system, which was supported by mitochondrial quality control induced by PBA (Figure 7). The fact that PBA promoted acetylation of a PGC-1α promoter is informative for the future application of PBA in the clinical setting and for the possible adaptation of PBA treatment to the other medical fields, although further studies are required. Acknowledgments We thank all the members of Laboratory of Retinal Cell Biology (RCB Lab) and of the Department of Ophthalmology of Keio University School of Medicine for kind assistance. We also appreciate Krzysztof Palczewski and Sanae Sakami (Case-Western University, Cleveland, OH, USA) for providing the P23H rhodopsin knock-in mice. Supplementary Materials The following are available online at: https://www.mdpi.com/article/10.3390/cells11091497/s1, Figure S1: Spider graphs showing the number of ONL nuclei and IS area at each part of the retina. Figure S2: Number of photoreceptors in the P23H knock-in heterozygotes treated with PBA five times a week from 2 weeks until 4 weeks of age was comparable to those treated with vehicle, Table S1: Primers for RT-PCR, and Table S2: Primers for ChIP-qPCR. Click here for additional data file. Author Contributions Conceptualization, Y.O.; methodology, Y.O., K.H.; validation, Y.O., K.H.; formal analysis, E.T., K.H.; investigation, E.T., K.H., H.O. (Hideto Osada); writing—original draft preparation, Y.O.; writing—review and editing, N.N., H.O. (Hideyuki Okano); supervision, K.T., H.O. (Hideyuki Okano); funding acquisition, Y.O. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement All animal experiments were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and the guidelines of the Animal Care Committee of Keio University (Approval number; 09203). Informed Consent Statement Not applicable. Data Availability Statement All data generated or analyzed during this study are included in this published article and its Supplementary Information. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Photoreceptor survival by PBA in the P23H knock-in heterozygotes (P23H RP models) (A) P23H knock-in heterozygotes (P23H RP models) started daily treatment with PBA five times a week from 2 weeks of age until the time point of analyses. (B–E) Hematoxylin-eosin staining and immunohistochemistry showed that a greater number of photoreceptors (B–D) (rhodopsin in pink), and area of IS marked by mitochondria marker, Tom20 (green) (C,E) were reduced in the P23H RP models compared with wild-type mice at 10 weeks of age; however, the levels were significantly higher in the models treated with PBA at the same time point. (D,E) Data at 200 μm distance from the optic nerve in the superior retina. RP, retinitis pigmentosa; IS, inner segment. Data are shown as mean ± standard deviation. n for WT treated with vehicle, 3; P23H RP models treated with vehicle, 8–9; P23H treated with PBA, 8. ONL, outer nuclear layer; IS, inner segments. * p < 0.05, ** p < 0.01. Scale bar, 20 μm. Figure 2 Preservation of visual function by PBA in the P23H RP models (A–H) ERG recorded at 10 weeks of age. (A) Representative wave forms of scotopic ERG from individual mice at a stimulus intensity from −2.1 to 2.9 log cd s/m2 in the P23H RP models with or without PBA treatment. Amplitudes of a-waves representing rod photoreceptor function (B) and amplitudes (D) and implicit times (E) of b-waves representing subsequent retinal cell function showed that PBA played a vision-protective role in rod photoreceptors. Representative wave forms of photopic ERG from individual mice at a stimulus intensity of 0.4 and 1.4 log cd s/m2 (F) showed that implicit time of b-wave representing cone photoreceptor function was preserved by PBA treatment (H). ERG, electroretinography. Data are shown as mean ± standard deviation. (A–E) n = 10–11. (F–H) n = 5–6. * p < 0.05. Figure 3 Protective effect of PBA in the photoreceptors of P23H RP models detected before photoreceptor loss (A,A’) Immunohistochemistry for Tom20 (green) and rhodopsin (pink) in the retina of P23H RP models at the age of 4 weeks with or without PBA treatment. Nuclei were counterstained with DAPI (blue). (A’) Magnified images of IS and OS. IS area (B), and mRNA levels of rhodopsin (C) and crx (D) in the retina were already different with or without PBA treatment. Data are shown as mean ± standard deviation. Respective n for P23H RP models treated with vehicle and PBA, (B) 4 and 5; (C) 5 and 4; (D) 10 and 10. ONL, outer nuclear layer; IS, inner segments; OS, outer segments. * p < 0.05. Scale bar, (A) 20 μm; (B) 10 μm. Figure 4 Induction of ERAD and mitochondrial marker mRNAs by PBA in the retina of P23H RP models (A–I) Real-time RT-PCR of the retinal samples obtained at the age of 4 weeks with or without PBA treatment. ERAD-related (A–C), mitochondrial kinetic (D,F,G) and autophagy (E) markers, pgc-1α (H), and tfam (I) were upregulated by PBA treatment. ERAD, ER-associated degradation. Data are shown as mean ± standard deviation. n for P23H RP models treated with vehicle and PBA, (A–C) 4–5; (D–I) 7–10. * p < 0.05, ** p < 0.01. Figure 5 OXPHOS activation by PBA (A–F) HEK 293 cells incubated with or without PBA for 24 h. RT-PCR showed that pgc-1α (A) and tfam (B) increased by PBA in a dose-dependent manner. Mitochondrial membrane potential measured using TMRE, a fluorescent dye (C) showed that PBA increased the membrane potential (C,D). BAM15, a mitochondria protonophore uncoupler, was added to cancel the membrane potential, and the subtracted values of the fluorescence intensity representing the potential were shown in the graph (D). CoxIV (CcO) activity (E) and ATP levels (F) also increased by PBA. Data are shown as mean ± standard deviation. n for vehicle and PBA groups, (A–E) 3–4; (F) 13–14. OXPHOS, oxidative phosphorylation, CcO; cytochrome c oxidase. * p < 0.05, ** p < 0.01. Scale bars; 50 μm for the images in the large panels, and 10 μm for the magnified images in the insets. Figure 6 Histone acetylation of PGC-1α promoter by PBA (A) Schematic illustration of upstream of human PGC-1α gene including promoters A and B. (B) ChIP-qPCR showed that amplified DNA levels were significantly increased with probes 1, 2, and 3, which are involved in promoter A. Data are shown as mean ± standard deviation. Respective n for control and PBA; 5 and 6. * p < 0.05, ** p < 0.01. Figure 7 Proposed roles of PBA in photoreceptor neuroprotection and visual protection P23H rhodopsin mutant protein causes misfolding and ER stress, and may have damaged mitochondria, thus weaken photoreceptors and impaired visual function. In contrast, constant PBA administration induced ERAD, and moreover, pgc-1α most likely through epigenetic change and may have affected mitochondrial quality control leading to supply ATP required to run the ERAD system, which protected photoreceptors and visual function. Protein structures (alphafold2; https://alphafold.ebi.ac.uk/ (accessed 14 February 2022) are shown at the top (left, wild-type rhodopsin; right, P23H mutant rhodopsin). IS, inner segment (where ER and mitochondria are densely distributed); OS, outer segment (where rhodopsin is concentrated). Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Morizane Y. Morimoto N. Fujiwara A. Kawasaki R. Yamashita H. Ogura Y. Shiraga F. Incidence and causes of visual impairment in Japan: The first nation-wide complete enumeration survey of newly certified visually impaired individuals Jpn. J. Ophthalmol. 2019 63 26 33 10.1007/s10384-018-0623-4 30255397 2. Hartong D.T. Berson E.L. Dryja T.P. Retinitis pigmentosa Lancet 2006 368 1795 1809 10.1016/S0140-6736(06)69740-7 17113430 3. Verbakel S.K. van Huet R.A.C. Boon C.J.F. den Hollander A.I. Collin R.W.J. Klaver C.C.W. Hoyng C.B. Roepman R. Klevering B.J. Non-syndromic retinitis pigmentosa Prog. Retin. Eye Res. 2018 66 157 186 10.1016/j.preteyeres.2018.03.005 29597005 4. Kaushal S. Effect of rapamycin on the fate of P23H opsin associated with retinitis pigmentosa (an American Ophthalmological Society thesis) Trans. Am. Ophthalmol. Soc. 2006 104 517 529 17471359 5. Yau K.W. Hardie R.C. Phototransduction motifs and variations Cell 2009 139 246 264 10.1016/j.cell.2009.09.029 19837030 6. Mendes H.F. van der Spuy J. Chapple J.P. Cheetham M.E. Mechanisms of cell death in rhodopsin retinitis pigmentosa: Implications for therapy Trends Mol. Med. 2005 11 177 185 10.1016/j.molmed.2005.02.007 15823756 7. Yoshida T. Ozawa Y. Suzuki K. Yuki K. Ohyama M. Akamatsu W. Matsuzaki Y. Shimmura S. Mitani K. Tsubota K. The use of induced pluripotent stem cells to reveal pathogenic gene mutations and explore treatments for retinitis pigmentosa Mol. Brain 2014 7 45 10.1186/1756-6606-7-45 24935155 8. Behnen P. Felline A. Comitato A. Di Salvo M.T. Raimondi F. Gulati S. Kahremany S. Palczewski K. Marigo V. Fanelli F. A Small Chaperone Improves Folding and Routing of Rhodopsin Mutants Linked to Inherited Blindness iScience 2018 4 1 19 10.1016/j.isci.2018.05.001 30240733 9. Gorbatyuk M.S. Knox T. LaVail M.M. Gorbatyuk O.S. Noorwez S.M. Hauswirth W.W. Lin J.H. Muzyczka N. Lewin A.S. Restoration of visual function in P23H rhodopsin transgenic rats by gene delivery of BiP/Grp78 Proc. Natl. Acad. Sci. USA 2010 107 5961 5966 10.1073/pnas.0911991107 20231467 10. Khan S. Komarya S.K. Jena G. Phenylbutyrate and beta-cell function: Contribution of histone deacetylases and ER stress inhibition Epigenomics 2017 9 711 720 10.2217/epi-2016-0160 28470097 11. Kolb P.S. Ayaub E.A. Zhou W. Yum V. Dickhout J.G. Ask K. The therapeutic effects of 4-phenylbutyric acid in maintaining proteostasis Int. J. Biochem. Cell Biol. 2015 61 45 52 10.1016/j.biocel.2015.01.015 25660369 12. Ozcan U. Yilmaz E. Ozcan L. Furuhashi M. Vaillancourt E. Smith R.O. Gorgun C.Z. Hotamisligil G.S. Chemical chaperones reduce ER stress and restore glucose homeostasis in a mouse model of type 2 diabetes Science 2006 313 1137 1140 10.1126/science.1128294 16931765 13. Qiu Y. Yao J. Jia L. Thompson D.A. Zacks D.N. Shifting the balance of autophagy and proteasome activation reduces proteotoxic cell death: A novel therapeutic approach for restoring photoreceptor homeostasis Cell Death Dis. 2019 10 547 10.1038/s41419-019-1780-1 31320609 14. Rubenstein R.C. Egan M.E. Zeitlin P.L. In vitro pharmacologic restoration of CFTR-mediated chloride transport with sodium 4-phenylbutyrate in cystic fibrosis epithelial cells containing delta F508-CFTR J. Clin. Investig. 1997 100 2457 2465 10.1172/JCI119788 9366560 15. Wang F.M. Chen Y.J. Ouyang H.J. Regulation of unfolded protein response modulator XBP1s by acetylation and deacetylation Biochem. J. 2011 433 245 252 10.1042/BJ20101293 20955178 16. Christianson J.C. Ye Y. Cleaning up in the endoplasmic reticulum: Ubiquitin in charge Nat. Struct. Mol. Biol. 2014 21 325 335 10.1038/nsmb.2793 24699081 17. Claessen J.H. Kundrat L. Ploegh H.L. Protein quality control in the ER: Balancing the ubiquitin checkbook Trends Cell Biol. 2012 22 22 32 10.1016/j.tcb.2011.09.010 22055166 18. Guzman Mendoza N.A. Homma K. Osada H. Toda E. Ban N. Nagai N. Negishi K. Tsubota K. Ozawa Y. Neuroprotective Effect of 4-Phenylbutyric Acid against Photo-Stress in the Retina Antioxidants 2021 10 1147 10.3390/antiox10071147 34356380 19. Chiang W.C. Kroeger H. Sakami S. Messah C. Yasumura D. Matthes M.T. Coppinger J.A. Palczewski K. LaVail M.M. Lin J.H. Robust Endoplasmic Reticulum-Associated Degradation of Rhodopsin Precedes Retinal Degeneration Mol. Neurobiol. 2015 52 679 695 10.1007/s12035-014-8881-8 25270370 20. Sakami S. Maeda T. Bereta G. Okano K. Golczak M. Sumaroka A. Roman A.J. Cideciyan A.V. Jacobson S.G. Palczewski K. Probing mechanisms of photoreceptor degeneration in a new mouse model of the common form of autosomal dominant retinitis pigmentosa due to P23H opsin mutations J. Biol. Chem. 2011 286 10551 10567 10.1074/jbc.M110.209759 21224384 21. Kawashima H. Ozawa Y. Toda E. Homma K. Osada H. Narimatsu T. Nagai N. Tsubota K. Neuroprotective and vision-protective effect of preserving ATP levels by AMPK activator FASEB J. 2020 34 5016 5026 10.1096/fj.201902387RR 32090372 22. Narimatsu T. Ozawa Y. Miyake S. Kubota S. Yuki K. Nagai N. Tsubota K. Biological effects of blocking blue and other visible light on the mouse retina Clin. Exp. Ophthalmol. 2014 42 555 563 10.1111/ceo.12253 24304494 23. Narimatsu T. Ozawa Y. Miyake S. Nagai N. Tsubota K. Angiotensin II type 1 receptor blockade suppresses light-induced neural damage in the mouse retina Free Radic. Biol. Med. 2014 71 176 185 10.1016/j.freeradbiomed.2014.03.020 24662196 24. Osada H. Okamoto T. Kawashima H. Toda E. Miyake S. Nagai N. Kobayashi S. Tsubota K. Ozawa Y. Neuroprotective effect of bilberry extract in a murine model of photo-stressed retina PLoS ONE 2017 12 e0178627 10.1371/journal.pone.0178627 28570634 25. Nagai N. Kawashima H. Toda E. Homma K. Osada H. Guzman N.A. Shibata S. Uchiyama Y. Okano H. Tsubota K. Renin-angiotensin system impairs macrophage lipid metabolism to promote age-related macular degeneration in mouse models Commun. Biol. 2020 3 767 10.1038/s42003-020-01483-2 33299105 26. Osada H. Toda E. Homma K. Guzman N.A. Nagai N. Ogawa M. Negishi K. Arita M. Tsubota K. Ozawa Y. ADIPOR1 deficiency-induced suppression of retinal ELOVL2 and docosahexaenoic acid levels during photoreceptor degeneration and visual loss Cell Death Dis. 2021 12 458 10.1038/s41419-021-03741-5 33963174 27. Ozawa Y. Toda E. Kawashima H. Homma K. Osada H. Nagai N. Abe Y. Yasui M. Tsubota K. Aquaporin 4 Suppresses Neural Hyperactivity and Synaptic Fatigue and Fine-Tunes Neurotransmission to Regulate Visual Function in the Mouse Retina Mol. Neurobiol. 2019 56 8124 8135 10.1007/s12035-019-01661-2 31190144 28. Gorbatyuk M.S. Gorbatyuk O.S. LaVail M.M. Lin J.H. Hauswirth W.W. Lewin A.S. Functional rescue of P23H rhodopsin photoreceptors by gene delivery Adv. Exp. Med. Biol. 2012 723 191 197 10.1007/978-1-4614-0631-0_26 22183333 29. Kroeger H. Messah C. Ahern K. Gee J. Joseph V. Matthes M.T. Yasumura D. Gorbatyuk M.S. Chiang W.C. LaVail M.M. Induction of endoplasmic reticulum stress genes, BiP and chop, in genetic and environmental models of retinal degeneration Invest. Ophthalmol. Vis. Sci. 2012 53 7590 7599 10.1167/iovs.12-10221 23074209 30. Ohgane K. Dodo K. Hashimoto Y. Retinobenzaldehydes as proper-trafficking inducers of folding-defective P23H rhodopsin mutant responsible for retinitis pigmentosa Bioorg. Med. Chem. 2010 18 7022 7028 10.1016/j.bmc.2010.08.014 20805032 31. Gragg M. Park P.S. Misfolded rhodopsin mutants display variable aggregation properties Biochim. Biophys. Acta. Mol. Basis Dis. 2018 1864 2938 2948 10.1016/j.bbadis.2018.06.004 29890221 32. Harlen K.M. Roush E.C. Clayton J.E. Martinka S. Hughes T.E. Live-Cell Assays for Cell Stress Responses Reveal New Patterns of Cell Signaling Caused by Mutations in Rhodopsin, alpha-Synuclein and TDP-43 Front. Cell Neurosci. 2019 13 535 10.3389/fncel.2019.00535 31920544 33. Parfitt D.A. Aguila M. McCulley C.H. Bevilacqua D. Mendes H.F. Athanasiou D. Novoselov S.S. Kanuga N. Munro P.M. Coffey P.J. The heat-shock response co-inducer arimoclomol protects against retinal degeneration in rhodopsin retinitis pigmentosa Cell Death Dis. 2014 5 e1236 10.1038/cddis.2014.214 24853414 34. Fenech E.J. Lari F. Charles P.D. Fischer R. Laetitia-Thezenas M. Bagola K. Paton A.W. Paton J.C. Gyrd-Hansen M. Kessler B.M. Interaction mapping of endoplasmic reticulum ubiquitin ligases identifies modulators of innate immune signalling eLife 2020 9 e57306 10.7554/eLife.57306 32614325 35. Ravanelli S. den Brave F. Hoppe T. Mitochondrial Quality Control Governed by Ubiquitin Front. Cell Dev. Biol. 2020 8 270 10.3389/fcell.2020.00270 32391359 36. Franz A. Kevei E. Hoppe T. Double-edged alliance: Mitochondrial surveillance by the UPS and autophagy Curr. Opin. Cell Biol. 2015 37 18 27 10.1016/j.ceb.2015.08.004 26343990 37. Farmer T. Naslavsky N. Caplan S. Tying trafficking to fusion and fission at the mighty mitochondria Traffic 2018 19 569 577 10.1111/tra.12573 29663589 38. Bertholet A.M. Delerue T. Millet A.M. Moulis M.F. David C. Daloyau M. Arnaune-Pelloquin L. Davezac N. Mils V. Miquel M.C. Mitochondrial fusion/fission dynamics in neurodegeneration and neuronal plasticity Neurobiol. Dis. 2016 90 3 19 10.1016/j.nbd.2015.10.011 26494254 39. Zhou H. Ren J. Toan S. Mui D. Role of mitochondrial quality surveillance in myocardial infarction: From bench to bedside Ageing Res. Rev. 2021 66 101250 10.1016/j.arr.2020.101250 33388396 40. Li P.A. Hou X. Hao S. Mitochondrial biogenesis in neurodegeneration J. Neurosci. Res. 2017 95 2025 2029 10.1002/jnr.24042 28301064 41. Kang I. Chu C.T. Kaufman B.A. The mitochondrial transcription factor TFAM in neurodegeneration: Emerging evidence and mechanisms FEBS Lett. 2018 592 793 811 10.1002/1873-3468.12989 29364506 42. Ozawa Y. Oxidative stress in the light-exposed retina and its implication in age-related macular degeneration Redox. Biol. 2020 37 101779 10.1016/j.redox.2020.101779 33172789 43. Hata M. Ikeda H.O. Modulation of valosin-containing protein by Kyoto University Substances (KUS) as a novel therapeutic strategy for ischemic neuronal diseases Neural. Regen. Res. 2017 12 1252 1255 10.4103/1673-5374.213540 28966635 44. Lochmann T.L. Thomas R.R. Bennett J.P. Jr. Taylor S.M. Epigenetic Modifications of the PGC-1alpha Promoter during Exercise Induced Expression in Mice PLoS ONE 2015 10 e0129647 10.1371/journal.pone.0129647 26053857 45. Tam B.M. Moritz O.L. Characterization of rhodopsin P23H-induced retinal degeneration in a Xenopus laevis model of retinitis pigmentosa Invest. Ophthalmol. Vis. Sci. 2006 47 3234 3241 10.1167/iovs.06-0213 16877386 46. Athanasiou D. Aguila M. Opefi C.A. South K. Bellingham J. Bevilacqua D. Munro P.M. Kanuga N. Mackenzie F.E. Dubis A.M. Rescue of mutant rhodopsin traffic by metformin-induced AMPK activation accelerates photoreceptor degeneration Hum. Mol. Genet. 2017 26 305 319 10.1093/hmg/ddw387 28065882 47. Adekeye A. Haeri M. Solessio E. Knox B.E. Ablation of the proapoptotic genes CHOP or Ask1 does not prevent or delay loss of visual function in a P23H transgenic mouse model of retinitis pigmentosa PLoS ONE 2014 9 e83871 10.1371/journal.pone.0083871 24523853 48. Athanasiou D. Aguila M. Bellingham J. Kanuga N. Adamson P. Cheetham M.E. The role of the ER stress-response protein PERK in rhodopsin retinitis pigmentosa Hum. Mol. Genet. 2017 26 4896 4905 10.1093/hmg/ddx370 29036441 49. Comitato A. Schiroli D. Montanari M. Marigo V. Calpain Activation Is the Major Cause of Cell Death in Photoreceptors Expressing a Rhodopsin Misfolding Mutation Mol. Neurobiol. 2020 57 589 599 10.1007/s12035-019-01723-5 31401765 50. Thangaraj A. Sil S. Tripathi A. Chivero E.T. Periyasamy P. Buch S. Targeting endoplasmic reticulum stress and autophagy as therapeutic approaches for neurological diseases Int. Rev. Cell Mol. Biol. 2020 350 285 325 10.1016/bs.ircmb.2019.11.001 32138902 51. Wen R.H. Loewen A.D. Vent-Schmidt R.Y.J. Moritz O.L. Autophagy Induction by HDAC Inhibitors Is Unlikely to be the Mechanism of Efficacy in Prevention of Retinal Degeneration Caused by P23H Rhodopsin Adv. Exp. Med. Biol. 2019 1185 401 405 10.1007/978-3-030-27378-1_66 31884645 52. Linke M. Fritsch S.D. Sukhbaatar N. Hengstschlager M. Weichhart T. mTORC1 and mTORC2 as regulators of cell metabolism in immunity FEBS Lett. 2017 591 3089 3103 10.1002/1873-3468.12711 28600802 53. Griciuc A. Roux M.J. Merl J. Giangrande A. Hauck S.M. Aron L. Ueffing M. Proteomic survey reveals altered energetic patterns and metabolic failure prior to retinal degeneration J. Neurosci. 2014 34 2797 2812 10.1523/JNEUROSCI.2982-13.2014 24553922 54. Nakano N. Ikeda H.O. Hasegawa T. Muraoka Y. Iwai S. Tsuruyama T. Nakano M. Fuchigami T. Shudo T. Kakizuka A. Neuroprotective effects of VCP modulators in mouse models of glaucoma Heliyon 2016 2 e00096 10.1016/j.heliyon.2016.e00096 27441270 55. Hata M. Ikeda H.O. Kikkawa C. Iwai S. Muraoka Y. Hasegawa T. Kakizuka A. Yoshimura N. KUS121, a VCP modulator, attenuates ischemic retinal cell death via suppressing endoplasmic reticulum stress Sci. Rep. 2017 7 44873 10.1038/srep44873 28317920 56. LeBleu V.S. O’Connell J.T. Gonzalez Herrera K.N. Wikman H. Pantel K. Haigis M.C. de Carvalho F.M. Damascena A. Domingos Chinen L.T. Rocha R.M. PGC-1alpha mediates mitochondrial biogenesis and oxidative phosphorylation in cancer cells to promote metastasis Nat. Cell Biol. 2014 16 992 1015 992–1003, 1001–1015 10.1038/ncb3039 25241037 57. Woo Y. Lee H.J. Jung Y.M. Jung Y.J. mTOR-Mediated Antioxidant Activation in Solid Tumor Radioresistance J. Oncol. 2019 2019 5956867 10.1155/2019/5956867 31929797 58. Parra M. Class IIa HDACs-new insights into their functions in physiology and pathology FEBS J. 2015 282 1736 1744 10.1111/febs.13061 25244360 59. Czubryt M.P. McAnally J. Fishman G.I. Olson E.N. Regulation of peroxisome proliferator-activated receptor gamma coactivator 1 alpha (PGC-1 alpha) and mitochondrial function by MEF2 and HDAC5 Proc. Natl. Acad. Sci. USA 2003 100 1711 1716 10.1073/pnas.0337639100 12578979 60. Xu W.S. Parmigiani R.B. Marks P.A. Histone deacetylase inhibitors: Molecular mechanisms of action Oncogene 2007 26 5541 5552 10.1038/sj.onc.1210620 17694093 61. Sancho-Pelluz J. Alavi M.V. Sahaboglu A. Kustermann S. Farinelli P. Azadi S. van Veen T. Romero F.J. Paquet-Durand F. Ekstrom P. Excessive HDAC activation is critical for neurodegeneration in the rd1 mouse Cell Death Dis. 2010 1 e24 10.1038/cddis.2010.4 21364632 62. Leveillard T. Mohand-Said S. Lorentz O. Hicks D. Fintz A.C. Clerin E. Simonutti M. Forster V. Cavusoglu N. Chalmel F. Identification and characterization of rod-derived cone viability factor Nat. Genet. 2004 36 755 759 10.1038/ng1386 15220920 63. Ait-Ali N. Fridlich R. Millet-Puel G. Clerin E. Delalande F. Jaillard C. Blond F. Perrocheau L. Reichman S. Byrne L.C. Rod-derived cone viability factor promotes cone survival by stimulating aerobic glycolysis Cell 2015 161 817 832 10.1016/j.cell.2015.03.023 25957687 64. Fridlich R. Delalande F. Jaillard C. Lu J. Poidevin L. Cronin T. Perrocheau L. Millet-Puel G. Niepon M.L. Poch O. The thioredoxin-like protein rod-derived cone viability factor (RdCVFL) interacts with TAU and inhibits its phosphorylation in the retina Mol. Cell Proteom. 2009 8 1206 1218 10.1074/mcp.M800406-MCP200 19279044 65. Yang Y. Mohand-Said S. Danan A. Simonutti M. Fontaine V. Clerin E. Picaud S. Leveillard T. Sahel J.A. Functional cone rescue by RdCVF protein in a dominant model of retinitis pigmentosa Mol. Ther. 2009 17 787 795 10.1038/mt.2009.28 19277021 66. Griciuc A. Aron L. Ueffing M. ER stress in retinal degeneration: A target for rational therapy? Trends Mol. Med. 2011 17 442 451 10.1016/j.molmed.2011.04.002 21620769 67. Sankrityayan H. Oza M.J. Kulkarni Y.A. Mulay S.R. Gaikwad A.B. ER stress response mediates diabetic microvascular complications Drug Discov. Today 2019 24 2247 2257 10.1016/j.drudis.2019.08.003 31430543

PMC009xxxxxx/PMC9099609.txt

==== Front Polymers (Basel) Polymers (Basel) polymers Polymers 2073-4360 MDPI 10.3390/polym14091781 polymers-14-01781 Article Calibration of the PA6 Short-Fiber Reinforced Material Model for 10% to 30% Carbon Mass Fraction Mechanical Characteristic Prediction https://orcid.org/0000-0002-0893-9878 Kurkin Evgenii * Spirina Mariia https://orcid.org/0000-0001-6538-7283 Espinosa Barcenas Oscar Ulises Kurkina Ekaterina Díez-Pascual Ana María Academic Editor Joint Russian-Slovenian Laboratory Composite Materials and Structures, Samara National Research University, 34 Moskovskoe Shosse, Samara 443086, Russia; marie.spirina@yandex.ru (M.S.); oscar.espinosa.barcenas@gmail.com (O.U.E.B.); ekaterina.kurkina@mail.ru (E.K.) * Correspondence: eugene.kurkin@mail.ru; Tel.: +7-960-831-9009 27 4 2022 5 2022 14 9 178131 3 2022 22 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Short-fiber reinforced composites are widely used for the mass production of high-resistance products with complex shapes. Efficient structural design requires consideration of plasticity and anisotropy. This paper presents a method for the calibration of a general material model for stress–strain curve prediction for short-fiber reinforced composites with different fiber mass fractions. A Mori–Tanaka homogenization scheme and the J2 plasticity model with layered defined fiber orientation were used. The hardening laws: power, exponential, and exponential and linear were compared. The models were calibrated using experimental results for melt front, orientation tensor analysis, fiber length, and diameter and tension according to ISO 527-2, for samples of PA6 which were either non-reinforced, or reinforced with 10%, 15%, 20%, and 30% carbon fiber mass fractions. The novelty of this study lies in the transition from the strain–stress space to the strain–stress–fiber fraction space in the approximation of the material model parameters. We found it necessary to significantly reduce the fiber aspect ratio for the correct prediction of the mechanical characteristics of a composite using the Mori–Tanaka scheme. This deviation was caused by the ideal solution of ellipsoidal inclusion in this homogenization scheme. The calculated strength limits using Tsai–Hill failure criteria, based on strain, could be used as a first approximation for failure prediction. composite mechanical characteristics material model short fibers polyamide 6 fiber mass fraction Russian Science Foundation19-79-10205 This research was funded by Russian Science Foundation, grant number 19-79-10205. ==== Body pmc1. Introduction Reinforced polyamide 6 (PA6), like many other reinforced thermoplastics, has found applications in durable goods, computer hardware, biomedical, automobile, and aerospace sectors [1,2,3,4]. For example, in the aerospace industry, short-fiber reinforced thermoplastics have been utilized in several components of A340-600 and C295 [5], A350 XWB [6], and V-22 tilt-rotor aircraft [7]. Reinforced thermoplastics are of interest for their high performance in structural applications, low-cost manufacturing process, ease of manipulation [1], and ability to meet waste and recycling regulations [8,9,10]. However, the difficulty of predicting the anisotropy throughout the design process leads to their implementation in secondary components more than in primary structures [11,12]. For example, accounting for the effect of anisotropy on the mechanical behavior of the components [13,14,15,16,17,18] requires knowledge of both the fiber length distribution and the fiber orientation distribution, which depend on the fiber content, the geometry of the mold, the processing conditions, and the injection gate [19,20]. Currently, anisotropy is considered in structural analysis by implementing a solution to the inclusion problem by introducing Eshelby’s tensor [21], and a micromechanical method such as the Mori–Tanaka model for predicting the effective properties of two-phase composites [22]. Contemporary investigation has focused on the mechanical behavior of short-fiber reinforced polymers (SFRP) with high fiber volume fractions [15,16], and the theory presented by Fu. S.-Y. on modeling SFRP [13,14] provides evidence of the importance of studying the influence of the fiber content on the mechanical characteristics of the polymers [23]. Furthermore, creating a more precise short-fiber model is a relevant topic that has been addressed by [24,25]. Although these studies increase the reliability of modeling for anisotropy, data for SFRP are generally provided for a specific fiber volume fraction, which limits the material model. Hence, selecting the material in the early stages of design requires the usage of an accurate material model with the ability to adjust the fiber weight fraction. This study presents a general material model for PA6 reinforced with different fiber mass fractions. To create the model, experimental studies were conducted using PA6 reinforced with different short carbon fiber weight fractions. The results of the experiments were used to obtain the mechanical characteristics as well as the fiber length and fiber orientation distributions, which were subsequently used for modeling the material using a second-order Mori–Tanaka homogenization scheme and three different hardening laws. The material models obtained are presented alongside their calibration data, which were obtained by applying curve fitting through optimization of the material parameters. 2. Materials and Methods The creation of a composite material model requires knowledge of the mechanical characteristics of its components. We started by obtaining the mechanical characteristics of non-reinforced PA6 using a tensile test. The fiber aspect ratio was then determined by burning a sample of reinforced PA6. The unidirectional short-fiber reinforced composite was modeled as a transversely isotropic material by application of the Mori–Tanaka homogenization scheme. Tucker’s procedure was applied to transform the material from a unidirectional transversely isotropic one into a composite, which depended on the fiber orientation. We assumed that the stress and strain of the composite depend on its percentage content fractions. The plasticity of the matrix was modeled using the J2 plasticity model. The following three different laws were used to model hardening: the power, the exponential, and the exponential and linear laws. 2.1. Plate Molding Experiment The filling process dictates the fiber orientation. The effect of the fiber mass fraction on the filling process and mechanical characteristics of short-fiber PA6 with different fiber mass fractions was investigated. The weight of each pellet was measured using an electronic balance (0.01 g resolution), and its size was measured using a Vernier caliper (0.1 mm resolution). PA6, and PA6 reinforced with carbon fiber, had a mass of 0.0102 g and 0.0048 g, respectively; a length of 3.3 ± 0.17 and 2.02 ± 0.11 mm, respectively; and a diameter of 2.3 ± 0.09 and 2.27 ± 0.14 mm, respectively. Figure 1 shows the pellets used for experimental investigation: non-reinforced PA6 (matrix material), short-fiber reinforced PA6 with 30% carbon fiber mass fraction (Gamma Plast UPA6—30 M), and a combination of both materials mixed in mass proportions of 1:2, 1:1, and 2:1 to fabricate the fiber mass fractions of 10%, 15%, and 20%, respectively. A standard cement mixer was used for mixing all materials until the mixture was homogeneous. The pellets were dried at a temperature of 90 C for 4 h in a plastic pellet dryer before injection. Tensile tests using 1B samples as specified by the ISO 527-2 standard were performed to obtain the mechanical characteristics of the material. The samples were cut from plates with the dimensions 200 mm × 150 mm × 4 mm (Figure 2a). The manufactured plates are shown in Figure 2b. Plates of different fiber compositions were injected using a Negri Bossi VE 210-1700 injection molding machine. The filling parameters were as follows: melt temperature 225 °C, mold temperature 80 °C, and flow rate 42 cm3/s. 2.2. Investigation of Fibers under Scanning Electronic Microscope Modeling the mechanical behavior of the short-fiber reinforced PA6 requires knowledge of the aspect ratio (AR) between the diameter and the length of the fibers [26,27,28]. Measuring the dimensions of the fibers can be achieved if the polymer containing the fibers is degraded at an elevated temperature (Figure 3). The technique used by [29] was taken as a basis for the fiber acquisition. Samples with fiber mass fractions of 15% (Figure 3a) and 30% were cut from the molded plates and burnt in an inert atmosphere oven (Figure 3b). The oven’s camera was filled with a nitrogen atmosphere to avoid fiber degradation, and maintained without a flow rate. The process was initiated at 20 °C with a heating rate of 5 °C/min and reached a temperature of 900 °C that was held for 20 min, before cooling at a rate of −5 °C/min. The fibers obtained were investigated under a Tescan Vega electronic microscope (Figure 3c). 2.3. The Effective Fiber Length For correct modeling and prediction of the fiber strength, the concept of the effective fiber length was introduced. It was first considered by Rosen [30] in 1965 while describing the mechanical characteristics of continuously reinforced composites. In [31], the effective fiber length is known as the debonding fraction due to the substitution of the ellipsoidal inclusion for the “equivalent debonded inclusion”, and in that study the fraction of the debonded interface surface was 0.26. In the present work, the effective fiber length was estimated quantitatively. This technique may be described using the following formula:(1) φ=ARmodelARexperimental 2.4. Calculation of Short-Reinforced Composite Tensile Curves The material modelling was performed in Digimat-MF, while the parameters were identified using Digimat-MX RVE. The microstructure consisted of two phases—the elastoplastic PA6 matrix, and the elastic short-carbon fibers modeled as elliptic inclusions. The models of the matrix were obtained after performing curve fitting of the tensile test results for the specimens without reinforcement. The fibers (inclusions) and the matrix were homogenized via Digimat-MF, using a second-order Mori–Tanaka homogenization scheme for the computation of the mechanical properties. The unidirectional short-fiber reinforced composite was modeled as a transversely isotropic material. The elastic moduli introduced by Tandon and Weng [32] were used to calculate the elastic coefficients:(2) E11Em=11+ϕf(A1+2υmA2)A6 (3) E22Em=11+ϕf[−2υmA3+(1−υm)A4+(1+υm)A5A6]2A6, where Em and υm are the Young’s modulus and Poisson ratio of the matrix, respectively. The volume fraction of the fiber is represented by ϕf. The parameters Ai are the functions of Eshelby’s tensor and can be found in [33]. In this study, we used Eshelby’s tensor for elliptical inclusion, and this depends on the fiber aspect ratio. The Tucker’s averaging procedure was used to determine the fiber orientation tensor, which is described as follows: (4) Cijkl=B1aijkl+B2(aijδkl+δijakl)+B3(aikδjl+ailδjk+ajlδik+ajkδil)+B4(δijδkl)+B5(δikδjl+δilδjl), where aijkl is the fourth-order fiber orientation tensor, δij is the second-order unit tensor, and the coefficients B are related to the components of the stiffness matrix of the transversely isotropic unidirectional composite [34]. The fourth-order tensor in Tucker’s averaging procedure was reduced to a second-order tensor by applying the orthotropic closure approximation presented by Cintra and Tucker in [35]. The approximation parameters of the fiber orientation tensors mainly influence the calculation of the stress–strain curves. The fiber direction from 0 to 90° was divided into 20 equal parts, with a tolerance interval in the fiber orientation tensor of 0.01. The composite stress and strain depend on the stress and strain of the matrix and fiber, proportional to their volume fractions:(5) ε=(1−ϕf)εm+ϕfεf, (6) σ=(1−ϕf)σm+ϕfσf. The stress–strain state of the matrix is described using the J2 plasticity model [36], based on the von Mises equivalent stress, σeq. When σeq exceeds the initial yield stress, the response becomes nonlinear and plastic deformation appears. Plastic strength is expressed as follows:(7) σeq=σY+R(εp), where σY is the initial yield stress; R(εp) is the isotropic strain hardening function; and εp is the accumulated plastic strain. Poisson’s ratio of the matrix, in the plastic range, is predicted through Lame parameters using spectral decomposition [37]; elastic bulk module K is taken as a constant, and shear moduli G will be (8) G=Ge(1−3Ge3Ge+d R(εp)d εp), where Ge, the elastic shear modulus, is defined using a Lame parameter, based on the given Young’s modulus and the Poisson’s ratio of the matrix in the elastic range. For all models so described, the matrix Poisson’s ratio in the elastic range is equal to the experimentally measured average value. Poisson’s ratio of the matrix in the elastic range had a slight effect on the stress–strain curves for reinforced PA6 when reverse-engineering the curves, and was not accurate, therefore, Poisson’s ratio of the matrix in the elastic range was excluded from the reverse-engineered parameters. A comparison between the three hardening stress functions is provided in [38]: Power law [39]:(9) R(εp)=kεpm; Exponential law:(10) R(εp)=R∞[1−e−mεp]; Exponential and linear law:(11) R(εp)=kεp+R∞[1−e−mεp]. where k is linear hardening modulus, MPa; m is hardening exponent; and R∞ is hardening modulus, MPa. 3. Results 3.1. Melt Front and Microstucture Experimental Investigation An experimental study of the melt front was conducted by studying partially filled molds (Figure 4). This study allowed us to verify the plastic injection molding models and showed that the fiber mass fraction had a small effect on the melt front of the plates. The fiber orientation determines the mechanical characteristics of the material to a great extent. To investigate the fiber orientation at the resultant fracture location, a sample was extracted from the 90° tensile test specimens with fiber mass fractions of 15% and 30% (Figure 2a). The shape and roughness of the fibers have a significant influence on the adhesion quality and strength of the composite material [40,41]. The morphology of the fracture surface is shown in Figure 5, it can be observed in the pulled-out fiber and defects in the background. Several factors characterize composite failure, including the adhesion between the matrix and the fiber. The failure surface of the 90° sample with 30 wt. % was analyzed by X-ray fluorescence under a scanning electron microscope (Figure 6). The spectrum on the center of the fiber surface (Figure 6a) consisted of 93.65% C and 6.35% O (Figure 6b), showing a low presence of PA6. The sample was examined in three places at the failure line; the images obtained under the microscope are shown in Figure 7. Qualitative evaluation of the fiber orientation tensor was performed by comparing the size of the layer against the fiber orientation tensor components. The average values of the specific skin layer, the core layer thickness, and the thickness of the samples with 15% fiber mass fraction, were 7.31%, 25.79%, and 3.98 mm, respectively. In Figure 7, in the shell layer (the layer between the skin and the core layers), the fibers are oriented along the x-axis, which corresponds to the fiber orientation tensor values of the component a11. In addition, the fiber orientation tensor predicts that the component a11 has the highest probability. The fibers presented a chaotic appearance within the core layer. The skin layer comprised approximately 3% of the thickness, while the core layers constituted 19%. The start and end of every layer were difficult to determine, especially at the transition from the skin layer to the shell layer. 3.2. Injection Molding and the Fiber Orientation Models Validation The injection molding simulation of PA6 reinforced with 30% mass fiber fraction was performed using Moldex3D R17 on a structured mesh of 4,499,918 elements. The number of elements in the plate (without runner and sprue) along the x-axis was 600, along the y-axis 296, and along the z-axis 24; this totaled 4,262,400 elements. A comparison of the simulated and experimental filling processes is shown in Figure 8a. The fiber orientation tensor is shown in Figure 8b and Table 1, and used in the models of the mechanical characteristics of the materials. Quantitative evaluation was performed by comparing the components of the fiber orientation tensor (Figure 8b) against the experimental data presented by Foss et al. [42], and we obtained a reasonable agreement. Microstructure analysis at the orientations 0 and 90° are shown in Figure 9. 3.3. Determination Length and Diameter of the Fibers To measure the size of the fibers, five images were recorded for each material using a scanning electronic microscope. From every image, 25 to 35 different fibers were measured. In the sample with a 15% fiber mass fraction, it was found that the diameter had a mean of 6.8 μm and a standard deviation of 1.04 μm, and the length had a mean of 167.41 μm and a standard deviation of 72.22 μm. In the sample with a 30% fiber mass fraction, was found that the diameter had a mean of 6.33 μm and a standard deviation of 1.14 μm, and the length had a mean of 135.66 μm and a standard deviation of 67.06 μm. The distribution of the diameter and length of the fibers is presented in the form of histograms in Figure 10. The AR of the fibres was calculated by dividing their measured lengths by their mean diameter (equal to 6.80 for the 15%, and 6.33 for the 30% fiber mass fractions). The AR distribution is presented in Figure 11. 3.4. Experimental Stress–strain Curves Tensile tests were performed for construction of the material models according to ISO 527-2 on 1B specimens from non-reinforced PA6 and composites with 10%, 15%, 20%, and 30%. For each material, nine samples were investigated, cut out from the plates at 0° (red lines), 45° (green lines), and 90° (blue lines) to the flow direction (Figure 2a). The test machine MTS 322.21, as well as the biaxial extensometer MTS 632.85F-14, were used for registering the force and displacement of the specimens. PA6 stress–strain curves are shown in Figure 12a; due to its isotropic composition a low variation in mechanical behavior existed between specimens (red lines—0°, green lines—45°, blue lines—90° specimens). The plastic behavior of the material is unreliable in determining the value of the elastic behavior limit. The offset yield point, σY, depends on the offset plastic strain, εp, and is shown in Figure 12b with a logarithmic scale. The dependence σY(εp) can be approximated as lg(σY)=p1lg(εp)+p2, where the coefficients (with a 95% confidence interval) are p1=0.383(0.3615, 0.4044) and p2=1.978(1.945, 2.012). Poisson’s ratio is shown in Figure 12c. The values for Poisson’s ratio were calculated in a small strain region, and ranged from 0.31 to 0.42, with a mean of 0.3721. The Poisson’s ratio obtained satisfactorily agrees with that presented by D.V. Rosato [43]. The Poisson’s ratio value of 0.3721 was used for the matrix of all material models. Stress–strain curves of short-fiber reinforced PA6 with different carbon-fiber mass fractions are shown in Figure 13. As expected, samples with a lower fiber content showed behavior closer to that of an isotropic material. Increasing the fiber mass fraction content increased the anisotropic behavior of the material. Moreover, the strength of the 0° samples increased more than the strength of 45° and 90° samples. Specimens, especially at 0° with higher fiber content, were more fragile—the value of ultimate strain decreased. The stress–strain curves of the samples with a 30% fiber mass fraction show the differences between them. 3.5. Determination of the Matrix Material Models Results of the ISO 527-2 tensile test for the PA6 matrix are presented in Figure 14 with dashed lines. The scatter of experimental stress measurements when strain equaled 0.03 had a standard deviation of 2.49 MPa, and a coefficient of variation of 3.71%. The identification of the parameters of the matrix models, including the parameters of the hardening laws in Equations (9)–(11), was established using the Digimat-MX RVE curve fitting module, and was based on the tensile tests presented in Figure 12a. To estimate the accuracy of the approximation, the averaged relative error for each hardening law was obtained [38]. The parameters of the matrix models and the relative errors of the approximations are provided in Table 2. Approximated curves are presented in Figure 14. 3.6. Comparison of the Effect of Considering the Distribution of Fiber Lengths and the Distribution of the Orientation Tensor on the Accuracy of Approximation of the Tensile Curves of a Short-Reinforced Composite During model construction, it was impossible to simultaneously consider the distribution of the AR of the fibers and the distribution of the fiber orientation tensor over the thickness of the sample. The stress–strain curves of short-fiber reinforced PA6 with a 30% carbon-fiber mass fraction, presented in Figure 15, were created using the exponential law. In Figure 15a, a comparison between single layer and multilayer (using Table 1) orientation tensor definitions, with a single AR value of 26.5 is presented. Figure 15b compares a fixed AR of 26.5 with the AR distribution according to Figure 11. It can be concluded that allowing for the layered orientation tensor had a more noticeable effect on the simulation of a 90° sample, whereas accounting for AR distribution could be replaced by an equivalent constant value. 3.7. Determination of the Parameters of the Composite Material Models, Common for Different Fiber Mass Identification of the model parameters was performed to find a material model suitable for describing the mechanical properties of short-carbon fiber reinforced PA6 with different percentages of fiber mass fraction (Figure 16). The stress–strain curves corresponding to calibrated material models based on experimentally obtained fiber and matrix properties are presented in the first row of Figure 16. The curves overestimated stiffness values for the materials, especially for samples at 0°. The most important influence on composite parameters was fiber size. The second row of Figure 16 shows the material models based on the mean reverse-engineered values of the AR of the fibers (Table 3). CV in Table 3, and below, is the coefficient of variation, which is the ratio of the standard deviation to the mean value of the corresponding quantities. The mean AR value for all reinforced PA6 material models was approximately 15, which constitutes 57% of the 26.5 value, equivalent to the experimentally observed distribution (Figure 10). Hence, the calculated effective fiber length (1) in this study was 0.57. The third row shows the models constructed with mean value parameters after approximation of all matrix elastoplasticity model parameters and fiber AR (Table 4). The particularity in the construction of Figure 16 is that the average values of the parameters were taken from every material model to build a mean model that describes the mechanical characteristics of the reinforced PA6 with a range of fiber mass fractions from 10 to 30%. The mean relative error of tension stress–strain curve fitting, using the mean models obtained, is presented in Table 5. The power, exponential, exponential and linear hardening laws had similar mean relative error values. In this study, the shape of the matrix stress–strain curve was more consistent with the exponential law. Adding a linear function to the exponential hardening law increased the stress–strain curve approximation accuracy only if the model parameters were calibrated in the composite tension experiment. The exponential and linear hardening law model, based on experimentally obtained matrix parameters, increased the composite curve approximation error due to the low precision in describing the slope at the end of the matrix stress–strain curve. Figure 17 shows the PA6 matrix stress–strain refined curves after curve fitting of the composites according to the mean parameters model (Table 4). It was found that, for a better description of the parameters of the composite material, the stiffness of the matrix was overestimated during the calibration process. 3.8. Failure Criterion Parameter Identification for Short-Fiber Reinforced Thermoplastic Composites For strength prediction, the Tsai–Hill 3D transversely isotropic strain-based failure criterion was applied to the reinforced PA6 (composite level) in the local finite element coordinate system (local axes) [38]. The critical fraction of failed pseudo-grains was 0.75, and multilayer failure occurred using the all-layer failure condition. The results of the identification of the model parameters are presented in Table 6. The mean strain limits for the ultimate tensile strength of short-fiber reinforced PA6 were close to those of the different plasticity models, and constituted axial tensile 0.02, in-plane tensile 0.032, and transverse shear 0.051 (Table 6). Figure 18 shows material models with different fiber mass fraction stress–strain curves, constructed using the mean parameters of the power, exponential and exponential and lineal hardening laws exposed in Table 4, and selected failure criterion. The mean relative error of the failure criteria for fracture strain is shown in Table 7. 4. Discussion A general model material for short-fiber reinforced polyamide 6 with carbon-fiber mass fractions from 10% to 30% was constructed using a second-order Mori–Tanaka homogenization scheme. The matrix was modeled using the J2 plasticity model along with power, exponential, and exponential and linear hardening laws. The mechanical characteristics of the matrix, the fiber aspect ratio, and stress–strain curves of the composites were obtained experimentally. From the examination of the three different hardening laws, it was found that the exponential law described the PA6 stress–strain curve with considerable accuracy. Furthermore, the addition of a linear term into the hardening law (exponential and lineal) increased the accuracy of model parameters that were previously reverse-engineered. The results from general calibration produced models for PA6 reinforced composite which could determine the mechanical characteristics of materials with different fiber mass fractions. The novelty of this approach lies in the transition from two-dimensional (stress–strain) to three-dimensional (stress–strain–fiber fraction) approximation space (Figure 19). It was found that the use of the experimental AR value in the construction of the models increased the strength and stiffness of 0° samples by a factor of 1.7. Such a deviation may have been caused by a failure to account for fibers with a low AR during the electron microscopy analysis; however, a more probable reason is the inadequate consideration of the influence of including the closed surface (ellipsoid) [32] and its ideal solution in the Mori–Tanaka homogenization scheme. Moreover, the introduction of an equivalent AR value was able to replace the AR distribution, simplifying the material model. The relative effective fiber length, i.e., the ratio between the equivalent and experimental ARs, was φ = 0.57, which corresponds to a virtual length reduction of the fiber, at both ends, of approximately 21.5%. The mean values for the material models obtained after performing curve fitting to the reinforced matrix’s parameters were used for the construction of the mean models. A mean model was produced that was capable of modeling reinforced PA6 across a wide range of fiber mass fraction, from 10 to 30%, and allowed the prediction of the mechanical characteristics with a mean relative error of 6.60%. The calibration of mean Tsai–Hill strength failure criteria based on strain, via a transversely isotropic statement, was performed, and produced a description of the strength of reinforced PA6 with 10 to 30% carbon fiber mass fractions. The calibrated strength failure criteria could be used only as a first approximation for the failure of carbon-reinforced PA6 structures because the fracture strain mean relative error value, the large dispersion in the experimental fracture strain values, and the question of failure criteria require further investigation. 5. Conclusions The Mori–Tanaka model allowed us to accurately predict the stiffness of a short-reinforced composite with an arbitrary fraction of reinforcing fibers, but required curve fitting for the whole composite because using fiber and matrix parameters obtained experimentally led to an erroneous overestimation of the composite’s mechanical characteristics. Fiber elongation had a decisive influence on the stiffness of the composite. It was found that for the correct prediction of the mechanical characteristics of a composite using the Mori–Tanaka model, it was necessary to significantly reduce the fiber aspect ratio. The exponential hardening law of the matrix provided a good description of the tensile diagram of the composite. Adding a linear term and switching to an exponential and linear model allowed an even more accurate demonstration of material behavior when approaching the strength limit, but required attention to the calibration of the slope of the tensile curve at high strain values. When dealing with different fiber fractions, adding a third dimension to the two-dimensional stress–strain space in the form of a fiber fraction allowed us to combine the flexibility of the Mori–Tanaka model with the accuracy of calibrated tension curves from a single fiber fraction composite material. This makes it possible to improve the accuracy of the selection of the fiber fraction in material to be used in the construction of structures during the early stages of design. In this work, averaging of the parameters of the tension curves calibrated for individual volume fractions was used to construct a three-dimensional, calibrated, Mori-Tanaka model. In the future, we plan to develop an algorithm and software that would allow simultaneous calibration of three-dimensional models over the entire volume of experimental data available. Acknowledgments The authors thank A.A Pavlov and S.E. Selivanov, for providing help to perform the tensile tests. The authors thank Pladep Ltd. for help in plate molding. The authors thank R.V. Shafigulin for help with burning of samples. Author Contributions Conceptualization, E.K. (Evgenii Kurkin); methodology, E.K. (Evgenii Kurkin); validation, E.K. (Evgenii Kurkin) and O.U.E.B.; formal analysis, M.S. and E.K. (Ekaterina Kurkina); investigation, E.K. (Evgenii Kurkin) and M.S.; data curation, O.U.E.B., M.S. and E.K. (Ekaterina Kurkina); writing—original draft preparation, E.K. (Evgenii Kurkin) and O.U.E.B.; writing—review and editing, E.K. (Evgenii Kurkin) and O.U.E.B.; visualization, O.U.E.B.; supervision, E.K. (Evgenii Kurkin); project administration, E.K. (Evgenii Kurkin); funding acquisition, E.K. (Evgenii Kurkin). All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement The raw data cannot be shared at this time as the data also forms part of an ongoing study. Conflicts of Interest The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results. Figure 1 Mixed pellets to obtain different fiber mass fraction. Figure 2 Plate molding: (a) Size and ISO 527-2 specimen layout; (b) Example manufactured plate. Figure 3 Short fiber extraction from composite: (a) Sample preparation; (b) burning samples in the oven; (c) fibers after matrix burning. Figure 4 Experimentally obtained plate melt front at molding composites with different fiber mass fraction: (a) 0%; (b) 10%; (c) 15%; (d) 20%; (e) 30%. Figure 5 Morphology of the fracture of short-fiber reinforced PA6 with 30% carbon-fiber mass fraction. Figure 6 Reinforced PA6 with 30 wt. % short-carbon fiber: (a) SEM; (b) XRF spectra. Figure 7 Fracture surface of short-fiber reinforced PA6 with 15% carbon-fiber mass fraction, ISO 527-2 sample, orientated at 90°. Dashed lines demark the size in micrometers of the skin and core layer. Figure 8 Comparison of simulated and experimental filling: (a) Melt front, blue lines—calculated isolines at each 2% of filling time, black and red lines—experimentally obtained melt fronts; (b) Orientation tensor components, solid lines—calculated values, dash lines—mean experimentally obtained zones sizes. Figure 9 The microstructure of short-fiber reinforced PA6 with 30% carbon-fiber mass fraction along the following orientations: (a) 0°; (b) 90°. Figure 10 Distribution of diameter and length in 30% and 15% fiber mass fraction samples. Figure 11 Measured fiber aspect ratio distribution for 15% (u15) and 30% (u30) fiber mass fractions. Figure 12 Tensile test of ISO 527-2 PA6 matrix samples: (a) Tension curves; (b) Yield stress; (c) Poisson’s ratio. Figure 13 Experimental stress–strain curves for ISO 527-2 samples of PA6 composites with different fiber mass fractions, cut at 0°, 45° and 90° angles to flow direction: (a) 10%; (b) 15%; (c) 20%; (d) 30%. Figure 14 The approximation of PA6 matrix ISO 527-2 tension curves, dashed lines—experiment, solid lines—model, with hardening laws: (a) power; (b) exponential; (c) exponential and linear. The solid lines show the optimized curves. Dashed lines are the experimental data. Figure 15 Stress–strain curves of short-fiber reinforced PA6 with 30% carbon fiber were modeled using the exponential law. Comparison between: (a) Single AR value and AR distribution; (b) Single layer and multilayer analysis. Figure 16 Stress–strain curves for the model material for PA6 with different short carbon fiber mass fractions, orientated in different directions. From left to right: 10%, 15%, 20%, and 30% fiber mass fraction. From top to bottom: non-calibrated parameters; calibrated fiber aspect ratio; all matrix parameters and fiber aspect ratio calibrated. Material model—color lines: 0°—red lines, 45°—green lines, 90°—blue lines. Experimental data—black dashed lines. Figure 17 Matrix models (solid lines) after the composite stiffness calibration with different hardening laws: (a) power; (b) exponential; (c) exponential and linear. Dashed lines—experiment. Figure 18 Stress–strain curves prepared from the mean obtained strain limits, and using Tsai–Hill 3D transversely isotropic strain-based values for 10%, 15%, 20%, and 30% fiber mass fractions. Material model—color lines: 0°—red lines, 45°—green lines, 90°—blue lines. Experimental data—black dashed lines. Figure 19 Common mass fraction composite material models for all fibers. Exponent and linear hardening law case. For angles between the flow and load directions: (a) 0°; (b) 45°; (c) 90°. polymers-14-01781-t001_Table 1 Table 1 Calculated fiber orientation tensor components. Layer Thickness, mm a11 a22 a33 a12 a13 a23 1 0.1667 0.603 0.304 0.0929 0.00104 0.00019 0.00011 2 1.5003 0.821 0.126 0.0533 0.00087 0.00001 0.00003 3 0.1667 0.711 0.244 0.0451 −0.00131 −0.00033 0.00008 4 0.3334 0.461 0.486 0.0526 −0.01870 −0.00113 −0.00495 5 0.1667 0.711 0.244 0.0451 −0.00131 −0.00033 0.00008 6 1.5003 0.821 0.126 0.0533 0.00087 0.00001 0.00003 7 0.1667 0.603 0.304 0.0929 0.00104 0.00019 0.00011 polymers-14-01781-t002_Table 2 Table 2 Calibrated pure PA6 tension curve parameters of elastoplastic matrix material models. Hardening Law Power Exponential Exponential and Linear Young’s modulus, MPa 3341 3547 3552 Yield stress, MPa 12.57 8.27 6.68 Hardening modulus, R∞ MPa - 60.59 55.41 Hardening Exponent m 0.21891 433.69 528.67 Linear hardening modulus k, MPa 158.12 - 623.17 Relative error, % 5.32 4.81 4.77 polymers-14-01781-t003_Table 3 Table 3 Fiber aspect ratio (AR) calibration for composites with different fiber mass fraction. Hardening Law Fiber Mass Fraction, % Mean AR Model Param. CV, % 10 15 20 30 Aspect ratio Power 14.00 16.53 14.47 14.82 15.00 7.35 Exponential 14.53 16.94 15.14 15.47 15.52 6.60 Exponential and linear 14.30 16.75 14.71 15.27 15.30 7.03 polymers-14-01781-t004_Table 4 Table 4 Calibration of all parameters of the matrix elastoplasticity model and fiber aspect ratio. Parameter Fiber Mass Fraction, % Mean Parameters Model Parameters CV, % 10 15 20 30 Power Law Young’s modulus, MPa 4383 4437 4162 3759 4185 7.4 Yield stress, MPa 12.3 12.6 10.2 9.9 11.2 12.3 Hardening modulus, MPa 147.6 161.9 124.6 137.0 142.8 11.1 Hardening exponent 0.1995 0.2238 0.2023 0.2475 0.2183 10.2 Fibers’ AR 8.84 12.23 14.19 16.79 13.01 25.8 Exponential Law Young’s modulus, MPa 4654 4615 4741 3994 4501 7.6 Yield stress, MPa 13.8 21.2 12.4 16.1 15.9 24.3 Hardening modulus, MPa 59.2 51.0 51.5 38.7 50.1 17.0 Hardening exponent 404.8 370.6 353.6 377.8 376.7 5.7 Fibers’ AR 9.02 12.25 13.37 16.62 12.8 24.4 Exponential and linear law Young’s modulus, MPa 4672 4842 4625 3994 4533 8.2 Yield stress, MPa 13.6 13.0 14.5 14.5 13.9 5.4 Hardening modulus, MPa 56.6 54.9 46.1 37.0 48.6 18.6 Hardening exponent 447.1 417.7 381.1 458.3 426.0 8.1 Hardening linear modulus, MPa 208.8 216.0 144.1 188.4 189.3 17.1 Fibers’ AR 8.82 12.36 13.80 16.54 12.9 24.9 polymers-14-01781-t005_Table 5 Table 5 Mean relative error of stress–strain curve prediction of different fiber mass fraction composites with models based on fibers and matrix characteristics obtained experimentally, mean fiber aspect ratio calibrated models, and the mean parameter model. Hardening Law Fiber Mass Fraction, % Mean Level of Error, % 10 15 20 30 Models, based on fibers and matrix characteristics obtained experimentally Power 29.0 25.7 35.8 35.9 31.6 Exponential 27.8 25.2 33.7 34.7 30.4 Exponential and linear 28.6 25.9 36.1 35.7 31.6 Mean fiber aspect ratio calibrated models Power 11.7 12.1 12.0 11.1 11.7 Exponential 11.5 12.3 11.3 11.7 11.7 Exponential and linear 11.7 12.5 12.6 12.2 12.3 Mean with the calibration of all matrix and fiber aspect ratio parameters Power 7.4 7.6 4.9 7.7 6.9 Exponential 6.9 6.8 4.1 8.0 6.5 Exponential and linear 7.0 6.9 3.9 7.9 6.4 polymers-14-01781-t006_Table 6 Table 6 Failure criteria strain limits for different fiber mass fraction cases. Parameters Fiber Mass Fraction, % Mean Strain Limits Parameters CV, % 10 15 20 30 Power law Axial tensile strain limit, 10−2 2.009 2.957 1.944 1.880 2.197 23.16 Inplane tensile strain limit, 10−2 3.516 3.855 3.573 2.151 3.274 23.30 Transverse shear strain limit, 10−2 5.606 5.140 5.235 4.364 5.086 10.26 Exponential law Axial tensile strain limit, 10−2 1.566 2.353 1.832 2.031 1.945 17.05 Inplane tensile strain limit, 10−2 2.999 3.524 3.430 2.278 3.058 18.57 Transverse shear strain limit, 10−2 5.873 5.249 5.355 4.076 5.138 14.77 Exponential and linear law Axial tensile strain limit, 10−2 2.141 2.402 1.908 1.944 2.099 10.81 Inplane tensile strain limit, 10−2 3.684 3.573 3.559 2.012 3.207 24.90 Transverse shear strain limit, 10−2 5.607 5.140 5.140 4.798 5.171 6.42 polymers-14-01781-t007_Table 7 Table 7 Mean relative error (%) of ultimate strain prediction for different fiber mass fraction composites with models based on fiber and matrix characteristics obtained experimentally, mean fiber aspect ratio calibrated models, and the mean parameter model. Hardening Law Fiber Mass Fraction, % Mean Level of Error, % 10 15 20 30 Power 16.24 7.79 18.92 56.63 24.90 Exponential 17.30 8.87 19.73 55.04 25.24 Exponential and linear 16.26 7.79 19.02 56.36 24.86 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Chawla K.K. Composite Materials: Science and Engineering Springer New York, NY, USA 2016 2. Rana S. Fangueiro R. Advanced Composites in Aerospace Engineering Advanced Composite Materials for Aerospace Engineering Elsevier Amsterdam, The Netherlands 2016 1 15 10.1016/B978-0-08-100037-3.00001-8 3. Fu S.-Y. Lauke B. Mai Y.-W. Science and Engineering of Short Fibre Reinforced Polymers Composites Woodhead Publishing in Materials CRC Press Boca Raton, FL, USA 2009 4. Nakonieczny D.S. Kern F. Dufner L. Antonowicz M. Matus K. Alumina and Zirconia-Reinforced Polyamide PA-12 Composites for Biomedical Additive Manufacturing Materials 2021 14 6201 10.3390/ma14206201 34683792 5. Díaz J. Rubio L. Developments to Manufacture Structural Aeronautical Parts in Carbon Fibre Reinforced Thermoplastic Materials J. Mater. Processing Technol. 2003 143–144 342 346 10.1016/S0924-0136(03)00450-3 6. Miaris A. Edelmann K. von Hayek Boelingen M. A350 WXB: Thousands of Thermoplastic Composite Parts in an FRP Aircraft-SAMPE Available online: https://www.nasampe.org/store/viewproduct.aspx?id=4417947 (accessed on 25 March 2022) 7. Chang I.Y. Pratte J.F. LDF™ Thermoplastic Composites Technology J. Thermoplast. Compos. Mater. 1991 4 227 252 10.1177/089270579100400302 8. Eguemann N. Giger L. Roux M. Dransfeld C. Thiebaud F. Perreux D. Compression Moulding of Complex Parts for the Aerospace with Discontinuous Novel and Recycled Thermoplastic Composite Materials Proceedings of the 19th International Conference on Composite Materials Paris, France 28 July–2 August 2013 1 11 9. European Bioplastics Fact Sheet: Mechanical Recycling European Bioplastics eV Berlin, Germany 2010 10. Andrzejewski J. Aniśko J. Szulc J. A Comparative Study of Biocarbon Reinforced Polyoxymethylene and Polyamide: Materials Performance and Durability Compos. Part A Appl. Sci. Manuf. 2022 152 106715 10.1016/j.compositesa.2021.106715 11. Steffens M. Himmel N. Maier M. Design and Analysis of Discontinuous Long Fiber Reinforced Thermoplastic Structures for Car Seat Applications Trans. Eng. Sci. 1998 21 35 44 12. Kondapalli P.S. Grumm K. Cao Y. Laurent V. Application and CAE simulation of over molded short and continuous fiber thermoplastic composite: Part II. 12th Int LS-DYNA® Users Conf FEA Inf. Eng. J. 2013 2 1 10 13. Fu S. Effects of Fiber Length and Fiber Orientation Distributions on the Tensile Strength of Short-Fiber-Reinforced Polymers Compos. Sci. Technol. 1996 56 1179 1190 10.1016/S0266-3538(96)00072-3 14. Fu S.-Y. Hu X. Yue C.-Y. Effects of Fiber Length and Orientation Distributions on the Mechanical Properties of Short-Fiber-Reinforced Polymers J. Soc. Mater. Sci. Jpn. 1999 48 74 83 10.2472/jsms.48.6Appendix_74 15. Capela C. Oliveira S.E. Pestana J. Ferreira J.A.M. Effect of Fiber Length on the Mechanical Properties of High Dosage Carbon Reinforced Procedia Struct. Integr. 2017 5 539 546 10.1016/j.prostr.2017.07.159 16. Capela C. Oliveira S.E. Ferreira J.A.M. Mechanical Behavior of High Dosage Short Carbon Fiber Reinforced Epoxy Composites Fibers Polym. 2017 18 1200 1207 10.1007/s12221-017-7246-0 17. Karsli N.G. Aytac A. Deniz V. Effects of Initial Fiber Length and Fiber Length Distribution on the Properties of Carbon-Fiber-Reinforced-Polypropylene Composites J. Reinf. Plast. Compos. 2012 31 1053 1060 10.1177/0731684412452678 18. Kufel A. Para S. Kuciel S. Basalt/Glass Fiber Polypropylene Hybrid Composites: Mechanical Properties at Different Temperatures and under Cyclic Loading and Micromechanical Modelling Materials 2021 14 5574 10.3390/ma14195574 34639971 19. Moritzer E. Heiderich G. Hirsch A. Fiber Length Reduction during Injection Molding AIP Publishing Dresden, Germany 2019 070001 10.1063/1.5084845 20. Kurkin E.I. Lukyanov O.E. Chertykovtseva V.O. Espinosa Barcenas O.U. Molding Gate Optimization for Weld Line Location Away from Structures Loaded Area J. Phys. Conf. Ser. 2021 1925 012056 10.1088/1742-6596/1925/1/012056 21. Eshelby J. The Determination of the Elastic Field of an Ellipsoidal Inclusion, and Related Problems Proc. R. Soc. Lond. A 1957 241 376 396 10.1098/rspa.1957.0133 22. Mori T. Tanaka K. Average Stress in Matrix and Average Elastic Energy of Materials with Misfitting Inclusions Acta Metall. 1973 21 571 574 10.1016/0001-6160(73)90064-3 23. Su N. Pierce R.S. Rudd C. Liu X. Comprehensive Investigation of Reclaimed Carbon Fibre Reinforced Polyamide (RCF/PA) Filaments and FDM Printed Composites Compos. Part B Eng. 2022 233 109646 10.1016/j.compositesb.2022.109646 24. Mentges N. Dashtbozorg B. Mirkhalaf S.M. A Micromechanics-Based Artificial Neural Networks Model for Elastic Properties of Short Fiber Composites Compos. Part B Eng. 2021 213 108736 10.1016/j.compositesb.2021.108736 25. Pietrogrande R. Carraro P.A. De Monte M. Quaresimin M. A Novel Pseudo-Grain Approach for the Estimation of the Elastic Stress Distributions within the Matrix of Short Fiber-Reinforced Polymers Compos. Part B Eng. 2018 150 115 123 10.1016/j.compositesb.2018.05.038 26. Templeton P.A. Strength Predictions of Injection Molding Compounds J. Reinf. Plast. Compos. 1990 9 210 225 10.1177/073168449000900301 27. Chin W.-K. Liu H.-T. Lee Y.-D. Effects of Fiber Length and Orientation Distribution on the Elastic Modulus of Short Fiber Reinforced Thermoplastics Polym. Compos. 1988 9 27 35 10.1002/pc.750090105 28. Friedrich K. Microstructural efficiency and fracture toughness of short fiber/thermoplastic matrix composites Comp. Sci. Tech. 1985 22 43 74 10.1016/0266-3538(85)90090-9 29. Verdejo de Toro E. Coello Sobrino J. Martínez Martínez A. Miguel Eguía V. Ayllón Pérez J. Investigation of a Short Carbon Fibre-Reinforced Polyamide and Comparison of Two Manufacturing Processes: Fused Deposition Modelling (FDM) and Polymer Injection Moulding (PIM) Materials 2020 13 672 10.3390/ma13030672 32028619 30. Rosen B.W. Dow N.F. Hashin Z. Mechanical Properties of Fibrous Composites General Electric Co. Philadelphia, PA, USA 1964 31. Jain A. Abdin Y. Van Paepegem W. Verpoest I. Lomov S.V. Effective Anisotropic Stiffness of Inclusions with Debonded Interface for Eshelby-Based Models Compos. Struct. 2015 131 692 706 10.1016/j.compstruct.2015.06.007 32. Tandon G.P. Weng G.J. The Effect of Aspect Ratio of Inclusions on the Elastic Properties of Unidirectionally Aligned Composites Polym. Compos. 1984 5 327 333 10.1002/pc.750050413 33. Li H. Zare Y. Rhee K.Y. Mathematical Simplification of the Tandon–Weng Approach to the Mori–Tanaka Model for Estimating the Young’s Modulus of Clay/Polymer Nanocomposites JOM 2017 69 2819 2824 10.1007/s11837-017-2609-y 34. Gusev A.A. Finite Element Estimates of Viscoelastic Stiffness of Short Glass Fiber Reinforced Composites Compos. Struct. 2017 171 53 62 10.1016/j.compstruct.2017.03.021 35. Cintra J.S. Tucker C.L. Orthotropic Closure Approximations for Flow-induced Fiber Orientation J. Rheol. 1995 39 1095 1122 10.1122/1.550630 36. Borja R.I. Borja R.I. Plasticity: Modeling & Computation SpringerLink Springer Berlin/Heidelberg, Germany 2013 37. Castañeda P.P. Exact Second-Order Estimates for the Effective Mechanical Properties of Nonlinear Composite Materials J. Mech. Phys. Solids 1996 44 827 862 10.1016/0022-5096(96)00015-4 38. Digimat Documentation. MF and MX User’s Guide. Digimat 2022.1—Online Help (HTML) Available online: https://help.mscsoftware.com/bundle/digimat_2022.1/page/digimat_main.htm (accessed on 25 March 2022) 39. Hollomon J. Tensile Deformation AIME Trans 1945 32 268 290 40. Chen J. Zhao L. Zhou K. Improvement in the Mechanical Performance of Multi Jet Fusion–Printed Aramid Fiber/Polyamide 12 Composites by Fiber Surface Modification Addit. Manuf. 2022 51 102576 10.1016/j.addma.2021.102576 41. Xu Z. Yang T. Nakamura M. Yang Y. Hamada H. Effect of Carbon Powder Surface Treatment on Carbon Fiber Reinforced PA Composites Energy Procedia 2016 89 15 23 10.1016/j.egypro.2016.05.003 42. Foss P.H. Tseng H.-C. Snawerdt J. Chang Y.-J. Yang W.-H. Hsu C.-H. Prediction of Fiber Orientation Distribution in Injection Molded Parts Using Moldex3D Simulation Polym. Compos. 2014 35 671 680 10.1002/pc.22710 43. Rosato D.V. Rosato D.V. Plastics Engineered Product Design Elsevier Advanced Technology Inglaterra/Estados Unidos, UK 2003

PMC009xxxxxx/PMC9099610.txt

==== Front Nutrients Nutrients nutrients Nutrients 2072-6643 MDPI 10.3390/nu14091969 nutrients-14-01969 Review Poultry Meat and Eggs as an Alternative Source of n-3 Long-Chain Polyunsaturated Fatty Acids for Human Nutrition Cartoni Mancinelli Alice 1* https://orcid.org/0000-0001-5063-6785 Mattioli Simona 1 Twining Cornelia 2 Dal Bosco Alessandro 1 Donoghue Ann M. 3 Arsi Komala 3 Angelucci Elisa 1 Chiattelli Diletta 1 https://orcid.org/0000-0002-6134-0901 Castellini Cesare 1 Cornish Stephen Academic Editor 1 Department of Agricultural, Food and Environmental Science, University of Perugia, Borgo XX Giugno, 74, 06100 Perugia, Italy; simona.mattioli@unipg.it (S.M.); alessandro.dalbosco@unipg.it (A.D.B.); elisa.angelucci@unipg.it (E.A.); diletta.chiattelli@libero.it (D.C.); cesare.castellini@unipg.it (C.C.) 2 Department of Fish Ecology and Evolution, Eawag: Swiss Federal Institute of Aquatic and Technical Sciences, Seestrasse 79, 6047 Kastanienbaum, Switzerland; cornelia.twining@gmail.com 3 Poultry Production and Product Safety Research Unit, ARS, USDA, Fayetteville, AR 72701, USA; annie.donoghue@usda.gov (A.M.D.); karsi@uark.edu (K.A.) * Correspondence: alice.cartonimancinelli@unipg.it 08 5 2022 5 2022 14 9 196910 4 2022 04 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). The beneficial effects of n-3 long-chain polyunsaturated fatty acids (n-3 LC-PUFA) on human health are widely known. Humans are rather inefficient in synthesizing n-3 LC-PUFA; thus, these compounds should be supplemented in the diet. However, most Western human diets have unbalanced n-6/n-3 ratios resulting from eating habits and the fact that fish sources (rich in n-3 LC-PUFA) are not sufficient (worldwide deficit ~347,956 t/y) to meet the world requirements. In this context, it is necessary to find new and sustainable sources of n-3 LC-PUFA. Poultry products can provide humans n-3 LC-PUFA due to physiological characteristics and the wide consumption of meat and eggs. The present work aims to provide a general overview of the main strategies that should be adopted during rearing and postproduction to enrich and preserve n-3 LC-PUFA in poultry products. The strategies include dietary supplementation of α-Linolenic acid (ALA) or n-3 LC-PUFA, or enhancing n-3 LC-PUFA by improving the LA (Linoleic acid)/ALA ratio and antioxidant concentrations. Moreover, factors such as genotype, rearing system, transport, and cooking processes can impact the n-3 LC-PUFA in poultry products. The use of a multifactorial view in the entire production chain allows the relevant enrichment and preservation of n-3 LC-PUFA in poultry products. poultry genotype long-chain polyunsaturated fatty acids antioxidants PRIN20172017S229WC This research was funded by PRIN2017, grant number 2017S229WC. ==== Body pmc1. Introduction The beneficial functions of long-chain polyunsaturated fatty acids of the n-3 series (n-3 LC-PUFA) in human health are well-known, as well as the importance of having a lower n-6/n-3 fatty acids ratio in diets [1]. However, humans are rather inefficient in synthesizing LC-PUFA; consequently, a certain amount of these compounds need to be acquired directly from the diet [2]. The current eating patterns, mainly in Western countries, result in an excessive intake of n-6 with a consequent n-6/n-3 unbalanced ratio [3]. This has led to a need to increase the n-3 content of foods. However, the benefits of an increase in n-3 LC-PUFA consumption for human health are in contrast with the difficulty in finding sustainable sources of these fatty acids. Fish are one of the richest sources of n-3 LC-PUFA, but, due to the potential of low sustainability of both fishing and aquaculture, it is important to shift the focus to an n-3 source derived from terrestrial animals with higher nutritional value and possibly with lower environmental impacts. It is well-known that, by altering the composition of animal feed, it is possible to modify the fatty acid profiles of products (e.g., milk, egg, meat [4]). It is important to stress that, due to the hydrogenated activity of microorganisms of the rumen in polygastric animals, there are closer relationships between the fatty acid composition of feeds and animal products in monogastric species. For example, it has been reported [5] that rabbits fed a diet enriched with n-3 precursor (C18:3n-3, α-Linoleic acid, ALA) exhibited meat with higher levels of n-3 LC-PUFA. Progressive increases in the n-3 content of eggs have been shown in laying hens fed diets containing 10% and 20% of n-3 precursors [6]. This review aims to provide a general overview of the importance of n-3 LC-PUFA in human diets and focuses on ways to increase their content in terrestrial animal products. Due to its physiological characteristics (monogastric species and short rearing cycle) and the widespread consumption of meat and eggs, poultry is considered to provide suitable terrestrial sources of n-3 LC-PUFA. The various factors related to increasing and preserving n-3 LC-PUFA in chicken products throughout the production chain are also outlined. In particular, the effects of the nutritional strategies, genotype, and the adopted rearing system are discussed. Further, to preserve the obtained n-3 LC-PUFA, the transport of the birds to the slaughterhouse and the cooking processes are also addressed. 2. Evolution of Human Diet Early records of human diet date back to the Paleolithic period when the development of small utensils made plant resources, such as roots and tubers, more accessible for human consumption [7]. A key event that influenced the evolution of several current human diets was the domestication of animals and the cultivation of plants. Furthermore, the development of technological processes changed the nutrient components related to “wild” foods. Compared with their domesticated relatives, wild animals have consistently low fat content [8]. Most fat in domesticated animals is characterized by high levels of saturated fatty acids (SFA), some of which can negatively impact human health in terms of cardiovascular disease risk (CVD [9,10]), while polyunsaturated fatty acid (PUFA) and mainly n-3 LC-PUFA show numerous positive effects [11], such as preventing pathological disorders (see Section 3.1). For example, the current human diets in the US contain 12% of the total energy as saturated fats [12], while this value in the Paleolithic diet was only 7.5% [13,14]. However, the main difference between current versus Paleolithic diets is not in terms of the total fat intake (Paleolithic diet about 35% of total energy intake versus current recommendation of 20–35%) but mainly the ratio between different PUFA [12]. During the early evolutionary history of the genus Homo, there was a strict balance between n-6 and n-3 PUFA. In particular, a significant amount of n-3 fatty acids was present in the foods commonly utilized by prehistoric humans, including meat from hunting animals, wild plants, eggs, and fish. The n-3 LC-PUFA are the major structurally significant and biochemically active components of the brain in all mammalian species [15]. The phenomenon described as “encephalization” proposes that the increase in the brain size in primates and humans is probably due to the availability of n-3 LC-PUFA in the diet during early human evolution [16]. However, dietary changes in the last 100 to 150 years have influenced the type and the amount of PUFA, as well as other bioactive molecules (e.g., antioxidant, essential amino acids, micronutrients, etc.) in the food. The best available estimate of ancestral human intake showed a different ratio between n-6 and n-3 (about 2:1) [14,17]. In contrast, the estimated PUFA consumption for humans today in the USA is about 15 g/d, but the intake of n-6 LC-PUFA is 10-fold higher than n-3 LC-PUFA [18]. In EU countries, there is a wide variation in LC-PUFA intake [19], with the amount of Eicosapentaenoic (C20:5n-3, EPA) plus Docosahexaenoic (C22:6n-3, DHA) acid varying between 106 and 419 mg/d, which is below the daily recommendations (500 mg/d). In critical population groups (e.g., infants, children, adolescents, elderly, and pregnant/lactating women) the ALA intake is within the daily dose recommendation in 77% of the EU countries, whereas the n-3 LC-PUFA intake met the recommendations only in 26% of the European countries. These results indicate that the intake of n-3 and n-6 PUFA is suboptimal in specific population groups in Europe [20] and that their ratio is about 12 times higher than the recommendations [21] and has increased since 1961. Thus, in Western countries, human diets are unbalanced in terms of PUFA, with a significant increase in n-6 and a decrease in n-3 LC-PUFA. Overall, during the last 100 years, the ratio between n-6 and n-3 has significantly increased to 20:1, whereas the recommendations suggest that this ratio should be less than 4:1 (Figure 1). 3. Notes on Metabolism of Essential Fatty Acids and Desaturase Activity (Details of Synthesis) Linoleic acid (C18:2n-6, LA) and ALA are the precursors of n-6 and n-3 LC-PUFA, respectively. These precursors cannot be synthesized by mammals, including humans, and birds, and, for this reason, they are defined as essential fatty acids (EFA; [23]). The most physiologically important n-3 LC-PUFA are EPA and DHA, whereas Arachidonic acid (C20:4n-6, AA) is the most important n-6 LC-PUFA [23]. Both LA and ALA can be converted into long-chain metabolites through an elongation and desaturation metabolic process (Figure 2), which predominantly occurs in the liver. Other tissues (e.g., brain, testicles, epididymis, ovaries, muscles), although capable of synthesizing a certain amount of LC-PUFA, have lower metabolic efficiency [24,25,26]. The Δ5- and Δ6-desaturases enzymatic complex, controlled by the fatty acid desaturase 1 and 2 (FADS1 and FADS2) genes, introduces double bonds into the respective fatty acids (i.e., LA or ALA). The FADS2 acts twice, first at the C18 level and second after the conversion of the latter into C24 derivatives; accordingly, it is considered one of the main factors limiting LC-PUFA biosynthesis. The conversion of ALA into EPA is in competition with the conversion of LA into AA because the biosynthesis of both n-3 and n-6 LC-PUFA requires the same enzymes [23] (Figure 2). This competition for desaturases and elongases in n-3 and n-6 LC-PUFA synthesis affects their relative concentration in tissues. Thus, animals fed diets rich in n-6 produce more LA metabolites, such as AA, compared to ALA metabolites, such as EPA and DHA, whereas higher ALA intake results in increased synthesis of n-3 LC-PUFA [26]. These two PUFA families, ALA and EPA for the n-3 and AA for the n-6 series, are the precursors of the main compounds involved in the inflammatory response, such as prostaglandins (PGs), thromboxanes (TH), and leukotrienes (LT) [27]. Such molecules, called eicosanoids, have opposite functions: anti-inflammatory and anti-aggregating properties when synthesized by n-3 LC-PUFA, whereas pro-inflammatory and aggregating properties when derived by n-6 LC-PUFA. In particular, Cyclooxygenase (COX) is the key enzyme in the synthesis of PGs from AA, and it is present in two isoforms:COX-1, a constitutive enzyme widely expressed in most tissues because it controls the synthesis of PGs involved in the regulation of homeostatic function; COX-2, a specific enzyme, exerts its functions only during inflammatory processes; thus, PGs formed by COX-2 are principally involved as mediators of pain and inflammation [27,28]. The biosynthesis of PGs is triggered following the onset of extracellular stimuli. These stimuli, in turn, trigger the activity of phospholipase A2 (PLA2) and phospholipase C (PLC), which cleaves phospholipids from the plasma membrane and increases the availability of fatty acids for the PGs synthesis by the COX enzyme (Figure 3). The ALA and EPA can be processed by COX-1 to generate PGs of the 3-series (PG3, less inflammatory), while AA can be processed by COX-2 to generate prostaglandins of the 2-series (PG2) or by epoxygenase and lipoxygenase to form epoxyeicosatrienoic acids (EETs), thromboxanes, leukotrienes, and hydroxyeicosatetraenoic acids (di-HETEsM) [29]. As previously reported, the PG2 are highly inflammatory and responsible for the pathophysiological process of fever (e.g., prostaglandin E2, PGE2). The availability of ALA, EPA, or AA for the PGs synthesis is impacted by the composition of the membrane phospholipids, which, in turn, is influenced by the levels of n-6 and n-3 precursors in the diet. A significant body of literature demonstrates that increased ingestion of n-3 PUFA is associated with a decreased PG2 synthesis [3,30]. This is due to the replacement of AA with n-3 PUFA in the phospholipids, which, when released, attenuates the rate of PG2 formation (AA-derived). Accordingly, modification of n-3 and n-6 PUFA (especially AA) availability with different strategies (diet, drugs, etc.) strongly affects the immune and inflammatory responses of the body [31]. Thus, the fatty acid composition of tissues can influence their inflammatory responses. This process of inflammation represents a physiological defense mechanism protecting the body from infection and diseases; however, it must be well-regulated in order to maintain homeostasis (inflammation vs. anti-inflammatory). Because n-6 PUFA content is much greater than n-3 PUFA in typical Western diets, controlling dietary AA allows a down-regulation of PG2 synthesis and, consequently, anomalous inflammatory responses [32]. In this context, it is important to mention a new class of compounds, recently discovered, called isoprostanoids (IsoPs). The IsoPs are a series of PGs-like compounds produced by the free-radical-catalyzed peroxidation of fatty acids, independent of the COX [33,34]. On the basis of the LC-PUFA involved, IsoPs can be divided into different classes: F2-IsoPs derived from AA oxidation; F1, F3, and F4-ISoPs from ALA, EPA, and DHA, respectively [35]. Because of their specific derivation, IsoPs are considered a very sensitive marker of lipid peroxidation. Recent studies [36,37,38] have established that these molecules, besides being robust markers of oxidative damage, also exhibit a wide range of biological activities. F2-IsoPs exert a vasoconstriction function, especially at the renal level; the F4-ISoPs, also called neuroprostane (4-F4t-NeuroP), is mainly present in brain tissue and increases its concentration with the onset of neurodegenerative disease. F1-ISoPs, because of its origin from ALA, is considered the main oxidation biomarker in plants, enhancing its level under stress conditions. Due to issues with the imbalance ratio of n-6/n-3 in modern diets, it was recommended to increase the amount of n-3 in the human diet, and, subsequently, there has been an increased interest in the research towards the IsoPs derived from EPA and DHA. In fact, since n-3 LC-PUFA exert beneficial functions for human health, it is possible to hypothesize that their derivative compounds can also be beneficial. Many studies report that a high dietary level of n-3 PUFA leads to the formation of F3-IsoPs and F4-IsoPs from non-enzymatic oxidation of EPA and DHA, respectively [26]. In addition, their abundance is also affected by the PUFA composition of different organs; for example, the brain and spermatozoa are particularly rich in n-3 LC-PUFA and also in F4-IsoPs [26]. However, despite the importance of the IsoPs at the physiological level, further investigations are needed to better understand the different role and metabolic pathways involving such compounds and their relation with the LC-PUFA precursor. 3.1. Relevance of n-3 LC-PUFA in Human Nutrition In the last few decades, the LC-PUFA compounds have been largely investigated for their nutritional value and for the numerous biological actions and therapeutic functions in different organs. The n-3 LC-PUFA are particularly abundant in the brain, retina, and reproductive cells and play important roles in many metabolic pathways, preventing pathological disorders, such as cardiovascular disease, reproductive dysfunction, chronic inflammatory diseases, depression, and deficiencies in the immune system [39,40]. As previously mentioned, humans are rather inefficient in synthesizing LC-PUFA [41] and, accordingly, they should be consumed directly through the diet. Thus, experts, such as the European Food Safety Authority (EFSA), have established n-3 nutritional recommendations in relation to age, sex, and body condition. It has been suggested that 250 mg/d is the minimum quantity of EPA and DHA required for an adult [42]. For pregnancy and lactation, considering the increased demand of n-3 LC-PUFA, the suggested intake is 350 to 450 mg/d of DHA, while, for young children (<24 months), the adequate intake is 100 mg/d of DHA. DHA is the main LC-PUFA recommended as it is a key component of the membrane lipids of the nervous system and adequate DHA concentration is linked to optimal brain development that occurs in the first 2 years of life [1]. In the fetus and newborn, the supply of DHA depends on the maternal diet since studies [43,44] showed that a high n-3 LC-PUFA intake by pregnant or lactating women can promote mental development in babies. On the contrary, low n-3 LC-PUFA concentration during the fetal period affects the brain volume, reducing its dimensions in childhood [45]. Moreover, in the brain, n-3 LC-PUFA, besides the formation of the plasma membrane, exert other functions, such as an increase in cognitive activity [46] and development of synaptic functionality and plasticity [47]. Indeed, a recognized beneficial effect of n-3 LC-PUFA consumption concerns neurodegenerative disorders, such as Parkinson’s (PD) and Alzheimer’s disease (AD) [48,49]. Accordingly, a lower concentration of DHA was found in the serum of patients affected by AD as compared to healthy people [50], suggesting that n-3 LC-PUFA could represent a preventative strategy against AD, especially when consumed in the early stage of the disease [51]. Numerous clinical trials were carried out to investigate the effects of n-3 consumption on the reduction in CVD; however, an open debate is underway regarding the efficacy. Recent studies support the hypothesis that the intake of n-3 LC-PUFA reduces the risk of CVD [52,53], but not all the research papers agree with this outcome [54] as varying experimental designs and the heterogenicity of the results render it difficult to find clear conclusions [55]. The EFSA [56] has performed numerous studies to determine the tolerable upper intake (UI) of n-3, concluding that a higher intake did not induce any adverse effect on human health; therefore, individuals could safely increase their daily n-3 LC-PUFA consumption. The n-3 LC-PUFA represent about 30 to 50% of the membrane composition of sperm cells and are fundamental for reproductive activity by regulating the fluidity and the acrosomal responsiveness [57,58]. The high LC-PUFA concentration in the spermatozoa plasma membrane makes it vulnerable to lipid peroxidation [59]. Thus, the presence of oxidative stressors, such as obesity, sexually transmitted disease, alcohol, and tobacco use, could represent possible factors affecting male infertility [60,61]. The Western diet rich in saturated fat and n-6 PUFA induces obesity that is associated with many diseases, such as cancers, behavioral disorders, cardiovascular complications, and insulin resistance. Although these pathologies come from different causes and conditions, their onset is linked to the increase in n-6 PUFA and decrease in n-3 PUFA in the modern diet so that they are defined as diet-related chronic diseases. As previously reported, the overconsumption of n-6 PUFA increases the synthesis of PG2, with a pro-inflammatory effect. Moreover, it has been recently discovered that the AA-derived endocannabinoid compounds linking to their brain receptors are able to increase the appetite and food intake, worsening obesity status [62]. The AA is widely present in various cells and tissues so that it can be quickly converted into pro-inflammatory eicosanoids, increasing the chronic disorders [62]. The higher AA concentration with respect to EPA and DHA can lead to oxidative stress in the cell due to the increase in reactive oxygen species production [63]. Oxidative and inflammation status can negatively affect the reproductive function and elevate the risk of CVD, cancer, and other chronic diseases. Therefore, increasing the n-3 intake in humans in order to maintain a balanced n-6/n-3 ratio is essential for regulating body homeostasis. 3.2. Global Requirements for EPA and DHA As previously reported, EFSA [42] provides the recommended minimum daily intake of EPA and DHA for different classes of population. Combining these data with those related to the world population classes [64,65], it is possible to estimate the annual global requirement for n-3 LC-PUFA (Table 1). As reported in Table 1, the annual global requirement for n-3 LC-PUFA (mainly EPA + DHA) to satisfy the need of the whole world population is about 722,960 t/y. The main source of n-3 is represented by fish; however, the supply of n-3 LC-PUFA from fish and fish products is insufficient to meet the world n-3 LC-PUFA demand [66]. Consequently, the estimated n-3 LC-PUFA deficit is about 347,956 t/y. It is noteworthy that such estimates do not take into account the most vulnerable groups of population (i.e., elderly and patients affected by physiological and metabolic disorders); thus, the amount of n-3 LC-PUFA deficit could be even higher. Within this context, it is not clear how to satisfy this enormous n-3 LC-PUFA deficit with the current foods. Accordingly, it is vital to develop alternative strategies for increasing n-3 PUFA and mainly n-3 LC-PUFA availability in common foods. 4. Sources of n-3 LC-PUFA and Strategies for Enriching Terrestrial Food The importance of n-3 LC-PUFA in human health is evident, and they have been extensively described herein; however, the problem of how to increase the n-3 intake in the human diet is still unsolved. In this section, the principal n-3 LC-PUFA sources in the human diet (vegetable oil, fish, and terrestrial animal products) are described and new strategies to increase their availability in the food are discussed. 4.1. Vegetable Source The beneficial effects of n-3 LC-PUFA on human health, and the low presence of these compounds in common foods, promoted the development of foods “enriched” in these fatty acids and the development of alternative terrestrial n-3 LC-PUFA sources. The main sources of PUFA in human diets are vegetable oils. Many plants are able to synthesize ALA by successive desaturation of oleic acid and LA [67]. This process takes place in the plant leaves, roots, and seeds. Although vascular plants exhibit two distinct pathways for PUFA biosynthesis [68], fatty acid biosynthesis occurs almost exclusively in the plastids [69], and it is catalyzed by fatty acid synthase [70], which permits the synthesis of long-chain MUFA (monounsaturated fatty acids), such as erucic acid (C22:1), but the production of AA, EPA, and DHA is null or very scarce [71]. Most of the elongase enzymes involved in glycerolipid metabolism in plants have relatively broad substrate specificities capable of synthesis of fatty acids with 20 and 22 carbon atoms (C20 and C22 fatty acids [72]). Accordingly, LC-PUFA are only marginally synthesized because plants have a low desaturase activity on fatty acids with over 20 carbon atoms [72,73]. Some fungi, bryophytes (i.e., mosses and liverworts), and some marine and freshwater algae are capable of synthesizing LC-PUFA. However, the mechanisms of how these organisms are able to produce these compounds are not known [74,75,76,77]. In particular, these microorganisms, besides having the same desaturases present in the higher plants, show a higher affinity for the C20 and C22 fatty acids. Moreover, while some vascular plants are rich in fats, including ALA and other PUFA (e.g., flaxseed, canola), even when ALA constitutes a considerable fraction of fatty acids [78,79], the ALA content of most plants is low because they are low in fat [80,81,82]. 4.2. Fish and Fish Products Fish are the main source of n-3 LC-PUFA for humans, but the content in EPA and DHA varies by species and by how the fish are raised (i.e., wild or farm-raised, warm or cold water) [83]. For example, Pacific cod has higher EPA and DHA content compared to Atlantic cod (0.235 and 85 g/100g vs. 0.134 and 85 g/100g, respectively) [83]. Farm-raised and wild fish often contain similar amounts of EPA and DHA, but farmed fish are typically fed fish meal [84], which is not sustainable at a global scale because wild fish are used to produce fish meal [85]. In addition, the total SFA and PUFA content in farm-raised fish is higher compared to wild fish due to the higher n-6 concentrations in fish feeds [84]. Importantly, a number of marine fish of high commercial value for human consumption are unable to synthesize n-3 LC-PUFA. In fact, studies on hepatocytes from marine fish have shown very low ALA desaturation rates, without the production of EPA or DHA [86]. As previously mentioned, LC-PUFA synthesis depends on the desaturase and elongase activities. The relative inability of some marine fish to produce EPA and DHA can result from limited activities of either C18 or C20 elongases, as well as from low activity of Δ5 desaturase, which converts 20:4n-3 into EPA [87]. These enzymes appear to vary in their efficiency based upon the availability of PUFA in natural ecosystems [88,89]. Marine fish have large amounts of EPA and DHA in their diets, whereas freshwater fish consume diets that are more variable in n-3 LC-PUFA content. For example, carnivorous marine fish (e.g., tuna) lack functional desaturation and elongation enzymes, likely because they have little selective pressure to maintain synthesis in an environment where it is not strictly required [90]. Accordingly, many marine fish are only “accumulators” of n-3 LC-PUFA produced by lower trophic levels (i.e., marine phytoplankton). For example, phytoplankton rich in EPA and DHA include Bacillariophyceae, and Chrysophyceae; Cryptophyceae, Prasinophyceae, Rhodophyceae, Xanthophyceae, Glaucophyceae and Eustigmatophyceae (EPA sources), Dinophyceae, Prymnesiophyceae, and Euglenophyceae (DHA sources) [91,92]. These algae can also have high lipid contents, as well as high n-3 LC-PUFA concentrations (around 30–70% of DHA). Although all fish originally evolved from marine lineages, in less n-3 LC-PUFA-rich freshwater environments, fish evolved an increased ability to synthesize n-3, including multiple species and populations with relatively recent marine origins [93]. Even though many studies have been carried out to increase the sustainability of aquaculture and, in particular, for the replacement of fish meal with other sources [94,95], fish are becoming progressively scarce and the over-exploitation of fishing areas worldwide is unsustainable. In addition, aquaculture cannot be considered a very sustainable source because, paradoxically, the feed used contains large quantities of wild fish [96]. Beyond the problems of sustainability, there is a growing concern with the methylmercury and polychlorinated biphenyls (PCB) levels in some species of fish, such as swordfish (Xiphias gladius), mackerel (including different species of pelagic fish, mostly from the family Scombridae family), and shark (Selachimorpha). Hites et al. [97] reported that levels of mercury and PCBs are higher in farm-raised salmon compared to wild salmon. The risks of methylmercury and PCB exposure are even more common in fish fed fish meal [98]. For this reason, the US Environmental Protection Agency [99], the US National Academy of Sciences [100], and additional international medical institutions recommend limiting the consumption of some species of fish. Therefore, the choice of terrestrial animals as sources of n-3 LC-PUFA, and, thus, their ability to elongate and desaturate PUFA, must be carefully considered in order to find other sustainable, healthy, and safe products. 4.3. Terrestrial (Farmed) Animals Terrestrial animals generally have lower content of n-3 LC-PUFA in their body than fish and a modest ability to synthesize LC-PUFA. Because the development of an embryo and fetus largely depends on LC-PUFA availability [101], females have higher n-3 LC-PUFA content and synthesis capacity than males [102]. Accordingly, female rats showed a higher DHA anabolism than males [102] and, thus, a higher liver concentration of DHA [103]. The most direct relation between the fatty acid profile of feed and that obtained in the food produced by it (meat, egg) is obtained in monogastric species. On the contrary, in ruminants, ingested fatty acids are hydrogenated by microorganisms of the rumen [104]. Stearic acid (C18:0) is the end product of this reaction, and it passes from the rumen into the abomasum, where it is digested and absorbed [105]. Fatty acids from the rumen are precursors of plasma synthesis of triglycerides that are mainly incorporated into the lipids of milk and adipose tissue [104]. Because of the hydrogenation activity of bacteria, the triglycerides of ruminant plasma, milk, and body fat result in a low PUFA content [106]. In ruminants, in order to bypass the rumen, PUFA must be protected for absorption in the duodenum [107]. In this way, the concentration of n-3 PUFA in milk can be increased fourfold by the inclusion of rumen-protected tuna oil in the diet of cows [108]. Unlike ruminants, in monogastric species (poultry, pig, rabbit), the PUFA profile of feed results in corresponding levels in their products; therefore, several strategies for enriching their content can be used. Thanks to these dietary strategies, there are a number of animal products in the market that are enriched in n-3 fatty acids, such as [109]:Meat and poultry products (sausages, frankfurters, etc.); Eggs and egg products (mayonnaise, etc.); Milk and milk products (yoghurt, cheese, etc.). It should be noted that the current meat and egg supply from monogastric animals is estimated to produce around 75.632 t/n-3 LC-PUFA/year (Table 2), which represents about 21 to 22% of the LC-PUFA annual world requirements, whereas other foods (beef, lamb meat) can add only a minor amount of n-3 LC-PUFA (<1%) [110,111]. It is likely that the major variation in lipid intake between populations reflects the true underlying differences in intakes and types of lipids consumed. Since it is difficult to change dietary habits, it is very important to modify the lipid profile of foods. In this context, poultry is particularly interesting for the following reasons:monogastric animal; short breeding cycle; meat-type chickens have an age at slaughtering of about 40 days; there are no religious limitations for poultry meat (or not as many as for pork or beef/lamb); lower environmental impact than other livestock productive chain [112] due to the high efficiency in converting feed into food; it is the most-consumed meat in the world; eggs easily meet the EFSA recommendation for n-3-enriched foods. Poultry Meat and Eggs as Functional Foods Poultry meat is rich in protein and is very suitable for human nutrition due to its low-fat content, high unsaturation degree of fatty acid, and low cholesterol levels [113]. It can even be considered to be a “functional food” as poultry meat can be beneficial for human health because it contains bioactive substances, such as vitamins and antioxidants, and has a balance of n-6 to n-3 ratio [114] close to the recommended ratio of 4:1. It is also well-established that poultry eggs contain vitamins and minerals in addition to biologically active compounds with antimicrobial, immunomodulator, antioxidant, anti-cancer, or anti-hypertensive properties [115]. Owing to their high nutritional value and positive effects on human health, several of these compounds found in eggs are also selectively isolated and produced on an industrial scale [116]. Numerous studies have demonstrated that the quality of poultry meat and eggs can be further improved through various methods, including manipulating the diet to target certain functionalities, leading to the concept of the chicken as a bioreactor for the production of substances for humans [117,118,119]. 5. n-3 LC-PUFA in Poultry Meat and Eggs, Strategies of Enrichment Several studies have demonstrated that it is possible to increase the n-3 LC-PUFA content of animal origin products through different strategies, such as dietary supplementation, genetic selection, and rearing systems management. Moreover, pre- (transport to slaughterhouse) and post- (cooking) mortem factors can affect the preservation of the n-3 LC-PUFA enrichment in foods (see Section 8 and Section 9). 5.1. Dietary Strategies for Broilers and Laying Hens The n-3 LC-PUFA enrichment of livestock products is based on the dietary supplementation of n-3 PUFA precursors (ALA) from terrestrial sources or n-3 LC-PUFA from marine oils (Figure 4). The first strategy implies that ALA must be converted by animal metabolism into LC-PUFA, while, in the second case, the LC-PUFA is simply absorbed, transferred, and stored in different tissues. Although this second strategy (e.g., addition of fish oil) easily enriches food in n-3 LC-PUFA, it is highly dependent upon wild fish production in marine ecosystems. In this view, it is important to consider that fish oil is an economically and environmentally expensive additive and the feed represents the major animal production cost [120]. Moreover, the fish oil approach also has other drawbacks [121,122], demonstrated by the fact that its use in the chicken diet could negatively affect the sensory properties of the meat. Furthermore, it is important to point out that fish oil could be considered a constituent of the human diet [123]. In fact, the use of refined fish oil is much more metabolically efficient if administered directly to humans [124] without passing through the metabolism of livestock animals to produce food. Therefore, following these considerations, it is important to enforce the nutritional strategy based on the use of n-3 precursor. Further prospective could be represented by the introduction of insects to a poultry diet as a source of n-3 precursor or n-3 LC-PUFA. However, research still needs to be conducted to validate this aspect, and clarification in the regulations is needed to better understand how to manage the insects rearing (see Section 5.1.3). By manipulating the broiler and the laying hen diets, it is possible to improve the conversion efficiency of ALA into n-3 LC-PUFA by exploiting the bird metabolism. Despite broilers and laying hens exhibiting a different nutritional requirement, in order to efficiently administer ALA dietary supplementation, it is important to consider two main aspects (Figure 4):The n-6/n-3 ratio of the diet. In fact, due the involvement of the same enzymes, the n-3 synthesis is in competition with the synthesis of the n-6 one; thus, the higher LA diet presence could reduce the production of EPA and DHA by favoring the AA synthesis [125]; The antioxidant supplementation (vitamin E, vitamin C, Selenium, etc.). Due to their double bonds, PUFA are very susceptible to oxidation, resulting in reduced shelf life of feed as well as meat and eggs. This can lead to a poor acceptance of the feed by the animals but also a poor acceptance of n-3 LC-PUFA-enriched products by the final consumers due to the development of unattractive colors or unpleasant tastes and aromas. 5.1.1. Dietary Enrichment for Broilers In general, chicken meat represents a poor source of n-3 LC-PUFA; however, it is possible to increase their content by manipulating the broiler diet. In this context, several studies [126,127] were conducted in order to evaluate the use of micro- and macroalgae in poultry nutrition as an n-3 LC-PUFA source. In particular, the use of Spirulina algae in the broiler diet increases the n-3 LC-PUFA content in meat, particularly EPA and DHA, with a positive effect also in the n6/n3 ratio [128]. The same results were obtained by Costa et al. [129] through the enrichment of the broiler diet with Brown macroalgae (e.g., Laminaria digitata). However, it is important to consider that both micro- and macroalgae are characterized by the presence of cell walls resistant to degradation by digestive enzymes. Thus, a high level of algae in an animal diet can compromise the nutrient digestibility, with a negative effect on the growth performance of animals [129]. Consequently, the main source of n-3 PUFA (ALA) used in the nutrition field is represented by flaxseed that is administered to the birds in different products, such as seeds, oil, or extract. Feeding turkeys with a diet containing 2.5% flaxseed oil from 16 days to 3 weeks of age before slaughter resulted in the recommended n-6/n-3 polyunsaturated fatty acids ratio of 4:1 in the meat [130]. Chickens, through the hepatic elongase and desaturase enzymes, are able to produce n-3 LC-PUFA from ALA [131]. Several studies demonstrated that it is possible to increase the level of n-3 LC-PUFA in chicken meat through the dietary supplementation of ALA. Other researchers did not obtain the same results, suggesting that there are other factors affecting the n-3 LC-PUFA biosynthesis. Diets with 8% of ALA increased the level of n-3 LC-PUFA in the chicken meat nine times as compared to the control group [132]. In contrast, Lopez-Ferrera et al. [133] reported that administering 8% of ALA in the diet obtained only a 3.6 times increase in n-3 LPC in the breast meat. The discrepancies found in the different studies are potentially the result of different durations of feeding ALA, the LA level in the diet, and genetic strain of birds in the studies (see Section 7 and Section 8). Furthermore, the ALA sources can affect the organoleptic properties of chicken meat; for example, 10% flaxseed addition in the last 14 days of the rearing cycle did not affect its taste or aroma [134], while 7% of flaxseed oil addition, in the same period, produced a fishy odor and taste in chicken meat [135]. As previously affirmed, when the level of PUFA increases, a higher antioxidant protection should be obtained. Vitamin E is the main antioxidant used for fatty acids supplementation and is generally added as α-tocopheryl acetate in poultry diets. Vitamin E supplementation in broiler diets not only increases the total tocopherols concentration in the different tissues (liver > adipose tissue > dark meat > white meat) but also enhances the antioxidant defense of major tissues, decreasing lipid peroxidation [136]. A body of literature also underlines that, in poultry, the effects of the dietary treatments are tissue-specific [132]. In other words, is not possible to establish a linear correlation among the ALA level provided with the diet and the concentration of n-3 LC-PUFA found in the different muscles (mainly breast and drumstick) in poultry meat. Table 3 highlights some data from the literature. Although the studies were carried out in different experimental conditions, the same trend of n-3 LC-PUFA distribution in relation to the tissues was observed. In particular, when chickens are fed with ALA and EPA supplementation, a higher level of n-3 LC-PUFA was found in the liver compared to the breast [137] and drumstick [138]. Moreover, the drumstick data show a higher ALA content in respect to the breast; the latter exhibits a higher concentration of n-3 LC-PUFA (EPA and DHA) [139,140]. This is probably due to the different roles performed by the breast and drumstick muscle. It is well-known that the drumstick is involved in the movement and such activity consumes energy obtained through two main sources: carbohydrates and free fatty acids. The carbohydrates (mainly glycogen) are used for a fast and short contracting activity, whereas the free fatty acids are involved in the slow and prolonged exercise [141]. For this reason, the drumstick is rich in fats compared to the breast, and a portion is used (β-oxidation) for kinetic activity [142]. In order to obtain meat rich in n-3 LC-PUFA, it is fundamental to consider both birds’ ability to convert ALA into n-3 LC-PUFA and their storage efficiency of such compounds in the edible tissues [143]. 5.1.2. Dietary Enrichment for Laying Hens In poultry, as well as in other species [144,145], it has been demonstrated that the conversion of ALA is higher in females than in males [146]. As previously mentioned, sex has a significant effect on FADS expression and consequently on the n-3 LC-PUFA content of poultry products; adult female chickens have higher n-3 LC-PUFA synthesis ability due to needs of the chicken embryo. This is the reason why the efficiency of conversion of ALA into LC-PUFA varies between chicken meat and eggs. Accordingly, egg yolk is a good source of n-3 LC-PUFA, especially DHA; indeed, standard egg yolks contain 0.1% EPA, 0.7% DHA, and 0.8% ALA [147]. Over the past 20 years, the influence of the dietary supplementation (mainly flaxseed) on the productive performance of the hens and the characteristics of the eggs have been extensively studied; however, the results reported are highly variable. Several authors reported a decrease in feed consumption [148,149], while others an increase [150] when flaxseed was added to the diet. Moreover, concerning the egg production, some authors showed an increase [151], while others reported a decrease [148] or no change of deposition rate [152]. Similar discordance was observed for egg weight [148,152]. These contradictory results can be ascribed to differences in experimental conditions; in fact, it is known that the age of hens and the genotype influence productive performance [153,154]. However, in these studies, the most important factor is represented by the diet formulation and, in particular, by mechanical/chemical treatments of raw materials. Flaxseed and other seeds used as ALA source often contain antinutritional factors (ANFs) that negatively affect the palatability and the digestion efficiency of the diet [118,155]. The cyanogenic glycosides present in the flaxseed are ANFs, responsible for impaired respiration rate in laying hens. Moreover, other ANFs, such as phytic acid and trypsin inhibitors, increased the intestinal viscosity [156] by reducing nutrient bioavailability [157]. These effects reduce hen performance and affect the egg quality [158]. However, by mechanical processing (extruding, heating, or by enzyme supplementation) of the raw seeds, it is possible to reduce/eliminate the ANFs [159,160]. Recent studies reported that the flaxseed extrusion [159], or the use of oils or soluble ingredients as ALA source, did not affect the productive performance in laying hens [118,161]. Therefore, in order to increase the n-3 LC-PUFA content in the eggs, it is important to use mechanically processed sources of ALA to avoid the effect of ANFs. Regardless, laying diets rich in ALA, such as flaxseed, can increase the EPA and DHA content of their eggs [162,163]. Fraeye et al. [164] in their review concluded that dietary supplementation of flaxseed in laying hens proportionally increases the level of ALA in the yolk. Furthermore, amounts of DHA in yolk increase as well, but not in a linear way with respect to the level of flaxseed supplementation, suggesting that the LA/ALA ratio of the diet is one of the major factors affecting LC-PUFA synthesis. As previously mentioned, the common enzymatic pathway between n-3 and n-6 PUFA induces a competition for their synthesis; thus, the level of ALA and LA in the diet represents a crucial aspect. A recent meta-analysis [165] showed a linear relationship between the ALA levels in the diet and the amount of EPA, DHA, and total n-3 LC-PUFA in egg yolks, whereas a decrease in LA concentration was simultaneously observed. Authors reported that adding 100 g of ALA per kg in the diet resulted in 126 mg of DHA in the egg. Additionally, it was confirmed that, by increasing the LA content in the diet, and, consequently, the LA/ALA ratio, there was a linear decrease in the concentration of EPA and DHA in egg yolks. Given these relationships, diet composition (concentration of ALA, n-3 PUFA, and LA/ALA) can predict the content of EPA and DHA in the egg with a certain accuracy. However, it is important to consider that, even though numerous studies show that the decrease in LA and the increase of ALA in animal feed promotes the synthesis of DHA [166,167], other researchers reported that the continuous exposure to diets characterized by a high gap of LA to ALA decreased the egg weight [168]. Dong et al. [169] affirmed that hens fed fish oil supplementation over 16 weeks resulted in decreased egg weight. Thus, it is necessary to maintain a balanced LA/ALA ratio in the diet. Additionally, the age of hen is important and affects the n-3 LC-PUFA synthesis. Older hens characterized by higher liver dimensions are more efficient in metabolizing DHA from ALA compared to younger hens [164]. Flaxseed addition also affects other nutritional traits of egg. Mattioli et al. [117] show that the supplementation of the hen’s diet with flax and alfalfa sprouts reduces plasma and egg cholesterol and increases the n-3 PUFA, vitamins (α-tocopherol, α-γ-tocotrienol, retinol), carotenes (β-carotene, lutein, zeaxanthin) and phytoestrogens (daidzein, equol, isolariciresinol). It is important to underline that direct dietary alterations represent the main method for modifying LC-PUFA content in animal products; however, genetic [170] and rearing strategies [171] should not be ignored (see Section 6 and Section 7). 5.1.3. Use of Insect and Earthworms as a Future Prospective In recent years, many efforts have been made to find alternative and sustainable sources of feed for animals. One of the most promising sources is insects. Insects constitute more than 75% of the animal kingdom [172] and have potential as a sustainable source of food and feed. Most edible insects are rich in protein, lipids, and minerals. Insects and other invertebrates are natural protein sources for poultry and can potentially replace fish and soybean meal. Insects, due to their high reproductive potential, chemical characteristics, low water and space requirements, ability to use waste as feed, and low environmental impact, can be produced sustainably for livestock feed [173]. To increase sustainability, the diet for insects should consist of previously unused biomass, such as waste, low-value by-products, and non-traditional livestock feedstuffs. However, currently, there are limitations on the types of feed allowed in insect rearing. Since insects are considered livestock, the ingredients allowed in insects’ diets are subjected to EU regulations on feed hygiene, which restricts the use of food sources such as catering waste and processed animal protein [174]. In addition, the nutrient composition of insects varies with species, age, life stage, production, and processing conditions [175,176,177]. Even though significant variations exist in the nutrient composition of edible insects, many of these insects are high in monounsaturated and/or PUFA, mainly LA, ALA, and γ-linolenic acid [178], whereas LC-PUFA are scarce. The fatty acid composition of insects also depends on the environment where they develop [178]. Terrestrial species have lower LC-PUFA content, especially lower n-3 LC-PUFA, compared to species with an aquatic larval stage [179] due to the major differences in n-3 LC-PUFA availability in terrestrial plants versus aquatic primary producers such as algae. Theoretically, insects could contribute to n-3 PUFA and LC-PUFA requirements for humans either by direct consumption of insects rich in n-3 PUFA or indirectly through consumption of fish and poultry products fed on such insects [180]. Among edible insects, black soldier flies (BSF, Hermetia illucens) are one of the most-studied, easily reared, and widely approved insects for use in poultry feeds in the US and Europe. BSF larvae can be raised on a wide range of substrates, resulting in insect biomass suitable for feeding animals [181]. The FA composition of insects such as BSF is partly determined by the composition of their diet, which can be modified to achieve a favorable n-6/n-3 ratio for animal feed. However, Hoc et al. [182], investigating the long-chain metabolic activity of BSF larvae, found that such insects produce scarcely any LC-PUFA. For example, even when fed high levels of ALA from flax-enriched diets, the larvae bioaccumulated around 13% of this fatty acid and metabolized approximately two-thirds of it into saturated fatty acid, such as lauric or myristic acid. In a recent study, Rossi et al., [183] demonstrated an LC-PUFA enrichment in yellow mealworm (Tenebrio molitor) larvae only when they were reared on diets enriched in fish oil but not on sunflower or flaxseed oil diets. Such results suggest that these insects (i.e., T. molitor or BSF) are not able to convert ALA into n-3 LC-PUFA, and, thus, they are not a potential source of these compounds for consumption. According with the above-reported studies, both BSF and Tenebrio molitor showed a scarce efficiency to convert ALA into n-3 LC-PUFA. Therefore, they are mainly used as a source of protein [184,185] and lipids [186,187] as replacements of conventional ones. However, studies on other insects, such as mealworms, have demonstrated that adding a source of n-3 fatty acids to their diet can significantly increase the n-3 PUFA content of some insect meals [183,188,189]. For instance, a recent study showed that an inclusion of 4% flax seed oil in diet resulted in 10- to 20-fold increase in n-3 fatty acid content in house crickets, lesser in mealworms and BSF [190]. Other studies also indicated an increase in the n-3 LC-PUFA content in BSF larvae when raised on oil seed by-products, fish offals, or seaweed-based mediums [181,191,192]. Grass-fed poultry can naturally ingest various amounts of insects and earthworms. Earthworms are yet another alternative and little-explored source of protein and LC-PUFA for domestic animals. The use of earthworms presents a unique opportunity as earthworms can efficiently recycle the organic wastes and by-products from livestock operations into valuable feed sources for animals. Earthworms are high in essential amino acids and n-3 PUFA compared to other insects, and they can obtain EPA from their gut microflora instead of depending on dietary sources [193,194]. Earthworms are already part of a chicken’s natural diet and their effect on inclusion in poultry diets has been reported occasionally from various geographical regions across Asia [195,196,197,198,199,200]. Dietary supplementation of earthworm meal (0.2–0.6%) improved performance in broilers as well as layers, especially in terms of laying performance, egg quality, and n-6/n-3 ratio of FA in egg yolks [198]. The content of n-3 FA, mainly EPA and DHA, is also important in the earthworm flour. For instance, the flour of E. Andrei is recommended as a source of protein and lipid in fish feed because of its high protein and PUFA content [201]. The current regulations in Western countries (EU and USA) do not allow the use of earthworms as an animal feed when reared on wastes. For example, the current EU legislation only permits certain insect species, including BSF, common house fly (Musca domestica), the Coleopetran species yellow mealworm, lesser mealworm (Alphitobius diaperinus), house cricket (Acheta domesticus), banded cricket (Gryllodes sigillatus), and field cricket (Gryllus spp.), which have been reared on materials of vegetal origin in aquaculture feed, but prohibits raising insects or earthworms on catering or manure waste due to the risk of pathogen transmission [202]. However, there are fewer restrictions in Asia, Africa, and America in terms of insect species and rearing substrates. More research is needed to identify the best substrates to raise earthworms, the safety of the feed sources, and the identification of the best species of earthworms to be reared for use in poultry and aquaculture. 6. Poultry Genotype The current European regulations do not provide a clear classification of meat-type chicken genotypes, so most EU countries make this discrimination on the basis of daily weight gain (DWG). Therefore, broilers can be divided into three major groups based on their productivity:Fast-growing genotypes (FG) are represented by birds used in intensive rearing systems reaching commercial weight in a very short time and characterized by a high breast yield (> 25% live weight). The most common genotypes are selected for precocity; at about 40 days of age their weight is more than 2.5 kg; Medium-growing genotypes (MG), also known as slower-growing genotypes (SrG), are a recently recognized group and comprise some commercial chicken genotypes that are lower-performing then the FG ones, which is why the breed companies define them as SrG. These genotypes are less common in intensive rearing systems, but they are widely used in alternative rearing systems (e.g., free-range and organic); Slow-growing genotypes (SG) are represented by breeds that are very important for maintaining biodiversity and genetic variability but, due to their low productive performance (growth rate and breast yield), are not utilized in intensive rearing systems. Thus, they are mainly used for niche-production in small-scale farms [203]. In the alternative systems, the presence of outdoor runs renders necessary the use of suitable genotypes with specific features, such as: kinetic and foraging activity, thermo-tolerance, and immune response of the organism [204,205]. It is widely known that FG genotypes use most of their dietary energy towards body growth, while the SG genotypes spend most of their energy on metabolic functions, such as thermoregulation, movement, and foraging [206]. Sirri et al. [207], comparing organically reared SG, MG, and FG chickens, found higher n-6 and n-3 PUFA concentration in the breast of SG, suggesting differential expression of genes encoding for desaturating enzymes. However, since SG and MG birds are reported to eat more grass than FG genotypes [208], their higher dietary ALA intake could have also contributed to the higher degree of unsaturation in their meat. Additional papers [209] confirmed that MG, and particularly SG chickens, showed a greater expression of FADS2 and FADS1 genes and a higher Δ6 and Δ5 activity and, consequently, higher n-3 LC-PUFA content in breast meat compared to FG genotypes. These findings agree with resource allocation theory [206] because the synthesis of n-3 LC-PUFA, due to additional cycles of elongation and desaturation (Figure 2 Section 3) and a final β-oxidation, is more metabolically expensive than that of the n-6. This could explain the preference of the FG, compared to SG, for n-6 [210]. As previously discussed, the competition between the two metabolic pathways (n-3 and n-6) affects the relative synthesis of different LC-PUFA. Dal Bosco et al. [211] show that the estimated Δ5/Δ6-desaturase index was higher in the SG as compared to the FG genotypes and, consequently, the SG meat was characterized by a higher percentage of LC-PUFA (both n-3 and n-6) with respect to the FG meat. Accordingly, genetic selection for high performance unintentionally modified the expression of genes coding for enzymes involved in LC-PUFA synthesis as well as the relative enzymatic activity [212]. The relationship between genotype and desaturating ability has been demonstrated to have a significant impact on the PUFAs content of meat. A recent paper [143] underlines that the presence of LC-PUFA in chicken meat depends on two main factors: the liver desaturase activity of the bird (conversion of ALA into LC-PUFA) and the storage capability of the LC-PUFA synthesized in tissues (mainly muscle). This study confirms the higher desaturase ability of the SG compared to the FG genotype but also points out the higher muscle storage of the FG chickens with respect to the SG as a consequence of their higher body fat content. SG birds more efficiently synthesize LC-PUFA, leading to a higher percentage of n-3 and n-6, but FG birds have a higher storage capability and, consequently, higher muscle fatty acids concentrations (in terms of mg fatty/100 g tissue). Overall, the research suggests that genetic effects on the desaturase and elongase activities are responsible for variation in n-3 LC-PUFA synthesis, making the genetic mechanisms behind the synthesis very interesting for future research. A new challenge in genetic selection should be to find a genotype with an optimum equilibrium between LC-PUFA synthesis and storage ability in order to increase the LC-PUFA content in chicken products. The identification of this genotype might represent an important goal not only for the agro-industry but also for the improvement of human nutrition. For laying hens, genetic effects are less studied and probably less relevant because laying hens maintain high LC-PUFA conversion efficiency, mainly modulated by the reproductive efficiency (e.g., deposition rate) of their genetic strain [171]. 7. Rearing System The EU produces about 13.4 million tons of poultry meat [213], and 95% of this meat comes from intensive rearing systems [214]. These data show that poultry meat production from alternative systems (e.g., organic and free-range) is very limited. However, in recent years, there has been increasing demand for poultry meat and eggs grown utilizing these systems. In this context, France dominates the EU market by producing 16% of the total chicken meat with outdoor rearing systems [215]. In alternative poultry production, the presence of a pasture is crucial since the foraging birds spend a lot of time outdoors eating forage, pebbles, weeds, crop seeds, earthworms, and insects [204,216,217]. Many studies [171,217,218] have assessed the effects of pastures on poultry meat and egg quality. A natural pasture is rich in n-3 precursors, vitamins, and antioxidants that are transferred to chicken products, thus improving the fatty acid profile and the oxidative status of the meat and eggs with respect to those obtained from animals fed only commercial feedstuffs [216,219]. For example, the presence of a pasture reduced the n-6/n-3 ratio in SG chickens compared to the same strains when conventionally reared [220]. Dal Bosco et al. [216] also found a higher antioxidant intake in chickens reared outdoors compared to those reared indoors. Consequently, the antioxidant capacity of the plasma and the antioxidant levels of the meat were also greater in the outdoor groups than in the indoor ones. Hens with access to a pasture produce eggs with at least twice as much vitamin A and E and n-3 LC-PUFA compared to hens with no access to a pasture [221]. The forage intake of hens also positively influences the FA profile of egg yolks. Organic-plus hens (local breed with 10 m2/hen of organic pasture) showed eggs with higher content of n-3 PUFA and lower concentrations of n-6 PUFA as compared to organic (commercial hens with 4 m2/hen of organic pasture) and conventional (commercial hens indoor-reared) eggs (Figure 5, [216]). Moreover, SG chickens reared outdoors are able to express all of their behavioral repertoire, mainly characterized by the foraging activity that naturally enriches their products with LC-PUFA and antioxidants [205]. When alternative rearing systems are adopted as a strategy to increase the LC-PUFA in animal products, it is important to choose suitable genotypes for this type of rearing (Section 6). It is widely recognized that FG genotypes, selected for high productive performance, being static animals, reared in outdoor conditions, exhibit low grazing behavior without exploiting the beneficial effects of a pasture [204,214]. 8. Animals Transport As discussed previously, the nutritional quality of poultry meat is affected by many factors, such as diet, genetic strain, rearing system, and animal welfare during the rearing phase. Other crucial factors that affect meat quality include the transport conditions of poultry from the farm to the slaughterhouse [222]. Chickens are captured and then, during transport, they are caged and deprived of water and feed and can be subjected to variable environmental conditions (i.e., noise, vibrations, and differences in temperature and humidity). European rules set several parameters for the humane transfer of poultry to the slaughterhouse; however, birds may still be subjected to high stress. Authors [223] show that chickens transported for 4h (which could be considered an average transport time in commercial conditions) exhibit higher stress (determined by a higher Heterophils/Lymphocytes ratio, an indicator of stress) compared to non-transported chickens. Research suggests that the effect of stress could be different in FG and SG strains [224]. Berri et al. [225] reported that SG suffer more during the lag phase between catching and slaughter due to their high kinetic activity (i.e., wing flapping) during transport and slaughtering. Accordingly, SG chickens, being more active, seem more sensitive to stress during transport compared to FG [226]. Moreover, the higher energy expenditure due to the metabolism increase caused by the stress consumes PUFA via β-oxidation, reducing the concentration in body tissues [142]. Notably, the length of transport increases the proportion of saturated fatty acids of breast meat (mainly C16:0 and C18:0) and decreases PUFA (LA, AA, EPA, and DHA) content, likely due to greater formation of peroxides, as confirmed by higher TBARS values. In addition, oxidative stress, caused by the length of transport, reduces the in vivo antioxidant content of the body and, consequently, enhances the post-mortem susceptibility of PUFA to lipid oxidation. Cartoni Mancinelli et al. [227], through the mobile poultry processing unit (MPPU), have proposed a possible solution applicable to small-scale farmers to avoid/reduce the transport of the birds. The MPPU, the first in Europe, consists of a truck equipped with a small slaughterhouse able to reach the poultry farm. At the moment, positive conclusions exist concerning the effect of MPPUs on animal welfare and meat quality. Thus, to improve and preserve the nutritional quality of poultry meat, it is very important to pay particular attention both during the rearing period of animals (genotypes, breeding systems, diet, animal welfare, and health) and in the successive phases, such as slaughtering and cooking procedures. 9. Cooking Procedures Meat is a perishable food; it has been reported that lipid oxidation of meat occurs in the following order: fish > poultry > pork > beef > lamb [228]. This different susceptibility to oxidation is attributed to the level of unsaturated fatty acids, as ALA oxidation is 20 to 30 times higher than LA [229], and to the antioxidant availability, in particular vitamin E stored in muscle cells. The cooking temperature of meat can result in the development of volatile organic compounds (VOC), which is largely attributed to the autoxidation of PUFA. The secondary products of lipid oxidation are responsible for warmed-over flavor production (WOF) [230]. The organoleptic rancidity due to the oxidative deterioration negatively affects acceptability for consumers. Pearson et al. [231] show that the temperature of 70 to 80 °C damages muscle membranes, which, in turn, induces the interaction of lipid oxidation catalysts with unsaturated fatty acids and consequent production of free radicals. Among the free radicals, the thioradicals (e.g., oxidized proteins) are largely responsible for the WOF [232]. Poultry meat, due to its high content of LC-PUFA, is very sensitive to the oxidative process and, as previously reported, the genotype represents the main factor affecting the concentration of PUFA in meat. Indeed, it has been widely demonstrated that the SG chicken genotypes are more efficient at synthesizing n-3 LC-PUFA as compared to the FG strains. A recent study [233] compared the fatty acid, antioxidants, and VOC content of raw and cooked meat samples derived from four chicken genotypes with different growth rates. The study showed a 5.5 times increase in VOC production in cooked meat compared to raw meat across genotypes. However, the VOC production level was related to the LC-PUFA content in the raw meat and to the genotype. Consequently, the SG meat was considered “more vulnerable” due to the lower content of antioxidants as compared with the FG genotype [233]. In fact, the SG birds, due to their innate higher kinetic activity with respect to FG, consume the dietary antioxidants also to balance the oxidative process induced by movement [234]. In order to contain the losses of n-3 LC-PUFA, the meat antioxidants/LC-PUFA ratio should be considered as these could be altered by the cooking process. 10. Conclusions The health importance of n-3 LC-PUFA in humans, together with the inefficiency of n-3 LC-PUFA synthesis, has increased interest in enriching foods in these compounds. Poultry can be considered a suitable animal model for studying n-3 LC-PUFA enrichment strategies in terms of rearing cycle practices and post-mortem processes. The main method of dietary manipulation consists of supplementation of n-3 PUFA precursors and vitamin E without compromising the diet LA/ALA balance. The genotype choices also have an important role in determining the poultry n-3 LC-PUFA content since it is necessary to consider both the bird’s ability to convert ALA into n-3 LC-PUFA and the deposition of such compounds into the edible tissues of chickens. When alternative rearing systems are used as a strategy to increase the n-3 LC-PUFA in animal products, due to the presence of ALA in grass and insects, it is crucial to adopt an explorative genotype able to use the available pasture. In order to preserve the n-3 LC-PUFA accumulated by the chickens during the rearing cycle in poultry products, it is also important to pay attention to the next steps, such as animal transport to the slaughterhouse and cooking procedures. For both processes, it is essential to reduce the duration and ensure an adequate quantity of dietary antioxidants in order to preserve the products’ quality. For enriching and preserving the n-3 LC-PUFA in poultry products, a multifactorial approach should be adopted that encourages the use of multiple strategies throughout the entire production chain. Further research efforts are still needed to clearly define the storage efficiency of the different strategies for the enrichment of poultry meat and eggs. Certainly, in the current world context, with insufficient n-3 LC-PUFA supplies for human nutrition, it is necessary to apply responsible and sustainable approaches, including:avoiding livestock and human competition for n-3 LC-PUFA; developing livestock systems with the best conditions for bio-conversion of n-3 precursors into n-3 LC-PUFA. Author Contributions Conceptualization, A.C.M. and C.C.; methodology, A.C.M. and C.C.; validation, S.M., C.T., A.D.B., A.M.D. and K.A.; investigation, A.C.M., C.C., S.M. and A.D.B.; resources, C.C.; data curation, A.C.M. and C.C..; writing—original draft preparation, A.C.M.; writing—review and editing, C.T., A.M.D. and K.A; visualization, E.A. and D.C.; supervision, C.C. and A.M.D.; project administration, C.C.; funding acquisition, C.C., All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Estimated intake (mg/d) of n-3, n-6, saturated fatty acids, and total fats in relation to calories (%) provided by fat during human evolution. Data from Simopoulos, 2019 [22]. Figure 2 Metabolic pathways of n-6 and n-3 PUFA. The red arrows indicate the synthesis of n-6 and the green arrows the synthesis of n-3. The double blue arrows represent the enzymatic pathway. Figure 3 Schematic representation of PUFA metabolism, PGs, and IsoPs synthesis. The n-3 and n-6 precursors (ALA and LA) from the diet through desaturation and elongation processes form LC-PUFA: AA (n-6), EPA, and DHA (n-3). The LC-PUFA are incorporated into the phospholipids of plasma membrane. When external stimuli occur, the PLA2 and PLC cleave phospholipids to increase the cell availability of AA, EPA, ALA, and DHA. The AA is processed by the COX2 enzyme to form PG2, or by epoxygenase and lipoxygenase to form epoxyeicosatrienoic acids (EETs) and leukotrienes (LEUK). The EPA is involved in the PG3 synthesis by the COX1 enzyme. Free radical peroxidation of LC-PUFA generates different classes of IsoPs. F2-IsoPs derive from AA oxidation, while F1, F3, and F4-ISoPs from ALA, EPA, and DHA, respectively. Figure 4 Graphical representation of the main dietary strategies to increase n-3 LC-PUFA in poultry products (egg and meat). The current dietary strategies consist of: (1) providing to animals n-3 precursor (ALA source) or (2) enriching animal feed with n-3 LC-PUFA. (3) A future perspective could be represented by insects as a source of n-3 precursors or n-3 LC-PUFA. Figure 5 Comparison between (a) the weight (g) and (b) n-3 content (mg/100 g) of eggs from hens reared in conventional, organic, and organic–plus system. Conventional rearing system was characterized by commercial hens reared indoors; in the organic rearing system, commercial hens with 4 m2/hen of organic pasture were used, whereas organic-plus system consisted of local breed hens with 10 m2/hen of organic pasture. Data from Dal Bosco, 2016 [216]. nutrients-14-01969-t001_Table 1 Table 1 Estimated global requirements of n-3 LC-PUFA and their availability from fish. n-3 LC-PUFA REQUIREMENT Population classes n EPA and DHA (mg/d) EPA and DHA (t/y) Total world population 7,922,857,397 Adult individuals 7,903,361,753 250 356 Women in pregnancy and lactation 9747,822 400 1423 Young children (<24 months) 9,747,822 100 721,182 TOTAL ~722,960 AVAILABILITY Wild and farm-raised fish 100,000,000 50% of fish is suitable for human consumption 50,000,000 15% of the n-3 represent EPA and DHA 375,000 DEFICIT ~347,956 The table shows the estimation of the global annual deficit of n-3 LC-PUFA (t/y) calculated considering the n-3 LC-PUFA requirement in the main population classes and the amount of EPA and DHA provided by fish. nutrients-14-01969-t002_Table 2 Table 2 Estimated global availability of n-3 LC-PUFA from terrestrial sources [110,111]). t/y mg LC-PUFA/d t LCP/y Poultry 1.3 × 108 0.62 40,300 Eggs 7.7 × 107 0.35 26,845 Pork 9.4 × 107 0.18 8487 Total 75,632 The table shows the estimated annual availability of the n-3 LC-PUFA (mg/d or t/y) obtained from the monogastric livestock mostly consumed in the human diet. nutrients-14-01969-t003_Table 3 Table 3 n-3 fatty acid profile (mg/100 g of fresh tissue) in liver, breast, and drumstick tissues of broilers fed with different ALA and EPA sources. Tissues ALA+EPA g/kg of Diet Time Feeding (d) Genotype ALA (C18:3) EPA (C20:5) DHA (C22:6) TOTAL References Breast 15 44 Ross 308 18.0 3 10 31 Cortinas, 2004 [139] Drumstick 197 7 17 221 Breast 25 21 Ross 308 147 13.5 31.5 192.0 Rymer, 2006 [140] Drumstick 258 10.8 17.5 286.3 Liver 6.1 6 Cobb × Ross 308 67.1 280.0 120.1 467.2 Shin, 2012 [138] Breast 52.3 36.1 20.4 108.8 Drumstick 104.7 38.8 15.6 159.1 Breast 10 30 Ross 308 185 56 86 327 González-Ortiz, 2013 [137] Liver 407 275 335 1017 The table shows higher level of n-3 LC-PUFA in the liver compared to the other tissues. The drumstick exhibits a higher ALA content in respect to the breast. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Van Dael P. Role of N-3 Long-Chain Polyunsaturated Fatty Acids in Human Nutrition and Health: Review of Recent Studies and Recommendations Nutr. Res. Pract. 2021 15 137 10.4162/nrp.2021.15.2.137 33841720 2. Goyens P.L. Spilker M.E. Zock P.L. Katan M.B. Mensink R.P. Conversion of α-linolenic acid in humans is influenced by the absolute amounts of α-linolenic acid and linoleic acid in the diet and not by their ratio Am. J. Clin. Nutr. 2006 84 44 53 10.1093/ajcn/84.1.44 16825680 3. Mariamenatu A.H. Abdu E.M. Overconsumption of Omega-6 Polyunsaturated Fatty Acids (PUFAs) versus Deficiency of Omega-3 PUFAs in Modern-Day Diets: The Disturbing Factor for Their “Balanced Antagonistic Metabolic Functions” in the Human Body J. Lipids 2021 2021 8848161 10.1155/2021/8848161 33815845 4. Mollica M. Trinchese G. Cimmino F. Penna E. Cavaliere G. Tudisco R. Musco N. Manca C. Catapano A. Monda M. Milk Fatty Acid Profiles in Different Animal Species: Focus on the Potential Effect of Selected PUFAs on Metabolism and Brain Functions Nutrients 2021 13 1111 10.3390/nu13041111 33800688 5. Dal Bosco A. Castellini C. Bianchi L. Mugnai C. Effect of Dietary α-Linolenic Acid and Vitamin E on the Fatty Acid Composition, Storage Stability and Sensory Traits of Rabbit Meat Meat Sci. 2004 66 407 413 10.1016/S0309-1740(03)00127-X 22064144 6. Ferrier L.K. Caston L.J. Leeson S. Squires J. Weaver B.J. Holub B.J. Alpha-Linolenic Acid- and Docosahexaenoic Acid-Enriched Eggs from Hens Fed Flaxseed: Influence on Blood Lipids and Platelet Phospholipid Fatty Acids in Humans Am. J. Clin. Nutr. 1995 62 81 86 10.1093/ajcn/62.1.81 7598070 7. Simopoulos A.P. Evolutionary Aspects of Diet: The Omega-6/Omega-3 Ratio and the Brain Mol. Neurobiol. 2011 44 203 215 10.1007/s12035-010-8162-0 21279554 8. Naughton J.M. O’Dea K. Sinclair A.J. Animal Foods in Traditional Australian Aboriginal Diets: Polyunsaturated and Low in Fat Lipids 1986 21 684 690 10.1007/BF02537241 3796233 9. Jiang J. Xiong Y.L. Natural antioxidants as food and feed additives to promote health benefits and quality of meat products: A review Meat Sci. 2016 120 107 117 10.1016/j.meatsci.2016.04.005 27091079 10. Salter A.M. Dietary fatty acids and cardiovascular disease Animal 2013 7 163 171 10.1017/S1751731111002023 23031737 11. Calder P.C. Marine Omega-3 Fatty Acids and Inflammatory Processes: Effects, Mechanisms and Clinical Relevance Biochim. Biophys. Acta Mol. Cell Biol. Lipids 2015 1851 469 484 10.1016/j.bbalip.2014.08.010 12. Lupton J.R. Brooks J. Butte N.F. Caballero B. Flatt J.P. Fried S.K. Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids National Academy Press Washington, DC, USA 2002 Volume 5 589 768 13. Cordain L. Miller J.B. Eaton S.B. Mann N. Holt S.H.A. Speth J.D. Plant-Animal Subsistence Ratios and Macronutrient Energy Estimations in Worldwide Hunter-Gatherer Diets Am. J. Clin. Nutr. 2000 71 682 692 10.1093/ajcn/71.3.682 10702160 14. Eaton S.B. Eaton S.B. Sinclair A.J. Cordain L. Mann N.J. Dietary Intake of Long-Chain Polyunsaturated Fatty Acids during the Paleolithic World Rev. Nutr. Diet 1998 83 12 23 9648501 15. Sinclair A.J. Incorporation of radioactive polyunsaturated fatty acids into liver and brain of developing rat Lipids 1975 10 175 184 10.1007/BF02534156 1128172 16. Crawford M.A. The Role of Dietary Fatty Acids in Biology: Their Place in the Evolution of the Human Brain Nutr. Rev. 1992 50 3 11 10.1111/j.1753-4887.1992.tb01283.x 17. Cordain L. Eaton S. Miller J.B. Mann N. Hill K. The paradoxical nature of hunter-gatherer diets: Meat-based, yet non-atherogenic Eur. J. Clin. Nutr. 2002 56 S42 S52 10.1038/sj.ejcn.1601353 11965522 18. Raatz S.K. Conrad Z. Jahns L. Trends in linoleic acid intake in the United States adult population: NHANES 1999–2014 Prostaglandins Leukot. Essent. Fat. Acids 2018 133 23 28 10.1016/j.plefa.2018.04.006 19. Harika R.K. Eilander A. Alssema M. Osendarp S.J. Zock P. Intake of Fatty Acids in General Populations Worldwide Does Not Meet Dietary Recommendations to Prevent Coronary Heart Disease: A Systematic Review of Data from 40 Countries Ann. Nutr. Metab. 2013 63 229 238 10.1159/000355437 24192557 20. Sioen I. Van Lieshout L. Eilander A. Fleith M. Lohner S. Szommer A. Petisca C. Eussen S. Forsyth S. Calder P.C. Systematic Review on N-3 and N-6 Polyunsaturated Fatty Acid Intake in European Countries in Light of the Current Recommendations—Focus on Specific Population Groups Ann. Nutr. Metab. 2017 70 39 50 10.1159/000456723 28190013 21. Schmidhuber J. Traill W.B. The changing structure of diets in the European Union in relation to healthy eating guidelines Public Health Nutr. 2006 9 584 595 10.1079/PHN2005844 16923290 22. Simopoulos A.P. Genetic Variation and Evolutionary Aspects of Diet Antioxidant Status, Diet, Nutrition, and Health 1st ed. Papas A. CRC Press Boca Raton, FL, USA 2019 64 89 23. Stark A.H. Crawford M.A. Reifen R. Update on alpha-linolenic acid Nutr. Rev. 2008 66 326 332 10.1111/j.1753-4887.2008.00040.x 18522621 24. Castellini C. Mattioli S. Moretti E. Cotozzolo E. Perini F. Dal Bosco A. Signorini C. Noto D. Belmonte G. Lasagna E. Expression of Genes and Localization of Enzymes Involved in Polyunsaturated Fatty Acid Synthesis in Rabbit Testis and Epididymis Sci. Rep. 2022 12 2637 10.1038/s41598-022-06700-y 35173269 25. Mattioli S. Dal Bosco A. Maranesi M. Petrucci L. Rebollar P.G. Castellini C. Dietary Fish Oil and Flaxseed for Rabbit Does: Fatty Acids Distribution and Δ6-Desaturase Enzyme Expression of Different Tissues Animal 2019 13 1934 1942 10.1017/S175173111900020X 30774049 26. Mattioli S. Collodel G. Signorini C. Cotozzolo E. Noto D. Cerretani D. Micheli L. Fiaschi A. Brecchia G. Menchetti L. Tissue Antioxidant Status and Lipid Peroxidation Are Related to Dietary Intake of n-3 Polyunsaturated Acids: A Rabbit Model Antioxidants 2021 10 681 10.3390/antiox10050681 33925444 27. Tapiero H. Ba G.N. Couvreur P. Tew K. Polyunsaturated fatty acids (PUFA) and eicosanoids in human health and pathologies Biomed. Pharmacother. 2002 56 215 222 10.1016/S0753-3322(02)00193-2 12199620 28. Khan A.A. Iadarola M. Yang H.-Y.T. Dionne R.A. Expression of COX-1 and COX-2 in a Clinical Model of Acute Inflammation J. Pain 2007 8 349 354 10.1016/j.jpain.2006.10.004 17270500 29. Mattos R. Staples C.R. Thatcher W.W. Effects of dietary fatty acids on reproduction in ruminants Rev. Reprod. 2000 5 38 45 10.1530/ror.0.0050038 10711734 30. Carneiro L.C. Williams E.J. Saut J.P.E. dos Santos R.M. Celeghini E.C.C. The Effect of N-3 Polyunsaturated Fatty Acid Supplementation Oimmune and Reproductive Parameters in Dairy Cows: A Review Braz. J. Vet. Res. Anim. Sci. 2021 58 e175224 10.11606/issn.1678-4456.bjvras.2021.175224 31. Kolobarić N. Drenjančević I. Matić A. Šušnjara P. Mihaljević Z. Mihalj M. Dietary Intake of N-3 Pufa-Enriched Hen Eggs Changes Inflammatory Markers’ Concentration and Treg/Th17 Cells Distribution in Blood of Young Healthy Adults—A Randomised Study Nutrients 2021 13 1851 10.3390/nu13061851 34071714 32. Calder P.C. Omega-3 Fatty Acids and Inflammatory Processes: From Molecules to Man Biochem. Soc. Trans. 2017 45 1105 1115 10.1042/BST20160474 28900017 33. Morrow J.D. Hill K.E. Burk R.F. Nammour T.M. Badr K.F. Roberts L.J. A Series of Prostaglandin F2-like Compounds Are Produced in Vivo in Humans by a Non-Cyclooxygenase, Free Radical-Catalyzed Mechanism Proc. Natl. Acad. Sci. USA 1990 87 9383 9387 10.1073/pnas.87.23.9383 2123555 34. Signorini C. Moretti E. Collodel G. Role of isoprostanes in human male infertility Syst. Biol. Reprod. Med. 2020 66 291 299 10.1080/19396368.2020.1793032 32842780 35. Milne G.L. Yin H. Morrow J.D. Human Biochemistry of the Isoprostane Pathway J. Biol. Chem. 2008 283 15533 15537 10.1074/jbc.R700047200 18285331 36. Galano J.M. Lee Y.Y. Oger C. Vigor C. Vercauteren J. Durand T. Giera M. Lee J.C.Y. Isoprostanes, Neuroprostanes and Phytoprostanes: An Overview of 25 Years of Research in Chemistry and Biology Prog. Lipid Res. 2017 68 83 108 10.1016/j.plipres.2017.09.004 28923590 37. Molnár P.J. Der B. Borsodi K. Balla H. Borbás Z. Molnár K. Ruisanchez É. Kenessey I. Horváth A. Keszthelyi A. Isoprostanes Evoke Contraction of the Murine and Human Detrusor Muscle via Activation of the Thromboxane Prostanoid TP Receptor and Rho Kinase Am. J. Physiol. Ren. Physiol. 2021 320 F537 F547 10.1152/ajprenal.00400.2020 38. Wu J. Fang S. Lu K.T. Wackman K. Schwartzman M.L. Dikalov S.I. Grobe J.L. Sigmund C.D. EP3 (E-Prostanoid 3) Receptor Mediates Impaired Vasodilation in a Mouse Model of Salt-Sensitive Hypertension Hypertension 2021 77 1399 1411 10.1161/HYPERTENSIONAHA.120.16518 33641369 39. Calder P.C. Very Long-Chain n-3 Fatty Acids and Human Health: Fact, Fiction and the Future Proc. Nutr. Soc. 2018 77 52 72 10.1017/S0029665117003950 29039280 40. Yannakopoulos A. Tserveni-Gousi A. Christaki E. Enhanced Egg Production in Practice: The Case of Bio-Omega-3 Egg Int. J. Poult. Sci. 2005 4 531 535 41. Tang A.B. Nishimura K.Y. Phinney S.D. Preferential Reduction in Adipose Tissue a.Linolenic Acid (18:3 3) During Very Low Calorie Dieting Despite Supplementation with 18:3 31 Lipids 1993 28 987 993 10.1007/BF02537119 7904040 42. Scientific Opinion on Dietary Reference Values for Fats, Including Saturated Fatty Acids, Polyunsaturated Fatty Acids, Monounsaturated Fatty Acids, Trans Fatty Acids, and Cholesterol EFSA J. 2010 8 1461 43. Helland I.B. Smith L. Saarem K. Saugstad O.D. Drevon C.A. Maternal Supplementation with Very-Long-Chain n-3 Fatty Acids During Pregnancy and Lactation Augments Children’s IQ at 4 Years of Age Pediatrics 2003 111 e39 e44 10.1542/peds.111.1.e39 12509593 44. Leyrolle Q. Decoeur F. Briere G. Amadieu C. Quadros A.R.A.A. Voytyuk I. Lacabanne C. Benmamar-Badel A. Bourel J. Aubert A. Maternal dietary omega-3 deficiency worsens the deleterious effects of prenatal inflammation on the gut-brain axis in the offspring across lifetime Neuropsychopharmacology 2021 46 579 602 10.1038/s41386-020-00793-7 32781459 45. Hoffman M.C. Freedman R. Law A.J. Clark A.M. Hunter S.K. Maternal nutrients and effects of gestational COVID-19 infection on fetal brain development Clin. Nutr. ESPEN 2021 43 1 8 10.1016/j.clnesp.2021.04.019 34024500 46. Castro-Gómez P. Garcia-Serrano A. Visioli F. Fontecha J. Relevance of dietary glycerophospholipids and sphingolipids to human health Prostaglandins Leukot. Essent. Fat. Acids 2015 101 41 51 10.1016/j.plefa.2015.07.004 26242691 47. Aryal S. Hussain S. Drevon C.A. Nagelhus E. Hvalby D. Jensen V. Walaas S.I. Davanger S. Omega-3 fatty acids regulate plasticity in distinct hippocampal glutamatergic synapses Eur. J. Neurosci. 2019 49 40 50 10.1111/ejn.14224 30367533 48. Butler M. Nelson V.A. Davila H. Ratner E. Fink H.A. Hemmy L.S. McCarten J.R. Barclay T.R. Brasure M. Kane R.L. Over-the-Counter Supplement Interventions to Prevent Cognitive Decline, Mild Cognitive Impairment, and Clinical Alzheimer-Type Dementia Ann. Intern. Med. 2018 168 52 62 10.7326/M17-1530 29255909 49. Moore K. Hughes C.F. Ward M. Hoey L. McNulty H. Diet, Nutrition and the Ageing Brain: Current Evidence and New Directions Proc. Nutr. Soc. 2018 77 152 163 10.1017/S0029665117004177 29316987 50. Chew E.Y. Clemons T.E. Agrón E. Launer L.J. Grodstein F. Bernstein P.S. Effect of Omega-3 Fatty Acids, Lutein/Zeaxanthin, or Other Nutrient Supplementation on Cognitive Function: The AREDS2 Randomized Clinical Trial JAMA J. Am. Med. Assoc. 2015 314 791 801 10.1001/jama.2015.9677 51. Wood A.H.R. Chappell H.F. Zulyniak M.A. Dietary and Supplemental Long-Chain Omega-3 Fatty Acids as Moderators of Cognitive Impairment and Alzheimer’s Disease Eur. J. Nutr. 2022 61 589 604 10.1007/s00394-021-02655-4 34392394 52. Bernasconi A.A. Wiest M.M. Lavie C.J. Milani R.V. Laukkanen J.A. Effect of Omega-3 Dosage on Cardiovascular Outcomes: An Updated Meta-Analysis and Meta-Regression of Interventional Trials Mayo Clin. Proc. 2021 96 304 313 10.1016/j.mayocp.2020.08.034 32951855 53. U.S. Food and Drug Administration (FDA) Advice about Eating Fish Available online: https://www.fda.gov/food/consumers/advice-about-eating-fish (accessed on 25 January 2022) 54. Meyer B.J. de Groot R.H.M. Effects of Omega-3 Long Chain Polyunsaturated Fatty Acid Supplementation on Cardiovascular Mortality: The Importance of the Dose of DHA Nutrients 2017 9 1305 10.3390/nu9121305 29189735 55. Van der Wurff I.S.M. Meyer B.J. de Groot R.H.M. Effect of Omega-3 Long Chain Polyunsaturated Fatty Acids (N-3 LCPUFA) Supplementation on Cognition in Children and Adolescents: A Systematic Literature Review with a Focus on n-3 LCPUFA Blood Values and Dose of DHA and EPA Nutrients 2020 12 3115 10.3390/nu12103115 33053843 56. Scientific Opinion on the Tolerable Upper Intake Level of Eicosapentaenoic Acid (EPA), Docosahexaenoic Acid (DHA) and Docosapentaenoic Acid (DPA) EFSA J. 2012 10 2815 57. Poulos A. Darin-Bennett A. White I. The phospholipid-bound fatty acids and aldehydes of mammalian spermatozoa Comp. Biochem. Physiol. Part B Comp. Biochem. 1973 46 541 549 10.1016/0305-0491(73)90094-1 58. Rodríguez M. Rebollar P.G. Mattioli S. Castellini C. N-3 PUFA Sources (Precursor/Products): A Review of Current Knowledge on Rabbit Animals 2019 9 806 10.3390/ani9100806 59. Signorini C. Moretti E. Noto D. Mattioli S. Castellini C. Pascarelli N.A. Durand T. Oger C. Galano J.M. de Felice C. F4-Neuroprostanes: A Role in Sperm Capacitation Life 2021 11 655 10.3390/life11070655 34357027 60. Yuxin L. Chen L. Xiaoxia L. Yue L. Junjie L. Youzhu L. Huiliang Z. Qicai L. Research Progress on the Relationship between Obesity-Inflammation-Aromatase Axis and Male Infertility Oxid. Med. Cell. Longev. 2021 2021 6612796 10.1155/2021/6612796 33628365 61. Agarwal A. Parekh N. Selvam M.K.P. Henkel R. Shah R. Homa S.T. Ramasamy R. Ko E. Tremellen K. Esteves S. Male Oxidative Stress Infertility (MOSI): Proposed Terminology and Clinical Practice Guidelines for Management of Idiopathic Male Infertility World J. Men’s Health 2019 37 296 312 10.5534/wjmh.190055 31081299 62. Simopoulos A.P. An Increase in the Omega-6/Omega-3 Fatty Acid Ratio Increases the Risk for Obesity Nutrients 2016 8 128 10.3390/nu8030128 26950145 63. Chaves H. Singh R.B. Khan S. Wilczynska A. Takahashi T. High Omega-6/Omega-3 Fatty Acid Ratio Diets and Risk of Noncommunicable Diseases: Is the Tissue, the Main Issue? The Role of Functional Food Security in Global Health 1st ed. Watson R.R. Singh R.B. Takahashi T. Academic Press London, UK 2018 217 259 64. World Population Prospect: The 2019 Revision—United Nations, Department of Economic and Social Affairs, Population Division (June 2019) Available online: https://population.un.org/wpp (accessed on 26 January 2022) 65. International Programs Center at the U.S. Census Bureau, Population Division Available online: https://www.census.gov/programs-surveys/international-programs.html (accessed on 29 January 2022) 66. Salem N. Eggersdorfer M. Is the World Supply of Omega-3 Fatty Acids Adequate for Optimal Human Nutrition? Curr. Opin. Clin. Nutr. Metab. Care 2015 18 147 154 10.1097/MCO.0000000000000145 25635599 67. Cherif A. Dubacq J. Mache R. Oursel A. Tremolieres A. Biosynthesis of α-Linolenic Acid by Desaturation of Oleic and Linoleic Acids in Several Organs of Higher and Lower Plants and in Algae Phytochemistry 1975 14 703 706 10.1016/0031-9422(75)83018-4 68. Anai T. Yamada T. Kinoshita T. Rahman S.M. Takagi Y. Identification of Corresponding Genes for Three Low-α-Linolenic Acid Mutants and Elucidation of Their Contribution to Fatty Acid Biosynthesis in Soybean Seed Plant Sci. 2005 168 1615 1623 10.1016/j.plantsci.2005.02.016 69. Somerville C. Browse J. Plant Lipids: Metabolism, Mutants, and Membranes Science 1991 252 80 87 10.1126/science.252.5002.80 17739077 70. Harwood J.L. Fatty acid metabolism Annu. Rev. Plant Physiol. Plant Mol. Biol. 1988 39 101 138 10.1146/annurev.pp.39.060188.000533 71. Harwood J.L. Recent advances in the biosynthesis of plant fatty acids Biochim. Biophys. Acta (BBA)-Lipids Lipid. Metab. 1996 1301 7 56 10.1016/0005-2760(95)00242-1 72. Millar A.A. Wrischer M. Kunst L. Accumulation of Very-Long-Chain Fatty Acids in Membrane Glycerolipids Is Associated with Dramatic Alterations in Plant Morphology Plant Cell 1998 10 1889 1902 10.1105/tpc.10.11.1889 9811796 73. Spychalla J.P. Kinney A.J. Browse J. Identification of an Animal-3 Fatty Acid Desaturase by Heterologous Expression in Arabidopsis Proc. Natl. Acad. Sci. USA 1997 94 1142 1147 10.1073/pnas.94.4.1142 9037020 74. Arao T. Yamada M. Biosynthesis of Polyunsaturated Fatty Acids in the Marine Diatom, Phaeodactylum Tricornutum Phytochemistry 1994 35 1177 1181 10.1016/S0031-9422(00)94817-9 75. Arao T. Sakaki T. Yamada M. Biosynthesis of Polyunsaturated Lipids in the Diatom, Phaeodactylum Tricornutum Phytochemistry 1994 36 629 635 10.1016/S0031-9422(00)89787-3 76. Bajpai P. Bajpai P.K. Eicosapentaenoic Acid (EPA) Production from Microorganisms: A Review J. Biotechnol. 1993 30 161 183 10.1016/0168-1656(93)90111-Y 7764033 77. Girke T. Schmidt H. Zä Hringer U. Reski R. Heinz E. Identification of a novel D6-acyl-group desaturase by targeted gene disruption in Physcomitrella patens Plant J. 1998 15 39 48 10.1046/j.1365-313X.1998.00178.x 9744093 78. Twining C.W. Brenna J.T. Hairston N.G. Flecker A.S. Highly Unsaturated Fatty Acids in Nature: What We Know and What We Need to Learn Oikos 2016 125 749 760 10.1111/oik.02910 79. Twining C.W. Parmar T.P. Mathieu-Resuge M. Kainz M.J. Shipley J.R. Martin-Creuzburg D. Use of Fatty Acids from Aquatic Prey Varies with Foraging Strategy Front. Ecol. Evol. 2021 9 735350 10.3389/fevo.2021.735350 80. Gill I. Valivety R. Polyunsaturated Fatty Acids, Part 1: Occurrence, Biological Activities and Applications Trends Biotechnol. 1997 15 401 409 10.1016/S0167-7799(97)01076-7 9351284 81. Gunstone F.D. Gammar linolenic acid occurrence and physical and chemical properties Prog. Lipid Res. 1992 31 145 161 10.1016/0163-7827(92)90007-6 1334265 82. Gunstone F. Movements towards tailor-made fats Prog. Lipid Res. 1998 37 277 305 10.1016/S0163-7827(98)00012-5 10209650 83. US Department of Agriculture National Nutrient Database Available online: http://www.nal.usda.gov/fnic/foodcomp/search/ (accessed on 25 January 2022) 84. Gebauer S.K. Psota T.L. Harris W.S. Kris-Etherton P.M. N-3 Fatty Acid Dietary Recommendations and Food Sources to Achieve Essentiality and Cardiovascular Benefits Am. J. Clin. Nutr. 2006 83 1526S 1535S 10.1093/ajcn/83.6.1526S 16841863 85. Hardy R.W. Utilization of Plant Proteins in Fish Diets: Effects of Global Demand and Supplies of Fishmeal Aquac. Res. 2010 41 770 776 10.1111/j.1365-2109.2009.02349.x 86. Tocher D.R. Metabolism and Functions of Lipids and Fatty Acids in Teleost Fish Rev. Fish. Sci. 2003 11 107 184 10.1080/713610925 87. Bell M.V. Tocher D.R. Biosynthesis of Polyunsaturated Fatty Acids in Aquatic Ecosystems: General Pathways and New Directions Lipids in Aquatic Ecosystems 1st ed. Kainz M. Brett M. Arts M. Springer New York, NY, USA 2009 211 236 88. Castro L.F.C. Tocher D.R. Monroig O. Long-chain polyunsaturated fatty acid biosynthesis in chordates: Insights into the evolution of Fads and Elovl gene repertoire Prog. Lipid Res. 2016 62 25 40 10.1016/j.plipres.2016.01.001 26769304 89. Lopes-Marques M. Kabeya N. Qian Y. Ruivo R. Santos M.M. Venkatesh B. Tocher D.R. Castro L.F.C. Monroig Ó. Retention of Fatty Acyl Desaturase 1 (Fads1) in Elopomorpha and Cyclostomata Provides Novel Insights into the Evolution of Long-Chain Polyunsaturated Fatty Acid Biosynthesis in Vertebrates BMC Evol. Biol. 2018 18 157 10.1186/s12862-018-1271-5 30340454 90. Betancor M.B. Oboh A. Ortega A. Mourente G. Navarro J.C. de la Gándara F. Tocher D.R. Monroig Ó. Molecular and Functional Characterisation of a Putative Elovl4 Gene and Its Expression in Response to Dietary Fatty Acid Profile in Atlantic Bluefin Tuna (Thunnus thynnus) Comp. Biochem. Physiol. Part B Biochem. Mol. Biol. 2020 240 110372 10.1016/j.cbpb.2019.110372 31669375 91. Hu Q. Sommerfeld M. Jarvis E. Ghirardi M. Posewitz M. Seibert M. Darzins A. Microalgal Triacylglycerols as Feedstocks for Biofuel Production: Perspectives and Advances Plant J. 2008 54 621 639 10.1111/j.1365-313X.2008.03492.x 18476868 92. Lang I. Hodac L. Friedl T. Feussner I. Fatty Acid Profiles and Their Distribution Patterns in Microalgae: A Comprehensive Analysis of More than 2000 Strains from the SAG Culture Collection BMC Plant Biol. 2011 11 124 10.1186/1471-2229-11-124 21896160 93. Ishikawa A. Kabeya N. Ikeya K. Kakioka R. Cech J.N. Osada N. Leal M.C. Inoue J. Kume M. Toyoda A. A Key Metabolic Gene for Recurrent Freshwater Colonization and Radiation in Fishes Science 2019 364 886 889 10.1126/science.aau5656 31147520 94. Kriton G. Dimitra K. Corraze G. Jaume P.-S. Adorjan A. Zsuzsanna J.S. Impact of Diets Containing Plant Raw Materials as Fish Meal and Fish Oil Replacement on Rainbow Trout (Oncorhynchus mykiss), Gilthead Sea Bream (Sparus aurata), and Common Carp (Cyprinus carpio) Freshness J. Food Qual. 2018 2018 1717465 10.1155/2018/1717465 95. Faudzi N.M. Yong A.S.K. Shapawi R. Senoo S. Biswas A. Takii K. Soy protein concentrate as an alternative in replacement of fish meal in the feeds of hybrid grouper, brown-marbled grouper (Epinephelus fuscoguttatus) × giant grouper (E. lanceolatus) juvenile Aquac. Res. 2017 49 431 441 10.1111/are.13474 96. Braña C.B.C. Cerbule K. Senff P. Stolz I.K. Towards Environmental Sustainability in Marine Finfish Aquaculture Front. Mar. Sci. 2021 8 10.3389/fmars.2021.666662 97. Hites R.A. Foran J.A. Carpenter D.O. Hamilton M.C. Knuth B.A. Schwager S.J. Global Assessment of Organic Contaminants in Farmed Salmon Science 2004 303 226 229 10.1126/science.1091447 14716013 98. Thomsen S.T. Assunção R. Afonso C. Boué G. Cardoso C. Cubadda F. Garre A. Kruisselbrink J.W. Mantovani A. Pitter J.G. Human Health Risk–Benefit Assessment of Fish and Other Seafood: A Scoping Review Crit. Rev. Food Sci. Nutr. 2021 1 22 10.1080/10408398.2021.1915240 99. Adams K.J. Drenner R.W. Chumchal M.M. Donato D.I. Disparity between State Fish Consumption Advisory Systems for Methylmercury and US Environmental Protection Agency Recommendations: A Case Study of the South Central United States Environ. Toxicol. Chem. 2016 35 247 251 10.1002/etc.3185 26605989 100. U.S. Environmental Protection Agency Guidance for assessing chemical contaminant data for use in fish advisories Risk Assess. Fish Consum. Limits 2000 2 101. Kuipers R.S. Luxwolda M.F. Offringa P.J. Boersma E.R. Dijck-Brouwer D.J. Muskiet F.A. Gestational age dependent changes of the fetal brain, liver and adipose tissue fatty acid compositions in a population with high fish intakes Prostaglandins Leukot. Essent. Fat. Acids 2012 86 189 199 10.1016/j.plefa.2012.02.007 22425685 102. Extier A. Langelier B. Perruchot M.-H. Guesnet P. Van Veldhoven P.P. Lavialle M. Alessandri J.-M. Gender affects liver desaturase expression in a rat model of n−3 fatty acid repletion J. Nutr. Biochem. 2010 21 180 187 10.1016/j.jnutbio.2008.10.008 19157821 103. Burdge G. Slater-Jefferies J. Grant R. Chung W.-S. West A. Lillycrop K. Hanson M. Calder P. Sex, but not maternal protein or folic acid intake, determines the fatty acid composition of hepatic phospholipids, but not of triacylglycerol, in adult rats Prostaglandins Leukot. Essent. Fat. Acids 2008 78 73 79 10.1016/j.plefa.2007.10.028 104. Toral P.G. Monahan F.J. Hervás G. Frutos P. Moloney A.P. Review: Modulating ruminal lipid metabolism to improve the fatty acid composition of meat and milk. Challenges and opportunities Animal 2018 12 s272 s281 10.1017/S1751731118001994 30139411 105. Ibrahim N.A. Alimon A.R. Yaakub H. Samsudin A.A. Candyrine S.C.L. Wan Mohamed W.N. Md Noh A. Fuat M.A. Mookiah S. Effects of Vegetable Oil Supplementation on Rumen Fermentation and Microbial Population in Ruminant: A Review Trop. Anim. Health Prod. 2021 53 422 10.1007/s11250-021-02863-4 34331142 106. Scislowski V. Bauchart D. Gruffat D. Laplaud P.M. Durand D. Effects of Dietary N-6 or n-3 Polyunsaturated Fatty Acids Protected or Not against Ruminal Hydrogenation on Plasma Lipids and Their Susceptibility to Peroxidation in Fattening Steers J. Anim. Sci. 2005 83 2162 2174 10.2527/2005.8392162x 16100072 107. Pereira G. Simões P. Bexiga R. Silva E. Mateus L. Fernandes T. Alves S.P. Bessa R.J.B. Lopes-da-Costa L. Effects of Feeding Rumen-Protected Linseed Fat to Postpartum Dairy Cows on Plasma n-3 Polyunsaturated Fatty Acid Concentrations and Metabolic and Reproductive Parameters J. Dairy Sci. 2022 105 361 374 10.3168/jds.2021-20674 34635360 108. Kitessa S.M. Gulati S.K. Simos G.C. Ashes J.R. Scott T.W. Fleck E. Wynn P.C. Supplementation of Grazing Dairy Cows with Rumen-Protected Tuna Oil Enriches Milk Fat with n -3 Fatty Acids without Affecting Milk Production or Sensory Characteristics Br. J. Nutr. 2004 91 271 277 10.1079/BJN20031050 14756913 109. Simopoulos A.P. New products from the agri-food industry: The return of n-3 fatty acids into the food supply Lipids 1999 34 S297 S301 10.1007/BF02562324 10419184 110. FAO (Food and Agriculture Organization of the United Nations) Fats and Fatty Acids in Human Nutrition: Report of an Expert Consultation FAO Food and Nutrition Geneva, Switzerland 2010 111. Givens D. Kliem K.E. Gibbs R.A. The role of meat as a source of n−3 polyunsaturated fatty acids in the human diet Meat Sci. 2006 74 209 218 10.1016/j.meatsci.2006.04.008 22062730 112. Williams A.G. Audsley E. Sandars D.L. Determining the Environmental Burdens and Resource Use in the Production of Agricultural and Horticultural Commoditites Defra Project Report IS0205 Available online: http://randd.defra.gov.uk/Default.aspx (accessed on 3 May 2022) 113. Cavani C. Petracci M. Trocino A. Xiccato G. Advances in Research on Poultry and Rabbit Meat Quality Ital. J. Anim. Sci. 2009 8 (Suppl. 2) 741 750 10.4081/ijas.2009.s2.741 114. Barroeta A. Nutritive value of poultry meat: Relationship between vitamin E and PUFA World’s Poult. Sci. J. 2007 63 277 284 10.1017/S0043933907001468 115. Abeyrathne E.D.N.S. Lee H.Y. Ahn D.U. Egg White Proteins and Their Potential Use in Food Processing or as Nutraceutical and Pharmaceutical Agents—A Review Poult. Sci. 2013 92 3292 3299 10.3382/ps.2013-03391 24235241 116. Marcet I. Sáez-Orviz S. Rendueles M. Díaz M. Egg Yolk Granules and Phosvitin. Recent Advances in Food Technology and Applications LWT 2022 153 112442 10.1016/j.lwt.2021.112442 117. Mattioli S. Dal Bosco A. Martino M. Ruggeri S. Marconi O. Sileoni V. Falcinelli B. Castellini C. Benincasa P. Alfalfa and Flax Sprouts Supplementation Enriches the Content of Bioactive Compounds and Lowers the Cholesterol in Hen Egg J. Funct. Foods 2016 22 454 462 10.1016/j.jff.2016.02.007 118. Lee S.H. Kim Y.B. Kim D.H. Lee D.W. Lee H.G. Jha R. Lee K.W. Dietary Soluble Flaxseed Oils as a Source of Omega-3 Polyunsaturated Fatty Acids for Laying Hens Poult. Sci. 2021 100 101276 10.1016/j.psj.2021.101276 34229219 119. Kwiecień M. Winiarska-Mieczan A. Danek-Majewska A. Kwiatkowska K. Krusiński R. Effects of Dietary Alfalfa Protein Concentrate on Lipid Metabolism and Antioxidative Status of Serum and Composition and Fatty Acid Profile and Antioxidative Status and Dietetic Value of Muscles in Broilers Poult. Sci. 2021 100 100974 10.1016/j.psj.2020.12.071 33588343 120. Wilkinson S.J. Big data for poultry—What is possible Proceedings of the 29th Annual Australian Poultry Science Symposium Sydney, Australia 4–7 February 2018 121. Schreiner M. Hulan H.W. Razzazi-Fazeli E. Böhm J. Moreira R.G. Effect of Different Sources of Dietary Omega-3 Fatty Acids on General Performance and Fatty Acid Profiles of Thigh, Breast, Liver and Portal Blood of Broilers J. Sci. Food Agric. 2005 85 219 226 10.1002/jsfa.1948 122. Chekani-Azar S. Shahriar H.A. Maheri-Sis N. Ahmadzadeh A.R. Vahdatpoor T. Omega-3 Fatty Acids Enrichment and Organoleptic Characteristics of Broiler Meat Asian J. Anim. Vet. Adv. 2008 3 62 69 10.3923/ajava.2008.62.69 123. Rizliya V. Mendis E. Biological, Physical, and Chemical Properties of Fish Oil and Industrial Applications Seafood Processing By-Products: Trends and Applications 1st ed. Kim S.K. Springer New York, NY, USA 2013 285 313 124. Harris W.S. Fish Oils and Plasma Lipid and Lipoprotein Metabolism in Humans: A Critical Review J. Lipid Res. 1989 30 785 807 10.1016/S0022-2275(20)38310-3 2677200 125. Sprecher H.W. Baykousheva S.P. Luthria D.L. Mohammed B.S. Differences in the Regulation of Biosynthesis of 20- versus 22-Carbon Polyunsaturated Fatty Acids Prostaglandins Leukot. Essent. Fat. Acids 1995 52 99 101 10.1016/0952-3278(95)90005-5 126. Bonos E. Kasapidou E. Kargopoulos A. Karampampas A. Christaki E. Florou-Paneri P. Nikolakakis I. Spirulina as a Functional Ingredient in Broiler Chicken Diets S. Afr. J. Anim. Sci. 2016 46 94 102 10.4314/sajas.v46i1.12 127. Alfaia C.M. Pestana J.M. Rodrigues M. Coelho D. Aires M.J. Ribeiro D.M. Major V.T. Martins C.F. Santos H. Lopes P.A. Influence of Dietary Chlorella Vulgaris and Carbohydrate-Active Enzymes on Growth Performance, Meat Quality and Lipid Composition of Broiler Chickens Poult. Sci. 2021 100 926 937 10.1016/j.psj.2020.11.034 33518146 128. El-Bahr S. Shousha S. Shehab A. Khattab W. Ahmed-Farid O. Sabike I. El-Garhy O. Albokhadaim I. Albosadah K. Effect of Dietary Microalgae on Growth Performance, Profiles of Amino and Fatty Acids, Antioxidant Status, and Meat Quality of Broiler Chickens Animals 2020 10 761 10.3390/ani10050761 129. Costa M.M. Pestana J.M. Osório D. Alfaia C.M. Martins C.F. Mourato M. Gueifão S. Rego A.M. Coelho I. Coelho D. Effect of Dietary Laminaria Digitata with Carbohydrases on Broiler Production Performance and Meat Quality, Lipid Profile, and Mineral Composition Animals 2022 12 1007 10.3390/ani12081007 35454252 130. Jankowski J. Zduńczyk Z. Mikulski D. Juśkiewicz J. Pomianowski J.F. Zduńczyk P. Fatty Acid Profile, Oxidative Stability and Sensory Quality of Breast Meat from Turkeys Fed Diets with Graded Levels of Flaxseed Oil for Different Periods of Time Anim. Prod. Sci. 2018 58 1164 1174 10.1071/AN16319 131. López-Ferrer S. Baucells M.D. Barroeta A.C. Galobart J. Grashorn M.A. N-3 Enrichment of Chicken Meat. 2. Use of Precursors of Long-Chain Polyunsaturated Fatty Acids: Linseed Oil Poult. Sci. 2001 80 753 761 10.1093/ps/80.6.753 11441842 132. Kartikasari L.R. Hughes R.J. Geier M.S. Makrides M. Gibson R.A. Dietary Alpha-Linolenic Acid Enhances Omega-3 Long Chain Polyunsaturated Fatty Acid Levels in Chicken Tissues Prostaglandins Leukot. Essent. Fat. Acids 2012 87 103 109 10.1016/j.plefa.2012.07.005 133. López-Ferrer S. Baucells M.D. Barroeta A.C. Grashorn M.A. N-3 Enrichment of Chicken Meat Using Fish Oil: Alternative Substitution with Rapeseed and Linseed Oils Poult. Sci. 1999 78 356 365 10.1093/ps/78.3.356 10090262 134. Gonzalez-Esquerra R. Leeson S. Effects of Menhaden Oil and Flaxseed in Broiler Diets on Sensory Quality and Lipid Composition of Poultry Meat Br. Poult. Sci. 2000 41 481 488 10.1080/713654967 11128390 135. Zelenka J. Jarošová A. Schneiderová D. Influence of N-3 and n-6 Polyunsaturated Fatty Acids on Sensory Characteristics of Chicken Meat Czech J. Anim. Sci. 2008 53 299 305 10.17221/356-CJAS 136. Surai P.F. Kochish I.I. Romanov M.N. Griffin D.K. Nutritional Modulation of the Antioxidant Capacities in Poultry: The Case of Vitamin E Poult. Sci. 2019 98 4030 4041 10.3382/ps/pez072 30805637 137. González-Ortiz G. Sala R. Cánovas E. Abed N. Barroeta A.C. Consumption of Dietary N-3 Fatty Acids Decreases Fat Deposition and Adipocyte Size, but Increases Oxidative Susceptibility in Broiler Chickens Lipids 2013 48 705 717 10.1007/s11745-013-3785-3 23529200 138. Shin D. Choi S. Go G. Park J. Narciso-Gaytán C. Morgan C. Smith S. Sánchez-Plata M. Ruiz-Feria C. Effects of dietary combination of n-3 and n-9 fatty acids on the deposition of linoleic and arachidonic acid in broiler chicken meats Poult. Sci. 2012 91 1009 1017 10.3382/ps.2011-01836 22399741 139. Cortinas L. Villaverde C. Galobart J. Baucells M.D. Codony R. Barroeta A.C. Fatty Acid Content in Chicken Thigh and Breast as Affected by Dietary Polyunsaturation Level Poult. Sci. 2004 83 1155 1164 10.1093/ps/83.7.1155 15285507 140. Rymer C. Givens D.I. Effect of Species and Genotype on the Efficiency of Enrichment of Poultry Meat with N-3 Polyunsaturated Fatty Acids Lipids 2006 41 445 451 10.1007/s11745-006-5118-2 16933789 141. Brooks G.A. Mercier J. Balance of Carbohydrate and Lipid Utilization during Exercise: The “crossover” Concept J. Appl. Physiol. 1994 76 2253 2261 10.1152/jappl.1994.76.6.2253 7928844 142. Schulz H. Beta oxidation of fatty acids Biochim. Et Biophys. Acta (BBA) Lipids Lipid Metab. 1991 1081 109 120 10.1016/0005-2760(91)90015-A 143. Cartoni Mancinelli A. di Veroli A. Mattioli S. Cruciani G. Dal Bosco A. Castellini C. Lipid Metabolism Analysis in Liver of Different Chicken Genotypes and Impact on Nutritionally Relevant Polyunsaturated Fatty Acids of Meat Sci. Rep. 2022 12 1888 10.1038/s41598-022-05986-2 35115659 144. Marra C.A. de Alaniz M.J.T. Influence of Testosterone Administration on the Biosynthesis of Unsaturated Fatty Acids in Male and Female Rats Lipids 1989 24 1014 1019 10.1007/BF02544071 2615569 145. Burdge G.C. Wootton S.A. Conversion of α-Linolenic Acid to Eicosapentaenoic, Docosapentaenoic and Docosahexaenoic Acids in Young Women Br. J. Nutr. 2002 88 411 420 10.1079/BJN2002689 12323090 146. Poureslami R. Raes K. Turchini G.M. Huyghebaert G. de Smet S. Effect of Diet, Sex and Age on Fatty Acid Metabolism in Broiler Chickens: N-3 and n-6 PUFA Br. J. Nutr. 2010 104 189 197 10.1017/S0007114510000395 20187998 147. Abedi E. Sahari M.A. Long-Chain Polyunsaturated Fatty Acid Sources and Evaluation of Their Nutritional and Functional Properties Food Sci. Nutr. 2014 2 443 463 10.1002/fsn3.121 25473503 148. Aymond W.M. VAN Elswyk M.E. Yolk Thiobarbituric Acid Reactive Substances and n-3 Fatty Acids in Response to Whole and Ground Flaxseed Poult. Sci. 1995 74 1388 1394 10.3382/ps.0741388 7479519 149. Hayat Z. Cherian G. Pasha T.N. Khattak F.M. Jabbar M.A. Effect of feeding flax and two types of antioxidants on egg production, egg quality, and lipid composition of eggs J. Appl. Poult. Res. 2009 18 541 551 10.3382/japr.2009-00008 150. Caston L.J. Squires E.J. Leeson S. Hen performance, egg quality, and the sensory evaluation of eggs from SCWL hens fed dietary flax Can. J. Anim. Sci. 1994 74 347 353 10.4141/cjas94-047 151. Rizzi L. Bochicchio D. Bargellini A. Parazza P. Simioli M. Effects of dietary microalgae, other lipid sources, inorganic selenium and iodine on yolk n-3 fatty acid composition, selenium content and quality of eggs in laying hens J. Sci. Food Agric. 2009 89 1775 1781 10.1002/jsfa.3655 152. Bean L. Leeson S. Long-term effects of feeding flaxseed on performance and egg fatty acid composition of brown and white hens Poult. Sci. 2003 82 388 394 10.1093/ps/82.3.388 12705398 153. Mattioli S. Ruggeri S. Sebastiani B. Brecchia G. Dal Bosco A. Cartoni Mancinelli A. Castellini C. Performance and Egg Quality of Laying Hens Fed Flaxseed: Highlights on n-3 Fatty Acids, Cholesterol, Lignans and Isoflavones Animal 2017 11 705 712 10.1017/S175173111600207X 27819218 154. Scheideler S.E. Jaroni D. Froning G. Strain and Age Effects on Egg Composition from Hens Fed Diets Rich in N-3 Fatty Acids Poult. Sci. 1998 77 192 196 10.1093/ps/77.2.192 9495477 155. Aguillón-Páez Y.J. Romero L.A. Diaz G.J. Effect of Full-Fat Sunflower or Flaxseed Seeds Dietary Inclusion on Performance, Egg Yolk Fatty Acid Profile and Egg Quality in Laying Hens Anim. Nutr. 2020 6 179 184 10.1016/j.aninu.2019.12.005 32542198 156. Alzueta C. Rodríguez M.L. Cutuli M.T. Rebolé A. Ortiz L.T. Centeno C. Treviño J. Effect of Whole and Demucilaged Linseed in Broiler Chicken Diets on Digesta Viscosity, Nutrient Utilisation and Intestinal Microflora Br. Poult. Sci. 2003 44 67 74 10.1080/0007166031000085337 12737228 157. Goyal A. Sharma V. Upadhyay N. Gill S. Sihag M. Flax and flaxseed oil: An ancient medicine & modern functional food J. Food Sci. Technol. 2014 51 1633 1653 10.1007/s13197-013-1247-9 25190822 158. Imran M. Anjum F.M. Nadeem M. Ahmad N. Khan M.K. Mushtaq Z. Hussain S. Production of Bio-Omega-3 Eggs through the Supplementation of Extruded Flaxseed Meal in Hen Diet Lipids Health Dis. 2015 14 126 10.1186/s12944-015-0127-x 26453429 159. Huang S. Baurhoo B. Mustafa A. Effects of Feeding Extruded Flaxseed on Layer Performance, Total Tract Nutrient Digestibility, and Fatty Acid Concentrations of Egg Yolk, Plasma and Liver J. Anim. Physiol. Anim. Nutr. 2020 104 1365 1374 10.1111/jpn.13364 160. Westbrook L.A. Cherian G. Egg Quality, Fatty-Acid Composition and Gastrointestinal Morphology of Layer Hens Fed Whole Flaxseed with Enzyme Supplementation Br. Poult. Sci. 2019 60 146 153 10.1080/00071668.2018.1556783 30525943 161. Kralik G. Kralik Z. Grčević M. Galović O. Hanžek D. Biazik E. Fatty Acid Profile of Eggs Produced by Laying Hens Fed Diets Containing Different Shares of Fish Oil Poult. Sci. 2021 100 101379 10.1016/j.psj.2021.101379 34411964 162. Beynen A.C. Fatty Acid Composition of Eggs Produced by Hens Fed Diets Containing Groundnut, Soya Bean or Linseed NJAS Wagening. J. Life Sci. 2004 52 3 10 10.1016/S1573-5214(04)80026-5 163. Jiang Z.R. Ahn D.U. Sim J.S. Effects of Feeding Flax and Two Types of Sunflower Seeds on Fatty Acid Compositions of Yolk Lipid Classes Poult. Sci. 1991 70 2467 2475 10.3382/ps.0702467 1784568 164. Fraeye I. Bruneel C. Lemahieu C. Buyse J. Muylaert K. Foubert I. Dietary Enrichment of Eggs with Omega-3 Fatty Acids: A Review Food Res. Int. 2012 48 961 969 10.1016/j.foodres.2012.03.014 165. Irawan A. Ningsih N. Hafizuddin Rusli R.K. Suprayogi W.P.S. Akhirini N. Hadi R.F. Setyono W. Jayanegara A. Supplementary N-3 Fatty Acids Sources on Performance and Formation of Omega-3 in Egg of Laying Hens: A Meta-Analysis Poult. Sci. 2022 101 101566 10.1016/j.psj.2021.101566 34823172 166. Huang S. Baurhoo B. Mustafa A. Effects of Extruded Flaxseed on Layer Performance, Nutrient Retention and Yolk Fatty Acid Composition Br. Poult. Sci. 2018 59 463 469 10.1080/00071668.2018.1476676 29764188 167. Omri B. Chalghoumi R. Izzo L. Ritieni A. Lucarini M. Durazzo A. Abdouli H. Santini A. Effect of Dietary Incorporation of Linseed Alone or Together with Tomato-Red Pepper Mix on Laying Hens’ Egg Yolk Fatty Acids Profile and Health Lipid Indexes Nutrients 2019 11 813 10.3390/nu11040813 30974860 168. Grobas S. Mateos G.G. Mendez J. Influence of Dietary Linoleic Acid on Production and Weight of Eggs and Egg Components in Young Brown Hens J. Appl. Poult. Res. 1999 8 177 184 10.1093/japr/8.2.177 169. Dong X. Liu S. Tong J. Comparative Effect of Dietary Soybean Oil, Fish Oil, and Coconut Oil on Performance, Egg Quality and Some Blood Parameters in Laying Hens Poult. Sci. 2018 97 2460 2472 10.3382/ps/pey094 29669020 170. Meluzzi A. Sirri F. Castellini C. Roncarati A. Melotti P. Franchini A. Influence of genotype and feeding on chemical composition of organic chicken meat Ital. J. Anim. Sci. 2009 8 766 768 10.4081/ijas.2009.s2.766 171. Mugnai C. Sossidou E.N. Bosco A.D. Ruggeri S. Mattioli S. Castellini C. The effects of husbandry system on the grass intake and egg nutritive characteristics of laying hens J. Sci. Food Agric. 2014 94 459 467 10.1002/jsfa.6269 23775487 172. Hallmann C.A. Sorg M. Jongejans E. Siepel H. Hofland N. Schwan H. Stenmans W. Müller A. Sumser H. Hörren T. More than 75 percent decline over 27 years in total flying insect biomass in protected areas PLoS ONE 2017 12 e0185809 10.1371/journal.pone.0185809 29045418 173. Fanatico A.C. Arsi K. Upadhyaya I. Ramos J.M. Donoghue D. Donoghue A.M. Sustainable Fish and Invertebrate Meals for Methionine and Protein Feeds in Organic Poultry Production J. Appl. Poult. Res. 2018 27 437 448 10.3382/japr/pfy037 174. van der Heide M.E. Stødkilde L. Nørgaard J.V. Studnitz M. The Potential of Locally-Sourced European Protein Sources for Organic Monogastric Production: A Review of Forage Crop Extracts, Seaweed, Starfish, Mussel, and Insects Sustainability 2021 13 2303 10.3390/su13042303 175. Bessa L.W. Pieterse E. Marais J. Hoffman L.C. Why for feed and not for human consumption? The black soldier fly larvae Compr. Rev. Food Sci. Food Saf. 2020 19 2747 2763 10.1111/1541-4337.12609 33336973 176. Caligiani A. Marseglia A. Sorci A. Bonzanini F. Lolli V. Maistrello L. Sforza S. Influence of the killing method of the black soldier fly on its lipid composition Food Res. Int. 2019 116 276 282 10.1016/j.foodres.2018.08.033 30716946 177. Liu X. Chen X. Wang H. Yang Q. Ur Rehman K. Li W. Cai M. Li Q. Mazza L. Zhang J. Dynamic Changes of Nutrient Composition throughout the Entire Life Cycle of Black Soldier Fly PLoS ONE 2017 12 e0182601 10.1371/journal.pone.0182601 28796830 178. Finke M.D. Oonincx D.G.A.B. Nutrient Content of Insects Insects as Food and Feed: From Production to Consumption van Huis A. Tomberlin J.K. Wageningen Academic Publishers Wageningen, The Netherlands 2016 290 316 179. Sánchez-Muros M.-J. Barroso F. Manzano-Agugliaro F. Insect meal as renewable source of food for animal feeding: A review J. Clean. Prod. 2014 65 16 27 10.1016/j.jclepro.2013.11.068 180. Bbosa T. Ndagire C.T. Mukisa I.M. Fiaboe K.K.M. Nakimbugwe D. Nutritional Characteristics of Selected Insects in Uganda for Use as Alternative Protein Sources in Food and Feed J. Insect Sci. 2019 19 23 10.1093/jisesa/iez124 181. Schreven S. Yener S. Van Valenberg H. Dicke M. Van Loon J. Life on a piece of cake: Performance and fatty acid profiles of black soldier fly larvae fed oilseed by-products J. Insects Food Feed 2021 7 35 49 10.3920/JIFF2020.0004 182. Hoc B. Genva M. Fauconnier M.-L. Lognay G. Francis F. Megido R.C. About lipid metabolism in Hermetia illucens (L. 1758): On the origin of fatty acids in prepupae Sci. Rep. 2020 10 11916 10.1038/s41598-020-68784-8 32680992 183. Rossi G. Mattioli S. Rondoni Rondoni G. Dal Bosco A. Servili M. Characterisation of Fatty Acid Profiles of Tenebrio Molitor Larvae Reared on Diets Enriched with Edible Oils J. Insects Food Feed. 2022 1 12 10.3920/JIFF2021.0164 184. Kierończyk B. Rawski M. Józefiak A. Mazurkiewicz J. Świątkiewicz S. Siwek M. Bednarczyk M. Szumacher-Strabel M. Cieślak A. Benzertiha A. Effects of Replacing Soybean Oil with Selected Insect Fats on Broilers Anim. Feed. Sci. Technol. 2018 240 170 183 10.1016/j.anifeedsci.2018.04.002 185. Schiavone A. Cullere M. de Marco M. Meneguz M. Biasato I. Bergagna S. Dezzutto D. Gai F. Dabbou S. Gasco L. Partial or Total Replacement of Soybean Oil by Black Soldier Fly Larvae (Hermetia Illucens L.) Fat in Broiler Diets: Effect on Growth Performances, Feed-Choice, Blood Traits, Carcass Characteristics and Meat Quality Ital. J. Anim. Sci. 2017 16 93 100 10.1080/1828051X.2016.1249968 186. Premalatha M. Abbasi T. Abbasi T. Abbasi S.A. Energy-Efficient Food Production to Reduce Global Warming and Ecodegradation: The Use of Edible Insects Renew. Sustain. Energy Rev. 2011 15 4357 4360 10.1016/j.rser.2011.07.115 187. Schiavone A. Dabbou S. de Marco M. Cullere M. Biasato I. Biasibetti E. Capucchio M.T. Bergagna S. Dezzutto D. Meneguz M. Black Soldier Fly Larva Fat Inclusion in Finisher Broiler Chicken Diet as an Alternative Fat Source Animal 2018 12 2032 2039 10.1017/S1751731117003743 29343316 188. Barroso F.G. Sánchez-Muros M.J. Segura M. Morote E. Torres A. Ramos R. Guil J.L. Insects as Food: Enrichment of Larvae of Hermetia Illucens with Omega 3 Fatty Acids by Means of Dietary Modifications J. Food Compos. Anal. 2017 62 8 13 10.1016/j.jfca.2017.04.008 189. Oonincx D.G.A.B. Laurent S. Veenenbos M.E. van Loon J.J.A. Dietary Enrichment of Edible Insects with Omega 3 Fatty Acids Insect Sci. 2020 27 500 509 10.1111/1744-7917.12669 30801963 190. Oonincx D.G.A.B. Finke M.D. Nutritional Value of Insects and Ways to Manipulate Their Composition J. Insects Food Feed. 2021 7 639 659 10.3920/JIFF2020.0050 191. Liland N.S. Biancarosa I. Araujo P. Biemans D. Bruckner C.G. Waagbø R. Torstensen B.E. Lock E.J. Modulation of Nutrient Composition of Black Soldier Fly (Hermetia illucens) Larvae by Feeding Seaweed-Enriched Media PLoS ONE 2017 12 e0183188 10.1371/journal.pone.0183188 28837591 192. St-Hilaire S. Cranfill K. McGuire M.A. Mosley E.E. Tomberlin J.K. Newton L. Sealey W. Sheppard C. Irving S. Fish Offal Recycling by the Black Soldier Fly Produces a Foodstuff High in Omega-3 Fatty Acids J. World Aquac. Soc. 2007 38 309 313 10.1111/j.1749-7345.2007.00101.x 193. Ding S. Lin X. He S. Earthworms: A source of protein J. Food Sci. Eng. 2019 9 159 170 194. Sampedro L. Jeannotte R. Whalen J.K. Trophic Transfer of Fatty Acids from Gut Microbiota to the Earthworm Lumbricus terrestris L. Soil Biol. Biochem. 2006 38 2188 2198 10.1016/j.soilbio.2006.02.001 195. Khan S. Naz S. Sultan A. Alhidary I. Abdelrahman M. Khan R. Khan N. Khan M. Ahmad S. Worm meal: A potential source of alternative protein in poultry feed World’s Poult. Sci. J. 2016 72 93 102 10.1017/S0043933915002627 196. Prayogi H.S. The Effect of Earthworm Meal Supplementation in the Diet on Quail’s Growth Performance in Attempt to Replace the Usage of Fish Meal Int. J. Poult. Sci. 2011 10 804 806 10.3923/ijps.2011.804.806 197. Rezaeipour V. Nejad O.A. Miri H.Y. Growth Performance, Blood Metabolites and Jejunum Morphology of Broiler Chickens Fed Diets Containing Earthworm (Eisenia foetida) Meal as a Source of Protein Int. J. Adv. Biol. Biomed. Res. 2014 2 2483 2494 198. Son J.H. The study on treatment of poultry waste by earthworms, and the effect of feeding earthworms meal on the performance of broilers and laying hens, and safety of meat and egg Korean J. Org. Agric. 2009 17 17 63 199. Son J.H. Jo I.H. Effects of Earthworm Meal Supplementation on the Performance of Broiler Chickens Korean J. Org. Agric. 2003 200. Zang Y.T. Bing S. Zhang Y.Z. Sheng X.W. Shu D.Q. Effects of Dietary Supplementation with Earthworm Powder on Production Performance, Blood Characteristics, and Heavy Metal Residues of Broiler Pullets J. Appl. Poult. Res. 2018 27 609 615 10.3382/japr/pfy024 201. Isea-León F. Acosta-Balbás V. Beatriz Rial-Betancoutd L. Luisa Medina-Gallardo A. Mélécony Célestin B. Evaluation of the Fatty Acid Composition of Earthworm Eisenia andrei Meal as an Alternative Lipid Source for Fish Feed J. Food Nutr. Res. 2019 7 696 700 10.12691/jfnr-7-10-2 202. EU Commission Commission Regulation (EU) No. 68/2013 on the Catalogue of Feed Materials Available online: https://www.ecolex.org/details/legislation/commission-regulation-eu-no-682013-on-the-catalogue-of-feed-materials-lex-faoc119700/ (accessed on 10 February 2022) 203. Cartoni Mancinelli A. Franzoni A. Dal Bosco A. Schiavone A. Mannelli F. Marzoni M. Castellini C. Distribution and Consistency of Ancona and Livorno Poultry Breed in Central Italy Ital. J. Anim. Sci. 2020 19 1297 1303 10.1080/1828051X.2020.1842814 204. Mancinelli A.C. Mattioli S. Menchetti L. Bosco A.D. Ciarelli C. Amato M.G. Castellini C. The Assessment of a Multifactorial Score for the Adaptability Evaluation of Six Poultry Genotypes to the Organic System Animals 2021 11 2992 10.3390/ani11102992 34680011 205. Mattioli S. Mancinelli A.C. Menchetti L. Bosco A.D. Madeo L. Amato M.G. Moscati L. Cotozzolo E. Ciarelli C. Angelucci E. How the kinetic behavior of organic chickens affects productive performance and blood and meat oxidative status: A study of six poultry genotypes Poult. Sci. 2021 100 101297 10.1016/j.psj.2021.101297 34280645 206. Van der Waaij E.H. A Resource Allocation Model Describing Consequences of Artificial Selection under Metabolic Stress J. Anim. Sci. 2004 82 973 981 10.2527/2004.824973x 15080316 207. Sirri F. Castellini C. Bianchi M. Petracci M. Meluzzi A. Franchini A. Effect of Fast-, Medium- and Slow-Growing Strains on Meat Quality of Chickens Reared under the Organic Farming Method Animal 2011 5 312 319 10.1017/S175173111000176X 22440776 208. Castellini C. Dal Bosco A. Mugnai C. Bernardini M. Performance and Behaviour of Chickens with Different Growing Rate Reared According to the Organic System Ital. J. Anim. Sci. 2003 6 561 573 10.4081/ijas.2002.291 209. Boschetti E. Bordoni A. Meluzzi A. Castellini C. Bosco A.D. Sirri F. Fatty acid composition of chicken breast meat is dependent on genotype-related variation of FADS1 and FADS2 gene expression and desaturating activity Animal 2016 10 700 708 10.1017/S1751731115002712 26670346 210. Castellini C. Dal Bosco A. Mattioli S. Davidescu M. Corazzi L. MacChioni L. Rimoldi S. Terova G. Activity, Expression, and Substrate Preference of the Δ6-Desaturase in Slow- or Fast-Growing Rabbit Genotypes J. Agric. Food Chem. 2016 64 792 800 10.1021/acs.jafc.5b05425 26745387 211. Dal Bosco A. Mugnai C. Ruggeri S. Mattioli S. Castellini C. Fatty Acid Composition of Meat and Estimated Indices of Lipid Metabolism in Different Poultry Genotypes Reared under Organic System Poult. Sci. 2012 91 2039 2045 10.3382/ps.2012-02228 22802202 212. Al-Hilal M. AlSaleh A. Maniou Z. Lewis F.J. Hall W.L. Sanders T.A.B. O’Dell S.D. Genetic Variation at the FADS1-FADS2 Gene Locus Influences Delta-5 Desaturase Activity and LC-PUFA Proportions after Fish Oil Supplement J. Lipid Res. 2013 54 542 551 10.1194/jlr.P032276 23160180 213. Poultry|European Commission Available online: https://https://ec.europa.eu/info/food-farming-fisheries/animals-and-animal-products/animal-products_en (accessed on 15 February 2022) 214. Bosco A.D. Mattioli S. Mancinelli A.C. Cotozzolo E. Castellini C. Extensive Rearing Systems in Poultry Production: The Right Chicken for the Right Farming System. A Review of Twenty Years of Scientific Research in Perugia University, Italy Animals 2021 11 1281 10.3390/ani11051281 33947001 215. Baéza E. Guillier L. Petracci M. Review: Production factors affecting poultry carcass and meat quality attributes Animal 2021 16 100331 10.1016/j.animal.2021.100331 34419417 216. Dal Bosco A. Mugnai C. Mattioli S. Rosati A. Ruggeri S. Ranucci D. Castellini C. Transfer of bioactive compounds from pasture to meat in organic free-range chickens Poult. Sci. 2016 95 2464 2471 10.3382/ps/pev383 26769274 217. Mancinelli A.C. Matiolli S. Bosco A.D. Ciarelli C. Castellini C. Grass intake and meat oxidative status of geese reared in three different agroforestry systems Acta Fytotech. Zootech. 2020 23 308 315 10.15414/afz.2020.23.mi-fpap.305-312 218. Bari S. Cohen-Barnhouse A.M. Campbell D.L.M. Early rearing enrichments influenced nest use and egg quality in free-range laying hens Animal 2020 14 1249 1257 10.1017/S1751731119003094 31907088 219. Mancinelli A.C. Mattioli S. Dal Bosco A. Piottoli L. Ranucci D. Branciari R. Cotozzolo E. Castellini C. Rearing Romagnola geese in vineyard: Pasture and antioxidant intake, performance, carcass and meat quality Ital. J. Anim. Sci. 2019 18 372 380 10.1080/1828051X.2018.1530960 220. Popova T. Petkov E. Ignatova M. Fatty acid composition of breast meat in two lines of slow-growing chickens reared conventionally or on pasture Food Sci. Appl. Biotechnol. 2018 1 70 76 10.30721/fsab2018.v1.i1.7 221. Mugnai C. Bosco A.D. Castellini C. Effect of rearing system and season on the performance and egg characteristics of Ancona laying hens Ital. J. Anim. Sci. 2009 8 175 188 10.4081/ijas.2009.175 222. Gou Z. Abouelezz K. Fan Q. Li L. Lin X. Wang Y. Cui X. Ye J. Masoud M. Jiang S. Physiological effects of transport duration on stress biomarkers and meat quality of medium-growing Yellow broiler chickens Animal 2020 15 100079 10.1016/j.animal.2020.100079 33573973 223. Mancinelli A.C. Mugnai C. Castellini C. Mattioli S. Moscati L. Piottoli L. Amato M.G. Doretti M. Bosco A.D. Cordovani E. Effect of transport length and genotype on tonic immobility, blood parameters and carcass contamination of free-range reared chickens Ital. J. Anim. Sci. 2018 17 557 564 10.1080/1828051X.2018.1423583 224. Rimoldi S. Lasagna E. Sarti F.M. Marelli S.P. Cozzi M.C. Bernardini G. Terova G. Expression profile of six stress-related genes and productive performances of fast and slow growing broiler strains reared under heat stress conditions Meta Gene 2015 6 17 25 10.1016/j.mgene.2015.08.003 26380816 225. Berri C. le Bihan-Duval E. Debut M. Santé-Lhoutellier V. Baéza E. Gigaud V. Jégo Y. Duclos M.J. Consequence of Muscle Hypertrophy on Characteristics of Pectoralis Major Muscle and Breast Meat Quality of Broiler Chickens J. Anim. Sci. 2007 85 2005 2011 10.2527/jas.2006-398 17431054 226. Castellini C. Mattioli S. Piottoli L. Mancinelli A.C. Ranucci D. Branciari R. Amato M.G. Dal Bosco A. Effect of Transport Length on in Vivo Oxidative Status and Breast Meat Characteristics in Outdoor-Reared Chicken Genotypes Ital. J. Anim. Sci. 2016 15 191 199 10.1080/1828051X.2016.1174082 227. Mancinelli A.C. Bosco A.D. Mattioli S. Ranucci D. Castellini C. Mobile Poultry Processing Unit as a Resource for Small Poultry Farms: Planning and Economic Efficiency, Animal Welfare, Meat Quality and Sanitary Implications Animals 2018 8 229 10.3390/ani8120229 30513677 228. Rhee K. Anderson L. Sams A. Lipid Oxidation Potential of Beef, Chicken, and Pork J. Food Sci. 1996 61 8 12 10.1111/j.1365-2621.1996.tb14714.x 229. Amaral A.B. da Silva M.V. Lannes S.C.D.S. Lipid oxidation in meat: Mechanisms and protective factors—A review Food Sci. Technol. 2018 38 1 15 10.1590/fst.32518 230. Bassam S.M. Noleto-Dias C. Farag M.A. Dissecting grilled red and white meat flavor: Its characteristics, production mechanisms, influencing factors and chemical hazards Food Chem. 2022 371 131139 10.1016/j.foodchem.2021.131139 34583172 231. Pearson A.M. Love J.D. Shorland F.B. “Warmed-over” Flavor in Meat, Poultry, and Fish Adv. Food Res. 1977 23 1 74 232. Byrne D.V. Bredie W.L.P. Mottram D.S. Martens M. Sensory and Chemical Investigations on the Effect of Oven Cooking on Warmed-over Flavour Development in Chicken Meat Meat Sci. 2002 61 127 139 10.1016/S0309-1740(01)00171-1 22064001 233. Mancinelli A.C. Silletti E. Mattioli S. Bosco A.D. Sebastiani B. Menchetti L. Koot A. van Ruth S. Castellini C. Fatty acid profile, oxidative status, and content of volatile organic compounds in raw and cooked meat of different chicken strains Poult. Sci. 2020 100 1273 1282 10.1016/j.psj.2020.10.030 33518084 234. Mattioli S. Dal Bosco A. Ruggeri S. Martino M. Moscati L. Pesca C. Castellini C. Adaptive Response to Exercise of Fast-Growing and Slow-Growing Chicken Strains: Blood Oxidative Status and Non-Enzymatic Antioxidant Defense Poult. Sci. 2017 96 4096 4102 10.3382/ps/pex203 29050412

PMC009xxxxxx/PMC9099611.txt

==== Front Molecules Molecules molecules Molecules 1420-3049 MDPI 10.3390/molecules27092729 molecules-27-02729 Article Monte Carlo Method and GA-MLR-Based QSAR Modeling of NS5A Inhibitors against the Hepatitis C Virus https://orcid.org/0000-0003-0670-5709 Liman Wissal 1 https://orcid.org/0000-0003-2465-0537 Oubahmane Mehdi 2 https://orcid.org/0000-0002-1435-5131 Hdoufane Ismail 2 https://orcid.org/0000-0002-4393-6867 Bjij Imane 3 https://orcid.org/0000-0002-6266-3817 Villemin Didier 4 Daoud Rachid 1 Cherqaoui Driss 2 https://orcid.org/0000-0002-4561-2161 El Allali Achraf 1* Lu Shaoyong Academic Editor 1 African Genome Center, Mohammed VI Polytechnic University, Ben Guerir 43150, Morocco; wissal.liman@um6p.ma (W.L.); rachid.daoud@um6p.ma (R.D.) 2 Department of Chemistry, Faculty of Sciences Semlalia, BP 2390, Marrakech 40000, Morocco; mehdi.oubahmane@ced.uca.ma (M.O.); i.hdoufane@uca.ac.ma (I.H.); cherqaoui@uca.ac.ma (D.C.) 3 Institut Supérieur des Professions Infirmières et Techniques de Santé (ISPITS), Dakhla 73000, Morocco; imane.bjij@gmail.com 4 Ecole Nationale Supérieure d’Ingénieurs (ENSICAEN) Laboratoire de Chimie Moléculaire et Thioorganique, UMR 6507 CNRS, INC3M, FR3038, Labex EMC3, Labex SynOrg ENSICAEN & Université de Caen, 14118 Caen, France; didier.villemin@ensicaen.fr * Correspondence: achraf.elallali@um6p.ma 23 4 2022 5 2022 27 9 272903 3 2022 21 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Hepatitis C virus (HCV) is a serious disease that threatens human health. Despite consistent efforts to inhibit the virus, it has infected more than 58 million people, with 300,000 deaths per year. The HCV nonstructural protein NS5A plays a critical role in the viral life cycle, as it is a major contributor to the viral replication and assembly processes. Therefore, its importance is evident in all currently approved HCV combination treatments. The present study identifies new potential compounds for possible medical use against HCV using the quantitative structure–activity relationship (QSAR). In this context, a set of 36 NS5A inhibitors was used to build QSAR models using genetic algorithm multiple linear regression (GA-MLR) and Monte Carlo optimization and were implemented in the software CORAL. The Monte Carlo method was used to build QSAR models using SMILES-based optimal descriptors. Four splits were performed and 24 QSAR models were developed and verified through internal and external validation. The model created for split 3 produced a higher value of the determination coefficients using the validation set (R2 = 0.991 and Q2 = 0.943). In addition, this model provides interesting information about the structural features responsible for the increase and decrease of inhibitory activity, which were used to develop eight novel NS5A inhibitors. The constructed GA-MLR model with satisfactory statistical parameters (R2 = 0.915 and Q2 = 0.941) confirmed the predicted inhibitory activity for these compounds. The Absorption, Distribution, Metabolism, Elimination, and Toxicity (ADMET) predictions showed that the newly designed compounds were nontoxic and exhibited acceptable pharmacological properties. These results could accelerate the process of discovering new drugs against HCV. chemoinformatics drug discovery molecular descriptors QSAR HCV NS5A ==== Body pmc1. Introduction Hepatitis C Virus (HCV) has significantly affected the lives of infected patients over the last century, as a small proportion of them shed the virus naturally. Most infected individuals develop a spectrum of liver diseases ranging from mild inflammation to extensive liver fibrosis, cirrhosis, chronic hepatitis C, and hepatocellular carcinoma [1,2]. According to the statistical report of the World Health Organization (WHO), an estimated 58 million people were infected with hepatitis C in 2019, and approximately 300,000 deaths were caused by HCV [3]. HCV belongs to the Flaviviridae family, the genus Hepacivirus. It is a single-stranded RNA virus encoded by 9600 nucleotide bases. The HCV genome consists of the open reading frames (ORF) between the 5’ and 3’ conserved untranslated regions encoding three structural proteins (C, E1 and E2) and seven non-structural proteins (NS1, NS2, NS3, NS4A, NS4B, NS5A and NS5B) [4]. HCV strains are classified into eight major genotypes, with 86 subtypes identified to date [5]. For the past two decades, the standard therapy for HCV infection has been based on peginterferon and the antiviral nucleoside analog ribavirin. To date, approximately half of patients achieved a lower sustained virologic response (SVR) and suffered from undesired harmful effects such as cardiac-related problems, leukopenia, and thrombocytopenia [6]. Recently, many direct-acting antiviral (DAA) drugs have been authorized for the treatment of HCV infection with higher SVR rates (>90%), shorter duration, and fewer adverse effects compared with older treatment therapies [7,8]. These innovative therapies have revolutionized HCV medicine and delivered significant insights into curing HCV patients. In 2016, an affordable combination treatment with the new drug Ravidasvir was shown to be safe and effective, with exceptionally elevated cure rates [9]. The fight against HCV infection is not fully covered due to the high costs associated with the therapies and the emergence of mutant strains resistant to DAA drugs. These treatments target three nonstructural proteins: NS3/4A protease, NS5B polymerase, and NS5A protein, which are involved in the replication and assembly processes of the virus [7]. The NS5A receptor is a 478-amino acid phosphoprotein containing three structural domains (I, II, and III) that terminate in four complementary functional zones (A, B, C, and D). The NS5A protein interacts with other important viral proteins (NS4B, NS5B, NS3) and host cell proteins (cyclophilin A, kinases and others) to regulate viral replication and assembly [10,11]. Due to its critical role in HCV replication, NS5A has emerged as a potential therapeutic target for treating chronic HCV infection. Recently, computational drug design has emerged as a powerful technique that plays a pivotal role in drug development. The quantitative structure–activity relationship (QSAR), which links the structural features of molecules to endpoints, is an important part of cheminformatics. The QSAR approach is widely used to predict biological activities and the development of new lead compounds. Thus, the biological activity of new structures based on the developed model can be easily determined using the QSAR method without the need for experimental synthesis and biological testing [12]. Due to its predictive power, the QSAR approach could also eliminate molecules with undesirable properties at an early stage. Therefore, it reduces the cost, time, and error rate in developing new drug molecules. Continuing our recent work on the development of new potent inhibitors targeting the NS4B receptor of HCV [13], we report here several QSAR models targeting NS5A. The current marketed anti-NS5A drugs have common structural features including C2 axial symmetry and the presence of methyl carbamates on both extremities. However, the symmetrical nature of anti-HCV agents is not essential for the inhibition of HCV, as reported by Nakamura et al. [14]. The QSAR models were built based on the structural features of asymmetrical NS5A derivatives with their potent inhibitory activity. The first model, based on Monte Carlo optimization, was applied to develop SMILES-based QSAR models that provide insights into the design of novel anti-HCV agents. The second model aims to confirm the prediction of inhibitory activity of the designed molecules using the genetic algorithm multiple linear regression (GA-MLR) technique. ADMET analysis was used to investigate and evaluate the drug-likeness properties of the newly designed inhibitors. 2. Results and Discussion 2.1. SMILES-Based QSAR Model In total, 24 QSAR models were developed from four random splits using two objective functions: TF1 without the IIC and TF2 with different values of the IIC. For TF2, different numerical values of WIIC were used, including 0.1, 0.3, 0.5, 0.7, and 0.9. The calculated statistical parameters for the created SMILES-based QSAR models show that the WIIC = 0.5 strengthens the influence of IIC on the Monte Carlo optimization (Supplementary materials, Spreadsheet). The statistical parameters calculated with WIIC = 0.5 for all splits are shown in Table 1. The experimental pEC50 values compared to the calculated values for the four splits are shown in Figure 1. Table 1 clearly shows the statistical reliability of all models and that they meet the criteria established by Tropsha et al. [15] and Ojha et al. [16]. The established QSAR model of split 3 provides the best statistical parameters (R2 = 0.991, CCC = 0.911, and Q2 = 0.943). The model equation of split 3 is given below:pEC50 = 0.532 (± 0.184) + 0.103 (± 0.003) × DCW (2,30)(1) An additional validation model for the Monte Carlo method was performed using The AD. We determined the theoretical range in which the predictions of the constructed SMILES-based QSAR model are accurate. In the case of TF1, without considering the influence of IIC on activity (pEC50), the number of outliers for split 3 was four (i.e., compounds No. 29, 32, 34, and 35). In the case of TF2, the number of outliers for split 3 was three (i.e., compounds No. 8, 9, and 16). 2.2. GA-MLR QSAR Model The GA-MLR method was performed on the training set and then evaluated against the test set based on the selected descriptors. In the GA-MLR model, the three selected descriptors from the entire set including RBN (i.e., No of rotatable bonds), MATS1e (i.e., Moran autocorrelation of lag 2 weighted by Sanderson electronegativity) and G(N..O) (i.e., Sum of geometrical distances between N..O), which contribute to the inhibition activity, were selected to build the QSAR model. The model created using the GA-MLR technique and its statistical parameters (Equation (2)) are shown below:pEC50 = 7.078 − 0.105 × (RBN) + 43.362 × (MATS1e) + 0.013 × (G(N..O))(2) Ntr = 26, Ntest = 10, Rtr2 = 0.915, RMSEtr = 0.491, Qloo2 = 0.880, Rext2 = 0.941, MAEext = 0.416, CCCext = 0.958, QF12= 0.914, QF22= 0.912, QF32= 0.920, F = 79.559, s = 0.534, Kxx = 0.271, ∆K = 0.189, RMSEcv = 0.585, RMSEAV Yscr = 1.588, R2Yscr = 0.117, Q2Yscr = −0.269 where Ntr is the total samples in training and CCC represents the concordance correlation coefficient [17]. QF12, QF22 and QF32 are external validation criteria [18]. The performance of the above parameters of the developed GA-MLR model meets the standard validation criteria according to the OECD guidelines. In addition, Figure 2a illustrates the experimental and the pEC50 endpoints predicted by the developed GA-MLR model, which shows a good correlation between the activity of concern and the three selected descriptors. To further validate the constructed model, the AD is used to evaluate the AD space of the leading model. The AD is performed with the leverage method as shown by the Williams plot in Figure 2b. The dashed lines show the cutoff value of ±3 s.d. and the warning line for the X outlier (h*) is 0.462. William plot show that all molecules are within the AD, with the exception of compound No. 21. 2.3. Mechanistic Interpretation A mechanistic interpretation is a crucial part of OECD. Molecular features responsible for increasing and decreasing an endpoint can be extracted and interpreted from such models. The mechanistic interpretation of the CORAL model can be obtained from multiple runs of Monte Carlo optimization. In three independent Monte Carlo optimization runs, the molecular features extracted from the SMILES attributes with positive CWs are found to be promoters of an increase in pEC50 activity, and the SMILES attributes with negative CWs are found to be promoters of a decrease in pEC50 activity. In contrast, the SMILES attributes with both positive and negative CWs are undefined. The main promoters leading to an increase or decrease in pEC50 values with their CWs for three independent runs of the built QSAR model for split 3 are shown in Table 2. Considering these data, the top-ranking fragments for increasing activity are: no. 1—combination of sp3 carbon with branching; no. 2—presence of sp3 oxygen surrounded by two sp3 carbons; no. 3—presence of oxygen; no. 4—combination of sp3 nitrogen with branching; no. 5—presence of sp3 carbon surrounded by sp3 oxygen and sp3 carbon; no. 6—combination of sp3 nitrogen and sp3 carbon in the aliphatic ring; no. 7—combination of sp3 nitrogen and sp3 carbon); no. 8—presence of sp3 nitrogen surrounded by two sp3 carbons; no. 9—presence of two sp3 carbon atoms; no. 10—presence of nitrogen; no. 11—presence of sp3 carbon surrounded by sp3 nitrogen and sp3 carbon; no. 12—maximum number of nitrogen is 8; and no. 13—maximum number of oxygen is 8. In contrast, the most ranking fragments for decreasing the activity are: no. 1—presence of one ring; no. 2—combination of oxygen, double bond, and branching; no. 3—presence of doubly bonded carbon; and no. 4—combination of sp3 oxygen and branching. Based on these considerations, the promoters of decrease were avoided. The promoters of propagation were exanimated in three different positions indicated by R1, L and R2 in the lead compound 25, which has the higher pEC50 value. The structures of all designed compounds with their pEC50 values are listed in Table S3 in the Supplemental Material. Consequently, eight novel HCV NS5A inhibitors were selected based on these promoters, which showed high activity among the designed NS5A inhibitors (Figure 3). All pEC50 values of the selected inhibitors predicted by SMILES-based QSAR and GA-MLR QSAR models were higher than that of the lead compound 25 (Figure 3). These newly designed hits with their chemical structure, promoters increase, and predicted pEC50 values are shown in Figure 3 and Table 3. 2.4. ADMET Study In silico ADMET analysis was performed using AdmetSAR and OSIRIS servers to evaluate the drug-likeness and pharmacokinetic characteristics of the newly designed compounds. The designed hit compounds do not present risks in terms of tumorigenic, irritant, mutagenic or reproductive effect profiles. Water solubility is important for drug formulation and the determination of the persistence of organic compounds in the environment. The results in Table 4 show that all the newly developed compounds are soluble (water solubility is expressed in log (mol/L)). In addition, the blood–brain barrier (BBB) is the major interface between the central nervous system and the bloodstream. The BBB is an important property because it controls whether drugs can pass through the brain barrier and exert their effects. It is believed that a molecule with a logBB > −1 is widely distributed in the brain. Consequently, the BBB permeability results in Table 4 clearly show the non-penetrating BBB for the new suggested compounds. Moreover, intestinal absorption in humans (HIA) is one of the most important ADME properties. A compound with an intestinal absorption value greater than 30% is considered to be highly absorbed. Consequently, all newly developed compounds can be expected to have good biological activity, drug-like features, and ADMET properties. 3. Materials and Methods 3.1. Data Preparation For this study, a dataset of 36 asymmetric inhibitors of HCV NS5A was used [14,19]. The chemical structures of these derivatives were drawn and were pre-optimized using the molecular mechanics’ force field MMFF94 of the ChemDraw package. Then, their geometries were optimized using the Gaussian 09 software [20], particularly the AM1 method in the gas phase. We calculated vibrational spectra to confirm the optimized structures to be the energy minima. The activity value of each molecule (half-maximal effective concentration, EC50) was converted to its negative logarithmic scale pEC50 = −log (EC50) and used as an independent variable to build QSAR models. Two QSAR models were created using Monte Carlo optimization and the GA-MLR technique. For the Monte Carlo method, the simplified molecular input line entry system (SMILES) was used to symbolize the chemical structure and to develop QSAR models. They were generated with ACD/ChemSketch software (File Version C35E41, Build 125843, 14 Jan 2022, Toronto, ON, Canada) [21]. For the GA-MLR model, the molecular descriptor values (0D–3D) of the 36 compounds were computed using OCHEM [22]. To avoid multicollinear variables in the QSAR model, the total number of variables generated was reduced by excluding descriptors that possessed more than 95% constant values and descriptor pairs with a correlation coefficient greater than 0.9. A final set of 625 descriptors was selected from the initial pool of 3085 descriptors. The molecular structures and their corresponding pEC50 data are listed in Table 5 (the SMILES notation can be found in the Supplemental Materials in Table S1). 3.2. SMILES-Based QSAR Model Construction The Monte Carlo optimization was used to create SMILES-based QSAR models using CORAL 2019 [23]. The SMILES attributes were used in this software to predict the endpoint using optimal descriptors (i.e., correlation weights (CWs)) and the balance-of-correlation method [24]. Four splits were created from the 36 compounds. Each split was randomly divided into 4 partitions: training (35%), invisible training (35%), calibration (15%), and validation (15%). Each set has a different task in constructing the QSAR model. The training set creates the QSAR model by calculating the correlation weight. The invisible training (inv. Train) set is assigned to evaluate the fitness of the molecules that are not included in the training set. The calibration set is used to identify the onset of overfitting, while the validation set is used to test the models for the compounds that are not included in the remaining sets [25,26,27]. Equation (3) describes the optimal descriptor of correlation weights:SMILESDCW (T, Nepoch) = ∑ CW(SK) + ∑ CW(SSK) + ∑ CW(SSSK) + CW (HARD) + CW (Cmax) + CW (Nmax) + CW (Omax)(3) SMILESDCW (T, Nepoch) combines SMILES-based attributes associated with a correlation weight (CW). A description of the optimal SMILES parameters is provided in Table 6. The linear regression approach was used to develop QSAR models after all CWs were calculated as shown in Equation (4). pEC50= C0 + C1 × SMILESDCW (T, Nepoch)(4) C0 is the intercept, while C1 is the slope of the regression equation. In the Monte Carlo method, we defined T as the threshold and Nepoch as the number of epochs. The T coefficient is used as a criterion to divide the SMILES attributes into two classes: an active class in which SMILES attributes are involved in model construction and a rare class (noise) that does not contain SMILES attributes. The T coefficient is used as a criterion to divide the SMILES attributes into two categories: an active class where SMILES attributes contribute to model construction and a rare class (noise) that contains no SMILES attributes. Overtraining can result from these rare attributes producing a good correlation during training and a poor correlation during validation. The Nepoch provides the best statistical quality during calibration [28]. To develop the QSAR models, two types of target functions (TF) are used. TF1 uses balance of correlation as described in Equation (5), while TF2 adds the Index of Ideality of Correlation (IIC) described in Equation (6) [29,30]. IIC (Equation (7)) was proposed as a criterion for evaluating of the predictive power of the developed QSAR models. Namely, it improves the accuracy of the model measured by the coefficient of determination (R2) and the mean absolute error (MAE). The value of the coefficient WIIC (The weight of IIC) can change the strength of the influence of IIC on Monte Carlo optimization. The preferred value of WIIC can be determined by two factors: molecular diversity and endpoint nature [31,32,33]. (5) TF1=Rtraining+Rinv.train−|Rtraining−Rinv.train|×Const (6) TF2=TF1+IIC×WIIC (7) IIC=Rset×min(−MAEset,+MAEset)max(−MAEset,+MAEset) Rtraining and Rinv.train are correlation coefficients between the experimental pEC50 and the calculated pEC50 for each respective set. The empirical Const is typically fixed. Moreover, Rset is the value of the correlation coefficient between the observed and predicted endpoint of a give set. MAE is the mean absolute error, calculated as follows:(8) MAEset=1N_∑k=1N_|Δk|    (Δk < 0, −N is the No. of Δk< 0) (9) +MAEset=1N+∑k=1N+|Δk|    (Δk ≥ 0, +N is the No. of Δk ≥ 0) (10) where, Δk=Observedk− Predictedk= pEC50k(obs)− pEC50k(pred) Δk is the accuracy for the kth substance from a set. A grid-search was used for the best values of T and Nepoch for the four splits (1 to 10 for T and 1 to 30 for Nepoch). The number of optimization probes was set to 3. 3.3. GA-MLR QSAR Construction The first step in QSAR analysis is to choose the most relevant descriptors from the entire pool of computed descriptors. For this purpose, the stepwise linear regression method was applied, and the value of the leave-one-out cross-validation coefficient was used as the fitness function. Thus, 3085 different molecular descriptors were calculated using the OCHEM server [22]. The calculated descriptors were first examined to remove the near-constant and constant variables to decrease the redundancy in the matrix of descriptors. The correlation between the calculated descriptors and inhibitory activity was examined to exclude the collinear descriptors. Finally, 625 molecular descriptors were filtered out from the original set of variables. Then, the stepwise-MLR method was used to select the most relevant descriptors. Finally, three molecular descriptors were selected from the whole set. Based on the selected molecular descriptors, the MLR method used the ordinary least squares (OLS) algorithm to establish a linear relationship between the pEC50 endpoints of NS5A inhibitors and their molecular descriptors. QSARINS software was used to create the GA-MLR model [34,35]. The data set was randomly split into training (26 molecules) and testing (10 molecules) sets with a percentage distribution of 70% and 30%, respectively. The default parameters were used to build the GA-MLR models, except for: subsets = 1 to 5, maximum generation = 10,000 and mutation probability = 0.05. 3.4. QSAR Models Validation The validation process is essential in QSAR to test the model’s suitability to make reliable forecasts of the modeled activity for new compounds with an unknown reaction. This process is considered one of the crucial steps to check the robustness, predictability and reliability of any QSAR model. Four steps are usually used to validate the constructed model, including (a) internal validation or cross-validation using the training set, (b) Y-randomization, (c) independent validation using the test set, and (d) applicability domain (AD) evaluation [36]. 3.4.1. Validation of GA-MLR QSAR Model In the GA-MLR, the validity of the generated QSAR model was confirmed based on: internal validation using leave-many-out (LMO) and leave-one-out (LOO) procedures, Y-randomization, independent validation, and finally by checking the model AD. Moreover, thorough fulfillment of the respective thresholds for the statistical metrics proposed in the literature was evaluated [37]: the determination coefficient Rtr2 ≥ 0.6, the Cross-validated Qloo2 ≥ 0.5, the determination coefficient obtained for the test set Rext2 ≥ 0.6, the root-mean square error RMSEtr < RMSEcv, the concordance correlation coefficient (CCC) ≥ 0.80, QFn2 ≥ 0.6, the Y-scramble correlation coefficient RYscr2 < 0.2, the the Y-scramble cross-validation coefficient QYscr2 < 0.2, QYscr2 < RYscr2, the root-mean-square of Y randomization RMSEAV Yscr and the mean absolute error (MAE) should be near to zero. 3.4.2. Validation of CORAL QSAR Model In Monte Carlo optimization, additional parameters were used to verify the quality of the predictions of the QSAR models. CRp2 is the deviation of the mean determination coefficient of the randomized models (Rr2) from the determination coefficient of the non-randomized models (R2). CRp2 should be greater than 0.5 for an acceptable QSAR model. Rm2 is a metric proposed by Roy et al. [38,39] to indicate the external predictability of QSAR models; the average Rm2 (AvgRm2) should be greater than 0.5, and ΔRm2 should be less than 0.2 (ΔRm2 = Rm2 (x,y) − Rm2 (y,x). x is the experimental value while y is the predicted value of endpoint). Any QSAR model that does not meet the above criteria is eliminated. The formulas for calculating these statistical parameters are listed in Supplementary Material Table S2. 3.5. Applicability Domain The applicability domain (AD) was proposed by the Organization for Economic Cooperation and Development (OECD) guidelines. AD allows the evaluation of the uncertainty in the prediction of a given molecule based on its similarity to the compounds used to develop the model. Compounds outside the AD are considered as outliers In CORAL QSAR models, the AD is determined by the calculated statistical defects d(A) of SMILES based on the distribution of available data among all sets (Equation (11)). The d(A) of the SMILES attribute is depicted as the difference between the probability of the attribute in the training set and that of the calibration set. Outliers are SMILES, whose SMILE error is higher than twice the average error over training set compounds. (11) (A)=|P(A)−P′(A)|N(A)−N′(A) Here, P(A) and P′(A) are the probabilities of the attributes (A) in the training and calibration set, respectively; N(A) and N′(A) are the numbers of times attribute (A) appears in the training and calibration set respectively. The statistical defect (D) for a particular molecule is the total of the statistical defects, d(A), of all the attributes accessible in the SMILES notation. (12) D=defect(SMILES)=∑k=1NAd(A) (13) A molecule is considered outlier when  D > 2 × D_ D_ is the average of the calculated D of training, inv. Train and calibration sets [40]. In the GA-MLR model, the William plot of standardized residual versus leverage was used to visualize the model AD. Reliable model predictions have leverage values framed between the critical leverage with ±3 standard deviations and lower than the warning leverage value h* of 0.48. Outliers are compounds that fall outside the horizontal reference lines on the plot. In contrast, the influential chemicals are compounds that have h > h* [41]. 3.6. ADMET Study ADMET assessment is critical in the early phase of drug discovery. A high-quality therapeutic agent is expected to have excellent efficacy against the target receptor and excellent ADMET properties at a therapeutic dose. Therefore, it is necessary to evaluate the pharmacokinetic profile of the Hit compounds to prevent subsequent drug failure [42]. Drug-likeness properties explain how a compound is distributed inside an organism and thus influence its pharmacological efficacy [43]. The ADMET predictions of the designed compounds were evaluated using AdmetSAR and Osiris property explorer [44,45]. 4. Conclusions Hepatitis C virus is a worldwide health problem that causes several life-threatening chronic liver diseases. Currently, there is no effective vaccine against hepatitis C, and treatment is still quite difficult. Computational methods have repeatedly proven useful in addressing the unique challenges of antiviral drug discovery. In this study, two QSAR models were developed to determine the quantitative relationship between anti-NS5A HCV biological activity and the molecular structure of a series of NS5A inhibitors. Two models were constructed using the GA-MLR and Monte Carlo optimization techniques. The results of the two models were in accordance with OECD guidelines. The model based on SMILES was used to evaluate the effects of the presence or absence of different molecular fragments on the biological activity studied. These results provided insights into the design of the eight novel NS5A inhibitors (against the NS5A target). The GA-MLR model confirmed the obtained inhibitory activities of the eight compounds. The ADMET study demonstrated that the designed molecules have advantageous chemical properties that provide promising inhibitory activity against NS5A. Acknowledgments The authors would like to thank Paola Gramatica for providing a copy of QSARINS Software. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/molecules27092729/s1. Table S1: SMILES notation for the 36 compounds and their experimental activity data; Table S2: The mathematical equation of various statistical parameters; Table S3: Chemical structures of designed NS5A inhibitors. Click here for additional data file. Author Contributions Conceptualization, W.L., D.C. and A.E.A.; methodology, W.L., M.O., I.H. and I.B.; formal analysis, W.L., D.C., R.D. and A.E.A.; resources, D.C. and A.E.A.; writing—original draft preparation, W.L, M.O. and I.H.; writing—review and editing, D.V., D.C., R.D. and A.E.A.; supervision, D.C., R.D. and A.E.A. All authors have read and agreed to the published version of the manuscript. Funding This research received no external funding. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Sample Availability Samples of the compounds are not available from the authors. Figure 1 Experimental versus calculated pEC50 values for the models (i.e., Four Splits). Figure 2 (a) Experimental vs. predicted pEC50 values computed by GA-MLR. (b) Williams plot. Figure 3 Chemical structures (25a–25h) of the newly designed compounds with favorable structural features. molecules-27-02729-t001_Table 1 Table 1 Statistical parameter of built QSAR models and their corresponding equations. Split Set N R 2 CCC IIC Q 2 QF12 QF22 QF32 s MAE F CRp2 AvgRm2 ΔRm2 Equation 1 Training 13 0.872 0.931 0.800 0.812 0.671 0.543 75 0.820 pEC50 = 0.655 (±0.344) + 0.101 (±0.004) × DCW(1,30) Inv.Train 13 0.899 0.921 0.565 0.867 0.580 0.469 98 0.845 Calibration 5 0.964 0.906 0.982 0.893 0.909 0.853 0.942 0.456 0.367 81 0.798 0.615 0.156 Validation 5 0.891 0.853 0.824 0.710 0.685 0.592 0.194 2 Training 13 0.889 0.941 0.314 0.856 0.561 0.388 81 0.843 pEC50 = −0.094 (±0.234) + 0.084 (±0.002) × DCW(1,30) Inv.Train 13 0.937 0.962 0.519 0.915 0.517 0.392 166 0.905 Calibration 5 0.989 0.963 0.994 0.975 0.927 0.927 0.950 0.438 0.342 290 0.858 0.695 0.094 Validation 5 0.860 0.904 0.691 0.566 0.592 0.767 0.132 3 Training 13 0.873 0.932 0.801 0.837 0.634 0.498 76 0.848 pEC50 = 0.532 (±0.184) + 0.103 (±0.003) × DCW(2,30) Inv.Train 13 0.865 0.871 0.581 0.829 0.685 0.510 71 0.805 Calibration 5 0.975 0.960 0.987 0.909 0.937 0.935 0.949 0.428 0.315 120 0.851 0.755 0.072 Validation 5 0.990 0.911 0.719 0.942 0.532 0.728 0.076 4 Training 13 0.941 0.970 0.831 0.923 0.390 0.301 178 0.894 pEC50 = 0.457 (±0.157) + 0.149 (±0.003) × DCW(1,30) Inv.Train 13 0.843 0.914 0.601 0.801 0.650 0.483 59 0.807 Calibration 5 0.924 0.945 0.961 0.637 0.908 0.906 0.897 0.568 0.412 37 0.657 0.752 0.101 Validation 5 0.943 0.892 0.317 0.399 0.647 0.777 0.081 N—number of samples; s—standard error of estimation; F—Fischer ratio. molecules-27-02729-t002_Table 2 Table 2 Promoters of increase and decrease of pEC50 endpoint from split 3. No. Sak CWs Probe 1 CWs Probe 2 CWs Probe 3 NT a NiT b NC c Defect [SAk] d Promoter of endpoint increase 1 C...(....... 0.137 0.188 0.464 13 13 5 0.000 2 C...O...C... 2.041 1.674 2.516 6 8 3 0.015 3 O........... 1.351 1.397 1.810 13 13 5 0.000 4 N...(....... 0.054 0.063 0.110 13 13 5 0.000 5 O...C...C... 0.093 0.324 0.338 6 11 4 0.034 6 N...C...1... 0.222 0.101 0.043 7 6 3 0.006 7 N...C....... 0.423 0.479 0.573 13 13 5 0.000 8 C...N...C... 0.648 0.821 0.356 9 9 3 0.008 9 C...C....... 0.325 0.382 0.273 13 13 5 0.000 10 N........... 0.330 0.111 0.147 13 13 5 0.000 11 N...C...C... 0.798 0.643 0.829 11 13 5 0.009 12 Nmax.8...... 0.798 0.510 0.590 2 2 0 1.000 13 Omax.6...... 0.169 0.527 0.871 2 5 2 0.061 Promoter of endpoint decrease 1 1........... −0.1812 −0.0120 −0.0134 13 13 5 0.000 2 =...O...(... −0.0579 −0.1326 −0.2212 13 13 5 0.000 3 C...=....... −0.2642 −0.0515 −0.1487 13 13 5 0.000 4 O...(....... −1.1001 −0.8219 −1.2695 13 13 5 0.000 NT, NiT, and Nc are the numbers of SMILES (samples) that include a given attribute (SAk) in the training set a, inv.Training set b, and calibration set c. d Defect [SAk] is the difference of probabilities of SAk in the training and calibration sets, divided by the sum of total numbers of the SAk in the training and calibration sets. molecules-27-02729-t003_Table 3 Table 3 The newly designed compounds and their predicted pEC50 using the Monte Carlo optimization and the GA-MLR models. Designed Compound Promoters of Endpoint Increase pEC50 (CORAL) pEC50 (GA-MLR) 25 9.68 10.01 25a - Combination of sp3 carbon with branching - Maximum number of oxygen is 6 9.88 10.15 25b - Combination of two sp3 carbons - Maximum number of oxygen is 6 11.78 10.13 25c - Presence of sp3 oxygen surrounded by two sp3 carbons 12.18 11.18 25d - Presence of sp3 carbon surrounded by sp3 oxygen and sp3 carbon 12.27 11.23 25e - Combination of sp3 nitrogen and sp3 carbon in aliphatic ring - Maximum number of oxygen is 6 - Maximum number of nitrogen is 8 11.95 10.71 25f - Presence of sp3 carbon surrounded by sp3 nitrogen and sp3 carbon - Maximum number of oxygen is 6 - Maximum number of nitrogen is 8 12.05 12.28 25g - Combination of two sp3 carbons - Combination of sp3 nitrogen and sp3 carbon in aliphatic ring - Maximum number of oxygen is 6 - Maximum number of nitrogen is 8 11.69 10.04 25h - Presence of sp3 carbon surrounded by sp3 oxygen and sp3 carbon - Combination of sp3 nitrogen and sp3 carbon in aliphatic ring - Maximum number of nitrogen is 8 12.19 11.44 molecules-27-02729-t004_Table 4 Table 4 Pharmacokinetic and ADME properties of the designed molecules and the lead compound evaluated using AdmetSAR and Osiris property explorer. Pharmacokinetic Properties MW (g·mol−1) Lipophilicity (logP) Solubility log(mol/L) TPSA (Å2) HBA HBD BBB HIA 25 835.50 6.71 −2.88 157.99 13 4 0.012 0.010 25a 849.52 6.67 −3.11 157.99 13 4 0.018 0.010 25b 863.53 6.95 −3.20 157.99 13 4 0.015 0.009 25c 865.51 5.37 −3.32 167.22 14 4 0.009 0.012 25d 879.53 5.68 −3.49 167.22 14 4 0.009 0.007 25e 864.53 4.73 −3.12 170.02 14 5 0.013 0.189 25f 878.54 4.85 −3.06 170.02 14 5 0.010 0.078 25g 878.54 5.39 −3.06 184.01 14 6 0.023 0.049 25h 894.54 4.26 −3.28 193.24 15 6 0.015 0.035 Molecular weight (MW), blood–brain barrier (BBB), total polar surface area (TPSA), hydrogen bond acceptor (HBA), hydrogen bond donor (HBD), human intestinal absorption (HIA). molecules-27-02729-t005_Table 5 Table 5 Chemical structures and the studied biological activity data. Compound P pEC50 1 t 10.01 2 5.55 Compound R1 pEC50 3 t 5.62 4 t 5.89 5 9.16 6 7.25 7 6.21 8 t 7.36 9 6.88 10 6.61 11 7.00 12 t 7.09 13 8.96 14 6.45 15 6.45 Compound R1 R2 pEC50 16 7.47 17 8.46 18 8.09 19 8.64 20 10.41 Compound P pEC50 21 5 22 5.55 Compound R1 L R2 pEC50 23 t 9.82 24 9.77 25 10.41 26 9.46 27 9.01 Compound R1 R3 pEC50 28 t 8.47 29 9.41 30 t 8.68 31 9.57 32 9.74 33 t 9.89 34 t 9.92 35 10.23 36 (Daclatasvir) 10.30 t: test set. *:Ramification position molecules-27-02729-t006_Table 6 Table 6 Description of the SMILES attributes. SMILES Notation Description SK One symbol or two symbols that cannot be examined separately SSK Combination of two SMILES atoms SSSK Combination of three SMILES atoms HARD Existence of some chemical element Cmax Number of rings Nmax Number of nitrogen atoms Omax Number of oxygen atoms Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. De Oliveria Andrade L. D’oliveira A. Melo R.C. De Souza E.C. Silva C.A.C. Parana R. Association between hepatitis C and hepatocellular carcinoma J. Glob. Infect. Dis. 2009 1 33 10.4103/0974-777X.52979 20300384 2. Davis G. Projecting future complications of chronic hepatitis C in the United States Liver Transpl. 2003 9 331 338 10.1053/jlts.2003.50073 12682882 3. Hepatitis C. Available online: https://www.who.int/news-room/fact-sheets/detail/hepatitis-c (accessed on 10 December 2021) 4. Moradpour D. Penin F. Hepatitis C Virus Proteins: From Structure to Function Curr. Top. Microbiol. Immunol. 2013 369 113 142 10.1007/978-3-642-27340-7_5 23463199 5. Gudi S.K. Chhabra M. Rashid M. Genotype-8: A Modern Family Member of Hepatitis C Virus Indian J. Pharm. Educ. Res. 2019 53 s518 s521 10.5530/ijper.53.4s.146 6. Venkatesan A. Prabhu Dass J F. Review on chemogenomic approaches towards hepatitis C viral targets J. Cell. Biochem. 2019 120 12167 12181 10.1002/jcb.28581 30887580 7. Baumert T.F. Berg T. Lim J.K. Nelson D.R. Status of Direct-Acting Antiviral Therapy for Hepatitis C Virus Infection and Remaining Challenges Gastroenterology 2019 156 431 445 10.1053/j.gastro.2018.10.024 30342035 8. Das D. Pandya M. Recent Advancement of Direct-acting Antiviral Agents (DAAs) in Hepatitis C Therapy Mini-Rev. Med. Chem. 2018 18 584 596 10.2174/1389557517666170913111930 28901852 9. Zhong M. Peng E. Huang N. Huang Q. Huq A. Lau M. Colonno R. Li L. Discovery of ravidasvir (PPI-668) as a potent pan-genotypic HCV NS5A inhibitor Bioorganic Med. Chem. Lett. 2016 26 4508 4512 10.1016/j.bmcl.2016.07.066 10. Pawlotsky J.-M. NS5A inhibitors in the treatment of hepatitis C J. Hepatol. 2013 59 375 382 10.1016/j.jhep.2013.03.030 23567084 11. Nakamoto S. Kanda T. Shirasawa H. Yokosuka O. Antiviral therapies for chronic hepatitis C virus infection with cirrhosis World J. Hepatol. 2015 7 1133 10.4254/wjh.v7.i8.1133 26052402 12. Kobayashi Y. Yoshida K. Development of QSAR models for prediction of fish bioconcentration factors using physicochemical properties and molecular descriptors with machine learning algorithms Ecol. Inform. 2021 63 101285 10.1016/j.ecoinf.2021.101285 13. Hdoufane I. Bjij I. Oubahmane M. Soliman M.E. Villemin D. Cherqaoui D. In silico design and analysis of NS4B inhibitors against hepatitis C virus J. Biomol. Struct. Dyn. 2020 40 1915 1929 10.1080/07391102.2020.1839561 33118481 14. Nakamura H. Akagi Y. Terui T. Fujioka S. Komoda Y. Kinoshita W. Maeda K. Ukaji Y. Inaba T. Discovery of a novel unsymmetrical structural class of HCV NS5A inhibitors with low picomolar antiviral activity Bioorganic Med. Chem. Lett. 2020 30 126932 10.1016/j.bmcl.2019.126932 31952964 15. Golbraikh A. Tropsha A. Beware of q2! J. Mol. Graph. Model. 2002 20 269 276 10.1016/S1093-3263(01)00123-1 11858635 16. Ojha P.K. Mitra I. Das R.N. Roy K. Further exploring rm2 metrics for validation of QSPR models Chemom. Intell. Lab. Syst. 2011 107 194 205 10.1016/j.chemolab.2011.03.011 17. Chirico N. Gramatica P. Real external predictivity of QSAR models: How to evaluate It? Comparison of different validation criteria and proposal of using the concordance correlation coefficient J. Chem. Inf. Modeling 2011 51 2320 2335 10.1021/CI200211N 18. Chirico N. Gramatica P. Real external predictivity of QSAR models. Part 2. New intercomparable thresholds for different validation criteria and the need for scatter plot inspection J. Chem. Inf. Modeling 2012 52 2044 2058 10.1021/ci300084j 19. Nakamura H. Fujioka S. Terui T. Okuda S. Kondo K. Tamatani Y. Akagi Y. Komoda Y. Kinoshita W. Ito S. Orally bioavailable HCV NS5A inhibitors of unsymmetrical structural class Bioorganic Med. Chem. Lett. 2020 30 127361 10.1016/j.bmcl.2020.127361 20. Frisch M.J. Trucks G.W. Schlegel H.B. Scuseria G.E. Robb M.A. Cheeseman J.R. Scalmani G. Barone V. Mennucci B. Petersson G.A. Fox, Gaussian 09, Wallingford C.T 2009 Available online: http://gaussian.com/ (accessed on 18 February 2022) 21. Chemical Structure Drawing Software |ACD/ChemSketch Available online: https://www.acdlabs.com/products/draw_nom/draw/chemsketch/ (accessed on 15 February 2022) 22. Sushko I. Novotarskyi S. Körner R. Pandey A.K. Rupp M. Teetz W. Brandmaier S. Abdelaziz A. Prokopenko V.V. Tanchuk V.Y. Online chemical modeling environment (OCHEM): Web platform for data storage, model development and publishing of chemical information J. Comput. Aided. Mol. Des. 2011 25 533 554 10.1007/s10822-011-9440-2 21660515 23. CORAL/Free Software for QSAR and nanoQSAR Available online: http://www.insilico.eu/coral/ (accessed on 9 December 2021) 24. Kumar P. Kumar A. Sindhu J. Lal S. QSAR Models for Nitrogen Containing Monophosphonate and Bisphosphonate Derivatives as Human Farnesyl Pyrophosphate Synthase Inhibitors Based on Monte Carlo Method Drug Res. 2019 69 159 167 10.1055/a-0652-5290 25. Duhan M. Singh R. Devi M. Sindhu J. Bhatia R. Kumar A. Kumar P. Synthesis, molecular docking and QSAR study of thiazole clubbed pyrazole hybrid as α-amylase inhibitor J. Biomol. Struct. Dyn. 2021 39 91 107 10.1080/07391102.2019.1704885 31838966 26. Kumar P. Kumar A. Monte Carlo Method Based QSAR Studies of Mer Kinase Inhibitors in Compliance with OECD Principles Drug Res. 2018 68 189 195 10.1055/s-0043-119288 28992659 27. Achary P. Toropova A. Toropov A. Combinations of graph invariants and attributes of simplified molecular input-line entry system (SMILES) to build up models for sweetness Food Res. Int. 2019 122 40 46 10.1016/j.foodres.2019.03.067 31229093 28. Kumar P. Kumar A. CORAL: QSAR models of CB1 cannabinoid receptor inhibitors based on local and global SMILES attributes with the index of ideality of correlation and the correlation contradiction index Chemom. Intell. Lab. Syst. 2020 200 103982 10.1016/j.chemolab.2020.103982 29. Duhan M. Kumar P. Sindhu J. Singh R. Devi M. Kumar A. Kumar R. Lal S. Exploring biological efficacy of novel benzothiazole linked 2,5-disubstituted-1,3,4-oxadiazole hybrids as efficient α-amylase inhibitors: Synthesis, characterization, inhibition, molecular docking, molecular dynamics and Monte Carlo based QSAR studies Comput. Biol. Med. 2021 138 104876 10.1016/j.compbiomed.2021.104876 34598068 30. Toropov A. Raška I. Toropova A.P. Raškova M. Veselinović A.M. Veselinović J.B. The study of the index of ideality of correlation as a new criterion of predictive potential of QSPR/QSAR-models Sci. Total Environ. 2019 659 1387 1394 10.1016/j.scitotenv.2018.12.439 31096349 31. Toropova A.P. Toropov A.A. The index of ideality of correlation: Improvement of models for toxicity to algae Nat. Prod. Res. 2019 33 2200 2207 10.1080/14786419.2018.1493591 30317875 32. Toropova A.P. Toropov A. The index of ideality of correlation: A criterion of predictability of QSAR models for skin permeability? Sci. Total Environ. 2017 586 466 472 10.1016/j.scitotenv.2017.01.198 28196626 33. Kumar P. Kumar A. Sindhu J. Design and development of novel focal adhesion kinase (FAK) inhibitors using Monte Carlo method with index of ideality of correlation to validate QSAR SAR QSAR Environ. Res. 2019 30 63 80 10.1080/1062936X.2018.1564067 30793981 34. Gramatica P. Chirico N. Papa E. Cassani S. Kovarich S. QSARINS: A new software for the development, analysis, and validation of QSAR MLR models J. Comput. Chem. 2013 34 2121 2132 10.1002/jcc.23361 35. Gramatica P. Cassani S. Chirico N. QSARINS-chem: Insubria datasets and new QSAR/QSPR models for environmental pollutants in QSARINS J. Comput. Chem. 2014 35 1036 1044 10.1002/jcc.23576 24599647 36. Gramatica P. Principles of QSAR models validation: Internal and external QSAR Comb. Sci. 2007 26 694 701 10.1002/qsar.200610151 37. Veerasamy R. Rajak H. Jain A. Sivadasan S. Varghese C.P. Agrawal R.K. Validation of QSAR Models—Strategies and Importance Int. J. Drug Des. Discov. 2011 2 511 519 38. Roy P.P. Paul S. Mitra I. Roy K. On Two Novel Parameters for Validation of Predictive QSAR Models Molecules 2009 14 1660 1701 10.3390/molecules14051660 19471190 39. Roy K. Kar S. The rm2 metrics and regression through origin approach: Reliable and useful validation tools for predictive QSAR models (Commentary on ’Is regression through origin useful in external validation of QSAR models?’) Eur. J. Pharm. Sci. 2014 62 111 114 10.1016/j.ejps.2014.05.019 24881556 40. Toropov A. Toropova A. Lombardo A. Roncaglioni A. Benfenati E. Gini G. CORAL: Building up the model for bioconcentration factor and defining it’s applicability domain Eur. J. Med. Chem. 2011 46 1400 1403 10.1016/j.ejmech.2011.01.018 21295893 41. Dimitrov S. Dimitrova G. Pavlov T. Dimitrova N. Patlewicz G. Niemela J. Mekenyan O. A stepwise approach for defining the applicability domain of SAR and QSAR models J. Chem. Inf. Modeling 2005 45 839 849 10.1021/ci0500381 42. Guan L. Yang H. Cai Y. Sun L. Di P. Li W. Liu G. Tang Y. ADMET-score—A comprehensive scoring function for evaluation of chemical drug-likeness Medchemcomm 2019 10 148 157 10.1039/C8MD00472B 30774861 43. Tian S. Wang J. Li Y. Li D. Xu L. Hou T. The application of in silico drug-likeness predictions in pharmaceutical research Adv. Drug Deliv. Rev. 2015 86 2 10 10.1016/j.addr.2015.01.009 25666163 44. Yang H. Lou C. Sun L. Li J. Cai Y. Wang Z. Li W. Liu G. Tang Y. AdmetSAR 2.0: Web-service for prediction and optimization of chemical ADMET properties Bioinformatics 2019 35 1067 1069 10.1093/bioinformatics/bty707 30165565 45. Molecular Properties Prediction—Osiris Property Explorer Available online: https://www.organic-chemistry.org/prog/peo/ (accessed on 10 December 2021)

PMC009xxxxxx/PMC9099612.txt

==== Front Int J Mol Sci Int J Mol Sci ijms International Journal of Molecular Sciences 1422-0067 MDPI 10.3390/ijms23094574 ijms-23-04574 Article Multi-Cohort Transcriptomic Subtyping of B-Cell Acute Lymphoblastic Leukemia https://orcid.org/0000-0002-7262-2656 Mäkinen Ville-Petteri 1234* https://orcid.org/0000-0001-5043-6943 Rehn Jacqueline 56 Breen James 678 Yeung David 56910 https://orcid.org/0000-0003-4844-333X White Deborah L. 5691112 Follo Matilde Yung Academic Editor 1 Computational and Systems Biology Program, Precision Medicine Theme, South Australian Health and Medical Research Institute, Adelaide, SA 5000, Australia 2 Australian Centre for Precision Health, UniSA Clinical & Health Sciences, University of South Australia, Adelaide, SA 5000, Australia 3 Computational Medicine, Faculty of Medicine, University of Oulu, FI-90014 Oulu, Finland 4 Center for Life Course Health Research, Faculty of Medicine, University of Oulu, FI-90014 Oulu, Finland 5 Blood Cancer Program, Precision Medicine Theme, South Australian Health and Medical Research Institute, Adelaide, SA 5000, Australia; jacqueline.rehn@sahmri.com (J.R.); david.yeung@adelaide.edu.au (D.Y.); deborah.white@sahmri.com (D.L.W.) 6 Faculty of Health and Medical Sciences, University of Adelaide, Adelaide, SA 5005, Australia; jimmymbreen@gmail.com 7 South Australian Genomics Centre, South Australian Health and Medical Research Institute, Adelaide, SA 5000, Australia 8 Robinson Research Institute, University of Adelaide, Adelaide, SA 5005, Australia 9 Australian and New Zealand Children’s Oncology Group, Clayton, VIC 3168, Australia 10 Department of Haematology, Royal Adelaide Hospital and SA Pathology, Adelaide, SA 5000, Australia 11 Faculty of Sciences, University of Adelaide, Adelaide, SA 5005, Australia 12 Australian Genomics Health Alliance, Parkville, VIC 3052, Australia * Correspondence: ville-petteri.makinen@sahmri.com; Tel.: +61-8-8128-4054 20 4 2022 5 2022 23 9 457411 3 2022 19 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). RNA sequencing provides a snapshot of the functional consequences of genomic lesions that drive acute lymphoblastic leukemia (ALL). The aims of this study were to elucidate diagnostic associations (via machine learning) between mRNA-seq profiles, independently verify ALL lesions and develop easy-to-interpret transcriptome-wide biomarkers for ALL subtyping in the clinical setting. A training dataset of 1279 ALL patients from six North American cohorts was used for developing machine learning models. Results were validated in 767 patients from Australia with a quality control dataset across 31 tissues from 1160 non-ALL donors. A novel batch correction method was introduced and applied to adjust for cohort differences. Out of 18,503 genes with usable expression, 11,830 (64%) were confounded by cohort effects and excluded. Six ALL subtypes (ETV6::RUNX1, KMT2A, DUX4, PAX5 P80R, TCF3::PBX1, ZNF384) that covered 32% of patients were robustly detected by mRNA-seq (positive predictive value ≥ 87%). Five other frequent subtypes (CRLF2, hypodiploid, hyperdiploid, PAX5 alterations and Ph-positive) were distinguishable in 40% of patients at lower accuracy (52% ≤ positive predictive value ≤ 73%). Based on these findings, we introduce the Allspice R package to predict ALL subtypes and driver genes from unadjusted mRNA-seq read counts as encountered in real-world settings. Two examples of Allspice applied to previously unseen ALL patient samples with atypical lesions are included. acute lymphoblastic leukemia RNA-seq confounder adjustment machine learning ==== Body pmc1. Introduction Acute lymphoblastic leukemia (ALL) is characterized by abnormal differentiation and proliferation of malignant lymphoid precursors in blood and bone marrow [1,2]. Usually, the disease manifests as abnormal proliferation of B-cells while less than a quarter of patients present with a T-cell malignancy. The incidence rate is the highest in children under the age of 10 and in adults over the age of 65, with an average age-adjusted global annual incidence of 0.85 per 100,000 individuals [3]. Despite recent progress in molecular phenotyping, ALL remains a life-threatening disease, and advanced age (beyond pediatric) and certain subtypes are predictive of poor outcomes [1,2,4,5,6,7]. For these reasons, there is a compelling rationale for improving diagnostic tools and for pursuing deeper biological insight into ALL through new emerging technologies. Cytogenetic testing, immunophenotyping and molecular assays are essential for the diagnosis and further stratification of the disease into subtypes with different biological characteristics and prognoses [8,9,10]. Transcriptomic profiling of lymphoblastic cells is a recent addition to the diagnostic toolbox and has led to the detection of new ALL subtypes [11,12,13,14,15]. Standard cytogenetic testing (G-banded karyotype) with adjunctive fluorescence in situ hybridization (FISH) can detect aneuploidy and chromosomal translocations such as BCR::ABL1, ETV6::RUNX1 and TCF3::PBX1 fusions [8,16]. KMT2A lesions, intrachromosomal amplification of chromosome 21 (iAMP21) and IGH::CRLF2 and P2RY8::CRLF2 fusions are also detectable. Other subtypes are identified only through additional analyses such as real-time PCR, single nucleotide polymorphisms and RNA sequencing. Subtypes such as DUX4 rearrangements, ETV6::RUNX1-like and PAX5 alterations are examples where the underlying genomic alterations are difficult to detect using standard-of-care laboratory methods, and RNA-seq profiling has emerged as an important diagnostic support tool [11,13,17]. For these reasons, understanding the information carried by transcriptomes in relation to the genome alterations in ALL is important, particularly for those patients for whom a conclusive genomic driver cannot be determined by current molecular diagnostics. The high volume of RNA sequencing data per individual makes it necessary to employ machine learning techniques to process the raw information for the identification of subtypes [7,11,12,18,19,20]. In one such study, Gu et al. used clustering algorithms to characterize the transcriptional landscape of ALL [12]; the clusters were further investigated against identifiable genomic lesions to classify individual patients and refine the taxonomy of ALL. Using the taxonomy, the authors were able to assign subtypes for 94% of the study subjects. Recently, Schmidt et al. used the taxonomy to classify patients in multiple datasets [11,21]. Encouraged by these successes, we set out to leverage gene expression data within our transcriptomic analysis pipeline of B-cell ALL to improve diagnoses and gain biological insight into the transcriptional landscape of ALL. Several issues related to mRNA-seq profiling of ALL have not yet been addressed. Firstly, clustering analyses of mRNA-seq data should be subjected to the rigorous adjustment of biases that is standard practice in epidemiology [22,23,24]; instead, most studies opt to (mis)use methods such as surrogate variable analysis that may cause artifacts if the data batches and biological characteristics are correlated (as they tend to be in most multi-cohort collections). These artifacts will be further amplified by clustering algorithms and machine learning models. Cohort biases may explain why gene signatures of subtypes derived from gene expression clusters, such as Ph-like, have been difficult to consolidate between different studies without additional experiments and analyses [25,26,27]. Secondly, a taxonomy that is based on gene expression profiles should not be used when fitting a machine learning model to RNA-seq read counts. Under the worst-case scenario, a transcriptome dataset affected by cohort bias leads to artificial clustering; the same artificial clustering of gene expression is then captured by machine learning and passed on as a falsely distinct subtype. Instead, we propose that gene expression classifiers should be trained with directly observable sequence variants or otherwise independently distinguishable subtypes or with longitudinal data on clinical outcomes and treatment effects. The aim of this study is to introduce a reliable ALL classifier that can be integrated into current transcriptomic analyses with minimal additional resources and that can reliably classify ALL cases in an unbiased manner. To achieve reliability, we use new techniques to adjust for cohort and RNA-seq platform biases in a set of 1279 North American patients and then validate the predictions in an independent cohort of 767 Australian patients. To achieve easy integration into existing workflows, we introduce the Allspice R package with extensive documentation, small programmatic footprint and additional features for predicting genomic drivers and for confirming the tissue identity of the biological samples. The tool is also easy to train for any other disease or classification problem or to update with improved models of ALL in the future. 2. Results 2.1. Cohort Characteristics and Study Design The distribution of ALL subtypes across cohorts is shown in Table 1. Note the definition of genomic subtypes: we excluded categories such as Ph-like due to the technical reasons described in Section 4 and in Supplement Figure S1. The most common subtypes included CRLF2 (between 3.0% and 18.4% of cohort participants), ETV6::RUNX1 (≥ 9.9% of pediatric patients), hyperdiploid (≥ 9.8% of pediatric patients), KMT2A (between 2.2% and 13.4%) and BCR::ABL1 (between 2.4% and 21.5%). There were differences between the cohorts regarding age (p ≤ 7.6 × 10−11) and several subtypes, including pediatric CRLF2 (P = 8.4 × 10−12), pediatric ETV6::RUNX1 (p = 2.3 × 10−8), KMT2A across both age categories (p ≤ 3.2 × 10−7) and non-pediatric BCR::ABL1 (p = 0.00016). The percentage of undefined samples was between 8.7% in the St. Jude Children’s Research Hospital cohort and 33.1% in the Australian pediatric cohort. The RNA data were filtered and normalized separately for the North American and Australian samples (Figure 1A,B), and the training data were adjusted for technical and cohort confounders (Figure 1C). We also saved the unadjusted RNA data and compared it against the adjusted data to remove the most confounded genes (Figure 1D). In the next step, we created an internal training and testing set by splitting the North American data into two random subsets (Figure 1E). These two subsets were used as the initial material for selecting features and controlling for the complexity of machine learning models via hyperparameters. For the full models, we used all North American samples for training and the Australian dataset as an independent external validation set (Figure 1F). 2.2. Adjustments for Confounders Confounding factors were mitigated first by surrogate variable analysis within pre-defined cohort batches (details in Section 4 and in Supplement Figure S2), and batch differences between the cohorts were then adjusted with a new approach we recently developed for time-series metabolomics data (accepted for publication). The adjustments removed correlations between gene expression and RNA library format (Figure 2A,D) and between the two continents of origin (Figure 2B,E) but did not influence the correlations between gene expression and ALL subtypes (Figure 2C,F; Supplement Figures S3–S5). To verify that the data processing was technically sound, we constructed a visual layout of the individuals according to the North American data (Figure 3A) and then projected the unadjusted American and Australian datasets onto the layout using the same statistical model (Figure 3B). The use of unadjusted data is important here since it provides a more realistic picture of how new, previously unseen samples would behave in a diagnostic pipeline. Clusters of the most frequent ALL subtypes were observable, and there was a high degree of concordance between the training and validation cohorts. A visualization of all subtypes and undefined samples is available in Supplement Figure S6. For convenience, we organized the subtypes into a central supergroup (CRLF2, hyperdiploid, hypodiploid, PAX5 alterations and BCR::ABL1) and distinct subtypes on the periphery (DUX4, ETV6::RUNX1, KMT2A, PAX5 P80R, TCF3::PBX1 and ZNF384). 2.3. Classification of B-Cell ALL Subtypes Based on RNA-Seq Profiling Positive predictive values (PPV) of machine learning models are visualized in Figure 4, and complete performance metrics are available in Supplement Tables S1–S3. We used three different types of models (centroids, PLS and random forest, details in Methods) to exclude any artifacts that may be specific to a particular algorithm. Hyperparameters are listed in Supplement Table S4. We focused on the external validation set as the primary benchmark of accuracy. Furthermore, all performance metrics were calculated for standardized but unadjusted data to simulate a scenario where new samples are analyzed one at a time without the opportunity to adjust for cohort effects. Overall, differences between the three methods were negligible when considering the confidence intervals of the performance estimates (Figure 4). Accurate classification models were achieved for DUX4 (PPV ≥ 95% in the external validation cohort across the three methods), ETV6::RUNX1 (PPV ≥ 91%), KMT2A (PPV ≥ 84%), PAX5 P80R (PPV ≥ 85%), TCF3::PBX1 (PPV ≥ 92%) and ZNF384 (PPV ≥ 96%). Together, 186 individuals (24%) of the Australian participants had one of these genomic subtypes (Figure 4D). More varied PPVs were observed in the central supergroup, including the CRLF2 subtype (PPV ≥ 83%, Figure 4B), hyperdiploid (PPV ≥ 72%), hypodiploid (PPV ≥ 54%), PAX5 alterations (PPV ≥ 33%) and BCR::ABL1 (PPV ≥ 90%). Collectively, corresponding genomic subtypes were observed in 304 individuals or 40% of Australian participants (Figure 4D). Medium to high accuracy was achieved for rare subtypes such as IKZF1 N159Y (PPV = 100%, Figure 4C); however, the small number of cases resulted in substantial statistical uncertainty. 2.4. Allspice Classifier Based on the results, we concluded that the centroid model of 45 prioritized genes is the preferred choice as the practical classifier due to its comparable performance to the other methods and technical simplicity. The centroids were also robust against overfitting, as indicated by the flattening of internal training and testing performance when the number of inputs was increased (Supplement Figure S7). The robustness against overfitting enabled us to modify the study design to extract the maximum information from the available data (Supplement Figure S8). In the new design, the centroid model was fitted to the combined adjusted American and Australian data. We also added an extra step to account for sex and age that may carry important predictive information. The results for a BCR::ABL1 patient are depicted in Figure 5. The classifier identifies the subtype centroid that is the most similar to the observed RNA expression profile (Figure 5A). The display includes the frequency of the subtype in the training data versus any other subtype given the patient’s RNA profile (Figure 5B). Allspice also provides more detailed information on how the patient fits the transcriptional landscape of ALL (technical details in Section 4). The left panel shows the proximity of the patient to each subtype, respectively (Figure 5C). The ETV6::RUNX1 subtype is an example of a genomic lesion that manifests as a clearly observable signature (Supplement Figure S9). On the other hand, there is more overlap within the central supergroup and subtypes such as hypodiploid can manifest simultaneous transcriptional proximity to multiple subtypes (Supplement Figure S10). If this overlap is too great, i.e., if it is uncertain statistically which subtype is the closest match, the patient is classified as having an ambiguous transcriptional profile (Supplement Figure S11). The middle panel contains information on which combination of lesion-harboring genes best fits with the observed RNA profile (Figure 5D, see also Methods). In this case, the gene expression profile is compatible with the classical BCR::ABL1 fusion without other strong signals. In the release version of Allspice, we have included only mutually exclusive gene combinations with at least five cases in the training set since rarer combinations could be difficult to confirm statistically. Matching affected genes directly with RNA profiles may provide additional clues for samples where other diagnostic results are inconclusive (Supplement Figures S10 and S11). The right-hand panel shows the proximity of the RNA profile to typical B-cell ALL versus other cell types from public sources (Figure 5E). For example, the Australian cohorts include patients that were recruited from routine practice, some of whom had a low leukemia burden (Supplement Figure S12). For these individuals, the RNA data is closer to healthy blood and will be indicated by this panel. This feature is useful in circumstances where the sequence analyst has limited clinical information available. 2.5. Classifier Performance The overall classification results are shown in Figure 6 and in Table 2. Of 2046 transcriptomes, 483 (24%) were designated ambiguous, and 89 (4.4%) were not classified due to poor proximity to any subtype. Performance was estimated first for all samples, including those with undefined genomic subtype. Since not all B-cell ALL cases can currently be attributed to a specific genomic lesion, these numbers are conservative estimates for the accuracy of the gene expression profile as an indicator of the underlying genomic lesion. Distinct subtypes were detected with high confidence (PPV ≥ 87%), whereas there were more ambiguous and unclassified samples in the central supergroup (see Supplement Figure S13 for a detailed break-down). The second set of results was calculated for 1598 patients that had a confirmed genomic subtype. Strong performance was observed for distinct subtypes (PPV ≥ 97%) and moderate accuracy for the central supergroup (PPV ≥ 75%). This scenario captures the statistical accuracy of the classifier in ideal conditions. Thirdly, when samples that failed the proximity or exclusivity thresholds were excluded, further improvements in PPV were seen across subtypes (10 out of 18 subtypes showed PPV = 100%). This shows how assessing the sample quality will help to avoid the misclassification of borderline cases. Of note, a high proportion of samples (56%) that were classified as having BCL2/MYC gene expression subtype were also identified as not originating from ALL B-cells in the tissue classifier (Supplement Figure S14), which may explain why it was the most difficult subtype to predict. 3. Discussion B-cell ALL remains a life-threatening disease, particularly for adult patients of specific genomic subtypes [1,2,3,9]. Recently, rapid progress has been made in detecting ALL subtypes by RNA sequencing [12,13,14,15,19] and in subtype-specific treatments [6,20]. In this study, we present new data from a large Australian dataset and new findings from rigorous statistical and practical considerations to better leverage gene expression profiling in the diagnosis of ALL subtypes. We confirmed six genomically defined subtypes in one-third of patients that produce highly predictable mRNA profiles (PPV ≥ 89%). A further 40% of individuals were distinguishable by mRNA-seq expression levels, although the associations between specific genomic lesions and transcriptomic profiles were less certain. To dissect the biological ambiguity, we developed a proof-of-concept classifier that aggregates genome-wide mRNA-seq read counts into simplified RNA biomarker scores that indicate how well and where a patient’s RNA profile fits in the landscape of B-cell ALL subtypes. 3.1. Definition of ALL Subtypes We used simpler definitions of genomic ALL subtypes compared to some of the previous reports [12,21]. The streamlined presentation provided multiple benefits, although trade-offs were unavoidable. Firstly, it allowed for sufficient group sizes for robust statistics and a statistically meaningful overview of the transcriptional landscape (e.g., Figure 3). On the other hand, more granular information on the exact nature of sequence alterations may be of high clinical importance but not captured by our subtype definitions. To gain better mechanistic insight, the Allspice classifier includes a feature that indicates genes that may harbor a driver mutation (Figure 5D), and further development of this concept may enable accurate diagnostics for targetable gene expression abnormalities. We also designed Allspice to support subtype definitions that are not mutually exclusive, thus providing flexibility for future updates. The second rationale for the streamlined ALL taxonomy was to ensure a rigorous study design for machine learning. Previous studies have classified patients according to the way their RNA-seq expression profiles cluster (examples include the Ph-like and ETV6::RUNX1-like subtypes [12,13]). However, these definitions are problematic for the training of RNA profiling classifiers—the same gene expression levels should not be used to first define and then predict a subtype. Instead, we relied on observable sequence abnormalities or other information that was not derived from gene expression levels (except for DUX4). As a trade-off, the size of the training set decreased, and the number of individuals with undefined genomic subtypes increased; however, such uncertainty is to be expected in real-world datasets that manifest substantial biological variation in how genomic lesions drive altered gene expression profiles. 3.2. Classification Performance and Utility Overall, B-cell ALL subtypes that could be identified by fusion callers and cytogenetics had distinctive mRNA-seq read count profiles (Allspice classified 90% of samples with a defined genomic subtype correctly). In a recent study that used mostly the same datasets, the correct classification rate was between 82% and 93% [21] and similar rates have been reported in other machine learning studies of ALL [11,19,28,29,30]. Therefore, the performance of the Allspice tool is within the range of other similar classifiers, which demonstrates the rich biological information available from RNA-seq data and the stability of the predictions across multiple types and implementations of classifiers. If a patient tests positive by Allspice for one of the six distinct subtypes (DUX4, ETV6::RUNX1, KMT2A, PAX5 P80R, TCF3::PBX1 and ZNF384), our findings suggest that the subtype can be validated by deeper exploration of the sequencing reads in the same RNA-seq dataset or by molecular diagnostics for at least 89% and up to 99% of cases depending on the subtype. Both the sequencing and molecular analyses can be time-consuming and inconclusive, whereas mRNA expression levels (i.e., inputs to Allspice) can be reproducibly calculated using highly standardized algorithms. This will shorten the time to diagnosis for the vast majority of ALL cases with the aforementioned recurrent lesions, and significantly shorten the time to delivering care in the clinical setting. Identification of other lesions such as CRLF2, hyperdiploid, hypodiploid, PAX5 alterations and BCR::ABL1 are less definitive, though this may improve with further algorithm training and refinement. There may be clinical utility for the Allspice biomarker panels as RNA risk factors for adverse outcomes in patients that show abnormal karyotypes. A total of 448 patients lacked a definitive genomic subtype under the streamlined taxonomy used in this study. Diagnostics for this patient subpopulation is where we expect mRNA-seq profiling to provide the best added value. In this respect, Allspice is a unique tool since it also provides quantitative RNA biomarkers for the most likely driver lesions and for the deviation from the healthy blood transcriptome. These features are particularly useful when the ALL subtype is difficult to establish due to the absence of identifiable sequence alterations or inconclusive cytogenetic findings. Based on the RNA data, we assigned a transcriptionally compatible ALL subtype to 207 patients (41% of 448) —these are conceptually similar to the Ph-like and ETV6::RUNX1-like gene expression profiles from previous studies. The same concept can be extended to genomic drivers. For example, there were 90 (20%) patients with simultaneous alterations in CRLF2 and P2RY8 and 79 (18%) patients with alterations in both IGH and BCL2 among the undefined subpopulation. Given the sizes of these secondary subgroups, they may be considered as additions or replacements for historical subtypes as genomic ALL datasets to get larger and better phenotyped. The gene expression profiles of 57 (13%) patients could not be matched with any typical ALL subtype, and we suspect many of these were individuals with a low leukemia burden. 3.3. Strengths and Weaknesses The large sample size (statistical power), diversity of data sources, careful mitigation of potential confounding factors and comparisons between three classes of machine learning algorithms make this study strong from a methodological perspective. Notably, the Australian samples were collected from routine health care settings, which provides a realistic spread of sample quality and leukemia burden as encountered in clinical practice. Furthermore, we used additional datasets to help assess the quality of the samples and safety against misclassification, which is an important practical consideration outside research settings [31]. The data were obtained from three Western countries, and caution is warranted if applying the findings in a different ethnic or socioeconomic context. We included rare subtypes in Allspice; however, these predictions should be interpreted with caution since we could not determine classification performance accurately (Figure 4C). Furthermore, we excluded genes that were expressed only in a few samples, some of which may have been highly indicative of a rare subtype. Currently, the driver gene feature of Allspice is based on limited information about the most important lesions observed in an individual, and the results should be interpreted as suggestive rather than definitive regarding causality. Due to the careful analyses and robust performance, we anticipate that the classifier we created captured biological information that reflects the causal mechanisms of ALL and is, therefore, likely to work well for most patient populations. 3.4. Practical Considerations Allspice is open source, easy to install on the popular R programming environment via the Comprehensive R Archive Network and comes with extensive documentation. It accepts raw read count data as produced by standard RNA sequencing pipelines, which is an advantage in clinical settings that may lack a dedicated bioinformatician. The R environment already includes tools for visual clustering of transcriptomes using algorithms such as t-SNE [32], but clustering results can be difficult to interpret for individual cases. Rather than relying on visual proximity in a scatter plot, Allspice uses quantitative probabilistic metrics to indicate the certainty of the predicted subtype. Furthermore, the ability to analyze one sample at a time is important: t-SNE or hierarchical cluster analysis are designed for the research space where large cohorts of labeled samples are readily available, whereas Allspice was designed for a single sample from the beginning. We included two examples with unusual lesions where Allspice helped to assign a (transcriptional) subtype. The first case was an individual with an undefined genomic subtype that Allspice classified as having an RNA profile compatible with an ETV6::RUNX1 fusion. Uncommon ETV6 fusions were discovered by detailed investigations (Supplement Figure S15). Another individual was classified as having a ZNF384-like transcriptional profile, while the exact causal lesion remained uncertain (Supplement Figure S16). These examples highlight how the additional information from Allspice can guide diagnostic efforts for patients with unusual genomic lesions. 4. Materials and Methods 4.1. North American Participants A total of 649 males, 541 females and 89 patients without gender information were from St Jude Children’s Research Hospital (St Jude); Children’s Oncology Group (COG); ECOG-AGRIN Cancer Research Group (ECOG-AGRIN); MD Anderson Cancer Center (MDACC); the Alliance of Clinical Trials in Oncology, Cancer and Leukemia Group B (CALGB) and University of Toronto (Toronto). Detailed clinical information for each case and listings of clinical trial numbers have been previously published [12]. RNA-seq data files were obtained from the European Genome-Phenome Archive (EGAD00001004461 and EGAD00001004463). The patients who participated in this study have provided written informed consent, assent (as appropriate) or parental consent (as appropriate) as part of enrolment protocols for research, including genetic research. All relevant ethical regulations were followed during this study. 4.2. Australian Participants A total of 387 males, 278 females and 102 patients without gender information were investigated through the Australasian Leukaemia and Lymphoma Group (ALLG) National Blood Cancer Registry and the associated Regalia project (ACTRN12612000337875), Australian & New Zealand Children’s Haematology/Oncology Group Acute Lymphoblastic Leukaemia Study 8 (ACTRN12607000302459) and Study 9 (ACTRN12611001233910). All protocols had been approved by relevant human research ethics committees. 4.3. Supporting RNA Data RNA data were sourced from a previously published study of 660 lymphoblast cell lines [33] and from the Genotype-Tissue Expression (GTEx) project release 8 [34]. The GTEx includes RNA-seq data from 948 donors and 54 tissues. In this study, we organized the data into 31 organ groups, of which those that contained at least 500 samples were selected, including whole blood as the most relevant tissue type for ALL. 4.4. RNA Sequencing and Pre-Processing RNA analyses of the North American samples have been described previously [12]. Briefly, RNA-seq was performed using TruSeq library preparation and HiSeq 2000 and 2500 sequencers (Illumina Inc., San Diego, CA, USA). All sequence reads were paired-end and were obtained from total RNA and stranded RNA-seq (75 or 100 base-pair reads) and polyA-selected mRNA (50, 75 or 100 base-pair reads). In Australia, library preparation for mRNA sequencing was performed using either Truseq Stranded mRNA LT Kit (Illumina Inc., San Diego, CA, USA) or Universal Plus mRNA-Seq with NuQuant (Tecan Group Ltd., Männedorf, Switzerland) from total RNA as per the manufacturer’s instructions. Samples were sequenced by either HiSeq 2000 or NextSeq 500 platforms (Illumina Inc., San Diego, CA, USA) producing 75b length paired-end (PE) reads with a median read depth of 65M reads. Raw reads from all cohorts were aligned and mapped to the GRCh37 reference genome with the STAR software version 2.4.2a and above using the two-pass mode [35]. Raw gene counts were generated from BAM files using featureCounts [36]. We defined a gene to be usable as a potential biomarker if it had a read count of ≥ 100 in at least 1% of samples in both the North American and Australian datasets, respectively. In total, 18,923 genes were included in the study. Expression counts were normalized using the DESeq2 algorithm [37], and the normalized counts were transformed using the formula log2(count + 1) before statistical analyses. 4.5. Genomic Subtyping In the text, we use the term ‘genomic ALL subtype’ when the subtype was assigned according to a DNA or RNA sequence abnormality or a clinical biomarker independently of gene expression levels (except for DUX4, PAX5 alterations and CDX2). The detection of genomic alterations and subsequent subtyping of ALL cases were based on the previously published analyses of the North American samples [12] and a preliminary definition of a recently discovered rare CDX2 subtype [38]. We were not able to define the DUX4 subtype independently of mRNA expression levels due to its cryptic nature. PAX5 alterations were defined partly by gene expression data (North America) or by directly observable PAX5 alterations for samples that could not be otherwise classified (Australia). Genomic alterations in the Australian samples were detected as follows. FusionCatcher [39], SOAPfuse [40] and JAFFA [41] were utilized to identify clinically relevant gene fusions. Only fusions reported by multiple fusion calling algorithms were considered, with the exception of rearrangements involving the IGH locus (IGH-DUX4), which were confirmed by high levels of DUX4 expression [42] or by manual inspection of the fusion and accompanying expression data. Single nucleotide variants were identified with GATK-haplotype caller [43] following the best practices workflow. Copy number alterations were detected by multiplex ligation-dependent probe amplification using two SALSA Reference kits (P335 and P202, MRC-Holland, Amsterdam, Netherlands), according to the manufacturer’s instructions. To harmonize subtype definitions between the cohorts and to include only the most confident classifications, we relabeled subtypes for the statistical analyses (Supplement Table S1, Supplement Figure S1). In addition to the pre-defined subtypes, we determined every gene and combinations of genes that harbored an alteration for any individual. The combinations were sourced computationally from the metadata by first collecting all the gene symbols included in the reported rearrangements or mutations. The labels were then collected as a list, and unique combinations were labeled as pseudo-subtypes. For example, an individual with the gene fusion BCR::ABL1 without additional alterations received the label ‘ABL1,BCR’ to indicate the two affected genes. However, an individual with KMT2A::MLLT1 and ID2::IGK fusions would receive the label ‘KMT2A,MLLT1,ID2,IGK’; thus, the gene combinations should not be automatically interpreted as fusions. These mutational profiles were also used for creating genetically matched subsets to allow for more accurate batch adjustments (details below). 4.6. Adjustments for Confounders We divided the data into five batches according to country of origin and age of the patients (Table 1). We then used statistical adjustments to mitigate the potentially confounding associations between data sources and ALL subtypes. Firstly, we used Surrogate Variable Analysis [44] to reduce undesired variation in normalized read counts within each batch, respectively. Next, we used genetically matched subsets and the Numero R package [45] to remove potentially confounding variations between the batches (Supplement Figure S2). Genes that were perfectly aligned with a batch were excluded, which left 18,503 adjusted genes (98%) for statistical analyses. We then calculated correlations between unadjusted and adjusted versions of gene expression and excluded unstable genes that showed a Pearson correlation of R < 0.9 (Supplement Figure S17). A total of 6673 genes were considered stable and included as inputs to classification models. 4.7. Machine Learning We chose a random forest approach [46] as an example of a supervised non-linear machine learning technique that can predict multi-class outcomes from complex input data. We also created Projections to Latent Structures (PLS) for each ALL subtype separately as an example of a linear factorization method [47]. Separate PLS models were fitted to each subtype versus other samples. The third type of modeling was based on neighbor distances: we calculated mean RNA profiles (centroids) for every ALL subtype and classified individuals based on the nearest centroid in the data space. Unsupervised clustering was achieved with the Uniform Manifold Approximation and Projection (UMAP) algorithm [48] to gain qualitative visual insight into the RNA-based segregation of ALL subtypes. Input data were standardized the same way for each model using the default settings of the function ‘numero.prepare()’ in the Numero R package [45]. This is a refined version of the empirical Z-score with protections in place for outliers and skewed distributions. Normalization and standardization parameters were determined for the training dataset and applied independently to the external validation set to simulate a scenario where new unseen data are analyzed one sample at a time and thus must be pre-processed using pre-defined parameters. 4.8. Pruning of Correlated Input Features To improve the performance of UMAP and nearest centroids, the full list of detectable genes was pruned using an approach similar to clumping in genetics [49] and the pruned set of genes was used as input features. First, we calculated Welch’s t-statistics for each gene and ALL subtype and converted them to Z-scores using the cumulative t-distribution and inverse cumulative Normal distribution. Next, we calculated the variance of the Z-scores for each gene as an aggregate measure of how well a gene segregated between ALL subtypes. Genes were then sorted from large to small variance. In the final step, the sorted list was traversed while checking if the next gene was correlated (R ≤ −0.3 or R ≥ 0.3) with any of the already selected. The Australian data were excluded from the pruning procedure to ensure independent external validation. 4.9. Training, Testing, External Validation and Performance Metrics The division of data for the evaluation of classification performance is shown in Figure 1. To determine the optimal model complexity (hyperparameters), North American participants were divided into two randomized subsets that were used as internal training and testing sets. The randomization was completed separately for each subtype to ensure matching subtype frequencies. Next, models were fitted to one of the subsets (internal training set) and classification performance was evaluated in the other (internal testing set). Classification performance in the testing set was used to determine optimal hyperparameter settings (Figure 1E). Full models were trained with the full North American dataset and validated externally in the Australian dataset. We focused on positive predictive value (PPV) as the primary performance metric due to its suitability for the low case frequencies of most ALL subtypes. Negative predictive values, sensitivity, specificity and the area under the receiver operating characteristic curve were also calculated. 4.10. Proximity and Exclusivity The open-source classification tool Allspice is available in the Comprehensive R Archive network (URL: https://cran.r-project.org, 18 April 2022). It includes a visual display of the classification results (Supplement Figure S8, Box SF) and two quality measures to help decide if an RNA profile indicates a specific ALL subtype. The proximity measure is the output value from the probit regression step in the Allspice modeling design (Supplement Figure S8, Box SE) and represents the likelihood of observing the subtype in the training population when balanced for group sizes and given the observed RNA biomarker value and covariates. We chose 50% as the threshold for acceptable proximity. To identify samples with mixed subtype characteristics (designated as ‘Ambiguous’), we defined exclusivity as the difference in the proximity scores between the best and second-best matching subtype centroids. The χ2-distribution with one degree of freedom was used to convert the difference into a probability. We chose 50% as the default threshold for acceptable exclusivity. 5. Conclusions We observed strong associations between genomic alterations and lymphoblast transcriptome-wide expression profiles in pediatric and adult patients of B-cell ALL. For one-third of patients, these associations are unambiguous and provide diagnostic information that is often quicker and easier to obtain compared to fusion callers or cytogenetic tests. For the rest of the patients, gene expression analyses may provide insight that is not available from other methods. An RNA-based ALL biomarker can inform sequence analysts where to look for lesions manually if automatic fusion callers failed or read counts were too low for statistical certainty. In both scenarios, Allspice can help oncologists determine the most likely causal drivers with greater confidence and identify potential therapeutic targets in a shorter time frame. Acknowledgments We are grateful for the contributions of Australian leukemia patients and health care professionals for providing us with the original biological material, including the Australian and New Zealand Children’s Haematology/Oncology Group. We acknowledge the assistance of the staff of the ALL laboratory SAHMRI for preparing leukemia samples, analysing and reporting mRNA-seq results and the South Australian Genomics Centre for generating RNA-seq data. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ijms23094574/s1. Click here for additional data file. Author Contributions V.-P.M. developed the machine learning models and analyzed RNA read counts; J.R. determined genomic ALL subtypes; J.R. and J.B. obtained and processed raw mRNA-seq data; D.Y. and D.L.W. contributed to the study design. All authors have read and agreed to the published version of the manuscript. Funding D.L.W. was funded by the National Health and Medical Research Council of Australia Target Call for Research (APP1160833) and by Cancer Council SA Beat Cancer Project Principal Cancer Research Fellowship (PRF1618). The work was also supported by the Australasian Leukaemia and Lymphoma Group and Australian and New Zealand Children’s Haematology/Oncology Group. Institutional Review Board Statement Australian ethics approvals were obtained as part of the Australasian Leukaemia and Lymphoma Group (ALLG) National Blood Cancer Registry and the associated Regalia project (ACTRN12612000337875), Australian & New Zealand Children’s Haematology/Oncology Group Acute Lymphoblastic Leukaemia Study 8 (ACTRN12607000302459) and Study 9 (ACTRN12611001233910). Informed Consent Statement North American patients who participated in this study provided written informed consent, assent (as appropriate) or parental consent (as appropriate) as part of enrolment protocols for research, including genetic research [12]. Individual consent for publishing the RNA profiles of the Australian patient CHI_0809 was obtained by the Children’s Cancer Institute and patient CHII_0804 by the Australian Genomics Health Alliance. Data Availability Statement The Allspice software and source code are available via the Comprehensive R Archive Network (URL: https://cran.r-project.org, 18 April 2022). The North American mRNA-seq data files are available in the European Genome-Phenome Archive (EGAD00001004461 and EGAD00001004463). The Australian dataset is available from the authors upon request. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Study design. (A,B) RNA-seq pre-processing was applied separately for North American and Australian datasets. (C) Non-biological differences due to technical artifacts and cohort effects were adjusted according to genetically matched subsets (details in Methods). (D) Correlation coefficients were calculated between unadjusted and adjusted expression levels to exclude genes that were heavily influenced by confounders. A total of 6673 stable genes with R ≥ 0.9 were included for subtype modeling. (E) Internal training and testing sets were randomly chosen from the North American participants as initial material to evaluate machine learning models and to determine hyperparameters. The randomization was conducted separately for each subtype to ensure matching subtype frequencies, which explains the small difference in set sizes. (F) Final machine learning models were trained with the full North American dataset and validated in the Australian dataset. Figure 2 Impact of confounder adjustments. Pearson correlation coefficients were calculated between 18,503 log-transformed genes, and a technical or clinical variable before and after gene expression levels were adjusted as described in Methods. A wide histogram indicates substantial covariation across the transcriptome, while a narrow histogram indicates successful removal of covariance. (A,D) The North American cohorts included 204 (16%) RNA samples that were sequenced with an unstranded library. (B,E) Strandedness and other technical differences manifested as substantial covariation between the transcriptome and the continent of origin. (C,F) Covariation between gene expression and ALL subtypes, such as ETV6::RUNX1, was preserved. Figure 3 Transcriptional landscape of the most frequent ALL subtypes. Clustering structure in the North American data was modeled by the Uniform Manifold Approximation and Projection (UMAP) algorithm and using the same genes that were prioritized and pre-processed for the centroid classifier. (A) Standardized but unadjusted gene expression values were used for re-projecting the North American samples onto the UMAP layout. We use the unadjusted expression profiles here since, in practical settings where patients arrive one-by-one, adjustments for batch effects that would be available in research settings cannot be made. (B) Unadjusted Australian data were projected onto the same UMAP layout as an independent external validation set. Figure 4 Comparison of three machine learning algorithms. Each model was fit to the batch-corrected North American data (training set n = 1078) and then evaluated in unadjusted Australian data (external validation set n = 520). Samples with undefined genetic subtypes were excluded from the analyses. (A–C) The forest plots show 95% confidence intervals that reflect the statistical uncertainty due to finite category sizes. The percentages written in the plots indicate the prevalence of the genetic subtype in the Australian dataset. (D) Description of subtype supergroups. Figure 5 Example of a report card from the Allspice classifier for an adult male patient from North America with Philadelphia (BCR::ABL1) genetic B-cell ALL subtype. (A) Sample identifier and predicted subtype based on RNA data in top-left corner. (B) The report shows the frequency of the corresponding genetic subtype in the training data, given the observed gene expression profile. In this case, there was a 79% chance that sequence and cytogenetics analyses would confirm the presence of the Philadelphia chromosome. (C) Visualization of how similar the sample is to each ALL subtype profile. The display can be interpreted as a panel of RNA “biomarkers” that are specific to each subtype. In this case, the high value for Ph indicates that the gene expression profile is compatible with a typical patient with the Philadelphia chromosome. (D) Allspice also indicates how similar the sample is to the RNA profiles associated with the presence of genetic alterations in one or more genes in parallel. In this case, the gene expression profile matches the typical profile of patients with independently verified BCR::ABL1 fusion (i.e., both BCR and ABL1 are altered). (E) Samples with high leukemia burden will typically produce a strong B-cell ALL signal in the tissue panel. (F) Minimal example of how to generate the report in the R programming environment. Figure 6 Classification performance of the Allspice centroid classifier. (A) The bars show positive predictive values for all samples including those with undefined genetic subtypes. They represent conservative estimates on how likely it is that the subtype predicted by RNA-seq expression levels can be confirmed as a specific sequence alteration or is also indicated by cytogenetics. (B) Proportions of genetic subtypes in the dataset. ijms-23-04574-t001_Table 1 Table 1 Patient characteristics and frequencies (%) of ALL subtypes according to known genetic alterations. The participants were grouped primarily by the recruiting institute and secondarily by age (> 99% of participants in the pediatric cohorts were below 20 years of age). Mean and standard deviation are shown for age. p-values for cohort differences were calculated by the χ2-test. Pediatric Cohorts Adult and General Cohorts COG St Jude Australia p-Value Multiple * Australia p-Value Male 177 221 83 4.4 × 10−6 251 304 0.021 Female 120 188 73 0.016 233 205 0.013 Unknown 27 52 89 1.8 × 10−21 10 13 0.77 Age (years) 9.2 ± 5.7 6.4 ± 4.4 7.4 ± 4.0 7.6 × 10−11 45.1 ± 14.8 35.7 ± 22.6 3.6 × 10−13 BCL2/MYC 0.0 0.4 0.4 0.50 2.8 0.8 0.024 CDX2 hi-exp 0.3 0.0 0.0 0.34 0.6 0.6 1.0 CRLF2 7.4 3.0 18.4 8.4 × 10−12 13.2 8.4 0.020 DUX4 7.4 7.4 2.0 0.0096 4.9 6.9 0.21 ETV6::RUNX1 9.9 26.7 17.6 2.3 × 10−8 1.0 2.3 0.18 HLF 0.0 0.4 0.4 0.50 0.6 0.6 1.0 Hyperdiploid 21.0 21.3 9.8 0.00034 4.0 5.6 0.27 Hypodiploid 1.2 0.7 0.8 0.68 11.5 3.6 3.1 × 10−6 iAMP21 6.5 0.4 0.8 0.72 × 10−8 0.2 0.4 1.0 IKZF1 N159Y 0.3 0.0 1.2 0.043 0.4 0.6 1.0 KMT2A 2.2 10.8 2.9 2.0 × 10−7 13.8 4.4 3.2 × 10−7 MEF2D 2.2 1.1 0.0 0.058 2.0 1.1 0.39 NUTM1 0.3 1.1 0.4 0.36 0.0 0.2 1.0 PAX5 Alt 4.0 5.9 3.7 0.32 8.3 1.9 6.4 × 10−6 PAX5 P80R 0.9 1.1 0.8 0.94 3.4 2.9 0.74 BCR::ABL1 5.6 2.4 4.1 0.070 12.3 21.5 1.6 × 10−4 TCF3::PBX1 2.2 6.7 1.2 0.00023 3.0 2.3 0.59 ZNF384 1.9 2.0 2.0 0.99 2.8 4.4 0.24 Undefined 26.9 8.7 33.5 1.7 × 10−16 15.0 31.6 7.0 × 10−9 * ECOG-ACRIN, Toronto, MDACC and CALGB. ijms-23-04574-t002_Table 2 Table 2 Positive predictive values for correct classification into genetically defined B-cell ALL subtypes. The values are presented as the percentages of samples for which the best matching RNA centroid was the same as the genetic subtype if defined. Quality control was set at ≥ 50% proximity and ≥ 50% exclusivity (details in Methods). All Samples Defined Subtypes Defined Subtypes and QC Pass Number of samples 2046 1598 1292 DUX4 (%) 96.8 100.0 100.0 ETV6::RUNX1 (%) 86.6 97.7 98.6 KMT2A (%) 91.1 96.6 100.0 PAX5 P80R (%) 95.2 97.6 100.0 TCF3::PBX1 (%) 98.5 98.5 100.0 ZNF384 (%) 89.1 96.6 100.0 CRLF2 (%) 62.0 91.2 97.6 Hyperdiploid (%) 69.8 92.1 99.0 Hypodiploid (%) 64.2 79.0 89.2 PAX5 Alt (%) 52.1 74.8 92.7 BCR::ABL1 (%) 67.2 84.8 96.6 BCL2/MYC (%) 17.5 54.1 66.7 CDX2 hi-exp (%) 38.9 77.8 100.0 HLF (%) 64.3 75.0 90.0 iAMP21 (%) 27.9 66.7 100.0 IKZF1 N159Y (%) 64.3 90.0 100.0 MEF2D (%) 72.2 78.8 100.0 NUTM1 (%) 40.0 66.7 100.0 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Terwilliger T. Abdul-Hay M. Acute lymphoblastic leukemia: A comprehensive review and 2017 update Blood Cancer J. 2017 7 e577 10.1038/bcj.2017.53 28665419 2. Hunger S.P. Mullighan C.G. Acute Lymphoblastic Leukemia in Children N. Engl. J. Med. 2015 373 1541 1552 10.1056/NEJMra1400972 26465987 3. Yi M. Zhou L. Li A. Luo S. Wu K. Global burden and trend of acute lymphoblastic leukemia from 1990 to 2017 Aging 2020 12 22869 22891 10.18632/aging.103982 33203796 4. Cario G. Leoni V. Conter V. Attarbaschi A. Zaliova M. Sramkova L. Cazzaniga G. Fazio G. Sutton R. Elitzur S. Relapses and treatment-related events contributed equally to poor prognosis in children with ABL-class fusion positive B-cell acute lymphoblastic leukemia treated according to AIEOP-BFM protocols Haematologica 2020 105 1887 1894 10.3324/haematol.2019.231720 31601692 5. Passet M. Boissel N. Sigaux F. Saillard C. Bargetzi M. Ba I. Thomas X. Graux C. Chalandon Y. Leguay T. PAX5 P80R mutation identifies a novel subtype of B-cell precursor acute lymphoblastic leukemia with favorable outcome Blood 2019 133 280 284 10.1182/blood-2018-10-882142 30510083 6. Samra B. Jabbour E. Ravandi F. Kantarjian H. Short N.J. Evolving therapy of adult acute lymphoblastic leukemia: State-of-the-art treatment and future directions J. Hematol. Oncol. 2020 13 70 10.1186/s13045-020-00905-2 32503572 7. Karol S.E. Pui C.-H. Personalized therapy in pediatric high-risk B-cell acute lymphoblastic leukemia Ther. Adv. Hematol. 2020 11 204062072092757 10.1177/2040620720927575 8. Moorman A.V. Chilton L. Wilkinson J. Ensor H.M. Bown N. Proctor S.J. A population-based cytogenetic study of adults with acute lymphoblastic leukemia Blood 2010 115 206 214 10.1182/blood-2009-07-232124 19897583 9. Chiaretti S. Zini G. Bassan R. Diagnosis and subclassification of acute lymphoblastic leukemia Mediterr. J. Hematol. Infect. Dis. 2014 6 e2014073 10.4084/mjhid.2014.073 25408859 10. Arber D.A. Orazi A. Hasserjian R. Thiele J. Borowitz M.J. Le Beau M.M. Bloomfield C.D. Cazzola M. Vardiman J.W. The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia Blood 2016 127 2391 2405 10.1182/blood-2016-03-643544 27069254 11. Brown L.M. Lonsdale A. Zhu A. Davidson N.M. Schmidt B. Hawkins A. Wallach E. Martin M. Mechinaud F.M. Khaw S.L. The application of RNA sequencing for the diagnosis and genomic classification of pediatric acute lymphoblastic leukemia Blood Adv. 2020 4 930 942 10.1182/bloodadvances.2019001008 32150610 12. Gu Z. Churchman M.L. Roberts K.G. Moore I. Zhou X. Nakitandwe J. Hagiwara K. Pelletier S. Gingras S. Berns H. PAX5-driven subtypes of B-progenitor acute lymphoblastic leukemia Nat. Genet. 2019 51 296 307 10.1038/s41588-018-0315-5 30643249 13. Sánchez R. Ribera J. Morgades M. Ayala R. Onecha E. Ruiz-Heredia Y. Juárez-Rufián A. de Nicolás R. Sánchez-Pina J. Vives S. A novel targeted RNA-Seq panel identifies a subset of adult patients with acute lymphoblastic leukemia with BCR-ABL1-like characteristics Blood Cancer J. 2020 10 43 10.1038/s41408-020-0308-3 32332702 14. Marincevic-Zuniga Y. Dahlberg J. Nilsson S. Raine A. Nystedt S. Lindqvist C.M. Berglund E.C. Abrahamsson J. Cavelier L. Forestier E. Transcriptome sequencing in pediatric acute lymphoblastic leukemia identifies fusion genes associated with distinct DNA methylation profiles J. Hematol. Oncol. 2017 10 148 10.1186/s13045-017-0515-y 28806978 15. Li J.-F. Dai Y.-T. Lilljebjörn H. Shen S.-H. Cui B.-W. Bai L. Liu Y.-F. Qian M.-X. Kubota Y. Kiyoi H. Transcriptional landscape of B cell precursor acute lymphoblastic leukemia based on an international study of 1223 cases Proc. Natl. Acad. Sci. USA 2018 115 E11711 E11720 10.1073/pnas.1814397115 30487223 16. Braoudaki M. Tzortzatou-Stathopoulou F. Clinical Cytogenetics in Pediatric Acute Leukemia: An Update Clin. Lymphoma Myeloma Leuk. 2012 12 230 237 10.1016/j.clml.2012.04.004 22609262 17. Rehn J.A. O’Connor M.J. White D.L. Yeung D.T. DUX Hunting-Clinical Features and Diagnostic Challenges Associated with DUX4-Rearranged Leukaemia Cancers 2020 12 2815 10.3390/cancers12102815 18. Lee J. Cho S. Hong S.-E. Kang D. Choi H. Lee J.-M. Yoon J.-H. Cho B.-S. Lee S. Kim H.-J. Integrative Analysis of Gene Expression Data by RNA Sequencing for Differential Diagnosis of Acute Leukemia: Potential Application of Machine Learning Front. Oncol. 2021 11 717616 10.3389/fonc.2021.717616 34497767 19. Walter W. Shahswar R. Stengel A. Meggendorfer M. Kern W. Haferlach T. Haferlach C. Clinical application of whole transcriptome sequencing for the classification of patients with acute lymphoblastic leukemia BMC Cancer 2021 21 886 10.1186/s12885-021-08635-5 34340673 20. Pui C.-H. Precision medicine in acute lymphoblastic leukemia Front. Med. 2020 14 689 700 10.1007/s11684-020-0759-8 33074527 21. Schmidt B. Brown L.M. Ryland G.L. Lonsdale A. Kosasih H.J. Ludlow L.E. Majewski I.J. Blombery P. Ekert P.G. Davidson N.M. ALLSorts: A RNA-Seq Classifier for B-Cell Acute Lymphoblastic Leukemia Bioinformatics Bethesda, MD, USA 2021 22. Gianfrancesco M.A. Tamang S. Yazdany J. Schmajuk G. Potential Biases in Machine Learning Algorithms Using Electronic Health Record Data JAMA Intern. Med. 2018 178 1544 10.1001/jamainternmed.2018.3763 30128552 23. Ferrari E. Retico A. Bacciu D. Measuring the effects of confounders in medical supervised classification problems: The Confounding Index (CI) Artif. Intell. Med. 2020 103 101804 10.1016/j.artmed.2020.101804 32143800 24. Skelly A.C. Dettori J.R. Brodt E.D. Assessing bias: The importance of considering confounding Evid.-Based Spine-Care J. 2012 3 9 12 10.1055/s-0031-1298595 23236300 25. Boer J.M. Marchante J.R.M. Evans W.E. Horstmann M.A. Escherich G. Pieters R. Den Boer M.L. BCR-ABL1-like cases in pediatric acute lymphoblastic leukemia: A comparison between DCOG/Erasmus MC and COG/St. Jude signatures Haematologica 2015 100 e354 e357 10.3324/haematol.2015.124941 26045294 26. Roberts K.G. The biology of Philadelphia chromosome-like ALL Best Pract. Res. Clin. Haematol. 2017 30 212 221 10.1016/j.beha.2017.07.003 29050694 27. Harvey R.C. Kang H. Roberts K.G. Chen I.-M.L. Atlas S.R. Bedrick E.J. Gastier-Foster J.M. Zhang J. Gerhard D.S. Smith M.A. Development and Validation Of a Highly Sensitive and Specific Gene Expression Classifier To Prospectively Screen and Identify B-Precursor Acute Lymphoblastic Leukemia (ALL) Patients With a Philadelphia Chromosome-Like (“Ph-like” or “BCR-ABL1-Like”) Signature For Therapeutic Targeting and Clinical Intervention Blood 2013 122 826 10.1182/blood.V122.21.826.826 28. Kashef A. Khatibi T. Mehrvar A. Treatment outcome classification of pediatric Acute Lymphoblastic Leukemia patients with clinical and medical data using machine learning: A case study at MAHAK hospital Inform. Med. Unlocked 2020 20 100399 10.1016/j.imu.2020.100399 29. Pan L. Liu G. Lin F. Zhong S. Xia H. Sun X. Liang H. Machine learning applications for prediction of relapse in childhood acute lymphoblastic leukemia Sci. Rep. 2017 7 7402 10.1038/s41598-017-07408-0 28784991 30. Sun J. Wang L. Liu Q. Tárnok A. Su X. Deep learning-based light scattering microfluidic cytometry for label-free acute lymphocytic leukemia classification Biomed. Opt. Express 2020 11 6674 10.1364/BOE.405557 33282516 31. Reddy S. Rogers W. Makinen V.-P. Coiera E. Brown P. Wenzel M. Weicken E. Ansari S. Mathur P. Casey A. Evaluation framework to guide implementation of AI systems into healthcare settings BMJ Health Care Inform. 2021 28 e100444 10.1136/bmjhci-2021-100444 32. Van der Maaten L. Hinton G. Viualizing data using t-SNE J. Mach. Learn. Res. 2008 9 2579 2605 33. The Geuvadis Consortium Lappalainen T. Sammeth M. Friedländer M.R. Hoen P.A.C. Monlong J. Rivas M.A. Gonzàlez-Porta M. Kurbatova N. Griebel T. Transcriptome and genome sequencing uncovers functional variation in humans Nature 2013 501 506 511 10.1038/nature12531 24037378 34. Lonsdale J. Thomas J. Salvatore M. Phillips R. Lo E. Shad S. Hasz R. Walters G. Garcia F. Young N. The Genotype-Tissue Expression (GTEx) project Nat. Genet. 2013 45 580 585 10.1038/ng.2653 23715323 35. Dobin A. Gingeras T.R. Mapping RNA-seq Reads with STAR Curr. Protoc. Bioinform. 2015 51 11.14.1 11.14.19 10.1002/0471250953.bi1114s51 36. Liao Y. Smyth G.K. Shi W. featureCounts: An efficient general purpose program for assigning sequence reads to genomic features Bioinforma. Oxf. Engl. 2014 30 923 930 10.1093/bioinformatics/btt656 37. Love M.I. Huber W. Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2 Genome Biol. 2014 15 550 10.1186/s13059-014-0550-8 25516281 38. Yasuda T. Sanada M. Kawazu M. Kojima S. Tsuzuki S. Ueno H. Iwamoto E. Iijima-Yamashita Y. Yamada T. Kanamori T. Two novel high-risk adult B-cell acute lymphoblastic leukemia subtypes with high expression of CDX2 and IDH1/2 mutations Blood 2021 139 1850 1862 10.1182/blood.2021011921 39. Nicorici D. Satalan M. Edgren H. Kangaspeska S. Murumagi A. Kallioniemi O. Virtanen S. Kilkku O. FusionCatcher—A Tool for Finding Somatic Fusion Genes in Paired-End RNA-Sequencing Data Bioinformatics Bethesda, MD, USA 2014 40. Jia W. Qiu K. He M. Song P. Zhou Q. Zhou F. Yu Y. Zhu D. Nickerson M.L. Wan S. SOAPfuse: An algorithm for identifying fusion transcripts from paired-end RNA-Seq data Genome Biol. 2013 14 R12 10.1186/gb-2013-14-2-r12 23409703 41. Davidson N.M. Majewski I.J. Oshlack A. JAFFA: High sensitivity transcriptome-focused fusion gene detection Genome Med. 2015 7 43 10.1186/s13073-015-0167-x 26019724 42. Patro R. Duggal G. Love M.I. Irizarry R.A. Kingsford C. Salmon provides fast and bias-aware quantification of transcript expression Nat. Methods 2017 14 417 419 10.1038/nmeth.4197 28263959 43. Poplin R. Ruano-Rubio V. DePristo M.A. Fennell T.J. Carneiro M.O. Van der Auwera G.A. Kling D.E. Gauthier L.D. Levy-Moonshine A. Roazen D. Scaling Accurate Genetic Variant Discovery to Tens of Thousands of Samples Genomics Amsterdam, The Netherlands 2017 44. Leek J.T. Storey J.D. Capturing Heterogeneity in Gene Expression Studies by Surrogate Variable Analysis PLoS Genet. 2007 3 e161 10.1371/journal.pgen.0030161 17907809 45. Gao S. Mutter S. Casey A. Mäkinen V.-P. Numero: A statistical framework to define multivariable subgroups in complex population-based datasets Int. J. Epidemiol. 2019 48 369 374 10.1093/ije/dyy113 29947762 46. Breiman L. Random forests Mach. Learn. 2001 45 5 32 10.1023/A:1010933404324 47. Trygg J. Wold S. Orthogonal projections to latent structures (O-PLS) J. Chemom. 2002 16 119 128 10.1002/cem.695 48. McInnes L. Healy J. Saul N. Großberger L. UMAP: Uniform Manifold Approximation and Projection J. Open Source Softw. 2018 3 861 10.21105/joss.00861 49. Mäkinen V.-P. Civelek M. Meng Q. Zhang B. Zhu J. Levian C. Huan T. Segrè A.V. Ghosh S. Vivar J. Integrative Genomics Reveals Novel Molecular Pathways and Gene Networks for Coronary Artery Disease PLoS Genet. 2014 10 e1004502 10.1371/journal.pgen.1004502 25033284

PMC009xxxxxx/PMC9099613.txt

==== Front Nanomaterials (Basel) Nanomaterials (Basel) nanomaterials Nanomaterials 2079-4991 MDPI 10.3390/nano12091491 nanomaterials-12-01491 Review Progress in Transparent Nano-Ceramics and Their Potential Applications Ming Wuyi 1 Jiang Zhiwen 1 Luo Guofu 1* Xu Yingjie 1 He Wenbin 1 Xie Zhuobin 2 Shen Dili 3 Li Liwei 1 Balázsi Csaba Academic Editor 1 Henan Key Lab of Intelligent Manufacturing of Mechanical Equipment, Zhengzhou University of Light Industry, Zhengzhou 450002, China; mingwuyi@gmail.com (W.M.); jzw336699@163.com (Z.J.); xyj15203873137@163.com (Y.X.); hwb@zzuli.edu.cn (W.H.); liliwei@zzuli.edu.cn (L.L.) 2 Guangdong HUST Industrial Technology Research Institute, Guangdong Provincial Key Laboratory of Digital Manufacturing Equipment, Dongguan 523808, China; xzb18336947581@163.com 3 School of Mechanical-Electronic and Automobile Engineering, Zhengzhou Institute of Technology, Zhengzhou 450052, China; shendili@163.com * Correspondence: luoguofu@zzuli.edu.cn 27 4 2022 5 2022 12 9 149101 4 2022 25 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Transparent nano-ceramics have an important high-transmittance, material-integrating structure and function and a variety of potential applications, such as use in infrared windows, optical isolators, composite armors, intelligent terminal screens, and key materials of solid-state lasers. Transparent ceramics were originally developed to replace single crystals because of their low fabricating cost, controllable shape, and variable composition. Therefore, this study reviews and summarizes the development trends in transparent nano-ceramics and their potential applications. First, we review the research progress and application of laser nano-ceramic materials, focusing on the influence of controllable doping of rare earth ions on thermal conductivity and the realization of large-scale fabrication technology. Second, the latest research progress on magneto-optical transparent nano-ceramics, mainly including terbium gallium garnet (Tb3Ga5O12, TGG) ceramics and terbium aluminum garnet (Tb3Al5O12, TAG) ceramics, are summarized, and their performance is compared. Third, the research progress of transparent armor nano-ceramic materials, represented by MgAl2O3 and Aluminum oxynitride (AlON), are reviewed. Lastly, the progress in electro-optical transparent nano-ceramics and scintillation transparent nano-ceramics is reported, and the influence of the material-fabrication process on electro-optic effect or luminous intensity is compared. Moreover, the effect of particle diameter on fabrication, the relationship between nano powder and performance, and different sintering methods are discussed. In summary, this study provides a meaningful reference for low-cost and sustainable production in the future. transparent nano-ceramics nano powder microstructure optical transmittance IR transmittance magneto-optical material mechanical strength preparation method Local Innovative and Research Teams Project of Guangdong Pearl River Talents Program2017BT01G167 Science and Technology Research Project of Henan Province222102220011 Key Scientific Research Project for Henan Province Higher school of China21A460034 This research was funded by the Local Innovative and Research Teams Project of Guangdong Pearl River Talents Program (2017BT01G167), the Science and Technology Research Project of Henan Province (222102220011), and the Key Scientific Research Project for Henan Province Higher school of China (21A460034). ==== Body pmc1. Introduction Transparent materials are essential for human beings. Advanced transparent materials have gradually replaced traditional materials, namely glass, for relevant technical applications where conventional materials are not applicable. Traditional transparent materials include glass, polymer, and alkali metal hydride, which have weak mechanical strength and unstable chemical properties [1]. Optical transparency may also be seen in single crystals of several inorganic materials. They are more stable and robust than typical materials. Single crystal growth, on the other hand, is frequently referred to as “growth” since it is primarily governed by rather slow thermodynamic processes. In addition, the single crystal with large growth is often challenging, especially for oxide materials with a high melting point, and the single crystal during growth cannot be used directly [2]. With the rapid development of nanomaterial technology, the preparation of high-purity nano powder can be realized [3,4,5]. The optimization of raw material powder promotes the densification process of transparent nano-ceramics, reduces the impurity content, and is a great breakthrough in the optical quality and various properties of ceramics. Because ceramic is an inorganic polycrystalline material, it generally does not have transparency. This is because there are a lot of defects, such as pores and impurities in the ceramic, which cause scattering and refraction loss to the light incident into the ceramic so that the incident light cannot pass through the ceramic. Figure 1 demonstrates the schematic light transmission phenomena in a polycrystalline ceramic, which has more light scattering sources than a single crystal, such as grain boundaries, pores, impurities, and birefringence [6]. Initially, the incident intensity (I0) of light depends on the surface roughness and is composed of diffuse reflection (RD) and diffuse transmission (IDT). In addition, a certain amount of light is reflected on each surface of the material through specular reflection (RT). Light passing through polycrystalline materials can also interact with pores (residual pore scattering), grain boundaries (second phase scattering), impurities (inclusion absorption or scattering), and birefringence (birefringence of the non-cubic phase) [6]. The main factors affecting the optical quality of ceramics are light absorption, scattering of pores and the second phase, grain boundary reflection and refraction (birefringence), grain boundary scattering, and light scattering caused by surface roughness. Based on the above factors, the transmittance of ceramics can be calculated by Equation (1):(1) II0=(1−R)2e−mx where I0 is the incident intensity of light, I is the transmitted intensity of light, m is the light absorption coefficient, x is the sample thickness, and R is the light reflectivity [7]. The light absorption coefficient m can be given by Equation (2), (2) m=α+Sim+Sop where α is the absorption coefficient of electron transition, Sim is the scattering caused by structural heterogeneity (e.g., pores, second phase), and Sop is the scattering caused by optical anisotropy (e.g., hexagonal system and other non-cubic systems). According to Equations (1) and (2), the necessary conditions for preparing high-quality transparent nano-ceramics are [8]:(1) High relative density and residual porosity less than 0.01%; (2) No optical anisotropy; (3) The grain boundary is free of impurities and second phase; (4) The selective absorption of the crystal to the incident light is small; (5) The surface is flat and has low roughness. Therefore, polycrystalline transparent nano-ceramics with excellent optical properties can be fabricated by using high-purity raw material powder [9,10], especially the application of nano powder, reducing impurities and eliminating pores through certain processes, and reducing scattering caused by surface roughness through polishing and other means. Compared with common optical function transparent single crystal or glass materials, transparent nano-ceramic materials have obvious advantages in their optical function effect, mechanical and thermal properties, low cost, and large size (compared with single crystal materials). At the same time, being ceramic materials, they have high strength and hardness, high-temperature resistance, and corrosion resistance, which are better than general optical materials [11]. With the improvement of the performance of transparent nano-ceramics, therefore, their functional applications have also been further developed. Research in many fields is gradually becoming commercialized, and the market potential is larger in the future. For example, transparent nano-ceramics have been used in laser gain media, transparent armor, aerospace windows, solid-state lighting, magneto-optic material, and other fields [2,11], as shown in Figure 2. In this study, published literature was selected from databases such as Web of Science, Google Scholar, CNKI, and the Engineering Index, as well as publishers’ databases, such as Elsevier, IEEE Xplore, and Springer. A total of nearly 130 related documents were collected, and 100 documents are reviewed and cited in this study. This literature covers the applications of transparent nano-ceramics. Among them, Web of Science, IEEE Xplore, and other academic databases are rich in literature with a wide range of research. There are also many excellent Chinese journals and dissertations in the research results included by CNKI, so we also chose a small number of classic Chinese journals. Nano powder, microstructure, and preparation methods are all the key points that need to be considered in the study of transparent nano-ceramics. Therefore, when searching literature, keywords related to transparent nano-ceramics (such as “nano-ceramics”, “nano powder”, “microstructure”, and “preparation method”) and keywords related to material properties (such as “optical transmittance”, “IR transmittance”, “magneto-optical material”, and “mechanical strength”) were used. Categories and application areas of nanomaterials (e.g., “magneto-optical transparent nano-ceramics”, “armor transparent nano-ceramics”, “electro-optical transparent nano-ceramics”, and “scintillation transparent nano-ceramics”) were also indexed. Then, this review summarizes the different applications of transparent nano-ceramics, mainly including laser transparent nano-ceramics, magneto-optical transparent nano-ceramics, transparent armor nano-ceramics, electro-optical transparent nano-ceramics, and scintillation transparent nano-ceramics. Section 2 reviews the research progress of transparent nano-ceramics in solid-state laser applications, focusing on the influence of the controllable doping of rare earth ions on their performance and the realization of large-scale fabrication technology. In Section 3, we summarize the latest research progress of magneto-optical transparent nano-ceramics, mainly including terbium gallium garnet (Tb3Ga5O12, TGG) transparent nano-ceramics and terbium aluminum garnet (Tb3Al5O12, TAG) transparent nano-ceramics, and we compare their properties and predict their future potential applications. Section 4 mainly reviews the research progress of transparent armor nano-ceramic materials, represented by MgAl2O3 and AlON. The potential applications of transparent armor nano-ceramic are aerospace windows, missile fairings, bulletproof windows, and other fields, which require high hardness and strength, good wear resistance, and impact resistance. In Section 5, we report the progress in electro-optical transparent nano-ceramics and scintillation transparent nano-ceramics. The influence of the material-fabrication process on the electro-optic effect or luminous intensity is compared. In Section 6, the advantages and disadvantages of the fabrication method of transparent nanoceramic material are discussed. In summary, this study reviews the preparation and applications of various transparent nano-ceramics in recent years, which provides a reference for low-cost and sustainable production in the future. 2. Transparent Nano-Ceramics for Solid-State LASERs 2.1. Brief Introduction Solid-state lasers play a leading role in the field of laser application because they have the advantages of high peak power, high efficiency, long service life, safety, and reliability. In solid-state lasers, the gain medium has the most significant influence on the laser’s output performance. Doped yttrium aluminum garnet (YAG) has the advantages of high thermal conductivity, high melting point, stable chemical properties, high mechanical strength, and high creep resistance. Compared with a single crystal and glass, the main advantages of YAG transparent nano-ceramics are: (1) the process of high concentration doping is simple, which can easily improve its properties; (2) it is easy to prepare ceramics with a large size and complex shape; (3) the preparation cost is low and the cycle is short; (4) it is convenient to realize special structures and functions. Therefore, it is the focus of research in the academic community, and many countries have invested a lot of human and material resources. Nowadays, YAG transparent nano-ceramics are widely used in the gain medium of solid-state lasers. According to the literature, the performance of transparent nano-ceramics is related to the controllable doping of rare earth ions and the size of nano-powder, which are critically reviewed in this section to elaborate on the progress of transparent nano-ceramics for solid-state lasers. 2.2. Doped YAG Transparent Nano-Ceramics Nd3+-doped. In 1995, Nd:YAG transparent ceramics for solid-state lasers with continuous wave (CW) laser emissions were first reported by Ikesue and Kinoshita [12], and the nano powers of Y2O3, Al2O3, and Nd2O3 were used as starting materials with the average particle diameters of 60, 400, and 500 nm, respectively. The optical scattering loss of Nd:YAG was about 0.9%/cm. The experiment demonstrated that the performance of solid-state lasers could be obtained with an oscillation threshold of 309 mW and a slope efficiency of 28%, respectively. In 2002, Lu et al. [13] developed Y3Al5O12 optical ceramic materials based on highly transparent nanocrystalline YAG. The pore volume concentration of YAG transparent ceramics was 1 ppm, and the average diameter of particles was about 10 μm. The grain boundary width was only about 1 nm. The results showed that in the preliminary comparative laser experiment of Nd:YAG ceramic and single-crystal rods, the output power of 88 and 99 W were obtained, respectively. This means that it could be used in high-power, solid-state lasers. Compared with single-crystal Nd:YAG, the light-to-light efficiency of Nd:YAG transparent ceramics with nanocrystalline needs to be further improved. However, it will become a good substitute for the widely used Nd: YAG single crystal, due to its low manufacturing cost, for different types of solid-state lasers. In 2010, Suárez et al. [14] first obtained 1 at.% Nd:YAG nano powder with an average particle size of 100 nm by using a reverse-strike precipitation method. Then, the Nd:YAG transparent nano-ceramics were prepared by the hot isostatic pressing (HIP) method. They found that the optical properties were significantly different with different sintering and HIP parameters. The infrared transmittance of the fabricated sample was 80%, and its emission spectrum was the same as a 1 at.% Nd:YAG single crystal. In 2011, Stevenson et al. [15] sintered Nd:YAG transparent ceramics at 1600 °C with B2O3 and SiO2 double-sintering additives. They adopted the solid-state reaction method to prepare the Nd:YAG transparent ceramics and α-Al2O3 (>99.99%, 100–300 nm), Y2O3 (>99.999%, 50 nm), and used Nd2O3 (>99.99%, 200 nm) nano powders as the starting materials. Additionally, the B3+: Si4+ atomic ratio ranged from 0.5 to 2 while keeping the total doping level at 1.35 mol%. The results demonstrated that the relative density of the samples exceeded 99.9% and the transmittance in the visible band was as high as 84%. They also found that densification could be completed at about 100 °C lower than the normal sintering temperature since B3+ greatly improved the driving force of densification. In 2014, Yavetskiy et al. [16] also utilized the solid-state reaction method to fabricate a Nd:YAG transparent ceramic, and investigated its phase formation and densification mechanism in the sintering process. As depicted in Figure 3, the particle size of Al2O3, Y2O3 starting powders, as well as 2.88 Y2O3–0.12Nd2O3–5Al2O3 powder mixture and Y2O3 powders after planetary ball milling for 15 h, ranged from 80 to 800 nm. The results showed that using Y2O3 nano powder, under bimodal particle size distribution (D50 ≈ 160 nm and 400 nm), could make the shrinkage effect higher than the expansion effect in the formation of the YAG phase during sintering. Additionally, the transmittance of the prepared 4 at.% Nd:YAG sample (1 mm thick) at 650 nm was 80%, which was close to that of Nd:YAG single crystal. In addition, Zhang et al. [17] studied the effect of Nd dopant and LiF additive on the microwave dielectric and optical properties of transparent YAG ceramics in the spark plasma sintering (SPS) process in 2016. The SEM images demonstrated that the size of YAG nano powder was almost between 50 and 100 nm, and the infrared transmittance of the sample was 81.8% after sintering at 1360 °C. In 2021, Jia et al. [18] comparatively analyzed the influence of tetraethoxysilane (TEOS) additives on the sintering kinetics of Nd:YAG transparent ceramics. The vacuum sintering method was used to evaluate the densification process and sintering kinetics of Nd:YAG transparent ceramic samples. The densification rate of ceramic samples rose dramatically when the amount of TEOS was raised from 0 to 3.0 wt.%. The experiment showed that the transmittance of the 0.5 wt.% TEOS sample reached 75% in the near-infrared region. Ho-doped. Under direct pumping, the Ho3+ ion emits a quasi-three-level emission at 2.0 μm, exploited for efficient CW lasing [19,20]. Additionally, infrared lasers have also been made with Ho:YAG transparent ceramics. In 2015, Bagayev et al. [21] fabricated nano powders generated by laser ablation and then used two ways to make Ho:YAG transparent ceramics. The nano powders were made up of near-spherical particles with an average size of 8–14 nm and specific surface areas of 83.8 and 46.0 m2/g for the Al2O3 and Ho:Y2O3 particles, respectively. The results revealed that the transparent ceramics produced by their proposed method had better transmittance (82%) in the infrared band. Additionally, the slope efficiency of laser oscillations in the fabricated Ho:YAG transparent ceramic sample (1 mm thick) for pumping power was 40% (at 1.85 μm). In 2018, Zhao et al. [22] demonstrated a Ho:Y2O3 ceramic laser with high power, which fabricated Ho:Y2O3 ceramics by vacuum sintering and HIP methods. The in-band pumping method produced a 2117 nm laser with an output power of 24.6 W, nearly an order of magnitude higher than other ceramics. For high-power, solid-state lasers, therefore, Ho-doped sesquioxide ceramics are ideal materials. Er-doped. Er-doped YAG transparent ceramics have very low levels of quantum defects, and their laser behavior is IR transitions at 1.5 and 3 μm. In 2011, Zhang et al. [23] demonstrated a 0.5 at.% Er:YAG ceramic laser, which exhibited CW emission at 1617 nm and had a slope efficiency of 51.7%. In 2015, Zhang et al. [24] reported a passively Q-switched ceramic Er:YAG laser using a saturable absorber, which emitted 1617 nm. The experimental result confirmed that the laser could reach a peak power of 11.3 kW. In 2018, a laser adopted by 0.5 at.% Er:YAG transparent ceramics, with a resonantly pumped eye-safe, was developed by Bigotta et al. The fabricated ceramics adopted a two-step approach, combining SPS+HIP methods [25]. In their study, high-purity 0.5 at.% Er3+:YAG powder with a specific surface area of 7 m2/g and an average size of 271 nm was used. The experimental results confirmed that the light-light efficiency of this laser was 20%, and the maximum slope efficiency was 31%. Tm-doped. The Tm3+ concentration should be at least 6% to guarantee efficient down-conversion energy transfer [26]. Experiments showed that the transparent ceramics doped with Tm3+ have good light transmittance [27]. Zhang et al. [28] prepared highly transparent Tm:YAG ceramic by solid-phase reaction and vacuum sintering and studied its optical properties, microstructure and laser properties. Zou et al. [29] developed a high-efficiency, continuous-wave Tm:YAG transparent nano-ceramic laser pumped using a Ti:sapphire laser. Output power of up to 860 mW was produced with an absorbed pump power of 2.21 W at 785 nm, equating to a slope efficiency of 42.1% and a light-to-light efficiency of 22%. Zhan et al. presented a 2.7 mm long passively mode-locked laser based on 6 at.% Tm:YAG ceramics [30]. The pulse duration was 55 ps, and the highest output power was 116.5 mW at 2007 nm. Based on these findings, Tm:YAG transparent nano-ceramics looked to be promising candidates for ultrafast lasers with high power densities and high-efficiency output. Yb-doped. The Yb3+-doped ceramics’ spectral properties ensure nearly pure four-level lasing, which can be easily controlled by adjusting the ambient temperature or the temperature inside the pumped lasing medium. In 2008, Nakamura et al. [31] developed a CW laser based on Yb:YAG transparent ceramics. With a slope efficiency of 72%, a 6.8 W CW output power was obtained, and the transverse intensity distribution of the Yb:YAG ceramic laser beam was a Gaussian beam. In 2012, Luo et al. [32] used Yb:YAG ceramics and a 940 nm fiber-coupled laser diode to accomplish CW lasing at 1030 nm. The basic materials were commercial Al2O3 powder (99.99 percent purity, 250 nm) and co-precipitated Y2O3 and Yb2O3 powders (60–80 nm, 9.5–10.0 m2/g). For a 3 mm-thick mirror-polished Yb:YAG ceramics sample, in-line transmittances at 1300 nm and 400 nm were measured to be 83.6 and 81.8%, respectively. The slope effectiveness of this laser was 62.7% according to the testing data. Table 1 summarizes the doped YAG transparent nano-ceramics, which are described in the text grouped by doped type and publication year. It can be drawn that doped YAG laser transparent nano-ceramics have a short preparation period, low production cost, large-scale production, and high doping concentration. 2.3. Application The schematic diagram of YAG transparent nano-ceramics’ application in a laser diode pumping system is shown in Figure 4. As depicted in Figure 4, a symmetrical ring pump source was created using 32 groups of laser diodes (the highest output of an LD at 807 nm was 10 W) and a ϕ4 mm × 105 mm 0.6% Nd:YAG transparent nano-ceramic rod. Then, a high-power Nd:YAG ceramic laser with CW 1.46 kW was developed [13], and this was the first time that the output power of a ceramic laser exceeded the kilowatt level. The experimental findings showed that increasing the pump power to 290 W resulted in an 88 W multimode CW laser output. This meant that the light-to-light efficiency of YAG transparent nano ceramics was about 30%. In 2010, Marsh Corporation in the United States used multiple Nd:YAG transparent ceramic slabs with composite structures to achieve a laser output of more than 100 kW using direct pumping technology, of which the output power of a single Nd:YAG slab could reach 17 kW [33,34]. Nakamura [31] developed a high-power efficient transparent ceramic Yb:YAG laser with a Yb concentration of 9.8%, a pumping power of 13.8 W, a T = 10% output coupler, and a cavity length of 20 mm at a room temperature of 20 °C. At a maximum output power of 1.6 W, the ceramic Yb:YAG laser showed continuous tunability in the spectral region of 63.5 nm from 1020.1 to 1083.6 nm. A high-power passive Q-switched Ho:YAG ceramic laser was created by Yuan et al. [35]. The maximum pulse energy of this laser was 0.94 mJ, the pulse width was 28 ns, and the peak power was 33.5 kW at a pulse-repetition frequency of 28.8 kHz. 2.4. Summary Since it took Ikesue and Kinoshita [12] 31 years to use lasers for Nd:YAG transparent nano-ceramics in 1995, this was not a rapid development. Over the next seven years, advances in powder synthesis and ceramic sintering allowed the 1 kW output power threshold to be broken in 2002 [13], followed by another seven years until the 100 kW mark was crossed in 2009 [33,34]. In terms of powder-preparation methods, the most mature technologies are the solid-state reaction method and liquid-phase coprecipitation process. The solid-state reaction method has a simple process, but the commercial raw material powder used has low sintering activity, which is not conducive to the densification of transparent ceramics. In terms of doped YAG transparent ceramics’ sintering, vacuum sintering is currently the most commonly used sintering technology for fabricating them. Although vacuum sintering helps to eliminate pores and improve the density, it is nevertheless unable to entirely eradicate residual pores inside the ceramics, resulting in most sintered samples having a transmittance of less than 80%. In addition, according to the needs of solid-state lasers, rare earth ion-doped YAG transparent ceramics can be used to make laser materials with excellent performance, which are widely used in the field of solid-state lasers. It can be seen from the above experiments that after doping with Nd3+, the transmittance increases to 81.8%; the output power of Ho-doped sesquioxide ceramics is nearly an order of magnitude higher than that of other ceramics; Er-doped YAG transparent ceramics have very low levels of quantum defects, and their peak power can also reach 11.3 kW; Tm-doped ceramics can ensure an efficient step-down version of energy transfer; and Yb doping can adjust the environment and the temperature inside the pumping laser medium. In general, the transmittance and output power of ceramics doped with YAG were greatly improved, and the temperature could be controlled at the same time, which makes them a good candidate for ultrafast lasers with a high power density and high power output. However, most of the current research on Nd:YAG transparent ceramics is based on experimental results [36,37], and there is a lack of relevant theoretical simulation data. For example, there is still a lack of research on the relationship between structural defects (such as grain boundaries) and the photothermal damage of ceramics, as well as on the types and concentration distributions of doped rare earth ions. In addition, it is also necessary to study the occupancy mechanism and distribution of dopant ions of different types and concentrations inside the ceramic, as well as the influence of the surrounding crystal field. These breakthroughs in the mechanism of action need to be solved through effective theoretical models. 3. Magneto-Optical Transparent Nano-Ceramics 3.1. Brief Introduction Magneto-optical material is a new type of optical functional material that has a magneto-optical effect in the ultraviolet to infrared band. Optical devices, such as magneto-optical switches, magnetometers, and magneto-optical sensors, with various functions, are made by using the magneto-optical properties of these materials and the interaction and conversion of light, electricity, and magnetism. According to the type of materials, they can be divided into magneto-optical glass, magneto-optical crystal, magneto-optical transparent ceramics, etc. The thermal conductivity of magneto-optical ceramics is equivalent to that of magneto-optical crystals, and the thermal diffusion performance is good, which can effectively prevent thermal damage in the process of using lasers. Compared with crystals, magneto-optical ceramic materials can more easily obtain larger sizes and can be made into large-diameter magneto-optical components, with high fracture toughness and good thermal shock resistance [37]. This means that magneto-optical transparent ceramics is a new type of magneto-optical dielectric material in recent years with the advantages of a large size, Verdet constant, high laser damage threshold, and high thermal conductivity [38,39,40]. At present, the reported magneto-optical transparent ceramic materials mainly include TGG transparent nano-ceramics and TAG transparent nano- ceramics. 3.2. TGG In 2003, Khazanov [41] reported TGG transparent nano-ceramics for the first time. The results demonstrated that the emergence of high-quality nano-ceramics and the improved Faraday device made it possible to apply TGG transparent nano-ceramics in higher-power lasers. In 2007, Yasuhara et al. [42] reported the Faraday effect of TGG ceramics for the first time. They tested the Verdet constant of TGG transparent nano-ceramics samples at 1053 nm as a function of temperature and found that the Verdet constant was 36.4 rad/(T·m) at room temperature, which was 87 times greater than that at the temperature of liquid nitrogen (77 K). This showed that under the same magnetic field, the length of the Faraday material could be shortened to 1/87, which provided advantages for femtosecond short-pulse lasers. In addition, the test data also confirmed that the Verdet constant of TGG transparent nano-ceramics was similar to that of TGG single crystals. In 2011, Yoshida et al. [39] systematically studied the optical characteristics and Faraday effect of TGG ceramics at room temperature. As shown in Figure 5a, the grain size distribution of the prepared TGG transparent nano-ceramics was relatively uniform, and the grain size was between 300 nm and 3 μm, which was smaller than that of the YAG transparent nano-ceramics. At the same time, they tested the optical transmittance of TGG ceramic samples in the visible and near-infrared bands and compared them with TGG single crystals. It could be drawn from Figure 5b that in the 600–1400 nm band, the transmittances of TGG ceramics and single crystals were almost equal, with both higher than 80%. In 2019, Jin et al. [43] prepared the magneto-optical properties of TGG transparent nano-ceramics. The microstructures of the TGG powders showed that the average particle size was about 80 nm, and the powder agglomerated through small connections to the particle necks, which facilitated the densification of the ceramic body. The best microstructure of the nano-ceramics with average grain sizes of 5.32 μm, fabricated by co-precipitated method and vacuum sintering methods, could be obtained while the sample was sintered at 1550 °C (in Figure 6). Additionally, they found that the best optical transmittance of TGG ceramics was close to 80% in the region of 400–1500 nm, and that the Verdet constant of the fabricated TGG ceramics decreased linearly with an increase in temperature. The results demonstrated that TGG transparent nano-ceramics could meet the requirements of magneto-optical devices working the visible-near-infrared region. In 2021, Li et al. [44] investigated the fabrication and evaluated the performance of novel TGG transparent nano-ceramics, which were doped by rare earth (RE) of Pr, Tm, and Dy. The microstructure of the powder particles was several hundred nanometers (300~700 nm) in length and several nanometers in width. After two-step sintering, no second phase was detected in the microstructure, although residual pores in the ceramic could be noticed. Therefore, the prepared ceramics all had good optical quality, and the online transmittance at 1070 nm was greater than 80%. The Verdet constant of RE:TGG transparent nano-ceramic samples (−143 rad/T·m at 632.8 nm) was optimized by rare earth doping, which was about 5% higher than that of TGG transparent nano-ceramics. With the increase in power of the CW laser and pump laser, the thermal effect of optical elements in a laser system becomes more and more serious [45,46,47]. Thermally induced birefringence reduces the isolation ratio of the Faraday isolator, which limits its use in high-power lasers [39]. In 2014, Yasuhara et al. [48] developed a TGG Faraday rotator for a high-power (257 W) laser with an isolation ratio of 33 dB and studied its thermal depolarization effect and thermal lens effect under laser irradiation. They tested the thermally induced depolarization ratio of TGG transparent nano-ceramics under laser irradiation, as shown in Figure 7. The results showed that the depolarization ratio of TGG ceramics at 257 W laser power was 5.48 × 10−4 under the applied magnetic field (45° Faraday rotation), and the corresponding isolation ratio was 33 dB. This confirmed that the thermally induced depolarization of TGG ceramics was almost the same as that of a single crystal. The results showed that the Faraday isolator based on TGG transparent nano-ceramics had basically met the requirements of service under a high-power laser. 3.3. TAG TAG transparent nano-ceramics have the same garnet structure and similar optical and thermal properties as TGG transparent nano-ceramics, but their Verdet constant is about 30~50% higher than that of TGG transparent nano-ceramics. Therefore, TAG transparent nano-ceramics have better magneto-optical properties and can be used in Faraday magneto-optical materials in visible and near-infrared bands. In 2011, Lin et al. [49] prepared TAG transparent nanoceramics for the first time by solid-phase reaction and vacuum sintering (depicted in Figure 8a) and studied the optical quality and microstructure of the samples. The samples sintered at 1650 °C had relatively good optical transparency between 400 nm and 1600 nm (up to 70%, as shown in Figure 8b). The experimental data demonstrated that the thermal conductivity of the prepared TAG transparent nano-ceramics was 6.5 W/m∙K (at room temperature), and the Verdet constant could reach −172.72 rad/T∙m (at 632.8 nm) with the best quality, which was 28.9% larger than that (−134 rad/T∙m) of the TGG single crystal [50]. This indicates that TAG transparent nano-ceramics have better magneto-optical properties than TGG single crystals and have potential commercial values. In 2015, Chen et al. [51] found that by optimizing the sintering aid, TEOS combined with MgO as a sintering aid could improve the optical quality of TAG transparent ceramics but had no effect on the magneto-optical properties. The study found that when the addition of TEOS was 0.4 wt.% and the addition of MgO was 0.1 wt.%, the optical transmittance of the obtained TAG ceramics exceeded 80% in the 500–1500 nm band, and the optical quality was greatly improved. Moreover, in 2017, Duan et al. [52] adopted the reaction sintering method combining muffle furnace pre-sintering and HIP sintering and obtained TAG transparent ceramics with ideal optical quality for the first time without vacuum sintering. After ball milling, the raw material powders were uniformly mixed, wherein the particle size of Al2O3 was 260 nm, and the particle size of Tb4O7 was 1.15 μm. The experimental results confirmed that the optical transmittance of the prepared 0.4 wt.% TEOS:TAG transparent nano-ceramics in the visible and near-infrared regions could reach more than 80%, which was very close to the theoretical limit. In addition, some researchers found that the Verdet constant of TAG transparent nano-ceramics could be changed by adding rare earth elements [38,53,54]. In 2012, Chen et al. [53] adopted solid-state reaction and vacuum sintering methods to prepare Y3+ and Ce3+ rare earth element-doped TAG transparent nano-ceramics. They found that the Verdet constant of Y3+-doped TAG transparent nano-ceramics was −108.79 rad/T·m at 632.8 nm, which was smaller than that of TAG transparent ceramics. This indicated that diamagnetic ion (Y3+) doping would have an adverse effect on the magneto-optical properties of TAG transparent ceramics. However, the magneto-optical properties of the Ce3+-doped TAG transparent nano-ceramics were greatly improved, and its Verdet constant, measured as −199.55 rad/T·m at 632.8 nm, was about 16% higher than that of TAG transparent ceramics. A year later, Chen et al. [54] continued to study the relationship between the sintering process parameters and optical properties of Y3+-doped TAG transparent nano-ceramics, and they found that the samples sintered at 1680 °C showed the best optical properties, obtaining transmittance of 75% in the range of 900 to 1600 nm. Additionally, X-ray diffraction (XRD) results showed that samples had pure garnet crystal structure without secondary phases. 3.4. Others Some sesquioxides also have magneto-optical effects and generally have high thermal conductivity, which is also a potential application value, such as in high-energy lasers [38,55,56,57]. For example, at 1064 nm wavelength, the optical transmittances of fully doped and 10% doped Y3+:Y2O3 ceramics were 63.4 and 79.3%, respectively, but both were lower than their theoretical transmittances of 81.1% [55]. The Ho2O3 magneto-optical transparent nano-ceramic was successfully prepared by the SPS sintering method, and its transmittance at 1 μm was measured to be about 60%, but its Verdet constant at the wavelength of 1064 nm was −46.3 rad/(T·m), which was close to that of TAG, which was about 1.3 times that of TGG [56]. These studies [55,56,57] showed that yttrium oxide (Y3+:Y2O3), holmium oxide (Ho2O3) [58], dysprosium oxide (Dy2O3), and other nano-transparent ceramics still need to be further optimized to improve their optical properties. 3.5. Potential Application Transparent nano-ceramic is the core of modern magneto-optical materials, which provides a broad development prospect for the Faraday isolator. Furuse et al. [59] compared characteristics of Faraday isolators with different materials of magneto-optical medium, which are listed in Table 2. It can be seen from Table 2 that there is still a certain gap in the isolation between TGG nano-ceramic ceramics and TGG crystals. TGG crystals can obviously work effectively in higher-power lasers and achieve a stable isolation ratio. Although the isolation measurement power is lower for TAG and Ce:TAG transparent nano-ceramics, it can be estimated by calculation that the extinction ratio of the TAG ceramic samples can be maintained above 30 dB when the laser power reaches the kilowatt level. In 2021, Starobor et al. [60] fabricated a TGG/sapphire composite magneto-optical element Faraday isolator, whose maximum operating power with an isolation ratio higher than 30 dB was estimated to exceed 2 kW, which was almost three times that of a single TGG crystal. In addition, they also performed experimental (shown in Figure 9) and numerical studies on thermally induced depolarization and thermal lensing of composite elements, and the optimized structure could be successfully operated at a high radiation power. Experimental results showed that an isolation ratio of 34 dB in the composite elements was achieved at a laser power of 700 W, which was 5 dB higher than the classical single element. This indicated that TAG transparent nano-ceramics could be used to prepare kW Faraday isolators. 3.6. Summary With the continuous improvement of techniques, the optical quality of transparent nano-ceramics has been greatly improved and is almost comparable to that of its crystals [61,62]. Among the magneto-optical ceramic materials that have appeared so far, the research focus is mainly on TGG and TAG transparent nano-ceramics. As can be seen from the research results of the existing literature, most of the TGG and TAG transparent nano-ceramics have a light transmittance of 80% in their expected working band, and the Verdet constant meets the working requirements, which potentially gives them commercial value in large-scale applications. In addition, the magneto-optical properties of TAG transparent nano ceramics are changed by doping different rare earth elements. This can not only increase the Verdet constant but also could reduce the Verdet constant, which can regulate its magneto-optical characteristics by controlling the type and proportion of rare earth elements. This research also needs attention in the future study of TGG transparent nano-ceramics. 4. Transparent Armor Nano-Ceramics 4.1. Brief Introduction Transparent armor, such as face shields, windows of military vehicles, and lookout windows of aircraft, is one of the most important personnel protection technologies. Currently available transparent armor consists of several layers and is very thick to resist the ballistic impact of multiple hits. Today, the demands placed on these systems are increasing [63,64]. As a result, traditional glass-based armors have become impractical in terms of weight and thickness constraints, which has led to increased consideration of transparent ceramics for such applications. As shown in Figure 10, a versatile four-layer design (A—projectile erosion fragmentation layer; B—energy absorption, crack arrest layer, C—fragmentation protection layer; and D—adhesive layer) can be used for the development of light armor, where the first layer is the core layer in the overall four-layer design [65]. When transparent nano-ceramics such as MgAl2O3 or AlON are used for this functional layer, the weight and thickness of glass-made armor can be reduced by about 30–60% [66]. In addition, some weapon systems that need to withstand harsh environmental conditions also require relatively large windows and domes [67,68,69]. 4.2. MgAl2O4 Due to its high hardness, strong chemical resistance, and high transparency in the UV-Visible and mid-IR range, MgAl2O4 transparent nano-ceramics are expected to be used in optical components and defense applications [64], such as optical lenses, aircraft/vehicle windows, and missile domes [70]. Moreover, they can achieve a balance between optical performance and production cost [71]. In 1974, Bratton [72] fabricated translucent polycrystalline MgAl2O4 ceramics with co-precipitated spinel as raw material, doped with 0.25 wt.% CaO as a sintering aid. The experimental results showed that at a certain temperature, the relative density of sintered spinel could reach 99.7% (close to the theoretical value of 100%) and that the transmittance in the visible light region was between 67 and 78%. To obtain MgAl2O4 transparent nano-ceramics with high optical transparency, high-quality precursor powders (average particle size of 150 nm [73]) and optimized concentrations of sintering aids, such as CaO [72], B2O3 [73], and LiF [74], are required. In 2013, Esposito et al. [75] fabricated MgAl2O4 transparent nano-ceramics by hot pressing and investigated their characterization. They prepared two starting powder mixtures (with a purity level greater than or equal to 99.99%) made from commercial Al2O3 and MgO products, taking into account the stoichiometric ratio of MgAl2O4. Experimental results showed that the value of D50 of MgAl2O4 was close to 180 nm. The particle size of the starting powder had a slight effect on the sintering evolution and final microstructure but not uniformly on the final transmittance. Optical inspection experiments showed that up to 70% transmittance (the highest value of 78% in the 1100 nm band) was obtained in visible light. In addition, thermodynamic studies of the reactions of LiF, MgO, and Al2O3 could help to understand the densification mechanism that affected the transmittance of spinel. In 2015, Esposito et al. [76] investigated the effect of the pressure applied during sintering on the final optical properties of the MgAl2O4 transparent nano-ceramics in hot pressing. By establishing a thermodynamic model, the role of lithium fluoride as a sintering aid was clarified. The results of the study showed that transparency, close to the theoretical value, could only be achieved at very high pressures (200 MPa or higher) due to spinel destabilization. Boulesteixa et al. [77] prepared MgAl2O4 transparent nano-ceramics from pure aluminum and magnesium sulfate using colloidal chemistry. Figure 11a,c shows that the basic particles are between 40 and 80 nm in diameter and have an isotropic shape. The phase composition of the powder is shown in Figure 11b, while the particle size distribution of the as-received powder is shown in Figure 11d. In general, it is challenging to fabricate MgAl2O4 transparent nano-ceramics using traditional pressureless sintering techniques [73,78]. Therefore, complex sintering strategies, such as hot pressing, HIP, and SPS, are often employed to develop these ceramics [74,75,76,79]. In 2010, Meir et al. [80] synthesized and densified MgAl2O4 transparent nano-ceramics using the SPS method. The alumina particles in the starting powders were well separated and polygonal in shape with a particle size of about 100–500 nm, while the magnesium oxide powder was strongly agglomerated, with a size of about 10–20 μm and a relatively high specific surface area. The addition of a sintering aid of 1 wt.% LiF to the mixed powder promoted the synthesis of spinel. The experimental results indicated that LiF vapor played an important role in eliminating residual carbon contamination and achieving a fully dense state. The optical transmittance of the MgAl2O4 transparent nano-ceramics prepared by the SPS method reached 78% in the 400–800 nm band. In 2020, Liu et al. [81] successfully prepared MgAl2O4 transparent nano-ceramics by air pre-sintering and HIP methods. They found that a small amount of CaO (0.1 wt.%) was very effective for the densification of ceramics. They also found that the relative density of MgAl2O4 transparent nano-ceramics changed from 86.3% to 99.4% when the pre-fired temperature was increased from 1450 °C to 1550 °C. Optical experimental results demonstrated that the prepared large transparent nano-ceramics, with an average grain size of 1500 nm, exhibited high transmittances of 86.3% and 82.5% at 1100 nm and 600 nm, respectively. In 2020, Liu et al. [82] used high-purity spinel nano powders, with a particle size of 55 nm and a specific surface area of 30 m2/g, to successfully prepare MgAl2O4 transparent nano-ceramics by microwave sintering and HIP methods. They found that microwave-sintered MgAl2O4 transparent nano-ceramics could achieve higher densities at lower temperatures and in a shorter time than conventional pressureless sintering. Experimental results showed that the in-line transmittances of MgAl2O4 transparent nano-ceramics 1.5 mm thick that were obtained by microwave sintering (1400 °C, 80 min) and HIP (1650 °C, 180 MPa, 3 h) were 86.4% and 80.2% at 1064 nm and 400 nm, respectively. Therefore, MgAl2O4 transparent nano-ceramics have excellent light transmission ability [83]. 4.3. AlON AlON transparent nano-ceramics have excellent light transmittance in the near-ultraviolet to mid-infrared band (0.2–6 μm), and the theoretical transmittance is as high as 85.2% [84]. In addition, they also have excellent properties, such as high hardness, high strength, high-temperature resistance, friction resistance, and acid and alkali corrosion resistance. In 1959, Yamaguch and Yanagida [85] first reported cubic spinel-type aluminum oxynitride (AlON), which was produced from a compound between alumina and aluminum nitride in a reducing atmosphere above 1650 °C. They also studied the physical constants of AlON, such as crystal structure, density, refractive index, dielectric constant, and magnetic susceptibility. First, nano powder is required to fabricate AlON transparent nano-ceramics. After obtaining the AlON nano powder, it should be fabricated by sintering at a temperature above 1850 °C for at least 20 h in nitrogen. In 2011, Qi et al. [86] adopted a two-step approach with Al2O3 and aluminum nitride (AlN) nano powders as the starting materials to prepare AlON transparent nano-ceramics. In their studies, the average particle size of nano powders of α-Al2O3, γ-Al2O3, and AlN were 80, 20, and 20 nm, respectively. Experiments showed that the AlON Nano-ceramic samples were transparent after sintering at 1880 °C for more than 5 h. With the extension of the holding time, the grain size of the sample increased slightly, while the pore size and porosity decreased obviously, so the light transmittance increased. Additionally, the sample had a transmittance of 55% (near 5 μm band) at a hold time of 20 h. While the infrared transmittance of ceramics is promising, the visible transmittance is not high enough. In 2012, Jin et al. [87] fabricated a highly AlON transparent nano-ceramic that was sintered from nano powder without pressure. During carbothermal nitridation, a layer of amorphous carbon on the surface of Al2O3 particles effectively prevented agglomeration and grain growth, and the bimodal particle size distributions of the obtained AlON powders were concentrated at 200 and 700 nm, and their maximum particle sizes were both below 900 nm. The experimental results confirmed that the AlON transparent nano-ceramics with an average online transmittance above 80% in the visible light to infrared range were obtained by a pressureless sintering method. In 2018, Zhao et al. [88] carried out a detailed investigation of planetary ball-milling for coarsened AlON nano powder. Their results revealed that the weight ratio of balls to powder, rotational speed, and planetary milling time had a significant effect on the microscopic morphology, particle size distribution, and average particle size of the powders. Using fine and uniform AlON nano powder with an average particle size of less than 300 nm and excellent sintering properties, subsequently, a sample of AlON was successfully fabricated from the finely treated powder synthesized by the carbon thermal nitriding method at 1880 °C for 6 h. The optical test demonstrated that the high performance with an online transmittance of 84% at 2000 nm of the sample could be achieved. In 2022, Zhang et al. [89] thoroughly milled the AlON powder to fully mix the nanopowder with the sintering agent. Figure 12a shows an SEM image of the particle morphology after grinding; the particle size distribution is shown in Figure 12b,c, which shows that the high-temperature decarbonization and ball-milling process do not affect the purity of the AlON powder. In the preparation of AlON transparent nano-ceramics, sintering aids are generally used to control the growth of Al2O3 grains. MgO and Y2O3 were used as co-sintering aids by Yuan et al. [90] to fabricate AlON transparent nano-ceramics by reactive sintering. The densification of AlN transparent nano-ceramics was regulated by the sintering aids and sintering duration, allowing their optical performance to be efficiently tuned. Furthermore, co-doping with MgO and Y2O3 enhanced densification more effectively than either MgO or Y2O3 alone. The optical test demonstrated that 1 mm-thick fabricated samples of AlON transparent nano-ceramics, doped with 1 wt.% MgO and 0.08 wt.% Y2O3, reached the maximum of 60% in-line transmittance at 600 nm after sintering in N2 for 12 h at 1950 °C. In 2018, Shan et al. [91] adopted the pressureless sintering method to fabricate AlON transparent nano-ceramics with CaCO3 doping. For a sample with a thickness of 2 mm, the transmittance of AlON transparent nano-ceramics doped with a mass fraction of 0.3–0.4% CaCO3 reached 83–85% at around 3700 nm. In the wavelength range of 200–6000 nm, the transmittance of AlON doped with CaCO3 was always higher than that of AlON doped with an ideal amount of Y2O3. In 2021, Li et al. [92] prepared AlON transparent nano-ceramics by HIP assisted by the dissolution of gas inclusions. They investigated the influence of additive content, pre-sintering, HIP, and annealing parameters on the performance of AlON transparent nano-ceramics. The fabricated samples and transmittance curve are depicted in Figure 13, and the sample doped with 0.5 wt.% SiO2 exhibited the best performance. By comparing different sintering additives and pre-sintering atmospheres, the densification mechanism of the material was studied. The prepared AlON transparent nano-ceramics maintained a high transmittance of 85.8% at 2000 nm. As mentioned above, ALON ceramics have excellent optical and mechanical properties and extremely strong light transmittance (up to 85.2%), which makes them light-transmitting materials with excellent application prospects. Secondly, because of their good wear resistance; good scratch resistance; and light, thin, and strong properties, even if they are rubbed and damaged, the light transmittance will not be affected, so they can be used as a reinforcing material in transparent armor. In the civilian field, because of its high hardness and good durability, it can be used in the casing of precision instruments such as watches and goggles [93]. 4.4. Potential Application As depicted in Figure 14a, a large-scale MgAl2O4 transparent nano-ceramic, 12 mm thick, was fabricated by gelcasting and pre-sintering in air HIP methods, and the average grain size was 8 μm. In 1998, after realizing the mass production of powder, the research on AlON transparent ceramics progressed rapidly. At present, the largest size of AlON flat window that can be successfully fabricated is 880 × 45 × 12 mm [94]. In 2019, the Surmet company realized the production and engineering application of large-scale AlON transparent nano-ceramics in batches through pressureless sintering and HIP methods [95]. Then, a high optical quality window, fabricated by AlON transparent nano-ceramics of about 0.41 m2, was realized. The emergence of large-scale armored transparent ceramics makes these materials more frequently used in helicopter protection, personnel protection, infrared windows for reconnaissance, and other fields because their weight and thickness are only half of that of traditional bulletproof glass. Figure 14b shows the effect observed at a distance of 30 m from a γ-AlON transparent ceramic that was prepared by HIP post-treatment at 1800 °C for 2 h under 190 MPa [96]. Unfortunately, at present, the manufacturing cost of transparent armor nano-ceramics is relatively high. A comparison of transparent armor materials for STANAG 4569 was conducted by Benitez et al. in 2017 [65] and is depicted in Figure 15. Compared with other materials, transparent nano-ceramic materials have the lowest areal density, only half of glass–ceramic materials. However, the maximum thickness of transparent armor nano-ceramics is still difficult to achieve compared to that of other materials. In addition, the manufacturing cost is also 5 to 10 times higher than other types of materials. To promote transparent armor nano-ceramics, therefore, it is necessary to reduce the fabricating cost. This could be achieved by reducing manufacturing steps or finding other alternative materials. For manufacturing process optimization, one-step sintering methods need to be created in the future to obtain fully dense materials. 4.5. Summary At present, it is difficult for MgAl2O4 transparent nano-ceramics to achieve the theoretical density required for transparency by conventional sintering methods because they are very sensitive to powder size, agglomerates, impurities, and additives. The defects of the stoichiometric ratio, impurities, particle size, and agglomerates of the starting nano powders are generally difficult to improve by adjusting the process parameters, which affect the optical properties of the prepared samples. To obtain MgAl2O4 transparent nano-ceramics with high optical transparency, high-quality starting powders are indispensable. Meanwhile, it is also necessary to optimize the concentration of sintering aids or add some rare earth elements. Existing commercial MgAl2O4 transparent nano-ceramics are sintered by pressureless sintering/HIP or hot-pressing sintering/HIP. Thereafter, the manufacturing cost of these processes is still high, and it is difficult to produce large plates. The AlON powder synthesized in the liquid phase has relatively high purity, small particle size, uniform distribution, and relatively high chemical activity, but the synthesis process conditions are not suitable for large-scale production. The solid-phase pulverization method can ball-mill micron-sized AlON powder to the nano-scale; the preparation process is relatively simple, and it is easy to achieve large-scale production, but ball milling causes powder lattice distortion increases the defect concentration, and introduces impurities. Therefore, the sintered ceramics have special structures such as impurities, which affect the optical properties. As reviewed in Section 4.3, AlON transparent nano-ceramics usually need to be sintered for a long time above 1800 °C without external field assistance; otherwise, it is difficult to make these materials fully dense. Although preparation methods have undergone great progress in the past few decades, there are still some difficulties in the large-scale preparation process, such as shrinkage during drying and cracks forming inside the green body. In addition, the sintering of large-sized AlON transparent nano-ceramics also causes an uneven microstructure due to uneven temperatures, which in turn leads to stress birefringence and reduces their optical properties. 5. Other Transparent Nano-Ceramics 5.1. Electro-Optical Transparent Nano-Ceramics The rise of emerging electro-optical materials is in line with the arrival of modern electronic, optical, and laser technologies, which are increasingly demanding materials. Compared with other materials, they have fast speed, low consumption, high reliability, and strong interference suppression ability, which mainly used in transducers, actuators and sensors, and it has a great impact on space structure, electronic industry, etc. [97,98]. The new electro-optic ceramic material not only has the advantages of ordinary electro-optic materials but also has a fast response speed. The response time is generally only a few nanoseconds, and the electro-optic coefficient is larger than other materials, about 2 × 10−15~6.6 × 10−15 (m/V)2. At present, commonly used electro-optical transparent ceramic materials are mainly two types: lead lanthanum zirconate titanate ceramics (PLZT) and lead magnesium titanate–lead titanate ceramics (PMN-PT) [99]. PLZT ceramic is a kind of transparent ceramic with an ABO3 type perovskite structure. It has strong light transmittance in the visible light to infrared light band. The light transmittance reaches its peak at about 600 nm, 802 nm, and 88 = 0 nm. PLZT ceramics are more commonly used in the manufacture of multi-functional, low-loss optical devices, as well as in holographic storage technology and optical fiber sensing [100]. The PMN-PT transparent ceramic is an ABX3 type ceramic with a high dielectric constant, which can reach up to 600 at room temperature. Additionally, it has better optical transparency. At the visible light wavelength of 400~2000 nm, the transmittance increases from 0 to 70% and finally stabilizes at 70% when the temperature rises from 400 °C to 2000 °C. It has a wide range of applications in satellite communications, sensors, frequency converters, etc. [101]. In 1970, Haertling et al. [102] first used the hot-pressing sintering method to prepare opto-ceramics with a thickness of 1 mm. This method could improve the compactness of ceramic materials. The density could reach about 99% of the standard density, and the light transmittance could reach 80%. In 2016, Somwan et al. [103] added powders of Bi2O3 and CuO oxides in the process of preparing PLZT ceramics and fired particles with a diameter of 1 cm without agglomeration. It was found that the dielectric constant reached a peak value of 2000 when the temperature increased to 215 °C, and then the dielectric constant began to decrease with the increase in temperature. After adding the mixture, the sintering temperature was reduced by up to 50 °C. In 2018, Samanta et al. [104] doped Nb and Fe elements during the preparation of PLZT ceramic materials, in which the radius of Fe3+ was 69 pm, and the radius of Nb5+ was 78 pm. The results showed that the conductivity increased from 10−8 to 10−6 S/cm when the frequency was increased from 100 Hz to 1 MHz at room temperature. When the doping was 2%, the energy storage density reached a maximum of 140 mJ/cm3. In 2017, Zhang et al. [105] prepared PMN-PT ceramics with a thickness of 0.35 mm using a two-step hot-pressing method, in which the contents of lanthanum, PMN, and PT were 3%, 75%, and 25%, respectively. When the measured temperature continued to rise, the half-wave voltage gradually increased from 200 V to 400 V, and the electro-optic coefficient continued to increase. In 2018, Wang et al. [106] reported on the preparation of PMN-PT/CFO thin films by sol–gel spin-coating technology. When the ratio of CFO to PMN-PT was 4:1, the leakage amount reached its maximum. The film quality was improved, and when the temperature increased from 650 °C to 730 °C, the leakage current decreased from 97.54 to 40.59. This indicates that the ferroelectric performance improves. The coercive electric field and the polarization curve are similar, indicating that there is some interaction between the ferroelectric phase and the magnetoelectric phase. In 2021, Ze et al. [97] prepared PMN-PT materials using a two-step sintering method. The thickness of the small ceramic was 0.85 mm, and the diameter was 10 mm. The thickness of the large ceramic was 1 mm, and the diameter was also 10 mm. After testing, it was found that as the wavelength gradually increased to 900 nm, the transmittance gradually increased to 70%. When Sm was doped, the peak value of optical power would move back. However, the basic rule was that with the increase in electric field intensity, the optical power gradually increased to 100% and then continued to decrease. Pramanika et al. [98] prepared PMN-platinum piezoelectric ceramics using the solid-state reaction method and analyzed the microstructure of the four fractured and sintered samples by scanning electron microscopy and found that their average particle sizes were 2.7, 3.2, 3.8, and 4.3 μm, respectively (Figure 16). Table 3 summarizes the representative papers on the preparation of electro-optically transparent nano-ceramics, mainly including the powder, fabrication and remarks. At this stage, electro-optic transparent ceramics are mainly used in electro-optic modulators, high-speed electro-optic switches, and ultrasonic transducers. Next, we will analyze these devices in detail. Firstly, using the electronically controlled refraction effect in the electro-optical material, the electro-optical transparent ceramic is prepared as an electro-optical modulator, in which the basic parameters such as the frequency and phase of the light beam can be determined by the characteristics of the electro-optical material. This can provide higher working accuracy, greater working reliability, and stronger anti-interference ability. The use of this modulator can minimize the degradation of the optical fiber system and increase the life of the equipment [107]. In addition, electro-optic transparent ceramics can also be used in high-speed electro-optic switches. The main working principle is to use the electronically controlled refractive index of electro-optic materials to adjust and use optical signals to control the switch. Moreover, the electro-optic material has good light transmittance, a high electro-optic coefficient, and a low cost, which have a great guiding effect on the development of electrical components in the future. However, the LiNbO3 crystal is mainly used in the existing electro-optic switch. There are still some defects, and the process of selecting the crystal axis when transmitting light is more complicated [108]. Moreover, PLZT ceramic materials have small diameters, good dielectric and piezoelectric properties, and good electromechanical coupling coefficients, which can be used in medical imaging technology and ultrasonic transducers. After using the transparent optical fiber as the main component, the size of the material and the damage to the surrounding materials, such as laser cutting, can be reduced to a limited extent. Additionally, the use of this material can improve the resolution of the device and increase the working accuracy [109]. Furthermore, this study found that when PMN-PT and polyvinylidene fluoride (PVDF) materials were mixed, the piezoelectric coefficient was very high. The device’s perovskite structure with good crystallinity results in better self-powering performance, harvesting energy from the environment and reducing the cost required for the device. When subjected to external force, both PMN-PT and PVDF materials generate an electric potential to maintain the stability of the material and reduce the possibility of fracture [110]. 5.2. Scintillation Transparent Nano-Ceramics A scintillator is a device that can convert high-energy (X, γ) radiation or charged particles into new materials that emit visible light. According to the shape, composition, and structure of the scintillator, it can be divided into scintillating glass, scintillating ceramics, scintillating gas, scintillating crystal, scintillating plastic, etc. Among them, single-crystal and ceramic scintillators are the most widely used materials, mainly used in high-energy physics (precision electromagnetic energy), medicine (medical imaging), industrial applications (CT flaw detection), and stone well detection [111,112]. With the wide application of scintillating ceramics, scholars have carried out a series of research on its preparation method, light transmission, and scintillation performance. In 1895, a German physicist, Lunchen, stumbled upon a flashing light near a tube coated with barium platinum cyanide while conducting a cathode-ray experiment and named it an X-ray [113]. In 1896, CaWO4 was first used as an X-ray fluorescent powder for human body X-ray photography [114]. In 2017, Zhou et al. [115] synthesized Nd3+ activated SrF2 nanoparticles via the precipitation method and characterized the microstructure and morphology behavior of Nd3+-doped nanoparticles using X-ray diffraction (XRD), field emission scanning electron microscopy (FESEM), and an energy-dispersive X-ray spectrometer (EDS). They found that the synthesized powders had good sinterability, and transparent ceramics with a transmittance of about 80% at 1060 nm could be prepared by vacuum hot-pressing sintering at 800 °C for 2 h. In 2018, Yi et al. [116] prepared (Ca0.94Gd0.06) F2.06 transparent ceramics with Nd3+ content of 0.5–5.0 at.% using the vacuum hot-pressing sintering method and studied its structure, spectrum, and thermal properties. It was found that the characteristic absorption peak intensity of the Nd3+ ion increases linearly with the increase in its content. With an increase in Nd3+ content from 0.5 at.% to 5.0 at.%, the measured life dropped dramatically from 484.9 μs to 47.8 μs. In addition, they also found that the thermal properties of the ceramics were closely related to temperature and Nd3+ concentration. Chen et al. [117] prepared Ce, Mg:LuAG transparent ceramics (Ce concentration of 0.025–0.3 at.% and Mg concentration of 0.2 at.%) using the solid-phase method and studied the effect of Ce content on its light transmittance. As shown in Figure 17, Figure 17a represents the transmittance of six samples of varying content. They discovered that with a Ce content of 0.1 at.%, the transmittance of six Ce3+ samples sintered at 510 nm was close to 71%. Mg:LuAG ceramics had the best properties, as shown in Figure 17b. In Figure 17c, it can be observed that the absorption bands of Ce3+ increase gradually at 345 nm and 445 nm after annealing. In 2019, Hostasa et al. [118] investigated the effects of three sintering aids (MgO, CaO, and TEOS) on the densification and optical properties of Ce:GGAG transparent ceramics. The experimental results showed that the three additives tested increased the density of Ce:GGAG after sintering, but only TEOS provided sufficient densification to result in transparency. In 2020, Trofimova et al. [119] studied the radioluminescence (RL) properties of Lu3Al5O12:Ce(LuAG:Ce) single-crystal and transparent polycrystalline ceramics at a high temperature of 600 °C. The results showed that the CE3+ comprehensive RL strength of a single crystal increased by 1.4 times from RT-450 °C. Polycrystalline ceramics increased by 1.9 times from room temperature to 300 °C, i.e., LuAG:Ce scintillators can be used effectively over a wide temperature range. Bartosiewicz et al. [120] studied the growth process, crystal composition, and optical and scintillation properties of the Lu3Al5O12:La (La = 0–0.45%) single crystal. They found that La doping significantly reduced the flashing afterglow of the LuAG:La crystal and increased the flashing output. Representative studies on transparent scintillating ceramics from recent years are summarized in Table 4. Compared with those from a single crystal, LuAG-based scintillating ceramics have the advantages of low preparation temperature, simple process, and low cost, and have important application prospects and development potential. Many scholars have continuously advanced their research work on the preparation, processing technology, and parameters of Lu3Al5O12(LuAG) doped with luminescent ions to provide an important reference basis for the design and preparation of new components of LuAG-based scintillating ceramics. At present, rare earth ion doping (Ce3+, Pr3+, etc.) is a kind of scintillating material with active research. This kind of material is mainly prepared by the precipitation method, and there is much research on its performance indexes (optical yield, energy resolution, radiation damage resistance, etc.). However, the amount of sinter used in the preparation of transparent ceramics and the amount of doped luminous ions are seldom studied, which should be the focus of future research work. 5.3. Summary With the in-depth research on transparent nano-ceramics, electro-optic and scintillation transparent nano-ceramics have also been found to have great application prospects. The hot-pressing two-step approach and sol–gel spin-coating method are two commonly utilized preparation procedures for the former. The compactness of ceramic materials can be considerably improved by sintering mixed oxide during the preparation process. The temperature and electric field strength, for example, have an effect on the transmittance of optoceramics, according to the current study. However, one of the next important research topics will be how to manage these parameters to change the transmittance of optoceramic materials. Scintillating transparent nano-ceramics, on the other hand, have a wide range of applications in the medical field, physics, and industry due to their scintillation function. The majority of scintillation transparent ceramics research now focuses on the preparation method, powder size, material characteristics, and so on. Following a review of the literature, it was discovered that the majority of scintillation transparent ceramics are made using the co-precipitation process, with scintillation qualities increasing via doping with scintillation ions (mostly rare earth ions Ce3+, Pr3+, etc.). Furthermore, the specific application of scintillation transparent ceramics and how to improve the transparency of ceramics (through a reasonable sintering method) and light output (through the doping of luminescent ions) will be the research direction and technical difficulties of the future. 6. Discussion 6.1. Effect of the Particle Diameter on Fabrication Figure 18 shows the average particle diameter of nano powders in the reviewed transparent nano-ceramics. As depicted in Figure 18, the average particle diameter ranges from 50 to 1000 nm, and the average is close to 150 nm. At present, most of the existing powders are of nano scale. Further reducing the particle diameter means that the manufacturing cost of powder is increased. For example, by laser reprocessing, the particle diameter of commercial powder can be further reduced, resulting in the desired specific surface areas. This is because polycrystalline transparent ceramics need to reach a relative density of more than 99.9%; that is, the porosity should be less than 1/10,000. It is difficult to eliminate these micropores via the conventional sintering process because the pores trapped in the crystal during grain growth are very difficult to discharge by diffusion at the end of sintering. Therefore, it is necessary to use ultra-fine nano powder with high purity and high activity and to adopt a multistage sintering process or sintering for a long time under low temperatures and a vacuum to eliminate pores. 6.2. Relationship between Nano Powders and Performance According to the types of materials, transparent nano-ceramic materials are mainly divided into metal oxides and non-metal oxides [2]. Transparent nano-ceramics based on metal oxide materials mainly include those based on alumina (Al2O3), magnesia (MgO), zirconia (ZrO2), yttria (Y2O3), lutetia (Lu2O3), and other oxides. Transparent nano-ceramics based on non-metal oxide materials mainly include those based on AlON and AlN, sialon and silicon nitride (Si3N4), and fluoride. To meet the needs of different scenarios, the required nano-transparent ceramic functional materials can be prepared by selecting different nano powder. For example, in order to solve lighting applications, alumina (Al2O3)-based nano-powders can be selected for sintering. To meet the requirements of the gain medium of high-powered lasers, doped YAG transparent nano-ceramics are generally selected. Therefore, by selecting different types of nano powders for preparation, we can obtain different types of transparent nano-ceramics. For example, some need to have high-temperature resistance, and some need to have a magneto-optical effect. Judging from the existing research on transparent nano-ceramics, those based on alumina (Al2O3) are popular research topics and have also been used in commercial applications that are closely related to the chemical and physical properties of Al. In addition, the doping of rare earth elements can significantly improve the performance of transparent nano-ceramics. The most typical example is the doping of rare earth elements in YAG transparent nano-ceramics, which can significantly improve their performance. Additionally, the Verdet constant of TAG transparent nano-ceramics can be changed by adding rare earth elements. 6.3. Comparison of Different Sintering Methods At present, the main sintering methods for preparing transparent nano-ceramics include HP, vacuum sintering, HIP, SPS, and microwave sintering. HP sintering is a high-pressure, low-strain-rate powder metallurgy process in which the creep process of sintering is controlled by applying heat and pressure. Due to the simultaneous application of force and heat, transparent nano-ceramics can thus be prepared at relatively low temperatures and achieve the desired density. High pressure can inhibit grain growth and induce plastic deformation to eliminate pores in grains, so the sintering mechanism under high pressure is completely different from that under normal pressure. Vacuum sintering belongs to the method of pressureless sintering. In a vacuum environment, a pressure difference is formed between the inside of the ceramic and the outside world, which helps the discharge of pores, reduces the porosity of the ceramic, makes the grains grow, and forms a high-density ceramic material. Vacuum sintering has the advantages of simple operation, low cost, and high production efficiency. It is currently the most widely used transparent ceramic sintering technology. Some oxide-based ceramic materials, such as rare-earth-doped YAG, Y2O3 and Al2O3, and other transparent nano-ceramic ceramic materials, can be prepared by vacuum sintering. For the traditional sintering process, HIP is a key step in the preparation of transparent nano-ceramics with high light transmittance, which can reduce the porosity inside the material and thus maximize the material density. To reduce manufacturing costs, HIP is usually used as the last step in the two-step sintering methods (HP + HIP sintering; vacuum sintering + HIP sintering). The SPS method is a new technique for heating and sintering by directly passing a pulse current between the nano-powder particles. Compared with the HIP method or the HP method, it has the distinctive features of a fast heating rate, short sintering time, controllable structure, energy savings, and environmental protection. In addition, since the sintering of transparent nano-ceramic materials requires the consideration of various factors, the SPS method can also be easily combined with optimization, such as doping of sintering aids, which makes its application range larger. During the microwave-sintering process, the processing material heats up very quickly, which can be carried out at low sintering temperatures and short sintering times to obtain nano-ceramics with high transparency. In contrast, microwave sintering has the advantages of a short fabrication time and low processing cost. 6.4. Relationship between Nano Ceramic Materials and Biomedicine Nano-ceramics have been successfully employed in medical diagnoses (nano biosensors, nano bioimaging) and medical therapy (nano drug loading, nano biomedical materials, nano biocompatible organs) due to their unique sensing and biological properties. Especially, nanoceramics have great application prospects in the manufacture of biocompatible organs (artificial organs, artificial blood vessels, and artificial bones) [121]. Abe et al. [122] evaluated the biocompatibility of several nano-ceramic particles (TiO2, In2O3, ITO, Y2O3: Eu, and CuO) with bone cells, tumor cells, and hepatocytes. The results showed that the nanoparticles could be safely used in industrial and biomedical applications. Manonmani et al. [123] found that a novel nano triphasic bioceramic composite could effectively improve the corrosion resistance and surface cell activity of orthopedic implants. Traditional medical materials and artificial organs and tissues made of various conventional materials have limited compatibility with the patient’s body in clinical use, so it is difficult to fundamentally solve the disease in the patient’s body. Nano-scale materials have good biocompatibility, which means they fit the structure of cells in the human body to a high degree and can successfully avoid problems such as postoperative trauma and infection [124]. At present, transparent nano-ceramics biocompatibility research and development is still limited, and there is still a long way to go before they are widely used in clinical practice. Furthermore, nanoparticles’ high compatibility and degradability with blood tissue will be a long-term study focus in the future. 7. Outlook Transparent nano-ceramics are a new class of materials, and various preparation strategies have been developed, especially using nano-powders, to obtain them with various compositions and properties to meet the needs of different applications. Here, we reviewed the research progress and potential applications of mainstream transparent nano-ceramics. In the future, the development of transparent nano-ceramics and their potential applications are anticipated as follows:(1) For the preparation of transparent nano-ceramics, high-purity and high-quality nano powder is very important, especially the average particle diameter of nano powder. At present, in the existing literature, many studies use the nano powder of commercial companies to prepare them, and the average particle diameter ranges from 50 to 100 nm (the average of 150 nm). However, in order to prepare high-quality transparent nano-ceramics, it is necessary to pretreat the existing nano powders, such as further improving the sintering activity of nano powders and reducing the average diameter of particles through laser processing or other methods, so that the light transmittance can be improved after the subsequent fabricating process. (2) The preparation of transparent nano-ceramics via new sintering technologies, such as spark plasma sintering and laser sintering, is still in the exploratory stage. The existing research on these sintering technologies is mainly based on experiments, and there is a lack of theoretical research to clarify the action mechanism of micro defects in the sintering process, especially to establish the thermodynamics and kinetics of sintering reaction. In addition, with the background of carbon peak and carbon neutralization, reducing the sintering temperature and sintering in a lower temperature range to obtain transparent nano-ceramics with excellent properties is still an important research direction in the future. (3) The preparation of transparent nano-ceramics via new sintering technologies, such as spark plasma sintering and laser sintering, is still in the exploratory stage. New sintering processes, such as spark plasma sintering and laser sintering, are continuously being investigated for the preparation of transparent nano-ceramics. Especially, the laser-induced plasma process, with the advantages of high-power plasma and an ultrafast laser, has potential for the fabrication of transparent nano-ceramics [125]. However, there is a shortage of theoretical studies to explain the action mechanism of micro defects in the sintering process and especially to establish the thermodynamics and kinetics of the sintering reaction in current studies. Furthermore, in consideration of the carbon peak and carbon neutralization, lowering the sintering temperature and sintering in a lower temperature range to create transparent nano-ceramics with good characteristics appears to be an important research direction. (4) Transmittance is one of the most important indicators for transparent nano-ceramics, and the microstructure is the most important component determining it. Transparent nano-ceramic samples supplemented with sintering aids have low porosity, allowing the grains to tightly fill the space and achieve an excellent microstructure. Different varieties of transparent nano-ceramics, in general, need different sintering aids. Furthermore, the doping ratio of sintering aids impacts the sintering process, affecting the transmittance for the same type of transparent ceramics. In the future, not only will trials be used to determine the best sintering aid, but computer simulations will also be used to save time and cost. (5) The powder particle size, sintering temperature, sintering duration, sintering environment, sintering pressure, and sintering aids are the primary elements impacting the sintering of transparent nano-ceramics. Aside from the standard light transmittance and physical properties, the indicators for evaluating transparent nano-ceramics must also take into account the manufacturing cost and environmental impact. As a result, sintering transparent nano-ceramics is an MIMO (multi-input, multi-output) process [126,127]. Multi-objective optimization of its preparation process is necessary for the future, which will help to lower its production costs and environmental impact. (6) Finally, theoretical research on transparent nano-ceramics is still in its early stages. For example, molecular dynamic models are efficient tools for studying the mechanism of action at the microscopic scale [128,129], which have been used to describe the microstructures of transparent glass ceramics [130,131]. As a result, researchers must develop suitable theoretical models to guide and optimize the transparent nano-ceramic preparation process. Then, the development of transparent nano-ceramic technology may be more promising if theoretical analysis and experimental results are combined. Author Contributions G.L. and D.S. outlined the structure of the paper, W.M. and Z.J. wrote the paper, Y.X., W.H., L.L. and Z.X. revised the paper. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Sketch of light transmission in a polycrystalline ceramic. Figure 2 Performance requirements and applications of transparent nano-ceramics. Figure 3 SEM images of Al2O3 (a), Y2O3 (b) starting nano powders, as well as 2.88 Y2O3–0.12Nd2O3–5Al2O3 powder mixture (c) and Y2O3 powders after planetary ball milling for 15 h (d) [16]. Reprinted with permission from Ref. [16]. Copyright 2014, copyright ELSEVIER. Figure 4 Schematic diagram of YAG transparent nano-ceramics’ application in laser diode pumping system. Figure 5 Microstructure of TGG transparent nano-ceramic and its transmittance spectra [39]; (a) SEM image showing the microstructure of a typical TGG transparent nano-ceramics sample; (b) Comparison of transmission spectra between TGG nano-ceramics samples and TGG single crystal. Reprinted with permission from Ref. [39]. Copyright 2011, copyright IEEE Xplore. Figure 6 SEM morphology of TGG powders sintered at 1100 °C (a), the surface of sample #1 sintered at 1500 (b), sample #2 sintered at 1550 °C (c), and sample #3 sintered at 1600 °C (d) [43]. Reprinted with permission from Ref. [43]. Copyright 2019, copyright ELSEVIER. Figure 7 Experimental results of depolarization as a function of laser power (red and blue circles represent results for TGG ceramics with and without magnetic field, respectively; squares represent calculated results for TGG single crystals; solid lines represent theoretical curves) [48]. Reprinted with permission from Ref. [48]. Copyright 2014, copyright Optical Society of America. Figure 8 Optical quality of TAG transparent nano-ceramics; (a) samples and (b) optical transmittance of polished TAG transparent nano-ceramics [49]. Reprinted with permission from Ref. [49]. Copyright 2011, copyright ELSEVIER. Figure 9 Scheme of the experiment of (a) thermal depolarization measurement and (b) studying thermally induced lens. Figure 10 Scheme of functional layers in a transparent armor concept with a four-layer design [65]. Reprinted with permission from Ref. [65]. Copyright 2017, copyright ELSEVIER. Figure 11 Sample particle size distribution and XRD pattern; (a) FEG-SEM micrograph, (b) XRD pattern, (c) TEM micrograph, and (d) particle size distribution of as-received spinel powder SPI-P1 [77]. Reprinted with permission from Ref. [77]. Copyright 2021, copyright ELSEVIER. Figure 12 Sample size and XRD pattern; (a) SEM images of the morphologies, (b) particle size distribution, and (c) XRD pattern measured of the AlON powders after ball milling [89]. Reprinted with permission from Ref. [89]. Copyright 2022, copyright ScienceDirect. Figure 13 Samples and transmittance curve; (a) samples doped with different SiO2 contents and different temperatures in the process of HIP; (b) transmittance of AlON transparent nano-ceramics doped with different SiO2 contents at 1810 °C for 3 h [92]. Reprinted with permission from Ref. [92]. Copyright 2021, copyright ScienceDirect. Figure 14 Transparent armor nano-ceramics; (a) MgAl2O4 transparent nano-ceramics, 12 mm thick, created through pre-sintering in air and HIP methods [83]. Reprinted with permission from Ref. [83]. Copyright 2014, copyright ScienceDirect. (b) a photograph revealing the transmittance of the Mg-γ-AlON transparent ceramic after HIP treatment [96]. Reprinted with permission from Ref. [96]. Copyright 2019, copyright ScienceDirect. Figure 15 Comparison of transparent armor materials for STANAG 4569; (a) Level 2 and (b) Level 3 [65]. Reprinted with permission from Ref. [65]. Copyright 2017, copyright ELSEVIER. Figure 16 SEM images of PMN-PT polycrystals; (a) 1275 °C, 30 min: average grain size (2.7 µm); (b) 1275 °C, 60 min: average grain size (3.2 µm); (c) 1300 °C, 90 min: average grain size (3.8 µm); (d) 1300 °C, 120 min: average grain size (4.3 µm) [98]. Reprinted with permission from Ref. [98]. Copyright 2019, copyright ScienceDirect. Figure 17 The properties of Mg:LuAG scintillation ceramics with different Ce contents (0.025~0.3 at.%) after sintering. (a) Transmittance of samples with different Ce contents; (b) Sintered state of different Ce content transmittances of Mg:LuAG ceramics; (c) Transmittance of annealed Mg:LuAG ceramics with different Ce contents [117]. Reprinted with permission from Ref. [117]. Copyright 2018, copyright ScienceDirect. Figure 18 Average diameter of nano powders in the reviewed transparent nano-ceramics [12,14,15,17,43,52,81,83,86,88]. nanomaterials-12-01491-t001_Table 1 Table 1 Summary of doped YAG transparent nano-ceramics described in the text grouped by doped type and published year. Doping Type Year, Powder, and Fabrication Findings Performance Remarks Nd3+-doped 1995; the starting nano powders include Y2O3 (60 nm), Al2O3 (400 nm), and Nd2O3 (500 nm); solid-state reaction method (Czochralski method). The average grain size and relative density of the 1.1 at. % Nd:YAG ceramics were about 50 μm and 99.98%, respectively. The optical scattering loss of Nd:YAG was about 0.9%/cm. Oscillation threshold of 309 mW and a slope efficiency of 28% [12]. For the first time, polycrystalline ceramics were successfully used for effective laser cutting. 2002; raw nano powder of oxide of aluminum, yttrium, and neodymium; ball milling -> slip casting -> vacuum sintering -> YAG transparent nano-ceramics. The pore volume concentration was 1 ppm, and the average diameter of particles was about 10 μm. The grain boundary width was only about 1 nm [13]. In laser experiment of Nd:YAG ceramic and single-crystal rods, the output powers of 88 W and 99 W were obtained, respectively [13]. Compared with single-crystal Nd:YAG, the light-to-light efficiency of Nd:YAG nanocrystalline ceramics needs to be further improved. 2010; an average particle size of 100 nm by reverse-strike precipitation method; HIP method. Freeze drying was proven to be an effective method to avoid caking and produced a material nano size distribution with uniform particles [14]. The infrared transmittance of the sample was 80%, and its emission spectrum was the same as 1 at.% Nd:YAG single crystal [14]. Nd: YAG nano powder with an average size of about 100 nm was prepared for the first time. 2021; high-purity powder mixture; cold isostatic pressing. The addition of TEOS promoted the densification of transparent ceramics. The transmittance of the 0.5 wt.% TEOS sample reached 75% in the near-infrared region [18]. The densification rate of Nd: YAG transparent ceramics could be adjusted by adding different wt.% TEOS so as to improve its transmittance. Ho-doped 2015; the nano powders were made up of near-spherical particles; solid-state reaction involving a pre-calcining stage. The transmittance in the infrared region was 82% [21]. The slope efficiency of laser oscillations in the fabricated Ho:YAG transparent ceramic sample for pumping power was 40% (at 1.85 μm) [21]. Based on the nano-powders prepared by laser ablation, Ho:YAG optical ceramics with finer particle size were prepared. 2018; uniform ceramic grain; HIP method. The total absorption spectral width was about 16 nm and suitable for pumping of diode lasers or fiber lasers, and the light-to-light efficiency was 52%. The in-band pumping method produced a 2117 nm laser with an output power of 24.6 W [22]. Further development of large-scale, YAG transparent ceramics with low Ho3+ doping concentration is required to alleviate the thermal effect during the lasing process. Er-doped 2018; high-purity 0.5 at.% Er3+:YAG powder; SPS+HIP methods. At 400 and 1100 nm wavelengths, the on-line transmission values were 75.8% and 82.7%, respectively [25]. The light–light efficiency of laser was 20%, and the maximum slope efficiency was 31% [25]. Transmission values of the Er3+: YAG transparent ceramics were lower than that of Er3+:YAG single crystals, which requires further improvement of the fabrication process. Tm-doped 2010. Light-to-light efficiency was 22%. Under an absorbed pump power of 2.21 W at 785 nm, an output power of up to 860 mW was produced [29]. Tm:YAG ceramic is a promising laser working medium. Higher power and efficiency can be achieved by using an improved laser cavity and an optimized transmission optical path. Yb-doped 2008. The transverse intensity distribution of the Yb:YAG ceramic laser beam was a Gaussian beam [31]. The ceramic Yb:YAG laser exhibited a continuous tunability at a maximum output power of 1.6 W [31]. Except for crystal Yb:YAG investigations, this was the first study of the tunability of ceramic Yb:YAG lasers. nanomaterials-12-01491-t002_Table 2 Table 2 Comparison of characteristics of Faraday isolators with different magneto-optical media [38,56]. Reprinted with permission from Ref. [38]. Copyright 2018, copyright CNKI. Reprinted with permission from Ref. [56]. Copyright 2017, copyright OSA. Medium Isolation Ratio@ Laser Power Isolation Ratio@ Laser Power Water Cooling TGG crystal 30 dB@650 W 6.5 m@340 W Optional TGG transparent nano-ceramic 30 dB@340 W 6.5 m@340 W Optional TAG transparent nano-ceramic 38 dB@300 W 8 m@300 W Required Ce: TAG transparent nano-ceramic (0.1 at.%) 31 dB@300 W 3.8 m@300 W Required nanomaterials-12-01491-t003_Table 3 Table 3 Summary of electro-optical transparent nano-ceramics described in the text grouped by publication year. Year, Authors Powder and Fabrication Findings Performance Remarks 1970, Haertling et al. [102] Preparation of optoceramics with a thickness of 1 mm using a two-step hot-pressing method. The density reached up to 99% of the standard density and the transmittance reached up to 80%. When the wavelength increased, the light transmission performance also gradually increased to 80%. The hot-pressing firing method could improve the density of ceramics, but the light transmission performance needed to be improved. 2016, Somwan et al. [103] Mixture of Bi2O3 and CuO; vibrating grinding and sintering. After adding mixed oxides, the sintering temperature decreased by nearly 50 °C to 1200 °C. At 1200 °C, the induced strain of the enhanced electric field reached 0.0079%. Higher dielectric constants could be obtained at lower sintering temperatures. 2017, Zhang et al. [105] Mixture of 3% lanthanum, 75% PMN, and 25% PT; two-step hot-pressing method. As the temperature increased, the half-wave voltage increased from 200 to 400 V. The electro-optic coefficient increased with an increase in temperature. The ferroelectric preparation process and transmittance could be controlled by temperature. 2018, Samanta et al. [104] Mixture of 69 ppm Fe3+ and 78 ppm Nb5+; sol–gel. Conductivity increased by two orders of magnitude as the sample changed from 100 Hz to 1 MHz. Conductivity was proportional to frequency. The conductivity could be controlled by controlling the magnitude of the frequency. 2018, Wang et al. [106] PMN-PT/CFO thin films; sol–gel spin coating. The temperature rose from 650 to 730 degrees Celsius; the leakage current decreased from 97.54 to 40.59. The ferroelectric properties decreased with an increase in the ratio of CFO to PMN-PT and increased with an increase in temperature. The coupling effect between the ferroelectric phase and the ferromagnetic phase was observed, which will pave the way for the preparation of multifunctional crystals in the future. 2021, Ze et al. [97] PMN-PT ceramic materials doped with Sm; two-step sintering method. When the Sm doping amount increased from 0 to 2%, the PMN-PT decreased from 3.15 to 3.05 eV. When there was Sm doping, it affected the size of the electro-optic coefficient. The optical power could be controlled by doping Sm. nanomaterials-12-01491-t004_Table 4 Table 4 Summary of scintillation transparent nano-ceramics described in the text grouped by publication year. Year, Authors Powder and Fabrication Findings Performance Remarks 2017, Zhou et al. [115] 99.9% strontium nitrate hydrate, 99.9% neodymium nitrate hydrate, and 99.9% potassium fluoride hydrate; the chemical precipitation method. SrF2 nanoparticles with Nd3+ doping concentrations up to 2 mol% kept a single cubic fluoride structure. The synthesized powder could prepare transparent ceramics with a transmittance of about 80% at 1060 nm. Nd3+ was successfully introduced into the SrF2 lattice, making it possible to use this material to make transparent ceramics. 2018, Yi et al. [116] Nd:(Ca0.94Gd0.06)2.06 nano powder; deionized water coprecipitation As the Nd3+ content increased from 0.5 to 5.0, the measured lifetime dropped sharply from 484.9 μs to 47.8 μs. The transparent ceramic had high transparency and an almost non-porous microstructure. The thermal conductivity of Nd:(Ca0.94Gd0.06)F2.06 transparent ceramics was better than that of traditional laser glass, and transparent ceramic is a promising laser material. 2019, Hostaša et al. [118] Industrial oxide powder, 0.3% Ce: GGAG,Ce0.009Gd2.991Al2Ga3O12 ceramic sample. At 1250 °C, the formation of the GGAG phase could be observed. This corresponds to the increase in shrinkage observed above 1200 °C, with the optimum at 1430 °C. TEOS was determined to be the most suitable sintering aid in the tests, providing the highest sample density and clarity. The optimum amount of sintering aid and the corresponding sintering cycle should be further investigated. 2020, Trofimov et al. [119] High-purity (99.99%) starting material and 0.5% tetraethyl orthosilicate (TEOS); the coprecipitation method. The Ce3+ comprehensive RL strength of single crystal increased 1.4 times from RT-450 °C, while polycrystalline ceramics increased 1.9 times from room temperature to 300 °C. Both single-crystal and polycrystalline ceramics exhibited high optical transparency up to about 2.5 eV. The LuAG:Ce scintillator could adapt to a wide temperature range and could be applied to many occasions. 2020, Bartosiewicz et al. [120] Mixure of 4 N-purity Lu2O3, Al2O3, and La2O3 oxides; the μ-PD method using RF induction heating. With the increase in La content, the main luminescence in the UV region gradually moved from 330 nm to 295 nm. Lu3Al5O12 produced strong luminescence in the deep ultraviolet spectral range LuAG: La doping significantly reduced the scintillation afterglow of LuAG: La crystals. Therefore, it is possible to generate new scintillators in the deep ultraviolet range. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Wang S. Zhang J. Luo D. Gu G. Tang D. Dong Z. Tan G. Que W. Zhang T. Li S. Transparent ceramics: Processing, materials and applications Prog. Solid State Chem. 2013 41 20 54 10.1016/j.progsolidstchem.2012.12.002 2. Xiao Z. Yu S. Li Y. Ruan S. Kong L. Huang Q. Huang Z. Zhou K. Su H. Yao Z. Materials development and potential applications of transparent ceramics: A review Mater. Sci. Eng. R Rep. 2020 139 100518 10.1016/j.mser.2019.100518 3. Pradhan A. Zhang K. Loutts G. Synthesis of neodymium added yttrium aluminum garnet (YAG) nanocrystalline powders leading to transparent ceramics Mater. Res. Bull. 2004 39 1291 1298 10.1016/j.materresbull.2004.04.005 4. Niihara K. New design concept of structural ceramics-ceramic nanocomposites J. Ceram. Soc. Jpn. 1991 99 974 982 10.2109/jcersj.99.974 5. Goldstein A. Krell A. Kleebe A. Transparent ceramics at 50: Progress made and further prospects J. Am. Ceram. Soc. 2016 99 3173 3197 10.1111/jace.14553 6. Merac M.R. Kleebe H.J. Müller M.M. Reimanis I.E. Fifty years of research and development coming to fruition; unraveling the complex interactions during processing of transparent magnesium aluminate (MgAl2O4) spinel J. Am. Ceram. Soc. 2013 96 3341 3365 10.1111/jace.12637 7. Cheng J. Dinesh A. Zhang Y. Rustum R. Microwave sintering of transparent alumina Mater. Lett. 2002 56 587 592 10.1016/S0167-577X(02)00557-8 8. Li S. Specialty Ceramic Technology J. Wuhan Univ. Technol. 2007 (In Chinese) 9. Drozdowski W. Brylew K. Witkowski M.E. Drewniak A. Masewicz A. Wojtowica A.J. Kisielewski J. Swirkowicz M. Effect of Lu-to-Y ratio and Mo coactivationon scintillation properties of LuYAG:Pr and LuAG:Pr,Mo crystals Opt. Mater. 2016 59 107 114 10.1016/j.optmat.2016.01.020 10. Jiang Y. Liu Q. Mao X. Su S. Li X. Liu X. Feng Y. Chen X. Li J. Influence of CaO on microstructure and properties of MgAl2O4 transparent ceramics Opt. Mater. 2020 111 110604 10.1016/j.optmat.2020.110604 11. Liu X. Chen F. Zhang F. Zhang H. Zhang Z. Wang J. Wang S. Huang Z. Hard transparent AlON ceramic for visible/IR windows Int. J. Refract. Met. Hard Mater. 2013 39 38 43 10.1016/j.ijrmhm.2013.01.021 12. Ikesue A. Kinoshita T. Kamata K. Yoshida K. Fabrication and Optical Properties of High-Performance Polycrystalline Nd:YAG Ceramics for Solid-State Lasers J. Am. Ceram. Soc. 1995 78 1033 1040 10.1111/j.1151-2916.1995.tb08433.x 13. Lu J. Ueda K. Yagi H. Yanagitani T. Akiyama Y. Kaminskii A.A. Alloys J. Neodymium doped yttrium aluminum garnet (Y3Al5O12) nanocrystalline ceramics-a new generation of solid state laser and optical materials J. Alloy. Compd. 2002 341 220 225 10.1016/S0925-8388(02)00083-X 14. Suárez M. Fernández A. Menéndez J. Nygren M. Torrecillas R. Zhao Z. Hot isostatic pressing of optically active Nd:YAG powders doped by a colloidal processing route J. Eur. Ceram. Soc. 2010 30 1489 1494 10.1016/j.jeurceramsoc.2009.12.006 15. Stevenson A. Kupp E. Messing G. Low temperature, transient liquid phase sintering of B2O3-SiO2-doped Nd:YAG transparent ceramics J. Mater. Res. 2011 26 1151 1158 10.1557/jmr.2011.45 16. Yavetskiy R. Baumer V. Doroshenko A. Kopylov Y. Kosyanova D. Kravchenkoc V. Parkhomenkoa S. Tolmacheva A. Phase formation and densification peculiarities of Y3Al5O12:Nd3+ during reactive sintering J. Cryst. Growth 2014 401 839 843 10.1016/j.jcrysgro.2014.01.034 17. Zhang X. Fan G. Lu W. Chen Y. Ruan X. Effect of the spark plasma sintering parameters, LiF additive, and Nd dopant on the microwave dielectric and optical properties of transparent YAG ceramics J. Eur. Ceram. Soc. 2016 36 2767 2772 10.1016/j.jeurceramsoc.2016.04.029 18. Jia W. Wei Q. Zhang H. Su C. Ren G. Zhao M. Ma C. Comparative analyses of the influence of TEOS additives on the sintering kinetics of Nd: YAG transparent ceramics J. Mater. Sci. Mater. Electron. 2021 32 19218 19229 10.1007/s10854-021-06443-y 19. Zhang W. Zhou J. Liu W. Li J. Wang L. Jiang B. Pan Y. Cheng X. Xu J. Fabrication, properties and laser performance of Ho:YAG transparent ceramic J. Alloy. Compd. 2010 506 745 748 10.1016/j.jallcom.2010.07.059 20. Yang H. Zhang J. Qin X. Luo D. Ma J. Tang D. Chen H. Shen D. Zhang Q. Polycrystalline Ho:YAG transparent ceramics for eye-safe solid state laser applications J. Am. Ceram. Soc. 2012 95 52 55 10.1111/j.1551-2916.2011.04953.x 21. Bagayev S. Osipov V. Vatnik S. Shitov V. Vedin I. Platonov V. Steinberg I. Maksimov R. Ho:YAG transparent ceramics based on nanopowders produced by laser ablation method: Fabrication, optical properties, and laser performance Opt. Mater. 2015 50 47 51 10.1016/j.optmat.2015.03.036 22. Zhao Y. Wang J. Yao W. Shao Z. Shen C. Yin D. Wang Y. Liu P. Zhou W. Tang D. High power Ho-doped sesquioxide ceramic laser in-band pumped by a Tm-doped all-fiber MOPA IEEE Photon. J. 2018 10 1502107 10.1109/JPHOT.2018.2820328 23. Zhang C. Shen D. Wang Y. Qian L. Zhang J. Qin X. Tang D. Yang X. Zhao T. High-power polycrystalline Er:YAG ceramic laser at 1617 nm Opt. Lett. 2011 36 4767 4769 10.1364/OL.36.004767 22179877 24. Zhang X. Shen D. Huang H. Liu J. Zhang J. Tang D. Fan D. Passively Q-switched 1617-nm polycrystalline ceramic Er:YAG laser using a Cr:ZnSe saturable absorber Appl. Phys. B 2015 120 305 309 10.1007/s00340-015-6137-6 25. Bigotta S. Galecki L. Katz A. Bohmler J. Lemonnier S. Barraud E. Leriche A. Eichhorn M. Resonantly pumped eye-safe Er3+:YAG SPS-HIP ceramic laser Opt. Express 2018 26 3435 3442 10.1364/OE.26.003435 29401871 26. Zou Y. Wei Z. Wang Q. Zhan M. Li D. Zhang Z. Zhang J. Tang D. High-efficiency diode-pumped Tm:YAG ceramic laser Opt. Mater. 2013 35 804 806 10.1016/j.optmat.2012.05.015 27. Liu P. Jin L. Liu X. Huang H. Zhang J. Tang D. Shen D. A diode-pumped dual-wavelength Tm, Ho: YAG ceramic laser IEEE Photon. J. 2016 8 1504007 10.1109/JPHOT.2016.2608661 28. Zhang W. Pan Y. Zhou J. Liu W. Li J. Jiang B. Cheng X. Xu J. Diode-pumped Tm:YAG ceramic laser J. Am. Ceram. Soc. 2009 92 2434 2437 10.1111/j.1551-2916.2009.03220.x 29. Zou Y. Zhang Y. Zhong X. Wei Z. Zhang W. Jiang B. Pan Y. Efficient Tm:YAG ceramic laser at 2 μm Chin. Phys. Lett. 2010 27 074213 30. Zhan M. Zou Y. Lin Q. Wang Z. Han H. Lu L. Wei Z. Zhang J. Tang D. Ti:saphire pumped passively mode-locked Tm:YAG ceramic laser Acta Phys. Sin. 2014 63 014205 10.7498/aps.63.014205 31. Nakamura S. Yoshioka H. Matsubara Y. Ogawa T. Wada S. Efficient tunable Yb:YAG ceramic laser Opt. Commun. 2008 281 4411 4414 10.1016/j.optcom.2008.05.009 32. Luo D. Jian Z. Xu C. Qin X. Tang D. Ma J. Fabrication and laser properties of transparent Yb:YAG ceramics Opt. Mater. 2012 34 936 939 10.1016/j.optmat.2011.04.017 33. Sanghera J. Kim W. Guillermo V. Brandon S. Colin B. Ceramic laser materials Opt. Mater. 2012 5 258 277 10.3390/ma5020258 28817044 34. Mandl A. Klimek D. Textron’s J-HPSSL 100 kW ThinZag laser program Conference on Lasers and Electro-Optics Optical Society of America San Jose, CA, USA 2010 35. Yuan J. Duan X. Yao B. Li J. Cui Z. Shen Y. Dai T. Ju Y. Li C. Kou H. Dual-end-pumped high-power Cr2+:ZnS passively Q-switched Ho:YAG ceramic laser Appl. Phys. 2015 119 381 385 10.1007/s00340-015-6086-0 36. Feizbakhsh M. Yazd N.S. Keshavarzi A. Doosti A. Anisotropic crystallization of YAG on the surface of glass by CO2 laser irradiation Crystengcomm 2022 24 2753 2758 10.1039/D1CE01292D 37. Li T. Zhou T. Cao Y. Cai Z. Zhao C. Yuan M. Zheng X. Huang G. Wang Z. Zhang L. Optical properties and energy transfer performances in high quality Cr,Nd:YAG transparent laser ceramics for solar pumped lasers Opt. Express 2022 30 8762 8776 10.1364/OE.454080 35299322 38. Li J. Dai J. Pan Y. Research progress on magneto-optical transparent ceramics J. Inorg. Mater. 2018 33 1 8 (In Chinese) 39. Yoshida H. Tsubakimoto K. Fujimoto Y. Mikami K. Kinoshita H. Optical properties and Faraday effect of ceramic terbium gallium garnet for a room temperature Faraday rotator Opt. Express 2011 19 15181 15187 10.1364/OE.19.015181 21934880 40. Wu Y. Sun Z. Feng G. Wang S. Xu L. Wu S. Preparation and properties of novel Tb3Sc2Al3O12 (TSAG) magneto-optical transparent ceramic J. Eur. Ceram. Soc. 2021 41 195 201 10.1016/j.jeurceramsoc.2021.09.006 41. Khazanov E. Investigation of faraday isolator and faraday mirror designs for multi-kilowatt power lasers High-Power Lasers Appl. 2003 4968 115 126 42. Yasuhara R. Tokita S. Kawanaka J. Kawashima T. Kan H. Yagi H. Nozawa H. Yanagitani T. Fujimoto Y. Yoshida H. Cryogenic temperature characteristics of Verdet constant on terbium gallium garnet ceramics Opt. Express 2007 15 11255 11261 10.1364/OE.15.011255 19547482 43. Jin W. Gai L. Chen J. Lin H. Li C. Su L. Wu A. Zeng F. Fabrication and magneto-optical properties of TGG transparent ceramics Phys. B 2019 555 96 98 10.1016/j.physb.2018.11.053 44. Li X. Snetkov I. Yakovlev A. Liu Q. Liu X. Liu Z. Chen P. Zhu D. Wu L. Yang Z. Fabrication and performance evaluation of novel transparent ceramics RE:Tb3Ga5O12 (RE = Pr, Tm, Dy) toward magneto-optical application J. Adv. Ceram. 2021 10 271 278 10.1007/s40145-020-0437-y 45. Moriana A.D. Zhang S. Determining the effects of BaTiO3 template alignment on template grain growth of Pb(Mg1/3Nb2/3)O3 -PbTiO3 and effects on piezoelectric properties J. Eur. Ceram. Soc. 2022 42 2752 2763 10.1016/j.jeurceramsoc.2022.01.062 46. Zhang H. Gao Y. Huang C. Chen Y. Dou R. Zhang Q. Electronic structure, optical dispersion and luminescence properties of terbium gallium garnet crystal Crystengcomm 2021 24 877 885 10.1039/D1CE01030A 47. Li X. Liu Q. Liu X. Chen X. Wu L. Xie T. Shi Y. Chen H. Li J. Novel (Tb0.99Ce0.01)3Ga5O12 magneto-optical ceramics for Faraday isolators Scr. Mater. 2019 177 137 140 10.1016/j.scriptamat.2019.10.023 48. Yasuhara R. Snetkov I. Starobor A. Zheleznov D. Palashov O. Khazanov E. Nozawa H. Yanagitani T. Terbium gallium garnet ceramic Faraday rotator for high-power laser application Opt. Lett. 2014 39 1145 1148 10.1364/OL.39.001145 24690692 49. Lin H. Zhou S. Teng H. Synthesis of Tb3Al5O12 (TAG) transparent ceramics for potential magneto-optical applications Opt. Mater. 2011 33 1833 1836 10.1016/j.optmat.2011.06.017 50. Kaminskii A. Eichler H. Reiche P. Uecker R. SRS risk potential in Faraday rotator Tb3Ga5O12 crystals for high-peak power lasers Laser Phys. Lett. 2005 2 489 492 10.1002/lapl.200510033 51. Chen C. Yi X. Zhang S. Feng Y. Tang Y. Lin H. Zhou S. Vacuum sintering of Tb3Al5O12 transparent ceramics with combined TEOS+MgO sintering aids Ceram. Int. 2015 41 12823 12827 10.1016/j.ceramint.2015.06.118 52. Duan P. Liu P. Xu X. Wang W. Wan Z. Zhang S. Wang Y. Zhang J. Fabrication of transparent Tb3Al5O12 ceramics by hot isostatic pressing sintering J. Am. Ceram. Soc. 2017 100 2893 2900 10.1111/jace.14810 53. Chen C. Zhou S. Lin H. Yi Q. Fabrication and performance optimization of the magneto-optical (Tb1-xRx)3Al5O12 (R = Y, Ce) transparent ceramics Appl. Phys. Lett. 2012 101 131908 10.1063/1.4756789 54. Chen C. Ni Y. Zhou S. Hui L. Yi X. Preparation of (Tb0.8Y0.2)3Al5O12 transparent ceramic as novel magneto-optical isolator material Chin. Opt. Lett. 2013 11 021601 10.3788/COL201311.021601 55. Snetkov I. Permin D. Balabanov S. Palashov O. Wavelength dependence of Verdet constant of Tb3+: Y2O3 ceramics Appl. Phys. Lett. 2016 108 161905 10.1063/1.4947432 56. Furuse H. Yasuhara R. Magneto-optical characteristics of holmium oxide (Ho2O3) ceramics Opt. Mater. Express 2017 7 827 833 10.1364/OME.7.000827 57. Morales J. Amos N. Khizroev S. Garay J. Magneto-optical Faraday effect in nanocrystalline oxides J. Appl. Phys. 2011 109 093110 10.1063/1.3583654 58. Hu D. Li X. Snetkov I. Yakovlev A. Balabanov S. Ivanov M. Liu X. Liu Z. Tian F. Xie T. Fabrication, microstructure and optical characterizations of holmium oxide (Ho2O3) transparent ceramics J. Non-Cryst. Solids 2021 41 759 767 10.1016/j.jeurceramsoc.2020.08.008 59. Furuse H. Yasuhara R. Hiraga K. Zhou S. High Verdet constant of Ti-doped terbium aluminum garnet (TAG) ceramics Opt. Mater. Express 2016 6 191 196 10.1364/OME.6.000191 60. Starobor A.V. Kuznetsov I.K. Palashov O.V. Pestov A.E. Chkhalo N.I. Faraday Isolator with composite magneto-optical TGG-sapphire elements IEEE J. Quantum Electron. 2021 57 7000208 10.1109/JQE.2021.3116984 61. Lin H. Jia H. Zhou L. Li N. Liu B. He J. Yao G. Li S. Zhou Y. Li C. Magneto-optical and fluorescence properties of Tb3+ doped glass-ceramics containing AlPO4 J. Non-Cryst. Solids 2022 579 121377 10.1016/j.jnoncrysol.2021.121377 62. Zhang L. Li X. Hu D. Liu Z. Xie T. Wu L. Yang Z. Li J. Fabrication and properties of transparentmagneto-optical ceramics J. Eur. Ceram. Soc. 2021 41 7208 7214 10.1016/j.jeurceramsoc.2021.07.024 63. Shbool M. Al-Abdallat Y. Jawwad A.K.A. Alsharairi S. Abu-Ghannam L. Abu-Khajil E.E. Badwan A. Examining the effect of nano-additions of rare earth elements on the hardness of body armor ceramic Indian J. Eng. Mater. S 2021 28 64 72 64. Yang X. Lai Y. Zeng Y. Yang F. Huang F. Li B. Wang F. Wu C. Su H. Spinel-type solid solution ceramic MgAl2O4-Mg2TiO4 with excellent microwave dielectric properties J. Alloy. Compd. 2022 898 162905 10.1016/j.jallcom.2021.162905 65. Tobias B. Sergio Y.G. de Oliveira A.P.N. Travitzky N. Hotza D. Transparent ceramic and glass-ceramic materials for armor applications—ScienceDirect Ceram. Int. 2017 43 13031 13046 66. Strassburger E. Ballistic testing of transparent armour ceramics J. Eur. Ceram. Soc. 2009 29 267 273 10.1016/j.jeurceramsoc.2008.03.049 67. Grujicic M. Bell W. Pandurangan B. Design and material selection guidelines and strategies for transparent armor systems Mater. Des. 2012 34 808 819 10.1016/j.matdes.2011.07.007 68. Subhash G. Transparent Armor Materials Exp. Mech. 2013 53 1 2 10.1007/s11340-012-9689-y 69. Salem J. Transparent armor ceramics as spacecraft windows transparent armor ceramics as spacecraft windows J. Am. Ceram. Soc. 2019 96 281 289 10.1111/jace.12089 70. Sepulveda J. Loutfy R. Chang S. Ibrahim S. High Performance Spinel Ceramics for IR Windows and Domes Proc. Spie Int. Soc. Opt. Eng. 2011 8016 801604 71. Gilde G. Patel P. Sands J. Patterson P. Blodgett D. Duncan D. Hahn D. Evaluation of hot isostatic pressing parameters on the optical and ballistic properties of spinel for transparent armor Am. Ceram. Soc. 2006 88 2747 2751 10.1111/j.1551-2916.2005.00527.x 72. Bratton R. Translucent sintered MgAl2O4 J. Am. Ceram. Soc. 1974 5 48 10.1111/j.1151-2916.1974.tb10901.x 73. Tsukuma K. Transparent MgAl2O4 spinel ceramics produced by HIP post-sintering J. Ceram. Soc. Jpn. 2006 114 802 806 10.2109/jcersj.114.802 74. Villalobos G. Sanghera J. Aggarwal I. Degradation of magnesium aluminum spinel by lithium fluoride sintering aid J. Am. Ceram. Soc. 2010 88 1321 1322 10.1111/j.1551-2916.2005.00209.x 75. Esposito L. Piancastelli A. Martelli S. Production and characterization of transparent MgAl2O4 prepared by hot pressing J. Eur. Ceram. Soc. 2013 33 737 747 10.1016/j.jeurceramsoc.2012.10.013 76. Esposito L. Piancastelli A. Miceli P. A thermodynamic approach to obtaining transparent spinel (MgAl2O4) by hot pressing J. Eur. Ceram. Soc. 2015 35 651 661 10.1016/j.jeurceramsoc.2014.09.005 77. Boulesteixa R. Goldsteinb A. Perrièrea C. Maîtrea A. Katzb M. Coureauc C. Salléd C. Transparent ceramics green-microstructure optimization by pressure slip-casting: Cases of YAG and MgAl2O4 J. Eur. Ceram. Soc. 2021 41 2085 2095 10.1016/j.jeurceramsoc.2020.11.003 78. Kagawa A. Effect of grain boundary microcracking on the light transmittance of sintered transparent MgAl2O4 J. Eur. Ceram. Soc. 2003 23 951 999 79. Zou Y. He D. Wei X. Yu R. Lu T. Chang X. Wang S. Li L. Nanosintering mechanism of MgAl2O4 transparent ceramics under high pressure Mater. Chem. Phys. 2010 123 529 533 10.1016/j.matchemphys.2010.05.009 80. Meir S. Kalabukhov S. Froumin N. Dariel M. Frage N. Synthesis and densification of transparent magnesium aluminate spinel by SPS processing J. Am. Ceram. Soc. 2010 92 358 364 10.1111/j.1551-2916.2008.02893.x 81. Liu Q. Jing Y. Su S. Li X. Liu X. Feng Y. Chen X. Li J. Microstructure and properties of MgAl2O4 transparent ceramics fabricated by hot isostatic pressing Opt. Mater. 2020 104 109938 10.1016/j.optmat.2020.109938 82. Liu Y. Zhu J. Dai B. Transparent MgAl2O4 ceramics prepared by microwave sintering and hot isostatic pressing—ScienceDirect Ceram. Int. 2020 46 25738 25740 10.1016/j.ceramint.2020.07.051 83. Krell A. Hutzler T. Klimke J. Defect strategies for an improved optical quality of transparent ceramics Opt. Mater. 2014 38 61 74 10.1016/j.optmat.2014.09.030 84. Mccauley J.W. Corbin N.D. High temperature reactions and microstructures in the Al2O3-AlN system Progress in Nitrogen Ceramics Springer Dordrecht, The Netherlands 1983 111 118 85. Yamaguchi G. Yanagida H. Study on the reductive Spinel-A new spinel formula AlN–Al2O3 instead of the previous one Al3O4 Bull. Chem. Soc. Jpn. 1959 32 1264 1265 10.1246/bcsj.32.1264 86. Qi J. Wang Y. Lu T. Yu Y. Pan L. Wei N. Wang J. Preparation and light transmission properties of AlON ceramics by the two-step method with nanosized Al2O3 and AlN Metall. Mater. Trans. A 2011 42 4075 4079 10.1007/s11661-011-0809-6 87. Jin X. Gao L. Sun J. Liu Y. Gui L. Highly transparent AlON pressurelessly sintered from powder synthesized by a novel carbothermal nitridation method J. Am. Ceram. Soc. 2012 95 2801 2807 10.1111/j.1551-2916.2012.05253.x 88. Zhao F. Qi J. Huang X. Guo X. Yu Y. Cao X. Wang Y. Wu D. Meng C. Lu T. Planetary ball-milling of AlON powder for highly transparent ceramics J. Am. Ceram. Soc. 2018 102 2377 2389 89. Zhang Y. Wu H. Jia B. Zhang D. Wang Y. Zhang Z. Qu X. Qin M. Transparent AlON ceramics by nitriding combustion synthesis precursors and pressureless sintering method Ceram. Int. 2022 Available online: 10.1016/j.ceramint.2022.03.075 (accessed on 31 March 2022) 10.1016/j.ceramint.2022.03.075 90. Yuan X. Zhang F. Liu X. Zhang Z. Wang S. Fabrication of transparent AlON ceramics by solid-state reaction sintering J. Inorg. Mater. 2011 26 499 502 (In Chinese) 10.3724/SP.J.1077.2011.00499 91. Shan Y. Sun X. Ren B. Wu H. Wei X. Olevsky E.A. Xu J. Li J. Pressureless sintering of highly transparent AlON ceramics with CaCO3 doping Scr. Mater. 2018 157 148 151 10.1016/j.scriptamat.2018.08.023 92. Li J. Zhang B. Tian R. Mao X. Zhang J. Wang S. Hot isostatic pressing of transparent AlON ceramics assisted by dissolution of gas inclusions J. Eur. Ceram. Soc. 2021 41 4327 4336 10.1016/j.jeurceramsoc.2021.02.035 93. Cheng X. Fu L. Su J. Liu W. Li Y. Mao P. Zhang L. Hui X. Research status and development trend of ALON ceramics Mater. Rep. 2020 S02 117 122 94. Mccauley J.W. Patel P. Chen M. Gilde G. Strassburger E. Paliwal B. Ramesh K.T. Dandekar D.P. AlON: A brief history of its emergence and evolution J. Eur. Ceram. Soc. 2009 29 223 236 10.1016/j.jeurceramsoc.2008.03.046 95. Goldman L.M. Smith M. Ramisetty M. Kashalikar U. Jha S. Sastri S. Scale up of large ALON windows Window and Dome Technologies and Materials XVI SPIE Bellingham, WA, USA 2019 Volume 10985 109850F 96. Liu L. Zhang C. Takahashia K. Nishimurac T. Segawad H. Hirosakia N. Xie R. Uniform and fine Mg-γ-AlON powders prepared from MgAl2O4: A promising precursor material for highly-transparent Mg-γ-AlON ceramics J. Eur. Ceram. Soc. 2019 39 928 933 10.1016/j.jeurceramsoc.2018.10.028 97. Ze F. Jiang X. Tian X. Zheng F. Cheng M. Zhao E. Ye W. Qin Y. Zhang Y. Ultratransparent PMN-PT electro-optic ceramics and its application in optical communication Adv. Opt. Mater. 2021 9 2002139 98. Pramanika R. Sahukarb M.K. Mohanc Y. Praveenkumard B. Sangaward S.R. Arockiarajana A. Effect of grain size on piezoelectric, ferroelectric and dielectric properties of PMN-PT ceramics Ceram. Int. 2019 45 5731 5742 10.1016/j.ceramint.2018.12.039 99. Slodczyk A. Colomban P. Probing the nanodomain origin and phase transition mechanisms in (Un)Poled PMN-PT single crystals and textured ceramics Materials 2010 32 5007 5028 10.3390/ma3125007 100. Lu G. Exploration of Optical Energy Storage and Nir Wavelength Conversion in Rare Earth Doped Relaxation Ceramics Harbin Institute of Technology Harbin, China 2013 O432.1 (In Chinese) 101. Wei Z. Optical characteristics of Er3+-doped PMN-PT transparent ceramics Ceram. Int. 2014 38 3397 3402 (In Chinese) 10.1016/j.ceramint.2011.12.051 102. Haertling G.H. Land C.E. Mckinney I.D. Ferroelectric Ceramic Optical Retardation Devices U.S. Patent No. 3512864 19 May 1970 3512864 103. Somwan S. Ngamjarurojana A. Limpichaipanit A. Dielectric, ferroelectric and induced strain behavior of PLZT 9/65/35 ceramics modified by Bi2O3 and CuO co-doping Ceram. Int. 2016 42 10690 10696 10.1016/j.ceramint.2016.03.181 104. Samanta S. Sankaranarayanan V. Sethupathi K. Effect of Nb and Fe co-doping on microstructure, dielectric response, ferroelectricity and energy storage density of PLZT J. Mater. Sci. Mater. Electron. 2018 29 20383 20394 10.1007/s10854-018-0173-z 105. Zhang X. Jin Y. Yao Z. Ye Q. Qu R. Cai H. Temperature characteristic of electro-optic effect of PMN-PT transparent ceramic Chin. J. Lasers 2017 44 1101007 10.3788/CJL201744.1101007 106. Wang A.-P. Feng M. Wang W. Li H. Zhao X. Xu H. Ke H. Zhou Y. Influence of composition ratio on ferroelectric, magnetic and magnetoelectric properties of PMN-PT/CFO composite thin films J. Mater. Sci. Mater. Electron. 2018 29 10164 10169 10.1007/s10854-018-9065-5 107. Zhang X. Ye Q. Cai H. Qu R. Polarization-independent electro-optic modulator based on PMNT electrically-controlled birefringence effect and Sagnac interferometer Opt. Laser Technol. 2014 57 5 8 10.1016/j.optlastec.2013.09.027 108. Qiao L. Ye Q. Gan J. Cai H. Qu R. Optical characteristics of transparent PMNT ceramic and its application at high speed electro-optic switch Opt. Commun. 2011 284 3886 3890 10.1016/j.optcom.2011.04.063 109. Chen X. Chen R. Chen Z. Chen J. Shung K. Zhou Q. Transparent lead lanthanum zirconate titanate (PLZT) ceramic fibers for high-frequency ultrasonic transducer applications Ceram. Int. 2016 42 18554 18559 10.1016/j.ceramint.2016.08.195 110. Li C. Luo W. Liu X. Xu D. He K. PMN-PT/PVDF nanocomposite for high output nanogenerator applications Nanomaterials 2016 6 67 10.3390/nano6040067 111. Nikl M. Mihokova E. Mares A. Vedda A. Blazek K. Traps and timing characteristics of LuAG:Ce3+ scintillator Phys. Status Solidi 2000 181 10 12 10.1002/1521-396X(200009)181:1<R10::AID-PSSA999910>3.0.CO;2-9 112. Zhou Z. Li W. Song J. Mei B. Yi G. Yang Y. Application of Judd-Ofelt theory in analyzing Nd3+ doped SrF2 and CaF2 transparent ceramics J. Eur. Ceram. Soc. 2019 39 2446 2452 10.1016/j.jeurceramsoc.2019.02.033 113. Wu T. Wang L. He H. Wang H. Shi Y. Research Progress of Lu3Al5O12-based Scintillation Ceramics Chin. J. Lumin 2021 41 917 937 10.37188/CJL.20210049 114. Blasse G. Scintillator materials Chem. Mater. 1994 6 1465 1475 10.1021/cm00045a002 115. Zhou Z. Li W. Song J. Yi G. Su L. Synthesis and characterization of Nd3+ doped SrF2 nanoparticles prepared by precipitation method Ceram. Int. 2017 44 4344 4350 10.1016/j.ceramint.2017.12.028 116. Yi G. Li W. Song J. Mei B. Zhou Z. Su L. Structural, spectroscopic and thermal properties of hot-pressed Nd:(Ca0.94 Gd0.06)F2.06 transparent ceramics J. Eur. Ceram. Soc. 2018 38 3240 3245 10.1016/j.jeurceramsoc.2018.03.016 117. Chen X. Hu Z. Cao M. Hu C. Liu S. Chen H. Shi Y. Kou H. Xie T. Vedda A. Influence of cerium doping concentration on the optical properties of Ce,Mg:LuAG scintillation ceramics J. Eur. Ceram. Soc. 2018 38 3246 3254 10.1016/j.jeurceramsoc.2018.03.023 118. Hostaša J. Cova F. Piancastelli A. Fasoli M. Zanelli C. Vedda A. Biasini V. Fabrication and luminescence of Ce-doped GGAG transparent ceramics, effect of sintering parameters and additives Ceram. Int. 2019 45 23283 23288 10.1016/j.ceramint.2019.08.025 119. Trofimova A. Jacobsohn L. Radioluminescence of Lu3Al5O12:Ce single crystal and transparent polycrystalline ceramic at high temperatures Ceram. Int. 2020 46 26335 26338 10.1016/j.ceramint.2019.12.247 120. Bartosiewicza K. Horiaib T. Yamajib A. Yoshikawab A. Kurosawab S. Yoshinob M. Zorenkoa Y. Effects of La doping on the crystal growth, phase stability and scintillation properties of Lu3Al5O12 single crystals Mater. Sci. Eng. B-Adv. 2020 261 114677 10.1016/j.mseb.2020.114677 121. Pagano S. Lombardo G. Costanzi E. Balloni S. Bruscoli S. Flamini S. Coniglio M. Valenti C. Cianetti S. Marinucci L. Morpho-functional effects of different universal dental adhesives on human gingival fibroblasts: An in vitro study Odontology 2021 109 524 539 10.1007/s10266-020-00569-x 33211211 122. Abe S. Seitoku E. Iwadera N. Hamba Y. Yamagata S. Akasaka T. Kusaka T. Inoue S. Yawaka Y. Iida J. Estimation of Biocompatibility of Nano-Sized Ceramic Particles with Osteoblasts, Osteosarcomas and Hepatocytes by Static and Time-Lapse Observation J. Biomed. Nanotechnol. 2016 12 472 480 10.1166/jbn.2016.2190 27280245 123. Manonmani R. Novel nano triphasic bioceramic composite coating on 316L SS by electrophoretic deposition process for enhanced corrosion resistance and cell proliferation J. Aust. Ceram. Soc. 2021 57 205 213 10.1007/s41779-020-00526-5 124. Leyva J.H.R. Vitale G. Hethnawi A. Hassan A. Perez-Zurita M.J. Ruiz-Esparza G.U. Nassar N.N. Mechanism of hierarchical porosity development in hexagonal boron nitride nanocrystalline microstructures for biomedical and industrial applications ACS Appl. Nano Mater. 2018 1 4491 4501 10.1021/acsanm.8b00765 125. Zhang Z. Zhang Y. Liu D. Zhang Y. Zhao J. Zhang G. Bubble behavior and its effect on surface integrity in laser induced plasma micro-machining silicon wafer ASME J. Manuf. Sci. Eng. 2022 accepted 10.1115/1.4054416 126. Ming W. Shen F. Zhang G. Liu G. Du J. Chen Z. Green machining: A framework for optimization of cutting parameters to minimize energy consumption and exhaust emissions during electrical discharge machining of Al 6061 and SKD 11 J. Clean. Prod. 2021 285 124889 10.1016/j.jclepro.2020.124889 127. Zhang Y. Zhang G. Zhang Z. Zhang Y. Huang Y. Effect of assisted transverse magnetic field on distortion behavior of thin-walled components in WEDM process Chin. J. Aeronaut. 2022 35 291 307 10.1016/j.cja.2020.10.034 128. Huang H. Zhigilei L.V. Computational study of laser fragmentation in liquid: Phase explosion, inverse Leidenfrost effect at the nanoscale, and evaporation in a nanobubble Sci. China Phys. Mech. 2022 online 10.1007/s11433-021-1881-8 129. Ming W. Zhang S. Zhang G. Du J. Ma J. He W. Cao C. Liu K. Progress in modeling of electrical discharge machining process Int. J. Heat Mass Transfer. 2022 187 122563 10.1016/j.ijheatmasstransfer.2022.122563 130. Mauroy V. Blanc W. Ude M. Trzesien S. Chaussedent S. Erbium-doped transparent glass ceramic optical fibres: Characterization using mass spectroscopy and molecular dynamics modeling Proceedings of the 2012 Photonics Global Conference Singapore 13–16 December 2012 IEEE Piscataway, NJ, USA 2013 (In Singapore) 131. Xu X. Zhao J. Chen X. Xu Q. Wang T. Yu S. Xu X. Qiao S. Du J. Fan X. Ca2+/Sr2+/Ba2+ dependent phase separation, nanocrystallization and photoluminescence in fluoroaluminosilicate glass J. Am. Ceram. Soc. 2020 103 5796 5807 10.1111/jace.17299

PMC009xxxxxx/PMC9099614.txt

==== Front Int J Environ Res Public Health Int J Environ Res Public Health ijerph International Journal of Environmental Research and Public Health 1661-7827 1660-4601 MDPI 10.3390/ijerph19095623 ijerph-19-05623 Article Use of Tobacco and Nicotine Products among Young People in Denmark—Status in Single and Dual Use https://orcid.org/0000-0003-0310-5871 Bast Lotus Sofie * Klitgaard Marie Borring https://orcid.org/0000-0001-6883-8764 Kjeld Simone Gad Jarlstrup Nanna Schneekloth Christensen Anne Illemann Tchounwou Paul B. Academic Editor National Institute of Public Health, University of Southern Denmark, Studiestraede 6, 1455 Copenhagen, Denmark; mboa@sdu.dk (M.B.K.); simk@sdu.dk (S.G.K.); nasc@sdu.dk (N.S.J.); anch@sdu.dk (A.I.C.) * Correspondence: loni@niph.dk 05 5 2022 5 2022 19 9 562304 3 2022 29 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Lots of new tobacco or nicotine products are being launched, e.g., e-cigarettes and smokeless tobacco, which appeal especially to the youngest part of the population. For example, the use of smokeless tobacco among Danish youth rose from approx. 2% in 2010 to 9% in 2020. Hence, there is an urgent need to follow and intervene against youth tobacco or nicotine product use. This study explored the current use of cigarettes, e-cigarettes, heated tobacco, and smokeless tobacco among Danish 15- to 29-year-olds. Further, we examined the concurrent use of two products or more. We used a nationwide survey conducted among 15- to 29-year-olds in February and March 2020. Overall, approx. 35,700 individuals received the questionnaire of which 35.5% responded (n = 13,315). One out of five (20.1%) smoked cigarettes, half of them daily, the other half occasionally. About one in twenty (3.9%) used e-cigarettes (daily or occasionally), and more than one in three (31.6%) had tried e-cigarettes. The use of heated tobacco among Danish youth is still relatively limited (0.3%). In comparison, about 9% used smokeless tobacco (daily or occasionally). Overall, 27.0% stated that they use at least one type of tobacco or nicotine product, while 5.6% used more than one product. Monitoring tobacco-related behavior in youth provides extremely important information for, e.g., policymakers and health professionals. tobacco cigarette nicotine product e-cigarette smokeless tobacco snus nicotine pouches heated tobacco dual use youth adolescents young adults tobacco prevention Danish Health Authority and the TrygFoundationNational Institute of Public Health, University of Southern DenmarkDanish Cancer Society, the Heart Association, and the Lung AssociationData collection and preparation of the publication is supported by the Danish Health Authority and the TrygFoundation. The study is anchored at the National Institute of Public Health, University of Southern Denmark, in collaboration with the Danish Cancer Society, the Heart Association, and the Lung Association. ==== Body pmc1. Introduction The global tobacco market has expanded widely in the last decade; a range of new tobacco- or nicotine-containing products have been introduced as alternatives to cigarettes—with increasing popularity, especially among youth in the Western world [1,2,3]. Moreover, international trends show a rise in the use of various tobacco or nicotine products simultaneously, i.e., dual and multiple use [4,5]. In Denmark, these current trends among youth have received significant attention from politicians and have raised public health concerns; Decades of decreasing smoking prevalence among the general population was in 2017 replaced by stagnation, and in certain groups of the Danish youth, even an increase; e.g., among 16- to 25-year-olds [6,7]. However, a recent Danish report showed that daily—but not occasional—smoking among 18–19-year-olds decreased again [8]—whether this decline is connected to a concurrent increase in the use of other tobacco or nicotine products has not yet been addressed. Internationally, in some high-income countries, the use of other tobacco or nicotine-containing products has exceeded the use of cigarettes among the youth population [3,9,10]. Further, the use of more than one tobacco or nicotine product is becoming prevalent. For example, among U.S. high school students in 2021, the most commonly cited product was e-cigarettes, which are currently used by 11.3% of high school students. Followed by cigarettes, cigars, smokeless tobacco, hookahs, nicotine pouches, heated tobacco products, and pipe tobacco. Overall, 13.4% were current (past 30 days) users of any product [10]. In a sample of adolescents (13- to 17-years) in Norway—a country similar to Denmark living wise, both culturally and economically—31.0% said that they had tried at least one tobacco product; among them, snus was twice as prevalent as both cigarettes and e-cigarettes [3]. Though the shift from cigarettes to other tobacco products may cause a decrease in cigarette smoking, the use of any tobacco product constitutes a threat to both the individual and the overall public health; the hazards of not only tobacco but also nicotine have become more evident in recent years, i.e., it causes increased risk in cardiovascular and respiratory illnesses, and impact the reproductive system [11]. Moreover, nicotine negatively impacts the development of the adolescent brain [12,13]. Further, experimenting with products such as e-cigarettes in youth may work as a gateway to cigarette smoking [14,15]. E-cigarettes, smokeless tobacco (i.e., snus, chewing tobacco, and nicotine pouches), and heated tobacco products are among the products that seem to gain increasing popularity among the youth worldwide [16]. A recent Danish report sought to map out the use of these products among the general population in Denmark [17]—however, the use among Danish youth is still not well-documented. Neither is the concurrent use of two or more of the products. Although, for years, tobacco regulations have been lenient in Denmark, a new law went into effect. By 1 April 2020, tobacco prices in Denmark increased from approx. DKK 40 to 55 for a package with 20 cigarettes (corresponding to a change from USD 6 to 8). In 2022, the prices will be increased by an additional DKK 5 (approx. USD 1) so that the overall increase in prices from 2020 to 2022 will be 50%. Furthermore, by 1 January 2021, a new law with multiple tobacco preventive initiatives became effective. The law comprises a range of initiatives; among them are a ban on additives flavors such as fruit, menthol, and mint flavor; a ban on promotion at the point of sale (POS); standardized tobacco packaging; health warnings on all nicotine-containing products; increased age control at POS; smoke-free school time; and an extended ban against advertising and sponsorship of tobacco, e-cigarettes, and nicotine-containing products. In response to the current concerns in terms of youth tobacco and nicotine product use, as well as the limited in-depth knowledge of concurrent product use, a nationwide Danish study entitled “§SMOKE—A study of tobacco, behavior and regulations” was initiated. The baseline survey was conducted before the implementation of any of the new regulations and hence will be used as a comparison when following trends in the years to come. The aims of the §SMOKE study are to follow the tobacco and nicotine product use among the Danish youth as well evaluate the effect of the stronger regulations with repeated cross-sectional surveys among 15-to-29-year-olds in the years 2020 until 2025. This study reports on current tobacco or nicotine product use (i.e., use of cigarettes, e-cigarettes, smokeless tobacco, and heated tobacco) and concurrent use of products in 2020. 2. Materials and Methods 2.1. Study Design This study is based on baseline data from “§SMOKE—A study of tobacco, behavior and regulations”. The baseline survey is part of a larger evaluation running from 2020 until 2025. The National Institute of Public Health is responsible for the data collection with funding from the Danish Health Authority and the TrygFoundation. The repeated yearly data collections in the years from 2021 until 2025 are carried out by the National Institute of Public Health in collaboration with The Danish Cancer Society, The Heart Association, and The Lung Association, with financial support from the TrygFoundation. 2.2. Participants and Sample Size Study participants were a nationally representative sample of Danish 15–29-year-olds with permanent residence in Denmark at the time of study. All Danish citizens have a unique identification number registered in The Danish Civil Registration System that was used to draw the random sample [18]. The current smoking prevalence among Danish youth at the time of study was 26% (daily or occasionally) [17]. To ensure that essential analyses could be made, including investigating changes in tobacco patterns over the study period as well as subgroup- and stratified analyses, the requested sample size was set to 15,000 respondents. With an expected response rate in this age group at 40%, we invited 37,500 persons to the study [19]. 2.3. Data Collection Data collection was conducted from February to March 2020. The majority of participants (98%) received a secure electronic e-mail with a link to the survey, while two percent of participants received a postal letter with a paper questionnaire, including a link to the web survey. Two reminders were sent to all non-responders. Topics in the questionnaire were related to the patterns of tobacco and nicotine product use as well as topics specifically related to three main components of the new tobacco law to reduce tobacco uptake among the youth in Denmark, especially with focus towards increased tobacco prices, POS display ban, and standardized packaging [14]. In total, 37,482 individuals received the questionnaire of which 13,315 returned valid responses (response rate = 35.5%), see Figure 1: Flow diagram. 2.4. Sociodemographic Characteristics and Weighting Procedure Weights were constructed using auxiliary information from Statistics Denmark’s registers to account for the possible selection bias in which participants responded to the survey. These weights were based on information on gender and age, which were the primary factors that differed between respondents and non-respondents (See also Table 1). Further, we examined the distribution of responses and non-responses according to region in Denmark and found that the distributions did not markedly differ. The weights ensured representability of the study sample and reduced the impact of non-responses (for more detailed information, see also [19]). 2.5. Measures We received data on gender and age from the Civil Registration System [18]. The remaining variables used for this study were obtained by participants’ self-reported answers to the questionnaire. In Table 2, an overview of the variables used for the current study is shown, including the item, response categories, and coding. 2.6. Analyses We examined possible differences in tobacco use (i.e., use of cigarettes, e-cigarettes, smokeless tobacco, and heated tobacco) according to age groups and gender using χ2-tests. A p-value of <0.05 was considered statistically significant. Moreover, we conducted descriptive analyses for dual use, including the proportion of the youth that daily or regularly use two or more of the included tobacco products (i.e., cigarettes, e-cigarettes, smokeless tobacco, and heated tobacco). Finally, we examined the type of tobacco or nicotine product used first stratified by gender and age groups. We used the Stata version 16 for all data analyses. Percentages presented in tables and figures are weighted, whereas the number of respondents is not. Further, numbers of five or fewer observations are reported as n/a. 3. Results One-fifth (20.1%) said that they smoked cigarettes, half of them daily (Table 3). Prevalences of cigarette smoking were higher among men compared to women and most prevalent among 18–24-year-olds. In the youngest age group (15–17 years), the majority occasionally smoked, while daily smoking was most common in the oldest group (25–29 years). Overall, 3.9% stated that they used e-cigarettes (daily or occasionally), with twice as many men using e-cigarettes compared with women. Using e-cigarettes was more prevalent in the two youngest age groups, e.g., more than one-third (37.5%) of the 18- to 24-year-olds stated that they had tried e-cigarettes. About 9% of respondents used smokeless tobacco (daily or occasionally). More men compared to women indicated to be current users, previous users, or having tried smokeless tobacco. Further, the use of smokeless tobacco was highest among the two youngest age groups (15–17 and 18–24-year-olds). The use of heated tobacco among participants was relatively limited, with less than a half percent stating to use this product. No gender differences were observed, but the use of heated tobacco was most prevalent among 18- to 24-year-olds. Overall, more than one-fourth (27.0%) said that they used at least one type of tobacco or nicotine product (Table 3). For men, the proportion was almost one in three (31.1%), while it was around one in fourth among women (22.8%). Further, the prevalence was highest among 18- to 24-year-olds (31.3%). Moreover, a high proportion had ever tried one or more tobacco products (71.3%), with the highest prevalences detected among the oldest age group (78.1%); however, as many as 50.4% of the 15–17-year-olds reported ever use of tobacco or nicotine product. The proportion of respondents currently using more than one product was 5.6% (data not shown). Among these, 87.2% were dual users and 12.8% were multiple product users (three or more products; Table 4). The most common combinations were cigarettes combined with e-cigarettes as well as cigarettes combined with smokeless tobacco. The combination of cigarettes and e-cigarettes was most prevalent among females, whereas the combined use of cigarettes and snus was mostly applied among men. Using more than two products was more prevalent among men compared to women (14.6 vs. 8.8%). Most respondents said that the first product they tried was cigarettes; more than 80% in both genders (Figure 2). Overall, about 15% stated that they first tried one of the other tobacco or nicotine products. Almost one in ten men tried e-cigarettes first, and 8.4% smokeless tobacco. For women, the numbers were a little lower. No women and almost no men (too few to report) tried heated tobacco first. Starting with e-cigarettes or smokeless tobacco was twice as prevalent among the 15- to 17-year-olds compared with the 18- to 24-year-olds (Figure 3), and almost none of the 25- to 29-year-olds started with these products (there were too few to report the specific numbers). 4. Discussion The results of this study indicate that the use of tobacco and nicotine products constitutes a significant public health issue in Denmark. Overall, cigarette smoking was most prevalent among Danish youth, followed by significant use of smokeless tobacco, fewer using e-cigarettes, and very few using heated tobacco. This contrasts with findings from the US, where e-cigarette use is more prevalent than smoking among youth [20], and findings from Norway, where smokeless tobacco use has exceeded the use of cigarettes [21,22]. As shown in this study, more than one out of four (27.0%) of Danish youth are current users of at least one tobacco or nicotine-containing product, and 5.4% use two or products or more. In comparison, there were 23.6% of the 14–18-year-olds currently using at least one product in the US in 2020 [20,23]. Men are more prone than women to use more than two tobacco- or nicotine products, and the combination of dual use differed between genders; women more often used e-cigarettes and cigarettes, whereas men used smokeless tobacco and cigarettes. Most respondents who currently use more than one product started with smoking cigarettes. However, especially among the youngest participants (15- to 17-year-olds)—a significant proportion reported first trying e-cigarettes (16.3%) and smokeless tobacco (16.2%). Among the 18- to 24-year-olds, there were also some who first initiated other products than cigarettes, but very few in the oldest age group (25- to 29-year-olds)—too few to report on actual prevalences. Obviously, the timing of introducing the different types of products influences which products are first used, i.e., e-cigarettes were just introduced to the market when respondents from the oldest age group initiated a tobacco use in their teenage years, and the smokeless product known as nicotine pouches have been on the market for only a couple of years. However, we cannot leave out the possibility of a trend towards starting with using other tobacco products than cigarettes. If youth tobacco habits in Denmark will follow international trends, we may detect declining smoking prevalences in the coming years but also increases in the use of other tobacco products. Recent numbers among Danish adolescents indicate a decline in daily cigarette smoking among 18–19-year-olds [8]. We should pay very close attention to the development of smoking in the coming years and not at least to mechanisms such as possible gateway effects of e-cigarette use leading to cigarette smoking [24]. The obvious risk is being addicted to nicotine—however, another important mechanism is that adolescents learn the habits and rituals of smoking through using e-cigarettes, i.e., the body language, the habits of taking smoking breaks, and handling a tobacco product, which may ease the path to cigarette smoking [24]. Further, research has found that e-cigarettes themselves constitute a significant health risk [25]. In Norway, the declining smoking prevalence in youth was accompanied by the increasing use of smokeless tobacco [21,22]. Whether this trend will be adopted by the Danish youth should be followed carefully in the years to come. Smokeless tobacco, such as snus and nicotine pouches, seem to be particularly popular in Scandinavian countries. A study from 2019 found that 14% of Norwegian adolescents used snus (regularly or occasionally), and 16% had tried snus [3]. Among Finnish adolescents, 11% were current users of snus, while 9% had tried using snus [26]. Among Swedish youth (<25 years), 12.0% were current users of smokeless tobacco. In another study across 17 European countries, 1.4% on average used smokeless tobacco [27]. In our study, 9.1% were currently using smokeless tobacco, while 27.3% had tried smokeless tobacco. Thus, a significant proportion of young people in Denmark have experience with smokeless tobacco (either as current users or as having tried using it) compared to the average of youth in European countries. The current use of smokeless tobacco in Danish youth seems to correspond with current use among youth in other Scandinavian countries, e.g., Norway, Finland, and Sweden. For Norwegian youth, the use of smokeless tobacco—with or without concurrent cigarette smoking—resulted in a higher risk of adult smoking as well as using smokeless tobacco in adulthood [21]. The use of heated tobacco products in the Danish youth seems to be quite similar to other countries. In a sample of middle- and high-school students in the US, the overall percentage of ever using heated tobacco was 2.4%, and current use was 1.6% [28]. Among Korean adolescents, ever use was 2.8% [29]. In comparison, we found that ever use was 3.2% and current use 0.3%. The products are marketed as less harmful than tobacco products with the risk that especially young people find them attractive and start using them [30,31]. In a study of 16- to 19-year-olds across Canada, England, and the USA, 7.0% reported awareness of heated tobacco, and 38.6% expressed interest in trying this product [32]. Youth, who were currently smoking or had previous experience with smoking, seemed to pay more attention to heated tobacco and more interest in trying the products. Further, males, cigarette smokers, and e-cigarette users had higher susceptibility to trying heated tobacco [28,32]. 4.1. Methodological Issues The strengths of this study are the large sample size drawn randomly from a national register and the questionnaire with multiple tobacco and nicotine-containing products, as well as other tobacco-related items enabling the study of patterns and trends—also within subgroups. The study sample size is large enough to examine subgroups in dual tobacco product use, which most other studies are not. We examined differences among participants and non-participants and used weighting for age and gender to account for possible bias in the responses due to these factors. Further, this study is the first in a row of cross-sectional studies in the §SMOKE study, which altogether contributes to the evaluation of the tobacco regulations in Denmark. The questionnaire was sent by secure electronic mail, which most people living in Denmark can receive. Moreover, persons without electronic mail received a postal letter. Two reminders were sent to the whole sample size to optimize the number of respondents. The response rate was 35.5%, which is in accordance with response rates among youth in other Danish population-based studies. In Denmark, we have experienced an overall trend of declining participation proportion in health surveys over the past 20–30 years, and the rather low participation might be a result of questionnaire fatigue among youth, which was also seen in other studies [33]. The reliability of self-reported survey data is based on confidence in the accuracy of the respondents’ recall as well as on their motivation to provide truthful information on the topic of interest. Youth behaviors were self-reported, with the risk of social desirability bias. However, previous research shows good correspondence between adolescent self-reported smoking status and biological measures [34,35]. Another general limitation of survey data is the cross-sectional design, which does not allow conclusions to be drawn on the direction of causality; however, the follow-ups planned in §SMOKE will allow for examinations of trends. 4.2. Implications Data from the §SMOKE study provides new important knowledge, with the possibility to monitor youth tobacco and nicotine product use over a period with multiple new regulations being implemented. The questionnaires cover a wide variety of topics not included in official statistical registers. For future research, the data derived from the §SMOKE surveys can be linked on an individual level to different official statistical registers (e.g., the Danish National Patient Register, the Danish Register of Causes of Death, The Danish National Prescription Register, and the Danish National Service Register) due to the unique personal registration numbers, which allows for analyses of the relationship between, e.g., risk factors and morbidity and mortality, social inequality in health, etc. 5. Conclusions Monitoring tobacco product use in youth provides extremely important information for policymakers and health professionals. This paper shows a continued need for regulation to prevent Danish youth from initiating the use of one or more of the addictive and health-damaging tobacco and nicotine-containing products. Acknowledgments We thank the Tryg Foundation for funding the §SMOKE study, and this manuscript. Further, we want to thank all participants for their answers to the questionnaire. Further, a thank to the Danish Cancer Society, Heart Association and Lung Association for qualified project group work. Author Contributions L.S.B. is the principal investigator of the §SMOKE study. L.S.B. participated in all phases of study preparation, data collection and writing the manuscript. M.B.K. participated in data collection, data work, analyses, and writing. S.G.K. participated in data collection and writing of the manuscript. N.S.J. participated in writing the manuscript, and A.I.C. participated in the study development, item selection, and writing the manuscript. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement There is no formal institution for ethical assessment and approval of questionnaire-based population studies in Denmark. However, the study is registered at the SDU Research and Innovation Organization. Informed Consent Statement Participants received an invitation letter for the study along with written information about the purposes of the study and how data were handled. Participants were also informed that completion of the questionnaire was voluntary, and their responses would be treated with confidence. Data Availability Statement Data not available. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Flow diagram over participants in the §SMOKE study. Figure 2 Type of tobacco or nicotine product used first, stratified by gender. Data for the use of heated tobacco among men is not shown due to too few respondents (n < 5). Figure 3 Type of tobacco or nicotine product used first, stratified by age groups. Data for the use of heated tobacco among 18–24-year-olds as well as data for the use of e-cigarettes, smokeless tobacco, and heated tobacco among 25–29-year-olds are not shown due to too few respondents (n < 5). ijerph-19-05623-t001_Table 1 Table 1 Sociodemographic characteristics of the study population. Respondents % (n) All Participants Invited to Study % (n) Gender     Men 44.4 (5914) 51.3 (19,222)     Women 55.6 (7401) 48.7 (18,260) Age     15–17 years 21.8 (2906) 18.0 (6754)     18–24 years 48.2 (6421) 47.0 (17,596)     25–29 years 30.0 (3988) 35.0 (13,132) Region of residence     Capital Region 32.2 (4288) 34.3 (12,857)     Region Zealand 11.5 (1530) 11.6 (4345)     Region of Southern Denmark 20.6 (2743) 19.7 (7390)     Central Denmark Region 25.5 (3396) 24.2 (9064)     North Denmark Region 10.2 (1358) 10.2 (3826) ijerph-19-05623-t002_Table 2 Table 2 Measures of current tobacco or nicotine product use, and first product used. Variable Item Response Categories Coding Cigarettes Do you smoke cigarettes? (1) Yes, everyday (2) Yes, at least once a week (3) Yes, less often than every week (4) No, but I have previously smoked/used (5) No, but I have tried smoking/using (6) No, I have never tried smoking/using Daily (1) Occasionally (2 + 3) Former (4) Tried (5) Never (6) E-cigarettes Do you use e-cigarettes? Smokeless tobacco Do you use any of these…?(a) Snus (b) Chewing tobacco (c) Nicotine bags Heated tobacco Do you smoke heated tobacco? (Also known as tobacco sticks, heat sticks etc.—used in a heating device) Product first used If you use multiple tobacco products: Which product did you try first? (1) Cigarettes (2) Smokeless tobacco (3) E-cigarettes (4) Heated tobacco ijerph-19-05623-t003_Table 3 Table 3 Use of tobacco or nicotine products by gender and age group, p-value for differences by gender and age group. Overall % (n) Gender % (n) Age Group % (n) Male Female p 15–17 Years 18–24 Years 25–29 Years p Cigarettes Daily 9.8 (1171) 10.4 (546) 9.2 (625) 0.002 4.1 (112) 10.7 (643) 11.5 (416) <0.001 Occasionally 10.3 (1278) 11.1 (606) 9.6 (672) 7.8 (215) 12.4 (741) 8.9 (322) Tried 33.8 (4154) 32.7 (1764) 34.8 (2390) 24.8 (675) 34.5 (2081) 37.4 (1398) Former 10.2 (1204) 10.0 (507) 10.5 (696) 2.6 (71) 9.2 (556) 15.5 (577) Never 35.9 (4643) 35.9 (2046) 35.9 (2597) 60.7 (1651) 33.2 (1993) 26.8 (999) E-cigarettes Daily 1.8 (212) 2.5 (130) 1.2 (82) <0.001 1.2 (30) 2.1 (119) 1.8 (63) <0.001 Occasionally 2.1 (255) 2.5 (141) 1.6 (114) 3.3 (85) 2.2 (128) 1.2 (42) Tried 31.6 (3838) 34.0 (1831) 29.1 (2007) 26.5 (701) 37.5 (2186) 26.6 (951) Former 6.8 (799) 8.9 (474) 4.8 (325) 4.3 (111) 8.3 (471) 6.2 (217) Never 57.7 (7085) 52.2 (2754) 63.3 (4331) 64.7 (1741) 49.9 (2979) 64.1 (2365) Smokeless tobacco Daily 4.3 (495) 6.7 (373) 1.7 (122) <0.001 3.6 (90) 6.0 (327) 2.4 (78) <0.001 Occasionally 4.8 (570) 6.5 (355) 3.0 (215) 5.1 (133) 5.9 (335) 3.1 (102) Tried 27.3 (3243) 28.4 (1462) 26.3 (1781) 17.5 (469) 28.9 (1701) 30.4 (1073) Former 3.6 (426) 5.0 (268) 2.2 (158) 3.4 (88) 4.5 (251) 2.6 (87) Never 60.0 (7439) 53.4 (2867) 66.7 (4572) 70.4 (1889) 54.8 (3258) 61.6 (2292) Heated tobacco Daily 0.1 (14) 0.1 (6) 0.1 (8) <0.001 n/a 0.2 (9) n/a <0.001 Occasionally 0.2 (24) 0.3 (16) 0.1 (8) n/a 0.3 (15) n/a Tried 3.2 (353) 3.8 (192) 2.5 (161) 1.1 (29) 2.8 (159) 4.7 (165) Former 0.4 (44) 0.6 (31) 0.2 (13) 0.2 (6) 0.4 (24) 0.4 (14) Never 96.1 (11,728) 95.2 (5073) 97.1 (6655) 98.4 (2614) 96.4 (5664) 98.4 (3450) Using at least one product 27.0 (3219) 31.1 (1657) 22.8 (1562) <0.001 18.1 (476) 31.3 (1833) 25.8 (910) <0.001 Ever used or tried any product 71.3 (8730) 72.4 (3901) 70.2 (4829) 0.007 50.4 (1370) 74.4 (4467) 78.1 (2893) <0.001 ijerph-19-05623-t004_Table 4 Table 4 Dual and multiple use of tobacco or nicotine-containing products. Overall % (n) Gender % (n) Age Groups % (n) Male Female 15–17 Years 18–24 Years 25–29 Years Dual use (Two products) 87.2 (597) 85.4 (370) 91.2 (227) 78.6 (121) 88.2 (374) 92.1 (102) Cigarettes + E-cigarettes 34.5 (208) 26.7 (95) 50.9 (113) 30.8 (38) 29.4 (115) 52.0 (55) + Smokeless tobacco 61.5 (365) 68.1 (255) 47.4 (110) 62.9 (76) 66.5 (244) 45.8 (45) + Heated tobacco 1.1 (6) 1.2 (4) 0.9 (2) 0 (0) 1.0 (4) 2.2 (2) E-cigarettes + Smokeless tobacco 2.6 (16) 3.5 (14) 0.8 (2) 5.3 (6) 2.8 (10) 0 (0) + Heated tobacco 0.33 (2) 0.5 (2) 0 (0) 0.9 (1) 0.3 (1) 0 (0) Smokeless tobacco + Heated tobacco 0.0 (0) 0.0 (0) 0.0 (0) 0.0 (0) 0.0 (0) 0.0 (0) Multiple use (Three or four products) 12.8 (89) 14.6 (65) 8.8 (24) 21.4 (33) 11.9 (48) 7.9 (8) Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Cruz T.B. McConnell R. Low B.W. Unger J.B. Pentz M.A. Urman R. Berhane K. Chou C.P. Liu F. Barrington-Trimis J.L. Tobacco marketing and subsequent use of cigarettes, e-cigarettes, and hookah in adolescents Nicotine Tob. Res. 2019 21 926 932 10.1093/ntr/nty107 29846704 2. Harrell P.T. Naqvi S.M.H. Plunk A.D. Ji M. Martins S.S. Patterns of youth tobacco and polytobacco usage: The shift to alternative tobacco products Am. J. Drug Alcohol Abus. 2017 43 694 702 10.1080/00952990.2016.1225072 27668320 3. Lundberg C.S. Kvaavik E. Tokle R. Nye bruksmønstre i et tobakksmarked i endring–kombinert bruk av sigaretter, snus og e-sigaretter blant ungdom Nord. Stud. Alcohol Drugs 2019 36 6 20 10.1177/1455072518797823 4. Johnson A.L. Collins L.K. Villanti A.C. Pearson J.L. Niaura R.S. Patterns of nicotine and tobacco product use in youth and young adults in the United States, 2011–2015 Nicotine Tob. Res. 2018 20 S48 S54 10.1093/ntr/nty018 30125012 5. Lee Y.O. Hebert C.J. Nonnemaker J.M. Kim A.E. Youth tobacco product use in the United States Pediatrics 2015 135 409 415 10.1542/peds.2014-3202 25647680 6. Hoffmann S.H. Schramm S. Jarlstrup N.S. Christensen A.I. Smoking Habits among Danish Citizens. The Development from 1994 to 2017 National Institute of Public Health Copenhagen, Denmark 2018 7. Holstein B.E. Andersen A. Damsgaard M.T. Due P. Bast L.S. Rasmussen M. Trends in socioeconomic differences in daily smoking among 15-year-old Danes 1991–2014 Scand. J. Public Health 2019 48 667 673 10.1177/1403494819848284 31074327 8. Brink A.-L. Stage M. The Youth Study of Smokefree Future 2017–2020 Danish Cancer Society and the TrygFoundation Copenhagen, Denmark 2021 9. Brown J. West R. Beard E. Michie S. Shahab L. McNeill A. Prevalence and characteristics of e-cigarette users in Great Britain: Findings from a general population survey of smokers Addict. Behav. 2014 39 1120 1125 10.1016/j.addbeh.2014.03.009 24679611 10. Wang T.W. Gentzke A.S. Creamer M.R. Cullen K.A. Holder-Hayes E. Sawdey M.D. Anic G.M. Portnoy D.B. Hu S. Homa D.M. Tobacco Product Use and Associated Factors among Middle and High School Students—National Youth Tobacco Survey, United States, 2021 MMWR Surveill. Summ. 2022 71 1 29 11. Mishra A. Chaturvedi P. Datta S. Sinukumar S. Joshi P. Garg A. Harmful effects of nicotine Indian J. Med. Paediatr. Oncol. 2015 36 24 31 10.4103/0971-5851.151771 25810571 12. Yuan M. Cross S.J. Loughlin S.E. Leslie F.M. Nicotine and the adolescent brain J. Physiol. 2015 593 3397 3412 10.1113/JP270492 26018031 13. Leslie F.M. Unique, long-term effects of nicotine on adolescent brain Pharmacol. Biochem. Behav. 2020 197 173010 10.1016/j.pbb.2020.173010 32738256 14. Owotomo O. Stritzel H. McCabe S.E. Boyd C.J. Maslowsky J. Smoking Intention and Progression from E-Cigarette Use to Cigarette Smoking Pediatrics 2020 146 e2020002881 10.1542/peds.2020-002881 33168672 15. O’Brien D. Long J. Quigley J. Lee C. McCarthy A. Kavanagh P. Association between electronic cigarette use and tobacco cigarette smoking initiation in adolescents: A systematic review and meta-analysis BMC Public Health 2021 21 954 10.1186/s12889-021-10935-1 34078351 16. Pacek L.R. Wiley J.L. McClernon F.J. A Conceptual Framework for Understanding Multiple Tobacco Product Use and the Impact of Regulatory Action Nicotine Tob. Res. 2019 21 268 277 10.1093/ntr/nty129 29931176 17. Danish Health Authority Smoking Habits among Danish Citizens 2019. Sub-Report 1: Nicotine Dependence Danish Health Authority Copenhagen, Denmark 2020 18. Pedersen C.B. The Danish civil registration system Scand. J. Public Health 2011 39 22 25 10.1177/1403494810387965 21775345 19. Klitgaard M. Jarlstrup N. Lund L. Brink A.-L. Knudsen A. Christensen A. Bast L. Evaluation of implementation of increased tobacco price, tobacco point-of-sale display ban and plain packaging in Denmark: A study protocol for monitoring the impact on 15–29-year-olds tobacco use, 2020–2025 Unpublished manuscript 20. Cornelius M.E. Wang T.W. Jamal A. Loretan C.G. Neff L.J. Tobacco Product Use Among Adults—United States, 2019 Morb. Mortal. Wkly. Rep. 2020 69 1736 10.15585/mmwr.mm6946a4 21. Grøtvedt L. Forsén L. Ariansen I. Graff-Iversen S. Holmen T.L. Impact of snus use in teenage boys on tobacco use in young adulthood; a cohort from the HUNT Study Norway BMC Public Health 2019 19 1265 10.1186/s12889-019-7584-5 31519157 22. Pedersen W. von Soest T. Tobacco use among N orwegian adolescents: From cigarettes to snus Addiction 2014 109 1154 1162 10.1111/add.12509 24521070 23. Gentzke A.S. Wang T.W. Jamal A. Park-Lee E. Ren C. Cullen K.A. Neff L. Tobacco Product Use among Middle and High School Students—United States, 2020 Morb. Mortal. Wkly. Rep. 2020 69 1881 10.15585/mmwr.mm6950a1 24. Chapman S. Bareham D. Maziak W. The gateway effect of e-cigarettes: Reflections on main criticisms Nicotine Tob. Res. 2019 21 695 698 10.1093/ntr/nty067 29660054 25. Glantz S.A. Bareham D.W. E-cigarettes: Use, effects on smoking, risks, and policy implications Annu. Rev. Public Health 2018 39 215 235 10.1146/annurev-publhealth-040617-013757 29323609 26. Tseveenjav B. Pesonen P. Virtanen J.I. Use of snus, its association with smoking and alcohol consumption, and related attitudes among adolescents: The Finnish National School Health Promotion Study Tob. Induc. Dis. 2015 13 34 10.1186/s12971-015-0058-3 26500472 27. Leon M.E. Lugo A. Boffetta P. Gilmore A. Ross H. Schüz J. La Vecchia C. Gallus S. Smokeless tobacco use in Sweden and other 17 European countries Eur. J. Public Health 2016 26 817 821 10.1093/eurpub/ckw032 27048433 28. Dai H. Heated tobacco product use and associated factors among US youth, 2019 Drug Alcohol Depend. 2020 214 108150 10.1016/j.drugalcdep.2020.108150 32645682 29. Kang H. Cho S.-I. Heated tobacco product use among Korean adolescents Tob. Control 2020 29 466 468 10.1136/tobaccocontrol-2019-054949 31164491 30. McKelvey K. Baiocchi M. Halpern-Felsher B. PMI’s heated tobacco products marketing claims of reduced risk and reduced exposure may entice youth to try and continue using these products Tob. Control 2020 29 e18 e24 10.1136/tobaccocontrol-2019-055318 32029537 31. McKelvey K. Popova L. Kim M. Chaffee B.W. Vijayaraghavan M. Ling P. Halpern-Felsher B. Heated tobacco products likely appeal to adolescents and young adults Tob. Control 2018 27 s41 s47 10.1136/tobaccocontrol-2018-054596 30352843 32. Czoli C.D. White C.M. Reid J.L. OConnor R.J. Hammond D. Awareness and interest in IQOS heated tobacco products among youth in Canada, England and the USA Tob. Control 2020 29 89 95 10.1136/tobaccocontrol-2018-054654 30696783 33. Pisinger V. Thorsted A. Jezek A. Jørgensen A. Christensen A. Thygesen L. Health and Wellbeing in High Schools Syddansk Universitet Odense, Denmark 2019 34. Kentala J. Utriainen P. Pahkala K. Mattila K.J.A.B. Verification of adolescent self-reported smoking Addict. Behav. 2004 29 405 411 10.1016/j.addbeh.2003.08.012 14732430 35. Caraballo R.S. Giovino G.A. Pechacek T.F. Self-reported cigarette smoking vs. serum cotinine among US adolescents Nicotine Tob. Res. 2004 6 19 25 10.1080/14622200310001656821 14982684

PMC009xxxxxx/PMC9099615.txt

==== Front Int J Mol Sci Int J Mol Sci ijms International Journal of Molecular Sciences 1422-0067 MDPI 10.3390/ijms23095102 ijms-23-05102 Article Suppression of the Proliferation of Huh7 Hepatoma Cells Involving the Downregulation of Mutant p53 Protein and Inactivation of the STAT 3 Pathway with Ailanthoidol Tseng Tsui-Hwa 12 Wang Chau-Jong 34 Lee Yean-Jang 5 Shao Yi-Chia 1 Shen Chien-Heng 6 Lee Ko-Chao 7 Tung Shui-Yi 68* https://orcid.org/0000-0002-8993-1615 Kuo Hsing-Chun 9101112* Chung Hwan-Suck Academic Editor 1 Department of Medical Applied Chemistry, Chung Shan Medical University, Taichung 40201, Taiwan; tht@csmu.edu.tw (T.-H.T.); a14253681@gmail.com (Y.-C.S.) 2 Department of Medical Education, Chung Shan Medical University Hospital, Taichung 40201, Taiwan 3 Department of Health Diet and Industry Management, Chung Shan Medical University, Taichung 40201, Taiwan; wcj@csmu.edu.tw 4 Department of Medical Research, Chung Shan Medical University Hospital, Taichung 40201, Taiwan 5 Department of Chemistry, National Changhua University of Education, Changhua 50007, Taiwan; leeyj@cc.ncue.edu.tw 6 Department of Hepato-Gastroenterology, Chang Gung Memorial Hospital, Chiayi 61363, Taiwan; gi2216@adm.cghmh.org.tw 7 Division of Colorectal Surgery, Department of Surgery, Chang Gung Memorial Hospital, Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung 83301, Taiwan; kclee@cgmh.org.tw 8 Graduate Institute of Clinical Medical Sciences, College of Medicine, Chang Gung University, Chiayi 61363, Taiwan 9 Division of Basic Medical Sciences, Department of Nursing, Chang Gung University of Science and Technology, Chiayi 61363, Taiwan 10 Research Fellow, Chang Gung Memorial Hospital, Chiayi 61363, Taiwan 11 Research Center for Food and Cosmetic Safety, College of Human Ecology, Chang Gung University of Science and Technology, Taoyuan 33303, Taiwan 12 Chronic Diseases and Health Promotion Research Center, Chang Gung University of Science and Technology, Chiayi 61363, Taiwan * Correspondence: ma1898@yahoo.com (S.-Y.T.); guscsi@gmail.com (H.-C.K.) 04 5 2022 5 2022 23 9 510206 4 2022 02 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Ailanthoidol (ATD) has been isolated from the barks of Zanthoxylum ailanthoides and displays anti-inflammatory, antioxidant, antiadipogenic, and antitumor promotion activities. Recently, we found that ATD suppressed TGF-β1-induced migration and invasion of HepG2 cells. In this report, we found that ATD exhibited more potent cytotoxicity in Huh7 hepatoma cells (mutant p53: Y220C) than in HepG2 cells (wild-type p53). A trypan blue dye exclusion assay and colony assay showed ATD inhibited the growth of Huh7 cells. ATD also induced G1 arrest and reduced the expression of cyclin D1 and CDK2. Flow cytometry analysis with Annexin-V/PI staining demonstrated that ATD induced significant apoptosis in Huh7 cells. Moreover, ATD increased the expression of cleaved PARP and Bax and decreased the expression of procaspase 3/8 and Bcl-xL/Bcl-2. In addition, ATD decreased the expression of mutant p53 protein (mutp53), which is associated with cell proliferation with the exploration of p53 siRNA transfection. Furthermore, ATD suppressed the phosphorylation of the signal transducer and activator of transcription 3 (STAT3) and the expression of mevalonate kinase (MVK). Consistent with ATD, the administration of S3I201 (STAT 3 inhibitor) reduced the expression of Bcl-2/Bcl-xL, cyclin D1, mutp53, and MVK. These results demonstrated ATD’s selectivity against mutp53 hepatoma cells involving the downregulation of mutp53 and inactivation of STAT3. ailanthoidol hepatoma mutant p53 STAT3 apoptosis cell cycle Ministry of Science and TechnologyMOST 108-2320-B-040-016-MY3 MOST 110-2320-B-255-005-MY3 Ministry of Science and Technology, Ministry of Education, and Chung Shan Medical UniversityCSMU 108-OM-A-114 This research was supported by the Ministry of Science and Technology Grant (MOST 108-2320-B-040-016-MY3, MOST 110-2320-B-255-005-MY3), Taiwan. Digital Image Analyzer was performed in the Instrument Center of Chung Shan Medical University, which was supported by the Ministry of Science and Technology, Ministry of Education, and Chung Shan Medical University (CSMU 108-OM-A-114), Taiwan. ==== Body pmc1. Introduction Hepatocellular carcinoma (HCC), the most common primary malignant tumor in liver cancer cases, is a complex disease caused by a variety of risk factors. Conventional types of liver cancer treatment, including surgical resection, radiotherapy, and chemotherapy, have been either limited in application or ineffective [1]. Transplantation of the liver is believed to be the only viable treatment; however, it is not easy to find the proper donor. Although scientists have generated intense research efforts to explore cellular, molecular, and physiological mechanisms of the disease for developing prevention and therapy strategies [2], the mortality rate of HCC remains high. The transcription factor p53 is activated in response to various stresses including nutrient deprivation, DNA damage, oncogene activation, and hypoxia. p53 is a well-established tumor suppressor and guardian of the genome that induces apoptosis and cell cycle arrest by activating downstream target genes [3]. However, p53 is mutated in around half of all human cancers. It is generally believed that p53 loses its tumor suppressor function because of a mutation in p53. Certain types of p53 mutations are gain-of-function mutations, which have been shown to have oncogenic functions [4]. HCC is a lethal malignancy associated with poor prognosis and a high recurrence. Effective HCC therapeutics still await a molecular understanding of the mechanisms promoting the development of selective and precise agents. HCC has a high rate of mutation in tumor suppressor protein p53, leading to the loss of its tumor suppressor activity and, in certain cases, gain-of-function activities that promote cell proliferation, tumor progression, and drug resistance [5]. Thus, mutant p53 has become an important target for the development of anticancer agents in HCC. The signal transducer and activator of transcription 3 (STAT3) is a pivotal transcriptional factor of multiple promoting genes in cancer development and immune evasion [6]. Phosphorylated STAT3s dimerize each other and translocate into the nucleus before activating the downstream genes. Under a normal physiological state, STAT3 activation is usually transient in the continuous stimulation of cytokines and contributes to protecting normal hepatocytes from inflammatory insults. It has been reported that constitutive phosphorylation of STAT3 in tumor tissue is correlated with poor prognosis in HCC patients [7]. Thereafter, the inactivation of the STAT3 signal pathway is a promising strategy in anti-HCC treatment. Ailanthoidol (ATD), a neolignan, has been isolated from the bark of Zanthoxylum ailanthoides (Rutaceae), of which the dried fruit is used as a spice in Taiwan. Our previous study demonstrated that ATD displays antitumor promotion effects using the multistep skin cancer model induced by 12-o-tetradecanoylphobol-13-acetate [8]. Recently, we found that ATD suppresses TGF-β1-promoted migration and invasion in HepG2 cells [9]. Kim and Jun reported that ATD has in vitro and in vivo anti-inflammatory effects [10]. Park et al. found that ATD possesses antiadipogenic activities [11]. In addition, ATD is a benzofuran derivative and indicates diverse pharmacological activities, including anticancer activities [12]. As the anticancer properties of ATD have not been well clarified, this study investigated the antiproliferation effects and molecular mechanism of ATD in hepatoma cells. 2. Results 2.1. Effects of ATD on the Growth of Huh7 and HepG2 Cells To understand cell viability under ATD treatment on the hepatocellular carcinoma cells (HCCs), a range of concentrations (0–80 μM) was evaluated in the Huh7 and HepG2 cells with an MTT assay. As shown in Figure 1, ATD suppressed cell viability in Huh7 cells, with IC50 values of 45 μM and 22 μM at 24 h and 48 h, respectively, while the IC50 value in HepG2 cells was above 80 μM. In addition, to examine the effects of ATD on the growth of Huh7 cells, a trypan blue dye exclusion assay and colony assay were performed. The results demonstrated that ATD decreased the growth of Huh7 cells in a time-dependent manner (Figure 2A). Furthermore, a colony formation assay confirmed that ATD decreased the growth of Huh7 cells significantly (Figure 2B). Thereafter, we proceeded to study the antitumor potential and mode of action of ATD in Huh7 cells. 2.2. Effect of ATD on the Cell Cycle Distribution of Huh7 Cells To determine the cellular mechanism preventing cancer cell proliferation, we examined cell cycle profiles in Huh7 cells with or without ATD administration at various times, using flow cytometry. When the cells were administrated with ATD (10 μM), the proportion of the subG1 and G0/G1 phases tended to significantly increase, compared with the control (0 h), while the G2/M phase was decreased (Figure 3A,B). In addition, ATD downregulated the expression levels of the checkpoint proteins involved in the regulation of G1 phase transition, such as cyclin D1 and CDK2 (Figure 3C). 2.3. Induction Apoptosis by ATD in Huh7 Cells To examine the mechanism of ATD-induced cytotoxicity, apoptotic induction of ATD (10, 20, and 40 μM) was evaluated by flow cytometry analysis with Annexin V/PI double staining. While the percentage of early apoptotic cells in the control group was 3%, in the ATD-treated groups, it increased from 4.13% to 13.43% (Figure 4A). ATD significantly increased the percentage of total apoptotic cells (early plus late) in a dose-dependent manner, from 8.35% to 22.49%, while in the control group, it was 5.95% for Huh7 cells (Figure 4A,B). To further characterize the cell death process, we investigated the downstream expression of apoptotic associated proteins using a Western blot assay. ATD decreased the expression levels of procaspase 3, procaspase 8, Bcl-xL, and Bcl-2 but increased the levels of Bax and cleavage poly(ADP-ribose) polymerase (PARP) (Figure 4C). 2.4. Induction Apoptosis by ATD in Huh7 Cells As ATD exhibited a marked reduction in the IC50 value in Huh7 cells (mutant p53 Y220C), compared with HepG2 cells (wild-type p53), the effect of ATD on the p53 expression in Huh7 cells could be determined. The immunoblotting assay against the p53 antibody (DO-1), which is recommended for detection of wide-type and mutant p53, revealed that ATD reduced the expression of p53 in Huh7 cells in a dose-dependent manner (Figure 5A). In addition, according to the immunofluorescence analysis against the p53 antibody (PAb240), which is recommended for mutant p53 under non-denaturing conditions, ATD reduced the fluorescence of p53 in Huh7 and PLC/PRF/5 cells (mutant p53 R249S), compared with the positive control group, respectively, while the negative control of HepG2 (wild-type p53) did not exhibit green fluorescence (Figure 5B). In order to determine whether mutant p53 was involved in the ATD-induced antiproliferation, we conducted a p53 knockdown experiment using the transfection of p53 siRNA. Although the cellular levels of p53 in Huh7 cells transfected with p53 siRNA were not completely knocked down, a distinct downregulation of the cellular p53 levels was observed (Figure 6A). The CCK-8 assay indicated that p53 knockdown indeed decreased cell viability (Figure 6B). In addition, ATD treatment significantly enhanced the antiproliferation property in the p53 knockdown cells (Figure 6B), indicating that mutant p53 was involved in ATD-induced apoptosis and cell cycle arrest in Huh7 cells. 2.5. ATD-Induced Antiproliferation of Huh7 Cells by Suppressing the STAT3 Pathway STAT3 has recently emerged as a potential therapeutic target for HCC [7]. In addition, it has been demonstrated that STAT3 may sustain mutp53 levels due to its interplay with the mevalonate pathway, which increases its stability [13]. Thereafter, we determined the effect of ATD on the phosphorylation of STAT3 and the expression of mevalonate kinases (MVK), a downstream target gene product of the STAT3 pathway. As shown in Figure 7A, ATD decreased the level of phosphorylated STAT3 and MVK. Consistent with ATD, S3I201, an inhibitor of STAT3, decreased the expression of Bcl-xL/Bcl2, p53, and MVK (Figure 7B,C). For one other cell line, PLC/PRF/5 cells (mutp53 R249S), ATD decreased expression of p53 (DO-1), MVK and Bcl-2, Bcl-XL, cyclin D1 as well as and phosphorylation of Stat3 (Figure 7D,E). 3. Discussion HCC, which accounts for nearly 80% of all liver cancer cases, is a heterogeneous type of cancer caused by a variety of risk factors, including exposure to the hepatitis virus, food contaminated with Aflatoxin B1, heavy alcohol intake, and obesity [14,15]. The incidence of HCC is rising rapidly worldwide. In addition, since HCC is diagnosed at a late stage in most cases, surgical resection and liver transplantation are not practical therapy options. Metastasis and recurrence are quite common. Therefore, the development of a promising compound with target therapy potential is an urgent task. Plants are major food and pharmaceutical sources for humans. Some phytochemicals, such as alkaloids, diterpenoids, and sesquiterpenes, display therapeutic potential for cancer treatment [16]. However, these therapeutic phytochemicals are also associated with adverse side effects, such as cardiovascular diseases, vomiting, renal dysfunction, and myelotoxicity. Thereafter, scientists have dedicated themselves to developing phytochemicals with minimal side effects and good bioavailability. Lignans and neolignans may possess great potential for anticancer treatment and display good safety profiles [12,17,18]. In the present study, ATD, a neolignan isolated from the bark of Zanthoxylum ailanthoides [19], exhibited antiproliferation potential in Huh7 hepatoma cells, which was related to the induction of cell cycle arrest and the activation of apoptosis. Cell cycle arrest was mediated by the ATD-induced cyclin D1 and CDK2 expression, while apoptosis was activated by ATD-downregulated Bcl-xL/Bcl2 and augmented Bax, resulting in the activation of caspase 3. For a real application, animal studies of ATD are required in the future. The tumor suppressor p53 regulates the transcription of numerous downstream target genes involved in cell cycle arrest, apoptosis, and metabolism. Loss of p53 activity by gene deletion or mutations in normal cells causes uncontrolled cell proliferation, leading to immortalization and, ultimately, cancer. Additionally, mutant p53 shows oncogenic gain-of-function activities, such as enhanced tumor progression, metastasis potential, and drug resistance [20]. As a result, obtaining efficient inhibitors against mutant p53 cancer cells remains an urgent task for medicine development. Reactivation of the wild-type p53 function and expression or abrogation of mutant p53 protein may halt cancer progression [21]. Accumulation of mutant p53 is critical for the gain of function related to p53 mutation, including enhanced cell growth and tumor progression; however, the manner in which mutp53 is regulated and promotes cancer progression is not well understood [4]. Enzymes controlling p53 proteasomal degradation or stability and some microRNA have been considered to regulate mutant p53 levels [13,22]. In the present study, we found that ATD had more potent cytotoxicity in Huh7 cells (mutant p53) than in HepG2 cells (WT p53), which was associated with reducing the level of mutp53. According to our results, ATD blocked the STAT3 pathway and mediated the abrogation of mutp53. Whether ATD affects the miRNA or enzymes associated with proteasomal degradation requires further clarification. Our data implicated that ATD displayed potent anticancer potential in mutp53-based HCC by impairing the gain of function of mutant p53. Among the diverse signaling molecules, STAT3 is considered an oncogenic factor in HCC [7]. Under a normal physiological state, STAT3 activation is usually transient, even in the continuous stimulation of cytokines, and contributes to protecting normal hepatocytes from inflammatory and toxic insults. In HCC, the persistent activation of STAT3 changes the gene transcriptions associated with cell survival, proliferation, invasion, and angiogenesis. The pro-proliferative role of STAT3 is related to its antiapoptotic functions toward HCC via upregulating antiapoptotic proteins such as Bcl-xL. Furthermore, constitutive phosphorylation of STAT3 in tumor tissue is closely correlated with a poor prognosis in HCC patients [6]. Recently, it has been reported that STAT3 sustains mutp53 expression due to its interplay with the mevalonate pathway, which increases the stability of mutp53 and prevents its degradation from proteasome [13]. In the present study, ATD inhibited the p-STAT3, MVK, and mutp53 levels in Huh7 cells. According to Figure 7, we supposed that ATD blocked the STAT3 pathway mediating a reduction in mutp53 protein in Huh7 cells. Although we found that ATD reduced the level of MVK (the downstream target gene product of STAT3, which might affect the stability of mutp53), the real interplay between STAT3 and mutp53 needs further elucidation. In addition to reducing the gain-of-function activity of mutp53, ATD also triggered apoptosis by decreasing the expression of Bcl-xL and Bcl-2, which is associated with the inactivation of the STAT3 pathway. Additional studies are still needed to elucidate the action mechanisms of the ailanthoidol (ATD) as a chemopreventive and therapeutic agent in in vivo xenograft mouse models. 4. Materials and Methods 4.1. Materials Dulbecco’s modified Eagle’s medium (DMEM), phosphate-buffered saline (PBS), fetal bovine serum (FBS), penicillin–streptomycin–neomycin (PSN), and trypsin–EDTA were purchased from Gibco Ltd. (Grand Island, NY, USA). Primary antibodies against p53(DO-1), p53(Pab-240), CDK2, Bax, Bcl-2, Bcl-xL pro-caspase 3/8, STAT3, MVK, GADPH, and actin were obtained from Santa Cruz Biotechnology (St. Louis, MO, USA). Anti-cyclin D1, anti-c-PARP, and anti-p-STAT3 (Tyr750) were obtained from Cell Signaling Technology (Beverly, MA, USA). Alexa 488-labeled goat anti-mouse IgG antibody was from Thermo Fisher Scientific, Waltham, MA, USA. ATD was provided by Dr. Lee and synthesized from 5-bromo-2-hydroxy-3-methoxybenzaldehyde, as previously reported [23]. Tris base and all other materials were purchased from Sigma Chemical Co. (St. Louis, MO, USA). 4.2. Cells and Cell Culture The human liver cancer cell line Huh7 (p53 Y220C) was obtained from the Food Industry Research and Development Institute (Hsinchu, Taiwan) and cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco BRL, Grand Island, NY, USA), supplemented with 10% FBS, 1% PSN, 1% essential amino acid, and 1mM glutamine. HepG2 (p53 WT) cells were cultured in DMEM supplemented with 10% FBS, 1% PSN, 1% essential amino acid, 1% sodium pyruvate, and 1mM glutamine. PLC/PRF/5 (R249S) cells were cultured in MEM supplemented with 10% FBS and 1% PSN. The cell cultures were maintained at 37 °C in a humidified atmosphere of 5% CO2. 4.3. Cell Viability Assay Huh7 and HepG2 cells were placed in a 24-well plate at a density of 2 × 104 cells/well, respectively, and treated with various concentrations of ATD (10–80 μM) or solvent control (0.2% DMSO) for 24 h and 48 h. Cell viability was determined in the presence of 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT) dye solution for 4 h. The medium was removed, and formazan was solubilized in isopropanol and measured spectrophotometrically at 560 nm using a microplate reader. 4.4. Trypan Blue Dye Exclusion Assay Huh7 cells were placed in a 10 cm dish at a density of 4 × 104 cells/dish and treated with ATD (10 μM) or solvent control (0.2% DMSO) for 24, 48, and 72 h. After treatment, trypan blue was added to the cell suspension, and viable cells that excluded the dye were counted on a hemacytometer. 4.5. Colony Formation Assay Cells were plated in 6-well plates, at a density of 500 cells/well. On the next day, cells were treated with 0.2% DMSO (control) or ATD at the indicated concentration for 48 h, then cultured for 7 days. The colony was fixed with methanol for 15 min and stained with Giemsa. Cell colonies were photographed and counted. 4.6. Cell Cycle Analysis Cell cycle distribution was determined using a flow cytometer with propidium iodide (PI) staining. Briefly, 6 × 105 cells/dish were treated with 0.2% dimethyl sulfoxide (DMSO; control) or 10 μM ATD for indicated time. Then, cells were harvested, fixed with cold 75% alcohol, and stained with 50 μg/mL PI solution in darkness for 30 min on ice. The distribution of cells in different cell cycle phases was determined using flow cytometry (FACSCalibur, BD Biosciences, San Jose, CA, USA). In total, 10,000 cells per sample were counted, and DNA histograms were analyzed using Cell Quest software (BD Biosciences, San Jose, CA, USA) to calculate the percentage of cells in each peak. 4.7. Annexin V/PI Double Staining Assay For this assay, 6 × 105 cells were plated in a 10 cm culture dish. After attachment, cells were treated with DMSO or ATD at the indicated concentration for 48 h and then harvested and resuspended in PBS. Apoptotic cells were measured with a FITC-Annexin V Apoptosis Detection Kit (BD Biosciences, San Jose, CA, USA) according to the manufacturer’s protocol. Briefly, cells were stained with FITC annexin V and propidium iodide (PI) solution for 15 min at room temperature in darkness. In total, 10,000 cells were analyzed for each histogram. Flow cytometry demonstrated that the early apoptotic cells were in the lower-right quadrant, and the advanced apoptotic cells were in the upper-right quadrant. The apoptotic rate was the sum of the early and advanced apoptotic rates. 4.8. Western Immunoblotting Equal amounts of protein from total cell lysates were separated in 8–12% polyacrylamide gel and transferred onto the PVDF membrane. The blot was subsequently incubated in blocking buffer (5% nonfat milk in PBS) for 1 h and then probed with a corresponding antibody against a specific protein overnight at 4 °C and washed with tris-buffered saline; the membrane was then incubated with an appropriate peroxidase-conjugated secondary antibody for 1 h. Finally, antigen–antibody complex was developed by ECL detection system. The relative image density was quantitated with densitometry. 4.9. Immunofluorescence After ATD or DMSO treatment, Huh7, PLC/PRF/5, and HepG2 cells were washed with PBS and fixed with 4% paraformaldehyde for 10 min. The cells were permeated with 0.1% Triton X-100, then incubated at 4 °C overnight with a monoclonal anti-p53 (Pab-240) antibody, followed by a 1 h incubation with an Alexa 488-labeled goat anti-mouse IgG antibody (Thermo Fisher Scientific, Waltham, MA, USA). After washing with PBS containing 0.1% tween 20, the DAPI was added for 10 min. The cells were observed under a fluorescence microscope at 400× magnification. 4.10. Transfection with Small Interfering RNA (siRNA) p53 siRNAs (sense: 5′-AGA-CCU-AUG-GAA-ACU-ACU-Utt-3′) were purchased from GeneDireX, (QUANTUM BIOTECHNOLOGY, INC., Durham, NC, USA) [24]. For transfection, 3 × 103 Huh7 cells were seeded on 96-well dishes or 4 × 105 on 10 cm dishes. After overnight incubation, p53 siRNA or control siRNA (40 nM) (Santa Cruz Biotechnology, Santa Cruz, CA, USA) were transfected using a T-Pro NTR II transfection reagent, according to the manufacturer’s instructions. Following incubation for 48 h, the cells were treated with or without ATD for 24 h. After ATD treatment, viable-cell counting was performed using Cell Counting Kit-8 (CCK-8 kit), or the total cell lysate was prepared for immunoblotting analysis. 4.11. Cell Proliferation Assay Following the transfection, cell proliferation was assayed using a CCK-8, according to the manufacturer’s protocols. Briefly, after transfection and ATD treatment, the CCK-8 solution was added and incubated for 3 h. The optical density was measured at 450 nm using a microplate reader. 4.12. Statistical Analysis Data are expressed as means ± SD from three independent experiments. The statistical significance of differences throughout the study was analyzed by a one-way ANOVA test. A p value < 0.05 was considered to be statistically significant. 5. Conclusions This study demonstrated a novel mechanism in which ATD exhibited a more potent antiproliferation potential on mutp53 HCC than on wtp53 HCC cells due to the downregulation of mutp53 and blockage of the STAT3 pathway (Figure 8). Author Contributions Provision of study material, collection and assembly of data, and manuscript writing, T.-H.T.; conception, collection, and assembly of data, C.-J.W.; provision of study material, Y.-J.L.; creation of models, Y.-C.S.; provision of study material, collection, and assembly of data, C.-H.S. and K.-C.L.; conception and design, financial support, administrative support, manuscript writing, and final approval of the manuscript, S.-Y.T. and H.-C.K. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement All relevant data are within the paper. Conflicts of Interest The authors declare no conflict of interest. Figure 1 The effect of ailanthoidol (ATD) on the cell viability of Huh7 and HepG2 cells was determined with an MTT assay. Huh7 and HepG2 cells were treated with or without ATD under the indicated concentration for 24 h and 48 h. Cell viability was measured with an MTT assay, as described in the text. Data are represented as the means ± SD of three independent experiments. ** p < 0.01 and *** p < 0.001, compared with the control group (24 h) (0.2% DMSO) of the Huh 7 cells or HepG2 cells, respectively. # p < 0.05 and ### p < 0.001, compared with the control group (48 h) of the Huh 7 cells or HepG2 cells, respectively. Figure 2 The effect of ailanthoidol (ATD) on the growth of Huh7 cells with a trypan dye exclusion assay and colony formation assay: (A) viable cells were counted using the trypan blue dye exclusion assay after treatment with ATD (10 μM) at 24 h, 48 h, and 72 h. Values are the means ± SD (n = 3). @ p < 0.05, * p < 0.0 5, ## p < 0.01, compared with the control group (0.2% DMSO) at 24 h, 48 h, and 72 h, respectively; (B) a total of 500 cells were seeded in a six-well dish. After attachment, the cells were treated with or without ATD (2.5, 5, and 10 μM) for 48 h and then cultured for seven days. The cells were fixed with methanol and stained with Giemsa. The number of colonies was counted. Data are represented as the means ± SD (n = 3). * p < 0.05, ** p < 0.01, *** p < 0.001, compared with the control group (0.2% DMSO). Figure 3 Ailanthoidol induced cell cycle arrest in Huh7 cells: (A) Huh7 cells were treated with or without ATD (10 μM) for 0, 24, 48, and 72 h. The harvested cells were stained with PI, and the DNA content was analyzed using a flow cytometer. The histograms are from one out of three experiments; (B) values are presented as means ± SD (n = 3). ** p < 0.01, *** p < 0.001 vs. o h; (C) Huh7 cells were treated with or without ATD for 24 h. The cells were harvested and equal protein amounts of the whole-cell extracts were analyzed by Western blotting against the indicated antibodies. α- tubulin was used as the loading control. Figure 4 Ailanthoidol induced apoptosis in Huh7 cells. Huh7 cells were treated with or without ATD for 48 h: (A) the ATD-induced apoptosis in the Hun7 cells was determined by using a flow cytometer with Annexin V-FITC/PI staining, as described in the text. The cells in the lower-right quadrant (Annexin V+/PI-) represent the early apoptotic cells, and those in the upper-right quadrant (Annexin V+/PI+) represent the late apoptotic cells. A typical photograph from three independent experiments with similar results is shown; (B) data are presented as means ± SD (n = 3). *** p < 0.001 vs. the control; (C) Huh7 cells were treated with or without ATD for 24 h. The cells were harvested and equal protein amounts of the whole-cell extracts were analyzed with Western blotting against the indicated antibodies. β-actin or GADPH was used as the loading control. Figure 5 Mutant p53 involvement in ATD-induced antiproliferation in Huh 7 cells: (A) after treatment with various concentrations of ATD for 24 h, the level of p53 in Huh7 cells was determined with immunoblotting analysis against anti-p53 (DO1). GADPH was used as loading control; (B) after treatment with DMSO (solvent control) or ATD (20 μM) for 24 h in Huh7 cells or PLC/PRF/5 cells, and HepG2 cells as negative control, the p53 against anti-p53 (PAb240) was detected with immunofluorescence analysis, as described in the text. The nuclear was stained by DAPI (blue). Green fluorescence indicated mutp53. Figure 6 Antiproliferation effect of p53 siRNA transfection in Huh7 cells: (A) after 48 h of transfection with scrambled siRNA (C−siRNA) or siRNA against p53 (p53 siRNA), Huh7 cells were treated with or without ATD (40 μM) for 24 h. The cell lysates were prepared, and the p53 levels were analyzed with Western blotting against p53; GADPH was used as the loading control; (B) after transfection and ATD treatment, the viable cells were determined by using a CCK−8 kit, as described in the text. Values are means ± SD (n = 3). The asterisks indicate statistic changes (** p < 0.01, *** p < 0.001). Figure 7 ATD-induced apoptosis and cell cycle arrest by blocking the STAT3 pathway in Huh7 cells: (A) after treatment with various concentrations of ATD for 24 h in Huh7 cells, the protein levels of p-STAT3, STAT3, and MVK were determined with immunoblotting. In addition, after treatment with various concentrations of S3I201 for 24 h in Huh7 cells, the protein levels of Bcl-XL, Bcl-2, and cyclin D1 (B), as well as p53 and MVK (C), were determined with immunoblotting analysis. β-actin or GADPH were used as the loading control. After treatment with various concentrations of ATD in the PLC/PRF/5 cells (mutp53 R249S), the protein levels of p53 (DO-1), p-STAT3, STAT3, MVK (D), Bcl-2, Bcl-XL, and cyclin D1 (E) were determined with immunoblotting. Figure 8 Antiproliferation of ailanthoidol (ATD) in mutp53 hepatoma cells. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Villanueva A. Hernandez-Gea V. Llovet J.M. Medical therapies for hepatocellular carcinoma: A critical view of the evidence Nat. Rev. Gastroenterol. Hepatol. 2013 10 34 42 10.1038/nrgastro.2012.199 23147664 2. Yoon S.K. Molecular mechanism of hepatocellular carcinoma Hepatoma Res. 2018 4 42 10.20517/2394-5079.2018.23 3. Kastenhuber E.R. Lowe S.W. Putting p53 in Context Cell 2017 170 1062 1078 10.1016/j.cell.2017.08.028 28886379 4. Alvarado-Ortiz E. de la Cruz-Lopez K.G. Becerril-Rico J. Sarabia-Sanchez M.A. Ortiz-Sanchez E. Garcia-Carranca A. Mutant p53 Gain-of-Function: Role in Cancer Development, Progression, and Therapeutic Approaches Front. Cell. Dev. Biol. 2020 8 607670 10.3389/fcell.2020.607670 33644030 5. Caron de Fromentel C. Levrero M. p53 functional loss, stemness and hepatocellular carcinoma Hepatoma Res. 2020 6 80 10.20517/2394-5079.2020.77 6. Svinka J. Mikulits W. Eferl R. STAT3 in hepatocellular carcinoma: New perspectives Hepat. Oncol. 2014 1 107 120 10.2217/hep.13.7 30190945 7. Lee C. Cheung S.T. STAT3: An Emerging Therapeutic Target for Hepatocellular Carcinoma Cancers 2019 11 1646 10.3390/cancers11111646 8. Lee Y.J. Kao E.S. Chu C.Y. Lin W.L. Chiou Y.H. Tseng T.H. Inhibitory effect of ailanthoidol on 12-O-tetradecanoyl-phorbol-13-acetate-induced tumor promotion in mouse skin Oncol. Rep. 2006 16 921 927 10.3892/or.16.4.921 16969515 9. Tseng T.H. Lee H.J. Lee Y.J. Lee K.C. Shen C.H. Kuo H.C. Ailanthoidol, a Neolignan, Suppresses TGF-beta1-Induced HepG2 Hepatoblastoma Cell Progression Biomedicines 2021 9 1110 10.3390/biomedicines9091110 34572296 10. Kim J.K. Jun J.G. Ailanthoidol suppresses lipopolysaccharide-stimulated inflammatory reactions in RAW264.7 cells and endotoxin shock in mice J. Cell Biochem. 2011 112 3816 3823 10.1002/jcb.23312 21826708 11. Park J.H. Jun J.G. Kim J.K. Anti-Adipogenic Activity of Ailanthoidol on 3T3-L1 Adipocytes Biomed. Sci. Lett. 2014 20 62 69 12. Teponno R.B. Kusari S. Spiteller M. Recent advances in research on lignans and neolignans Nat. Prod. Rep. 2016 33 1044 1092 10.1039/C6NP00021E 27157413 13. Romeo M.A. Gilardini Montani M.S. Benedetti R. Santarelli R. D’Orazi G. Cirone M. STAT3 and mutp53 Engage a Positive Feedback Loop Involving HSP90 and the Mevalonate Pathway Front. Oncol. 2020 10 1102 10.3389/fonc.2020.01102 32754441 14. El-Serag H.B. Epidemiology of viral hepatitis and hepatocellular carcinoma Gastroenterology 2012 142 1264 1273.e1261 10.1053/j.gastro.2011.12.061 22537432 15. Aravalli R.N. Cressman E.N. Steer C.J. Cellular and molecular mechanisms of hepatocellular carcinoma: An update Arch. Toxicol. 2013 87 227 247 10.1007/s00204-012-0931-2 23007558 16. Desai A.G. Qazi G.N. Ganju R.K. El-Tamer M. Singh J. Saxena A.K. Bedi Y.S. Taneja S.C. Bhat H.K. Medicinal plants and cancer chemoprevention Curr. Drug Metab. 2008 9 581 591 10.2174/138920008785821657 18781909 17. Ranaware A.M. Banik K. Deshpande V. Padmavathi G. Roy N.K. Sethi G. Fan L. Kumar A.P. Kunnumakkara A.B. Magnolol: A Neolignan from the Magnolia Family for the Prevention and Treatment of Cancer Int. J. Mol. Sci. 2018 19 2362 10.3390/ijms19082362 18. Ong C.P. Lee W.L. Tang Y.Q. Yap W.H. Honokiol: A Review of Its Anticancer Potential and Mechanisms Cancers 2019 12 48 10.3390/cancers12010048 19. Sheen W.S. Tsai I.L. Teng C.M. Chen I.S. Nor-Neolignan and Phenyl Propanoid from Zanthoxylum Ailanthoides Phytochemistry 1994 36 213 215 20. Oren M. Rotter V. Mutant p53 gain-of-function in cancer Cold Spring Harb. Perspect. Biol. 2010 2 a001107 10.1101/cshperspect.a001107 20182618 21. Bykov V.J.N. Eriksson S.E. Bianchi J. Wiman K.G. Targeting mutant p53 for efficient cancer therapy Nat. Rev. Cancer 2018 18 89 102 10.1038/nrc.2017.109 29242642 22. Vijayakumaran R. Tan K.H. Miranda P.J. Haupt S. Haupt Y. Regulation of Mutant p53 Protein Expression Front. Oncol. 2015 5 284 10.3389/fonc.2015.00284 26734569 23. Lin S.Y. Chen C.L. Lee Y.J. Total synthesis of ailanthoidol and precursor XH14 by Stille coupling J. Org. Chem. 2003 68 2968 2971 10.1021/jo020653t 12662079 24. Iwao C. Shidoji Y. Induction of nuclear translocation of mutant cytoplasmic p53 by geranylgeranoic acid in a human hepatoma cell line Sci. Rep. 2014 4 4419 10.1038/srep04419 24658405

PMC009xxxxxx/PMC9099616.txt

==== Front J Clin Med J Clin Med jcm Journal of Clinical Medicine 2077-0383 MDPI 10.3390/jcm11092640 jcm-11-02640 Article Platelet Rich Plasma Injections for Knee Osteoarthritis Treatment: A Prospective Clinical Study Moretti Lorenzo 1† https://orcid.org/0000-0002-8596-2422 Maccagnano Giuseppe 2*† https://orcid.org/0000-0003-3585-1000 Coviello Michele 1 https://orcid.org/0000-0001-9212-1253 Cassano Giuseppe D. 1 Franchini Andrea 1 Laneve Andrea 2 Moretti Biagio 1 Saita Yoshitomo Academic Editor 1 Orthopaedic & Trauma Unit, Department of Basic Medical Sciences, Neuroscience and Sense Organs, School of Medicine, University of Bari Aldo Moro, AOU Consorziale Policlinico, 70124 Bari, Italy; lorenzo.moretti@libero.it (L.M.); michelecoviello91@gmail.com (M.C.); dancassanox@gmail.com (G.D.C.); andreafranchini1988@gmail.com (A.F.); biagio.moretti@uniba.it (B.M.) 2 Orthopaedics Unit, Department of Clinical and Experimental Medicine, Faculty of Medicine and Surgery, University of Foggia, Policlinico Riuniti di Foggia, 71122 Foggia, Italy; andreaa.laneve@gmail.com * Correspondence: g.maccagnano@gmail.com; Tel.: +39-3389652261 † These authors contributed equally to this work. 08 5 2022 5 2022 11 9 264016 4 2022 06 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Background: The aim of this prospective study was to evaluate the efficacy and safety of Platelet Rich Plasma (PRP) injections in patients affected by knee osteoarthritis (KOA). An autologous blood product containing a high percentage of various growth factors (GFs), cytokines and modulating factors as PRP has shown promising results in achieving this goal. Methods: One hundred and fifty-three patients (72 males, mean age 59.06 ± 8.78, range 40–81 years old) from January 2018 to January 2020 received three consecutive PRP injections and completed the follow ups. Western Ontario and McMaster University Osteoarthritis index (WOMAC), Knee society score (KSS) and Visual Analogic Scale (VAS) were evaluated before PRP injection (T0), one month (T1), three months (T2) and six months (T3) after the treatment. All patients underwent baseline and at 6 months MRI and X-ray evaluation. Results: A statistically significant VAS, KSS and WOMAC reduction emerged in the comparison between evaluations (p < 0.05), MRI demonstrated non-statistically significant improvement in cartilage thickness for both tibial plate and femoral plate (p = 0.46 and p = 0.33 respectively), and no radiographic changes could be seen in any patients. Conclusions: PRP injection represents a valid conservative treatment to reduce pain, improve quality of life and functional scores even at midterm of 6 months follow-up. knee osteoarthritis (KOA) cartilage plateled riched plasma (PRP) knee injection biologic therapy This research received no external funding. ==== Body pmc1. Introduction Cartilage structure modifications are responsible for several degenerative joint diseases, such as chondropathy and osteoarthritis. Osteoarthritis (OA) is one of the most common progressive and degenerative knee diseases, affecting the intra-articular, tibiofemoral, and patellofemoral cartilage together with the adjacent joints and structures [1,2]. Musculoskeletal pain and movement restriction are symptoms associated with OA, resulting in a reduction in daily performance [3,4]. According to Kurtz et al. [5] the percentage of knee OA is expected to increase over the next 10 years due to the growing rate of obesity and of the population’s average age. Additionally, a recent study demonstrated that thickness and volume of cartilages are significantly lower for 50-year-old patients, with a Body Mass Index (BMI) equal to and greater than 25 [6]. Several surgical and nonsurgical treatments have been suggested to treat the join pain [7,8]; the latest are especially recommended for young and middle-aged patients presenting earlier stages of OA [1]. Among the conservative treatments, the use of non-steroidal anti-inflammatory drugs (NSAIDs), the intra-articular injections with corticosteroid (CS) or hyaluronic acid (HA) and saline have been used to manage mild KOA for several years [4]. Various meta-analyses investigated the effectiveness of the Platelet-rich plasma (PRP) by comparing it with other procedures, the results highlighted a better pain relief and functional improvement observed at different times after injection. In particular, the PRP is an autologous blood product containing a high percentage of various growth factors (GFs), such as fibroblast growth factor, epidermal growth factor, vascular endothelial growth factor, transforming growth factor-β and platelet-derived growth factor [4]. A recent study suggested that these GFs and cytokines, released by platelets after being damaged by an injury or pathology, might be involved in modulating the inflammatory processes contributing to the tissue structures preservation or regeneration [9]. Different APCs (Autologous platelet concentrates) were used in regenerative and reparative medicine, such as plasma rich in growth factors (PRGF), advanced-platelet rich fibrin (A-PRF) and injectable-platelet rich fibrin (i-PRF). They demonstrated excellent results like PRP in different fields such as maxillofacial surgery and orthopaedic. [10,11,12] Moreover, the effect of PRP injections on MRI changes remains unclear [13]. The current study aims to assess the clinical effects of PRP injections in patients affected by KOA of grades from 1 to 3 (Kellgren-Lawrence (K-L) radiographic classification scale) at 6 months follow-up, with VAS reduction as a definite primary endpoint. 2. Materials and Methods We designed a prospective study which was approved by the local Ethics Committee (delib. 0104) and registered at ClinicalTrials.gov (accessed on 21 April 2021) (NCT04852380). The subjects enrolled were informed that data from the research would be submitted for publication and signed an informed consent form. Two hundred and ten patients referred to the Orthopaedic and Trauma Unit of the local University Hospital between January 2018 and January 2020 with knee osteoarthritis (KOA) were prospectively recruited. Twelve of them refused the procedures, twenty-one do not respect inclusion or exclusion criteria and fourteen were excluded because they were affected by SARS-COV2. Ten patients were lost at the follow up. Finally, 153 patients were enrolled in the study (Scheme 1). Inclusion criteria were: (1) Age between 40 and 81; (2) Body mass index (BMI) between 20 and 29.9; (3) Chronic history (for at least 4 months) of knee joint pain; (4) Radiographically documented knee osteoarthritis of grades 1 to 3 (Kellgren-Lawrence K-L radiographic classification scale). Exclusion criteria were: (1) Radiographically severe documented knee osteoarthritis of grade 4 (K-L radiographic classification scale); (2) Previous femur and tibia fractures; (3) Knee previous surgical treatment (e.g., arthroscopy); (4) Hyaluronic acid infiltration within the previous six months; (5) Hemoglobin levels < 10 g/dL; (6) History of oncohematological disease, infections, or immunodepression. All the patients were subjected to a clinical evaluation before starting the procedure at each visit, and routine blood tests were carried out before injection, including complete blood count and screening for transmittable diseases (e.g., HIV, HBV, HCV). All tests were performed in the same place, and the same researchers (two orthopaedic surgeons with more than ten years of experience in knee surgery) tested all patients. The evaluation times were: T0 (recruitment), T1 (one month after the last injection), T2 (three month after the last injection), T3 (six month after the last injection). The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), the Knee Score Society (KSS) score and the Visual Analogue Scale (VAS) were evaluated and recorded for each patient at each follow up. Their MRI and X-ray images at baseline and at 6 months were analyzed and included in this study. As a primary endpoint, pain was quantified using the VAS scale with scores ranging from 0 (no pain) to 10 (worst imaginable pain). The functional recovery as a secondary endpoint was monitored using the WOMAC index and KSS. The PRP concentrate required for injection was prepared at the Immunohematology Department of Bari University Hospital by apheresis of venous blood from each patient. All the patients were advised to fast for 10 h before the blood collection in an effort to avoid any effects of food intake on PRP concentrate, meanwhile water intake was not restricted. Venous blood drawn from each patient was centrifuged at room temperature with Arthrex Angel System (Arthrex®, Naples, FL, USA) in order to separate the blood, the plasma, the buffy coat and residual red blood cells (RBCs). The total amount of the PRP concentrate obtained was divided in three part (about 5 mL each), and then stored at −40 °C at Blood Bank of Bari University Hospital. The 5 mL 5% concentrate PRP was injected every week three times, starting from recruitment. The patient was in a supine position with a 90 degree knee flexion. All the procedures were carried out in an aseptic condition with a 21-gauge needle inserted into the antero-lateral soft spot area of the knee. The patient was observed for 30 min under medical care after the procedures and then they were discharged if no complications appear. A post-treatment therapy was prescribed to each patient consisting of an antibiotic cycle, functional rest for 24–48 h, paracetamol (in case of postprocedural pain) and local cryotherapy. No adverse event was observed in the treated data. The WOMAC index, ranging between 0 and 96, consists of five items for pain (score range 0–20), two for stiffness (score range 0–8), seventeen for functional limitations (score range 0–68) and WOMAC total index, with lower WOMAC value indicating benefits after treatment. The KSS is divided in two parts: Knee Score and Function Score. The first includes measure of pain (maximum 50 points), range of motion (maximum 25 points), and anteroposterior and mediolateral stability (maximum 25 points), and flexion contracture, extension range, and alignment were evaluated. The second part consists of walking and stair climbing skills (maximum 50 points each); at the same time walking aids were considered. Scores of 80 to 100 were considered as excellent, 70 to 79 good, 60 to 69 fair, and less than 60 poor [14]. Magnetic resonance imaging data were acquired on Siemens MAGNETOM® Essenza (Milan, Italy) 1.5 T, extremity coil, and used for image analysis: sagittal T1 spin echo. Patients remained in a supine position with a fully extended knee and the foot perpendicular to the MRI table. Femoro-tibial cartilage was divided into four plates by anatomic location: medial tibial, lateral tibial, medial femoral and lateral femoral. Three measurement points were taken for each plate: anterior, median and posterior. The mean cartilage thickness and standard deviation (SD) were calculated for tibial and femoral plates at baseline and at 24 weeks. A prospective clinical study was conducted. The data were collected and analyzed using SPSS (v 23; IBM® Inc., Armonk, NY, USA). Descriptive statistics were calculated for the overall sample and for follow-up and pathology. Categorical variables were presented as numbers or percentages. Continuous variables not normally distributed were presented as median and range normally distributed variables were presented as mean and standard deviation. The Shapiro-Wilk test was used to test for normality of the data. Mann-Whitney U tests or Kruskal-Wallis tests for group comparisons were conducted for follow-ups and pathologies, since the variables were not normally distributed. A p-value of <0.05 was considered statistically significant. 3. Results 3.1. Patient Characteristics One hundred and fifty-three patients (72 males) affected by knee osteoarthritis disorder were recruited in the current study. The mean age of the OA patient was 59.06 ± 8.78 SD with the patients’ age range between 40 and 81 years. The mean BMI of all treated patients was 25.4 ± 3.9 SD. Demographics of patients are shown in Table 1. 3.2. WOMAC Statistically significant differences (p ≤ 0.05) occurred in the 4-time WOMAC for the functional limitations, pain and total WOMAC index. The WOMAC functional limitations value demonstrated statistically significant reduction between T0 and T1 (21.61 ± 12.86 SD vs. 15.78 ± 9.67 SD, p ≤ 0.001), between T1 and T2 (15.78 ± 9.67 SD vs. 12.41 ± 7.36 SD, p ≤ 0.001, between T2 and T3 (12.41 ± 7.36 SD vs. 7.43 ± 5.28 SD, p ≤ 0.001) and between T0 and T3 (21.61 ± 12.86 SD vs. 7.43 ± 5.28 SD, p ≤ 0.001). The WOMAC pain value showed statistically significant reduction between T0 and T1 (6.53 ± 3.65 SD vs. 4.83 ± 3.2 SD, p ≤ 0.001), between T1 and T2 (6.53 ± 3.65 SD vs. 4.09 ± 2.63 SD, p ≤ 0.05), between T2 and T3 (4.09 ± 2.63 SD vs. 2.31 ± 2.17 SD, p ≤ 0.001) and between T0 and T3 (6.53 ± 3.65 SD vs. 2.31 ± 2.17 SD, p ≤ 0.001). The WOMAC total value showed statistically significant reduction between T0 and T1 (30.56 ± 17.55 SD vs. 22.44 ± 13.38 SD, p ≤ 0.001), between T1 and T2 (22.44 ± 13.38 SD vs. 18.10 ± 10.30 SD, p ≤ 0.001), between T2 and T3 (18.10 ± 10.30 SD vs. 10.69 ± 7.53 SD, p ≤ 0.001) and between T0 and T3 (30.56 ± 17.55 SD vs. 10.69 ± 7.53 SD, p ≤ 0.001). The WOMAC stiffness value showed statistically significant reduction between T0 and T1 (2.42 ± 2.12 SD vs. 1.82 ± 1.64 SD, p ≤ 0.05), between T2 and T3 (1.60 ± 1.37 SD vs. 0.93 ± 1.05 SD, p ≤ 0.001) and between T0 and T3 (2.42 ± 2.12 SD vs. 0.93 ± 1.05 SD, p ≤ 0.001). No statistically significant difference was shown between T1 and T2 (1.82 ± 1.64 SD vs. 1.60 ± 1.37 SD, p = 0.33). Results are illustrated in Figure 1. 3.3. KSS Statistically significant differences (p ≤ 0.05) occurred in the 4-time KSS for the knee score, and functional score, excepting for the value comparing T1 and T2 (p > 0.05) for functional score only. The Knee Score, part of KSS, showed a significant increase over time between T0 and T1(82.61 ± 16.24 SD vs. 87.50 ± 14.14 SD, p ≤ 0.001), between T1-T2 (87.50 ± 14.14 SD vs. 89.78 ± 12.41 SD, p ≤ 0.05), between T2-T3 (89.78 ± 12.41 SD vs. 91.01 ± 12.03 SD, p ≤ 0.05) and between T0-T3 (82.61 ± 16.24 SD vs. 91.01 ± 12.03 SD, p ≤ 0.001). The functional KSS score showed a statistically significant differences between T0 and T1(82.97 ± 17.59 SD vs. 88.75 ± 14.91 SD, p ≤ 0.001), between T2-T3 (90.89 ± 13.15 SD vs. 93.72 ± 11.60 SD, p ≤ 0.05) and between T0-T3 (82.97 ± 17.59 SD vs. 93.72 ± 11.60 SD, p ≤ 0.001). No statistically significant difference was shown between T1 and T2 (88.75 ± 14.91 SD vs. 90.89 ± 13.15 SD, p = 0.10). Results are illustrated in Figure 2. 3.4. VAS The VAS score improved statistically significantly between T0 and T1 (4.81 ± 2.11 SD vs. 3.25 ± 1.85 SD, p ≤ 0.001), between T2 and T3 (2.77 ± 1.51 SD vs. 1.79 ± 1.51 SD, p ≤ 0.001), and between T0 and T3 (4.81 ± 2.11 SD vs. 1.79 ± 1.51 SD, p ≤ 0.001). No statistically significant difference was shown comparing T1 and T2 (3.25 ± 1.85 SD vs. 2.77 ± 1.51 SD, p > 0.05). Results are illustrated in Figure 3. 3.5. MRI The tibial and femoral plates thickness improved non-statistically significantly between T0 and T3 (13.04 ± 2.64 SD vs. 13.23 ± 1.87 SD, p = 0.46 and 14.16 ± 2.56 SD vs. 14.40 ± 1.90 SD, p = 0.33, respectively) (Figure 4a,b). No X-ray changes have been demonstrated using Kellgren-Lawrence (K-L) radiographic classification scale (p > 0.05). 4. Discussion The results of this study highlight that PRP infiltrations represent a useful conservative treatment to reduce pain, improve quality of life and functional scores at the midterm of 6 months follow-up in patients with knee osteoarthritis. The PRP injection benefits in joint disease are demonstrated by several authors [4,15,16]. Positive effects were shown on patients with decreasing pain as indicated by WOMAC pain index, KSS index and VAS. The WOMAC and KSS also indicated that the physical function was improved, even if there was no statistical evidence comparing the KSS score at 1 month and 3 months of the follow-up. A statistically significant subjective and objective improvement of the three scales was demonstrated. The effects of PRP injection on pain reduction have been previously observed in other studies and several authors [9,17,18] have reported the analgesic properties of platelets. More recently, a meta-analysis [2] indicated that PRP reduces pain by influencing the expression of mediators (e.g., prostaglandin E2, substance P, dopamine, 5-hydroxy-tryptamine) and that the GFs, contained in the PRP concentrate, promote the synthesis of cartilage matrix, stimulating the growth of chondrocytes and the inhibition of the local inflammatory response [19]. Tang et al. in their meta-analysis found the differences between the treatment of osteoarthritis with PRP versus Hyaluronic acid [20]. Compared with acid hyaluronic, PRP has greater benefits in the conservative treatment of knee osteoarthritis, such as less long-term discomfort and better knee joint function. PRP poses no additional risks and can be used as a conservative treatment for osteoarthritis in the knee. Patients with knee osteoarthrosis who received PRP intra-articular injections had the best overall success when compared to steroids, hyaluronic acid, and placebo at 3, 6, and 12-month follow-ups [21]. Cavazos et al. [22] suggest that while both single and multiple PRP injections improved pain and there was no difference between the two, triple PRP injections were more effective than single injections in enhancing joint functionality in individuals with knee OA. Moreover, different APCs (Autologous platelet concentrates) were used in regenerative and reparative medicine. For instance, advanced platelet-rich fibrin+ (A-PRF+) and leucocyte platelet-rich fibrin protocol (L-PRF) demonstrated to control bleeding and to promote injured tissues healing [23]. Platelet concentrate without the use of anticoagulants (i-PRF) showed a high influence on osteoblast behavior by influencing human osteoblast migration, proliferation and differentiation [24]. All the scores referring to stiffness and physical functional showed an improvement overtime, agreeing with a previous study [25,26] which pointed out a decrease in the WOMAC index and an increase in the KSS total score, suggesting a positive influence of the treatment. However, it is worth noting the lack of statistical differences in the Knee Score part of KSS that refers to flexion contracture, the extension lag, the alignment, anteroposterior and mediolateral stability in contrast with the pain and range of motion. This aspect has never been pointed out by other authors, which referred only to the total KSS score, suggesting that further study is required to highlight the role of the pain score on the total KSS score. The literature remains unclear about the correlation between MRI and PRP injections [13]. Most of the studies do not show an improvement in both RX and MRI [27], in line with our results. Moreover, a lack of standardization of cartilage thickness measurement, which is different for each study, is highlighted. A standardized method is needed to make the studies homogeneous and consequently to analyze the results with higher level studies, such as reviews or metanalysis. This study has some limitations. First, a control group was not included in this study. Second, there is no comparison between single and multiple injections. Additionally, the follow-up period was short and long-term effectiveness was therefore not assessed. On the other hand, a strong point of this work is the selection of treated patients, respecting stringent inclusion and exclusion criteria and the sample size. 5. Conclusions The results confirm the efficacy of the PRP injections on the KOA, also suggesting that decreasing pain was obtained already after one month after injection with best results observed after 6 months. They suggests in further studies that data could be collected and recorded at only three different times: at recruitment, one, and six months after administration, leading to a time reduced follow-up protocol. Additionally, further studies are required to assess the long-term effects of this technique, testing the PRP injection on a large number of patients. Author Contributions Conceptualization, G.D.C.; data curation, M.C.; writing—original draft preparation, A.L. and A.F.; writing—review and editing, L.M. and G.M.; supervision, B.M. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of Policlinico di Bari Hospital (delib. 0104) and registered at ClinicalTrials.gov (accessed on 21 April 2021) (NCT04852380). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the patient(s) to publish this paper if applicable. Data Availability Statement The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy. Conflicts of Interest The authors declare no conflict of interest. Figures, Scheme and Table jcm-11-02640-sch001_Scheme 1 Scheme 1 Flow diagram for enrollment and analysis. Figure 1 Comparison between evaluation times of WOMAC score. (T0: Recruitment, T1: one month after the last injection, T2: three months after the last injection, T3: six months after the last injection, WOMAC: The Western Ontario and McMaster Universities Osteoarthritis Index). Figure 2 Comparison between evaluation times of KSS score. (T0: Recruitment, T1: one month after the last injection, T2: three months after the last injection, T3: six months after the last injection, KSS: The Knee Society Score). Figure 3 Comparison between evaluation times of VAS score. (T0: Recruitment, T1: one month after the last injection, T2: three months after the last injection, T3: six months after the last injection, VAS: The Visual Analogic Scale). Figure 4 Magnetic Resonance Imaging evaluation of knee at T0 and T3. (a): Baseline (T0) MRI T1-sagittal image of a 41 years old female. (b): Six months (T3) follow-up MRI T1-sagittal image of a 41 years old female. The arrows show a slight increase in cartilage thickness, although non-statistically significant. jcm-11-02640-t001_Table 1 Table 1 Baseline evaluation of study participants. Participant’s Variables Age (year)   Mean ± SD 59.06 ± 8.78   Range 40–81 Gender   Male. n (%) 72 (47%)   Female. n (%) 81 (53%) BMI (kg/m2)   Mean ± SD 25.4 ± 3.9   Range 21.5–29.3 Side   Left. n (%) 67 (43.8%)   Right. n (%) 86 (56.2%) Kellgren-Lawrence Grade   Grade 1. n (%) 27 (17.7%)   Grade 2. n (%) 58 (37.9%)   Grade 3. n (%) 68 (44.4%) No statistical differences emerged between recruitment and to the subsequent follow-ups about epidemiological (sex, age and BMI: Body Mass Index) data. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Dai W.-L. Zhou A.-G. Zhang H. Zhang J. Efficacy of Platelet-Rich Plasma in the Treatment of Knee Osteoarthritis: A Meta-analysis of Randomized Controlled Trials Arthroscopy 2017 33 659 670.e1 10.1016/j.arthro.2016.09.024 28012636 2. Ren H. Zhang S. Wang X. Li Z. Guo W. Role of platelet-rich plasma in the treatment of osteoarthritis: A meta-analysis J. Int. Med. Res. 2020 48 1 10 10.1177/0300060520964661 3. Paterson K.L. Nicholls M. Bennell K.L. Bates D. Intra-articular injection of photo-activated platelet-rich plasma in patients with knee osteoarthritis: A double-blind, randomized controlled pilot study BMC Musculoskelet. Disord. 2016 17 67 10.1186/s12891-016-0920-3 26861957 4. Ip H.L. Nath D.K. Sawleh S.H. Kabir H. Jahan N. Regenerative Medicine for Knee Osteoarthritis—The Efficacy and Safety of Intra-Articular Platelet-Rich Plasma and Mesenchymal Stem Cells Injections: A Literature Review Cureus 2020 12 e10575 10.7759/cureus.10575 33101819 5. Kurtz S. Ong K. Lau E. Mowat F. Halpern M. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030 J. Bone Jt. Surg.—Ser. A 2007 89 780 785 10.2106/00004623-200704000-00012 6. Matada M.S. Holi M.S. Raman R. Jayaramu Suvarna S.T. Visualization of Cartilage from Knee Joint Magnetic Resonance Images and Quantitative Assessment to Study the Effect of Age, Gender and Body Mass Index (BMI) in Progressive Osteoarthritis (OA) Curr. Med. Imaging 2019 15 565 572 10.2174/1573405614666181018123251 7. Carr A.J. Robertsson O. Graves S. Price A.J. Arden N.K. Judge A. Beard D.J. Knee replacement Lancet 2012 379 1331 1340 10.1016/S0140-6736(11)60752-6 22398175 8. Cattaneo G. De Caro A. Napoli F. Chiapale D. Trada P. Camera A. Micro-fragmented adipose tissue injection associated with arthroscopic procedures in patients with symptomatic knee osteoarthritis BMC Musculoskelet. Disord. 2018 19 176 10.1186/s12891-018-2105-8 29848328 9. Guillibert C. Charpin C. Raffray M. Benmenni A. Dehaut F.X. El Ghobeira G. Giorgi R. Magalon J. Arniaud D. Single injection of high volume of autologous pure PRP provides a significant improvement in knee osteoarthritis: A prospective routine care study Int. J. Mol. Sci. 2019 20 1327 10.3390/ijms20061327 10. Raeissadat S.A. Hosseini P.G. Bahrami M.H. Roghani R.S. Fathi M. Ahangar A.G. Darvish M. The comparison effects of intra-articular injection of Platelet Rich Plasma (PRP), Plasma Rich in Growth Factor (PRGF), Hyaluronic Acid (HA), and ozone in knee osteoarthritis; a one year randomized clinical trial BMC Musculoskelet. Disord. 2021 22 134 10.1186/s12891-021-04017-x 33536010 11. Antonelli A. Giudice A. Muraca D. Fortunato L. Usefulness of advanced-platelet rich fibrin (A-PRF) and injectable-platelet rich fibrin (i-PRF) in the management of a massive medication-related osteonecrosis of the jaw (MRONJ): A 5-years follow-up case report Indian J. Dent. Res. 2020 31 813 818 10.4103/ijdr.IJDR_689_19 33433526 12. Albilia J. Herrera-Vizcaíno C. Weisleder H. Choukroun J. Ghanaati S. Liquid platelet-rich fibrin injections as a treatment adjunct for painful temporomandibular joints: Preliminary results Cranio 2020 38 292 304 10.1080/08869634.2018.1516183 30231809 13. Samara O. Al-Ajlouni J. Al-Najar M. Saleh M. Al-Ryalat N. Gharaibeh A. Al Khayat O.W. Awidi M. Dweik M. Awidi A. Intra-Articular Autologous Platelet Lysates Produce Positive Mri Structural Changes in Early and Intermediate Knee Osteoarthrosis Pak. J. Radiol. 2017 27 14 18 14. Rahimzadeh P. Imani F. Faiz S.-H. Alebouyeh M.-R. Azad-Ehyaei D. Bahari L. Memarian A. Kim K.-H. Adding Intra-Articular Growth Hormone to Platelet Rich Plasma under Ultrasound Guidance in Knee Osteoarthritis: A Comparative Double-Blind Clinical Trial Anesthesiol. Pain Med. 2016 6 e41719 10.5812/aapm.41719 15. Cheng L. Chang S. Qian L. Wang Y. Yang M. Extracorporeal shock wave therapy for isokinetic muscle strength around the knee joint in athletes with patellar tendinopathy J. Sports Med. Phys. Fit. 2019 59 822 827 10.23736/S0022-4707.18.09023-0 16. Gobbi A. Lad D. Karnatzikos G. The effects of repeated intra-articular PRP injections on clinical outcomes of early osteoarthritis of the knee Knee Surg. Sports Traumatol. Arthrosc. 2015 23 2170 2177 10.1007/s00167-014-2987-4 24748286 17. Asfaha S. Cenac N. Houle S. Altier C. Papez M.D. Nguyen C. Steinhoff M. Chapman K. Zamponi G.W. Vergnolle N. Protease-activated receptor-4: A novel mechanism of inflammatory pain modulation J. Cereb. Blood Flow Metab. 2007 150 176 185 10.1038/sj.bjp.0706975 18. Moretti L. Notarnicola A. Ostuni A. Pesce V. Maccagnano G. Coviello M. Covelli I. Franchini A. Bianchi F.P. Moretti B. Umbilical cord blood platelet-rich plasma injections for epicondylitis treatment: A prospective clinical study J. Biol. Regul. Homeost. Agents 2021 35 1337 1342 19. Notarnicola A. Tamma R. Moretti L. Panella A. Dell’Endice S. Zallone A. Moretti B. Effect of shock wave treatment on platelet-rich plasma added to osteoblast cultures Ultrasound Med. Biol. 2011 37 160 168 10.1016/j.ultrasmedbio.2010.10.016 21084153 20. Tang J.Z. Nie M.J. Zhao J.Z. Zhang G.C. Zhang Q. Wang B. Platelet-rich plasma versus hyaluronic acid in the treatment of knee osteoarthritis: A meta-analysis J. Orthop. Surg. Res. 2020 15 403 10.1186/s13018-020-01919-9 32912243 21. Migliorini F. Driessen A. Quack V. Sippel N. Cooper B. El Mansy Y. Tingart M. Eschweiler J. Comparison between intra-articular infiltrations of placebo, steroids, hyaluronic and PRP for knee osteoarthritis: A Bayesian network meta-analysis Arch. Orthop. Trauma. Surg. 2020 141 1473 1490 10.1007/s00402-020-03551-y 32725315 22. Vilchez-Cavazos F. Millan-Alanis J.M. Blázquez-Saldaña J. Álvarez-Villalobos N. Peña-Martínez V.M. Acosta-Olivo C.A. Simental-Mendía M. Comparison of the Clinical Effectiveness of Single Versus Multiple Injections of Platelet-Rich Plasma in the Treatment of Knee Osteoarthritis: A Systematic Review and Meta-analysis Orthop. J. Sports Med. 2019 7 10.1177/2325967119887116 23. Brancaccio Y. Antonelli A. Barone S. Bennardo F. Fortunato L. Giudice A. Evaluation of local hemostatic efficacy after dental extractions in patients taking antiplatelet drugs: A randomized clinical trial Clin. Oral Investig. 2020 25 1159 1167 10.1007/s00784-020-03420-3 24. Wang X. Zhang Y. Choukroun J. Ghanaati S. Miron R.J. Effects of an injectable platelet-rich fibrin on osteoblast behavior and bone tissue formation in comparison to platelet-rich plasma Platelets 2017 29 48 55 10.1080/09537104.2017.1293807 28351189 25. Huang P.H. Wang C.J. Chou W.Y. Wang J.W. Ko J.Y. Short-term clinical results of intra-articular PRP injections for early osteoarthritis of the knee Int. J. Surg. 2017 42 117 122 10.1016/j.ijsu.2017.04.067 28476542 26. Shen L. Yuan T. Chen S. Xie X. Zhang C. The temporal effect of platelet-rich plasma on pain and physical function in the treatment of knee osteoarthritis: Systematic review and meta-analysis of randomized controlled trials J. Orthop. Surg. Res. 2017 12 16 10.1186/s13018-017-0521-3 28115016 27. Buendía-López D. Medina-Quirós M. Fernández-Villacañas Marín M.Á. Clinical and radiographic comparison of a single LP-PRP injection, a single hyaluronic acid injection and daily NSAID administration with a 52-week follow-up: A randomized controlled trial J. Orthop. Traumatol. 2018 19 3 10.1186/s10195-018-0501-3 30128934

PMC009xxxxxx/PMC9099617.txt

==== Front Cancers (Basel) Cancers (Basel) cancers Cancers 2072-6694 MDPI 10.3390/cancers14092249 cancers-14-02249 Article Surgery in Acute Metastatic Spinal Cord Compression: Timing and Functional Outcome https://orcid.org/0000-0002-4576-1239 Meyer Hanno S. *† https://orcid.org/0000-0001-9947-2240 Wagner Arthur † https://orcid.org/0000-0002-7592-8773 Raufer Alessandra Joerger Ann-Kathrin https://orcid.org/0000-0003-4123-4690 Gempt Jens https://orcid.org/0000-0001-6486-7955 Meyer Bernhard Rades Dirk Academic Editor Schild Steven E. Academic Editor Department of Neurosurgery, School of Medicine, Technical University of Munich, 81675 Munich, Germany; arthur.wagner@tum.de (A.W.); alessandra.raufer@t-online.de (A.R.); annkathrin.joerger@tum.de (A.-K.J.); jens.gempt@tum.de (J.G.); bernhard.meyer@tum.de (B.M.) * Correspondence: hanno.meyer@tum.de; Tel.: +49-89-4140-2151; Fax: +49-89-4140-4889 † These authors contributed equally to this work. 30 4 2022 5 2022 14 9 224905 3 2022 27 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Simple Summary Spinal metastases affect an exceptionally high number of cancer patients and thereby represent a common challenge for healthcare providers. Patients may suffer from debilitating symptoms, including excruciating back pain, immobility and even neurological dysfunction. An exceptionally acute clinical presentation is caused by the compression of the spinal cord through growth of a spinal metastasis within the spinal canal, which may leave the patient with acute spinal cord injury in need of rapid surgical treatment. In clinical practice and science, no true timeframe has yet been defined within which these patients need to undergo surgery, although it is generally understood that their recovery and functional rehabilitation correlate with the time to surgery after symptom onset. In our study, we analyzed a surgically treated cohort of patients with acute spinal cord injury by metastatic compression to investigate the correlation of the timing of surgery with neurological recovery. We were able to identify a subgroup of patients with significantly improved recovery, in whom surgery was initiated within 16 h after admission. Complication rates were not significantly more frequent in this subgroup compared to patients operated on after 16 h. Based on these findings, we conclude that striving for surgery as early as feasible is a warranted strategy in patients with acute neurological deterioration due to metastatic spinal cord compression. Abstract Background: Patients with metastatic spinal cord compression (MSCC) may experience long-term functional impairment. It has been established that surgical decompression improves neurological outcomes, but the effect of early surgery remains uncertain. Our objective was to evaluate the impact of early versus late surgery for acute MSCC due to spinal metastases (SM). Methods: We retrospectively reviewed a consecutive cohort of all patients undergoing surgery for SMs at our institution. We determined the prevalence of acute MSCC; the time between acute neurological deterioration as well as between admission and surgery (standard procedure: decompression and instrumentation); and neurological impairment graded by the ASIA scale upon presentation and discharge. Results: We screened 693 patients with surgery for spinal metastasis; 140 patients (21.7%) had acute MSCC, defined as neurological impairment corresponding to ASIA grade D or lower, acquired within 72 h before admission. Non-MSCC patients had surgery for SM-related cauda equina syndrome, radiculopathy and/or spinal instability. Most common locations of the SM in acute MSCC were the thoracic (77.9%) and cervical (10.7%) spine. Per standard of care, acute MSCC patients underwent surgery including decompression and instrumentation, and the median time from admission to surgery was 16 h (interquartile range 10–22 h). Within the group of patients with acute MSCC, those who underwent early surgery (i.e., before the median 16 h) had a significantly higher rate of ASIA improvement by at least one grade at discharge (26.5%) compared to those who had late surgery after 16 h (10.1%; p = 0.024). Except for a significantly higher sepsis rate in the late surgery group, complication rates did not differ between the late and early surgery subgroups. Conclusions: We report data on the largest cohort of patients with MSCC to date. Early surgery is pivotal in acute MSCC, substantially increasing the chance for neurological improvement without increasing complication rates. We found no significant impact when surgery was performed later than 24 h after admission. These findings will provide the framework for a much-needed prospective study. Until then, the treatment strategy should entail the earliest possible surgical intervention. functional outcome neurological function recovery spinal metastasis timing Department of NeurosurgeryThe study was fully financed by the Department of Neurosurgery. ==== Body pmc1. Introduction Spinal metastases are the most commonly encountered spinal tumors and have gained considerable epidemiological significance in recent decades. Advances in standards of care and targeted systemic therapies constituted a substantial increase in life expectancy, which in turn led to an increasing incidence of MSCC that has been reported to occur in 5–14% of all cancer patients [1,2]. Modern interdisciplinary treatment paradigms take cancer biology, the prospective survival, and the functional status of the patient into account [2,3]. Patients with MSCC frequently suffer not only from pain, but also from neurological deficits and impaired functional autonomy [4,5]. This negatively impacts their ability to undergo adjunct treatment and thus overall survival [6,7,8,9]. It has long been established that patients with symptomatic MSCC benefit from surgical decompression in addition to radiotherapy with regards to their functional outcome and possibly also to their life expectancy [4,9]. Based on the underlying pathophysiological concepts of spinal cord damage by a growing epidural mass with rapid neurological deterioration, that is, direct pressure and ischemia as well as tumor-related spinal instability, it is obvious that timing is critical, similar to the timing of surgery in spinal trauma. This has prompted most neurosurgical services to adopt a strategy aiming at timely surgery [3,10,11]. However, there are only few studies comparing early versus late surgery in acute symptomatic MSCC to date, and the actual timing of surgery in large centers is unclear. In this retrospective study, we aimed to determine the timing of surgery in a tertiary care spine center and the impact of early surgery on functional recovery in patients with MSCC. 2. Methods 2.1. Patient Population We retrospectively reviewed data of all 693 patients admitted to undergo surgery for SMs at our institution between 2007 and 2019 (Figure 1). Of these, 681 patients had available documentation on symptom onset, timing of surgery, functional status on admission and discharge as well as imaging data. One hundred and forty-eight of these patients (21.7%) had symptomatic MSCC with neurological impairment as per the American Spinal Cord Injury Association (ASIA) grade D or lower; the remaining 533 patients (78.3%) had surgery for one or multiple different indications (such as SM-related spinal instability, cauda equina syndrome or radiculopathy, i.e., patients not fulfilling the aforementioned criteria), and were designated our non-MSCC study group. Epidural compression was assessed on preoperative magnetic resonance imaging (MRI), computed tomography (CT) or, in select cases, on post-myelography CT. In patients with multiple metastases located in different parts (cervical, thoracic, lumbar, sacral) of the spine, cases were assigned to the spinal part that was affected by the lesion most relevant to the surgical indication (Table 1). Most common primary entities in the acute MSCC group were prostate (25.0%), breast (15.7%), lung (12.1%), gastrointestinal tract (12.1%) and renal cell (7.6%) cancers. The proportions were similar for the non-MSCC group with prostate (20.5%), breast (19.0%), lung (11.1%), gastrointestinal tract (9.6%) and renal cell (7.6%) cancers, without significant difference between groups (p = 0.401). In 8 of the 148 patients with symptomatic MSCC, the symptom onset was more than 72 h prior to admission. In the remaining 140 patients (20.6%), a new neurological impairment corresponding to ASIA A–D or a deterioration of a pre-existing impairment to ASIA A-D had developed within 72 h prior to admission; these represented our primary study cohort (acute symptomatic MSCC; Figure 1). 2.2. Surgical Treatment and Timing As per standard of care, surgical decompression of the spinal cord was carried out at the symptomatic spinal levels affected by MSCC, i.e., usually at least one or multiple tumor laminectomies and, if needed, additional osteotomies (such as, e.g., pediculectomy or vertebral body replacement). In addition, pedicle screw instrumentation (typically percutaneously) was performed, usually including two spinal segments above and below the decompressed levels. Additional instrumentation of the anterior column (e.g., vertebral body replacement) was performed if indicated based on preoperative imaging, pre- and post-decompression instability, and the general condition and oncologic prognosis of the patient, either immediately or as a staged second surgery. The extent of the decompression, instrumentation and all surgical techniques complied with international guidelines and decision frameworks [12,13]. Due to the time-sensitive nature of surgical treatment, these judgements were frequently not made within the context of an interdisciplinary board meeting. Decisions on further adjunct treatment regimens after completion of surgery were, however, again according to standard of care. The exact times of admission to our hospital, surgical incision and discharge were drawn from the records. The time of symptom onset was established from the patient’s history and records of referring hospitals. The median time between admission to our hospital and surgical incision was analyzed post hoc and represented the basis for further stratification. For the primary outcome in the acute MSCC cohort, we compared the ordinal change in ASIA score by at least one grade between admission and discharge for two predefined subgroups: (1) the early subgroup undergoing surgery within the median time between admission and surgery as established for the entire acute MSCC cohort (2) the late subgroup undergoing surgery after the median. Further cutoffs were defined pre hoc as secondary outcomes—12 h and 24 h, respectively. 2.3. Statistical Analysis For the primary outcome, we used chi-square testing to compare the proportions of improved ASIA grades between subgroups by discharge. Secondary analyses included comparison of occurrence rates of adverse events after surgery and Student’s t-test to compare metric items as well as Wilcoxon’s signed rank test to compare interval items between subgroups. We used IBM SPSS Statistics for Windows version 25.0 (Armonk, NY, USA 2017; https://www.ibm.com/products/spss-statistics (accessed on 1 February 2022)). The level of significance was defined a priori as α = 0.05. 2.4. Ethical Considerations All procedures were indicated and conducted in compliance with our department’s standards and the Declaration of Helsinki. The Ethics Committee Klinikum rechts der Isar of the Technical University Munich (Ethikkommission Klinikum rechts der Isar der Technischen Universität München) granted a positive vote (reference no. 96/19S) and waived the requirement for written informed consent. 3. Results 3.1. Patient Population and Timing of Surgery We screened all 693 patients who had surgery for SM at our institution (cf. Methods; Figure 1). One hundred and forty (21.7%) of these had acute MSCC, defined as neurological impairment corresponding to ASIA grade D or lower acquired within 72 h before admission. All acute MSCC patients underwent surgery including decompression and, in the vast majority of cases (81%), pedicle screw instrumentation; 6% had additional vertebral body replacement. Surgery began within a median time of 16 h (interquartile range 10–22 h admission to surgical incision). This was the basis for defining the early (within 16 h) and late (after 16 h) surgery subgroups. Baseline characteristics of both the non-MSCC group and the acute MSCC group (early vs. late surgery) are shown in Table 1. Non-MSCC patients had surgery much later than acute MSCC patients (147 h; interquartile range 59–312 h; p < 0.001). As expected, the most common locations of the SM in acute MSCC were the thoracic (77.9%) and cervical (10.7%) spine, whereas there were more lumbar/sacral metastases in non-MSCC patients. The most common primary entities of both subgroups were prostate (early: 24.3% vs. late: 25.7%), breast (20.0% vs. 12.9%) and lung (12.9% vs. 11.4%) without statistically significant difference (p = 0.099). There were significantly more male patients in the acute MSCC group compared to the non-MSCC group. In the acute MSCC group, 33.6% of patients presented with ASIA grades A or B. The early and late acute MSCC subgroups did not differ with regards to age, sex, location of SM or initial ASIA grade (Table 1). There was no significant difference in the frequency of additional instrumentation between the early and late subgroups (p = 0.764). The early subgroup had a median hospital stay of 17 days compared to 16 days for the late subgroup (p = 0.527). About half of the patients in the acute MSCC group were admitted within 24 h after symptom onset, the other half after 24–72 h (52% vs. 48%). In the entire cohort, the symptom onset was more than 28 days prior to admission in 31.4% (Table 2). When stratified by symptom onset to time of surgical incision, 16.4% of patients in the acute MSCC group underwent surgery within 24 h, 31.4% within 48 h and 52.1% within 72 h. 3.2. Correlation of Functional Recovery with Timing of Surgery In acute MSCC, patients operated on before the median of 16 h after admission had a significantly higher rate of improvement of ASIA grades by discharge compared to those operated on later than 16 h after admission (Table 3). The rates differed by a factor of about 2.5 (26.5% vs. 10.1% ASIA improvement). In the early subgroup, 16.2% improved by one ASIA grade and 10.3% by 2 grades, whereas in the late subgroup, 9.0% improved by 1 grade and only 1.1% by 2 grades. This benefit of early surgery increases when a cutoff of 12 h between admission and surgical incision is defined, for an improvement rate of 32.6% in the pre-12 h stratum compared to 11.0% after 12 h (p = 0.008; Table 3). With a cutoff of 24 h, no significant differences were found. A stratification by symptom duration also did not yield significant differences in the rates of ASIA grade changes (p = 0.271). 3.3. Correlation of Admission Time with Timing of Surgery When stratified by the time of day, most patients (47.1%) were admitted during regular working hours between 6 a.m. and 4 p.m. (Figure 2). The time of admission influenced the timing of surgery: patients admitted during the night shift, i.e., between 12 p.m. and 6 a.m., were significantly more likely to undergo early surgery (i.e., within the median; 92.3% vs. 7.7%; p < 0.001; Table 4). 3.4. Complications Complication rates including mortality are reported in Table 5. When patients with acute MSCC were stratified by the timing of surgery, the late subgroup experienced a significantly higher rate of sepsis (7.4% vs. 2.2%; p = 0.027); other complication rates did not differ between the early and late surgery subgroups. The time of admission did not significantly affect any of the complication or mortality rates. Compared to patients undergoing surgery for spinal metastasis who did not suffer from acute MSCC, however, both the reoperation rate for surgical complications during the index hospital stay (14.4% vs. 7.6%; p = 0.014) and the ICU admission rate (12.2% vs. 6.6%; p = 0.032) were almost twice as high in the acute MSCC group (Table 5). 4. Discussion 4.1. A Case for Early Surgery in Patients with Acute MSCC This study was carried out to investigate the impact of surgical timing in acute MSCC. We found that the rate of recovery from severe neurological impairment depends on the time spent until surgery. In fact, the proportion of improved patients was increased by a factor of 2.6 in the group that received early surgery within the median 16 h (26.5% vs. 10.1%). While a cutoff of 12 h separated improved patients from unchanged or deteriorated patients even better (32.6% vs. 11.0%), a cutoff at 24 h did not. This indicates that the time window permitting effective spinal cord decompression closes rapidly and it appears that “time is spinal cord” in these patients. This strongly argues for a treatment strategy aiming at the earliest possible surgery in acute MSCC 24 h, seven days a week, and the goal would always be to initiate surgery within 24 h. In our department, it has been standard practice to aim for early surgery including decompression and instrumentation in acute MSCC. This means that patients are operated on hours after admission at the latest. Patients getting ready for surgery during regular operating hours would usually be operated on the same day after regular surgeries are finished, and those arriving later during the day or at night would either be operated on during the night or during the regular operating hours of the next day, depending on anesthesia capacities during the night. This certainly leaves room for even earlier surgery in many cases. One might assume that such an “ultra-early surgery” treatment paradigm could entail side effects negatively affecting the patient, such as an increase in adverse events. However, in this study, complications and especially revision surgeries did not occur more frequently in the early subgroup. The fundamental paradigms underlying surgical treatment of SMs are symptom control, spinal stabilization and preservation of neurological function [12,14,15,16,17,18]. The guidelines available to treating specialists generally regard surgery as a first step in the oncological treatment regimen, laying the foundation for adjunct radiation and systemic treatment tailored to the burden of disease and the primary tumor entity, optimally through a network of oncological specialists [12,13,19]. Maintaining functional autonomy is crucial for patients; not only does it secure quality of life, but often it enables the feasibility of said adjunct treatment regimens in the first place [15,20,21,22,23]. It follows that patients with acute neurological impairment due to metastatic spinal cord compression are particularly prone to remain or to become unfit for further treatment and, given our findings, that early surgery may be even more important in these cases. 4.2. Instrumentation for Spinal Stabilization in Patients with Acute MSCC In our department, posterior pedicle screw instrumentation in addition to decompression is the standard procedure in acute MSCC. This treatment paradigm is based on several aspects, several of which have emerged recently. First, there have always been cases that require immediate spinal stabilization, e.g., when tumor-related instability has led to spinal deformity causing spinal cord compression. Secondly, as mentioned above, the life expectancy of patients with SMs has increased. This means that tumor-related as well as decompression-related instability that might not be relevant in the short term need to be addressed in order to ensure mid- and even long-term pain relief and functional independence. Spinal cord decompression via tumor laminectomy renders the affected segments of the spine less stable, promoting sagittal malalignment such as post-laminectomy kyphosis that is associated not only with pain, but also with neurological deterioration. This particularly holds true for junctional segments, where often not only posterior, but also anterior column stabilization may be warranted [24,25]. Thirdly, the advent of minimally invasive approaches including the routine use of navigational systems and specialized pedicle screw systems have drastically facilitated spinal instrumentation and improved procedure-related safety and efficacy [2,26,27]. These developments have rendered posterior pedicle screw instrumentation in addition to decompression the standard of care in many centers across a wide range of SM patients [3,12,19,28,29]. Naturally, the decision for additional instrumentation must be based on weighing these substantial benefits against potential disadvantages associated with more extensive surgery. The patient’s individual performance status, general condition and burden of the systemic disease must always be taken into account [8,12,18,30]. A staged approach after initial pedicle screw instrumentation and decompression of MSCC may for instance comprise the initiation of adjunct radiation therapy before conducting delayed anterior instrumentation or addressing asymptomatic but nonetheless unstable lesions [7,15,28]. In our department, we have refrained from immediate instrumentation in cases that are unlikely to become unstable even with tumor laminectomies (e.g., in patients with osteoblastic metastases, or with sufficient fusion already present). In patients with extremely poor general condition requiring the least invasive surgery possible, we typically have aimed at staged instrumentation when tumor-related or decompression-related instability was present. At any rate, it is crucial to adopt a standard of treatment guideline for clinical decision-making in such scenarios. Surgical treatment should be integrated into an overarching interdisciplinary oncological treatment regimen whenever possible, as one of the primary goals is to preserve or restore functional competence of patients in order to be able to undergo radiation and systemic therapies [8,9,12,19,31]. Several guidelines are available to aid in these decision-making processes, such as the NOMS framework [14]. Of note, because of the lack of reliable data, these eschew providing clear recommendations on the timing of surgical intervention. 4.3. Comparison with Previous Studies and Outlook In contrast to traumatic spinal cord injury, which has repeatedly been demonstrated to require timely surgical intervention [11,32,33], there is little evidence concerning the timing of surgery in acute MSCC, and what is available originates from retrospective investigations [1,34,35,36]. This is mirrored in vague recommendations in current practice guidelines that eschew any explicit recommendation regarding the timing of surgical intervention [37]. More recently, prospective cohort studies were able to demonstrate superior neurological outcomes including the ability to ambulate with early surgery, although the evidence level remains low due to very limited cohort sizes [35,38]. Van Tol et al. were able to analyze a sizable series and found that timely surgery may lead to better quality of life and higher survival rates; their study, however, is limited by tremendous variation in the allocation to their delayed and early arms based on a loose definition at the discretion of the attending surgeon [17]. Even though the present study is also based on a monocentric retrospective analysis, which is obviously its most significant limitation, we are confident that our findings add to the evidence supporting early surgery in acute MSCC given that we report the largest series to date and that we are able to delineate, for the first time, a reasonable time cutoff (the impact of surgery was even more pronounced with 12 h than with the median 16 h after admission, and there was no difference with a cutoff of 24 h after admission any more). This will provide the framework for a much-needed prospective study that may eventually inform practice guidelines and lead to implications for both in-house treatment strategies as well as patient referral strategies within the hospital network aiming at early recognition and transfer of patients in need of rapid decompressive surgery. The potential benefit may represent the difference between a patient being able to walk on their own and a patient being bedridden for the rest of their life. This in itself harbors a multitude of implications for the patients’ ability to pursue activities of daily life, their functional independence, health-related quality of life and fitness to undergo adjunct systemic oncological treatment and, eventually, patient survival [17,39]. 5. Study Limitations By nature, retrospective analyses may be limited by imprecise data, particularly concerning time periods. This likely explains why we were not able to find a significant impact of symptom duration, which is, at the resolution of hours, much more difficult to narrow down than the time between admission and surgery in a retrospective series. Consequently, prospective studies are needed to confirm our findings. Until then, the significant outcome difference between the early and late surgery groups in the present study warrants a treatment strategy aiming at surgery as fast as feasible and possible in acute MSCC, without jeopardizing the patients’ safety. 6. Conclusions In this largest cohort study of patients with acute MSCC to date, early surgery within 16 h or even 12 h appears to substantially increase the chance of functional recovery without increasing the risk of serious peri- and postoperative complications. Prospective studies are needed to establish a distinct cutoff and high-quality evidence regarding the impact of early surgery on both neurological outcome as well as adverse events. Until then, surgical treatment strategies should aim at early intervention in these patients given consistent findings in retrospective series. Author Contributions Conceptualization—H.S.M.; data curation—A.R., A.W., A.-K.J. and J.G.; formal analysis—A.W. and H.S.M.; project administration—B.M., J.G. and H.S.M.; supervision—H.S.M. and B.M.; writing: original draft—A.W. and H.S.M.; writing: review and editing—H.S.M., A.W. and B.M. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee Klinikum Rechts der Isar of the Technical University of Munich, reference number: 96/19S. Informed Consent Statement The need for informed consent was waived by the Ethics Committee. Data Availability Statement Data is available in the text. Conflicts of Interest The authors declare that they have no conflict of interests affecting this study. Figure 1 Flow chart illustrating patient groups. Bold type denotes subgroup with acute MSCC used for primary analyses. SM, spinal metastasis; MSCC, metastatic spinal cord compression. Figure 2 Proportions of patients in the acute MSCC subgroup admitted at different times of day. cancers-14-02249-t001_Table 1 Table 1 Baseline characteristics of both the non-MSCC as well as the acute MSCC groups. Early and late subgroups within acute MSCC defined by surgery within or after the median time from admission to surgery; N—number; p—level of significance; bold typeface denotes statistically significant difference. Variation Acute MSCC p Early/Late Non-MSCC n = 533 p MSCC/Others Early n = 70 Late n = 70 Mean age in years (range) 65.2 (24.3–94.3) 69.1 (15.9–93.6) 0.637 65.7 (15–94) 0.255 Sex (n; % of subgroup) Male 49 50 0.702 315 0.007 (70.0%) (71.4%) (59.1) Localization (n; % of subgroup) Cervical 6 9 0.476 40 <0.001 (8.6%) (12.9%) (7.4) Thoracic 57 52 204 (81.4%) (74.3%) (37.7) Lumbar 0 2 163 (0%) (2.9%) (30.1) Cervicothoracic 3 3 51 (4.3%) (4.3%) (9.4) Thoracolumbar 4 4 41 (5.7%) (5.7%) (7.6) Lumbosacral 0 0 42 (0) (0) (7.7) Mean hours admission–cut (hours) 9.6 27.8 <0.001 245.7 <0.001 ASIA on admission (n; % of subgroup) A 10 12 0.752 - - (14.3%) (17.1%) B 14 11 (20.0%) (15.7%) C 16 18 (22.9%) (25.7%) D 30 29 (42.9%) (41.4%) E 0 0 (0.0%) (0.0%) cancers-14-02249-t002_Table 2 Table 2 Times of symptom onset prior to admission for acute MSCC subgroup and entire cohort. Symptom Onset Acute MSCC All n % n % Unknown 38 5.6 <6 h 23 16.4 32 4.7 6–24 h 44 31.4 64 9.4 24 h–3 d 73 52.2 95 14.0 3 d–7 d 57 8.4 7 d–28 d 181 26.6 >28 d 214 31.4 cancers-14-02249-t003_Table 3 Table 3 Changes in ASIA grades by discharge in patients with acute MSCC, stratified by different surgical timing cutoffs: left column, by median time interval from admission to surgery (i.e., within vs. after 16 h); middle column, within vs. after 12 h; right column, within vs. after 24 h. Bold typeface denotes statistically significant difference. Variation Early n = 70 Late n = 70 p <12 h n = 46 >12 h n = 94 p <24 h n = 125 >24 h n = 15 p ASIA change by discharge Same (%) 70.6 81.2 0.026 63.0 83.0 0.006 77.6 66.7 0.617 Worse (%) 2.9 8.7 4.3 6.4 5.6 6.7 Better (%) 26.5 10.1 32.6 10.6 16.8 26.7 ASIA grades by discharge A (%) 13.8 21.4 0.010 23.9 8.5 0.011 12.8 6.7 0.491 B (%) 15.7 9.6 8.7 10.6 11.2 0.0 C (%) 20.0 31.4 8.7 6.4 8.0 13.3 D (%) 37.1 34.3 10.9 31.9 25.6 20.0 E (%) 14.3 4.3 47.8 42.6 42.4 60.0 cancers-14-02249-t004_Table 4 Table 4 Patients with acute MSCC: Timing of surgery in relation to time of admission to hospital. Bold typeface denotes statistically significant difference. Variation Time of Admission (Hours) p 6–16 16–24 24–6 Timing of surgery Early  (n; % of column) 26 (39.4%) 32 (52.5%) 12 (92.3%) <0.001 Late  (n; % of column) 40 (60.6%) 29 (47.5%) 1 (7.7%) cancers-14-02249-t005_Table 5 Table 5 Complication rates after surgery, stratified by timing of surgery, time of admission and for acute MSCC versus non-MSCC patients. ICU—intensive care unit; bold typeface denotes statistically significant difference. Variation Timing of Surgery Time of Admission MSCC vs. Others Early Late p 6:00–16:00 16:00–24:00 24:00–6:00 p Acute MSCC Non-MSCC p Reoperation ( %) 13.0 15.7 0.654 16.7 13.3 7.7 0.668 14.4 7.6 0.014 Pneumonia ( %) 10.9 7.4 0.511 13.0 5.8 8.3 0.441 9.3 5.3 0.095 Thromboembolic event ( %) 1.6 1.9 0.903 3.7 - - 0.300 1.7 5.5 0.084 Sepsis ( %) 2.2 7.4 0.027 1.9 5.8 - 0.425 3.4 1.0 0.053 ICU stay ( %) 12.9 11.5 0.810 8.2 17.5 7.7 0.262 12.2 6.6 0.032 Mortality ( %) 3.0 6.6 0.340 4.9 5.6 - 0.692 4.7 2.7 0.251 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Cole J.S. Patchell R.A. Metastatic epidural spinal cord compression Lancet Neurol. 2008 7 459 466 10.1016/S1474-4422(08)70089-9 18420159 2. Sciubba D.M. Petteys R.J. Dekutoski M.B. Fisher C.G. Fehlings M.G. Ondra S.L. Rhines L.D. Gokaslan Z.L. Diagnosis and management of metastatic spine disease. A review J. Neurosurg. Spine 2010 13 94 108 10.3171/2010.3.SPINE09202 20594024 3. Nater A. Martin A.R. Sahgal A. Choi D. Fehlings M.G. Symptomatic spinal metastasis: A systematic literature review of the preoperative prognostic factors for survival, neurological, functional and quality of life in surgically treated patients and methodological recommendations for prognostic studies PLoS ONE 2017 12 e0171507 10.1371/journal.pone.0171507 28225772 4. Patchell R.A. Tibbs P.A. Regine W.F. Payne R. Saris S. Kryscio R.J. Mohiuddin M. Young B. Direct decompressive surgical resection in the treatment of spinal cord compression caused by metastatic cancer: A randomised trial Lancet 2005 366 643 648 10.1016/S0140-6736(05)66954-1 16112300 5. Vanek P. Bradac O. Trebicky F. Saur K. de Lacy P. Benes V. Influence of the Preoperative Neurological Status on Survival after the Surgical Treatment of Symptomatic Spinal Metastases with Spinal Cord Compression Spine 2015 40 1824 1830 10.1097/BRS.0000000000001141 26536440 6. Ju D.G. Zadnik P.L. Groves M.L. Hwang L. Kaloostian P.E. Wolinksy J.-P. Witham T.F. Bydon A. Gokaslan Z.L. Sciubba D.M. Factors associated with improved outcomes following decompressive surgery for prostate cancer metastatic to the spine Neurosurgery 2013 73 657 666 10.1227/NEU.0000000000000070 23839521 7. Wagner A. Haag E. Joerger A.K. Jost P. Combs S.E. Wostrack M. Gempt J. Meyer B. Comprehensive surgical treatment strategy for spinal metastases Sci. Rep. 2021 11 7988 10.1038/s41598-021-87121-1 33846484 8. Douglas S. Huttenlocher S. Bajrovic A. Rudat V. Schild S.E. Rades D. Prognostic factors for different outcomes in patients with metastatic spinal cord compression from cancer of unknown primary BMC Cancer 2012 12 261 10.1186/1471-2407-12-261 22720880 9. Douglas S. Schild S.E. Rades D. A new score predicting the survival of patients with spinal cord compression from myeloma BMC Cancer 2012 12 425 10.1186/1471-2407-12-425 23009630 10. Fehlings M.G. Nater A. Tetreault L. Kopjar B. Arnold P. Dekutoski M. Finkelstein J. Fisher C. France J. Gokaslan Z. Survival and Clinical Outcomes in Surgically Treated Patients with Metastatic Epidural Spinal Cord Compression: Results of the Prospective Multicenter AOSpine Study J. Clin. Oncol. 2015 34 268 276 10.1200/JCO.2015.61.9338 26598751 11. Fehlings M.G. Vaccaro A. Wilson J.R. Singh A. Cadotte D.W. Harrop J.S. Aarabi B. Shaffrey C. Dvorak M. Fisher C. Early versus Delayed Decompression for Traumatic Cervical Spinal Cord Injury: Results of the Surgical Timing in Acute Spinal Cord Injury Study (STASCIS) PLoS ONE 2012 7 e32037 10.1371/journal.pone.0032037 22384132 12. Barzilai O. Fisher C.G. Bilsky M.H. State of the Art Treatment of Spinal Metastatic Disease Neurosurgery 2018 82 757 769 10.1093/neuros/nyx567 29481645 13. Laufer I. Rubin D.G. Lis E. Cox B.W. Stubblefield M.D. Yamada Y. Bilsky M.H. The NOMS framework: Approach to the treatment of spinal metastatic tumors Oncologist 2013 18 744 751 10.1634/theoncologist.2012-0293 23709750 14. Masuda K. Ebata K. Yasuhara Y. Enomoto A. Saito T. Outcomes and Prognosis of Neurological Decompression and Stabilization for Spinal Metastasis: Is Assessment with the Spinal Instability Neoplastic Score Useful for Predicting Surgical Results? Asian Spine J. 2018 12 846 853 10.31616/asj.2018.12.5.846 30213167 15. Nater A. Sahgal A. Fehlings M. Management—Spinal metastases Handb. Clin. Neurol. 2018 149 239 255 10.1016/b978-0-12-811161-1.00016-5 29307356 16. Choi D. Ricciardi F. Arts M. Buchowski J.M. Bunger C. Chung C.K. Coppes M. Depreitere B. Fehlings M. Kawahara N. Prediction Accuracy of Common Prognostic Scoring Systems for Metastatic Spine Disease: Results of a Prospective International Multicentre Study of 1469 Patients Spine 2018 43 1678 1684 10.1097/BRS.0000000000002576 30422958 17. van Tol F.R. Suijkerbuijk K.P.M. Choi D. Verkooijen H.M. Oner F.C. Verlaan J.J. The importance of timely treatment for quality of life and survival in patients with symptomatic spinal metastases Eur. Spine J. 2020 29 3170 3178 10.1007/s00586-020-06599-x 32948899 18. Cofano F. Di Perna G. Alberti A. Baldassarre B.M. Ajello M. Marengo N. Tartara F. Zenga F. Garbossa D. Neurological outcomes after surgery for spinal metastases in symptomatic patients: Does the type of decompression play a role? A comparison between different strategies in a 10-year experience J. Bone Oncol. 2021 26 100340 10.1016/j.jbo.2020.100340 33294319 19. Laufer I. Bilsky M.H. Advances in the treatment of metastatic spine tumors: The future is not what it used to be J. Neurosurg. Spine 2019 30 299 307 10.3171/2018.11.SPINE18709 30835704 20. Padalkar P. Tow B. Predictors of survival in surgically treated patients of spinal metastasis Indian J. Orthop. 2011 45 307 313 10.4103/0019-5413.82333 21772622 21. Paton G.R. Frangou E. Fourney D.R. Contemporary treatment strategy for spinal metastasis: The “LMNOP” system Can. J. Neurol. Sci. 2011 38 396 403 10.1017/S031716710001177X 21515496 22. Kim C.H. Chung C.K. Jahng T.A. Kim H.J. Resumption of ambulatory status after surgery for nonambulatory patients with epidural spinal metastasis Spine J. 2011 11 1015 1023 10.1016/j.spinee.2011.09.007 22000724 23. Cho W. Chang U.K. Neurological and survival outcomes after surgical management of subaxial cervical spine metastases Spine 2012 37 E969 E977 10.1097/BRS.0b013e31824ee1c2 22343276 24. Hubertus V. Gempt J. Mariño M. Sommer B. Eicker S.O. Stangenberg M. Dreimann M. Janssen I. Wipplinger C. Wagner A. Surgical management of spinal metastases involving the cervicothoracic junction: Results of a multicenter, European observational study Neurosurg. Focus 2021 50 E7 10.3171/2021.2.FOCUS201067 33932937 25. Kanda Y. Kakutani K. Sakai Y. Zhang Z. Yurube T. Miyazaki S. Kakiuchi Y. Takeoka Y. Tsujimoto R. Miyazaki K. Surgical outcomes and risk factors for poor outcomes in patients with cervical spine metastasis: A prospective study J. Orthop. Surg. Res. 2021 16 423 10.1186/s13018-021-02562-8 34217343 26. Versteeg A.L. Verlaan J.-J. de Baat P. Jiya T.U. Stadhouder A. Diekerhof C.H. van Solinge G.B. Oner F.C. Complications After Percutaneous Pedicle Screw Fixation for the Treatment of Unstable Spinal Metastases Ann. Surg. Oncol. 2016 23 2343 2349 10.1245/s10434-016-5156-9 26965698 27. Depreitere B. Ricciardi F. Arts M. Balabaud L. Bunger C. Buchowski J.M. Chung C.K. Coppes M.H. Fehlings M.G. Kawahara N. How good are the outcomes of instrumented debulking operations for symptomatic spinal metastases and how long do they stand? A subgroup analysis in the global spine tumor study group database Acta Neurochir. 2020 162 943 950 10.1007/s00701-019-04197-5 31953690 28. Rothrock R. Pennington Z. Ehresman J. Bilsky M.H. Barzilai O. Szerlip N.J. Sciubba D.M. Hybrid Therapy for Spinal Metastases Neurosurg. Clin. N. Am. 2020 31 191 200 10.1016/j.nec.2019.11.001 32147010 29. Hamad A. Vachtsevanos L. Cattell A. Ockendon M. Balain B. Minimally invasive spinal surgery for the management of symptomatic spinal metastasis Br. J. Neurosurg. 2017 31 526 530 10.1080/02688697.2017.1297374 28264589 30. Rades D. Fehlauer F. Veninga T. Stalpers L.J. Basic H. Hoskin P.J. Rudat V. Karstens J.H. Schild S.E. Dunst J. Functional outcome and survival after radiotherapy of metastatic spinal cord compression in patients with cancer of unknown primary Int. J. Radiat. Oncol. Biol. Phys. 2007 67 532 537 10.1016/j.ijrobp.2006.08.047 17084539 31. Rades D. Lange M. Veninga T. Stalpers L.J.A. Bajrovic A. Adamietz I.A. Rudat V. Schild S.E. Final results of a prospective study comparing the local control of short-course and long-course radiotherapy for metastatic spinal cord compression Int. J. Radiat. Oncol. Biol. Phys. 2011 79 524 530 10.1016/j.ijrobp.2009.10.073 20452136 32. Kirshblum S. Waring W. 3rd Updates for the International Standards for Neurological Classification of Spinal Cord Injury Phys. Med. Rehabil. Clin. N. Am. 2014 25 505 517 10.1016/j.pmr.2014.04.001 25064785 33. Lynch J. Cahalan R. The impact of spinal cord injury on the quality of life of primary family caregivers: A literature review Spinal Cord. 2017 55 964 978 10.1038/sc.2017.56 28653672 34. Fan Y. Zhou X. Wang H. Jiang P. Cai S. Zhang J. Liu Y. The timing of surgical intervention in the treatment of complete motor paralysis in patients with spinal metastasis Eur. Spine J. 2016 25 4060 4066 10.1007/s00586-016-4406-7 26831534 35. Lo W.Y. Yang S.H. Metastatic spinal cord compression (MSCC) treated with palliative decompression: Surgical timing and survival rate PLoS ONE 2017 12 e0190342 10.1371/journal.pone.0190342 29287117 36. Lee R.S. Batke J. Weir L. Dea N. Fisher C.G. Timing of surgery and radiotherapy in the management of metastatic spine disease: Expert opinion J. Spine Surg. 2018 4 368 373 10.21037/jss.2018.05.05 30069530 37. Hirnmetastasen und Meningeosis Neoplastica Kommission Leitlinien der Deutschen Gesellschaft für Neurologie AWMF Berlin, Germany 2014 38. Quraishi N.A. Rajagopal T.S. Manoharan S.R. Elsayed S. Edwards K.L. Boszczyk B.M. Effect of timing of surgery on neurological outcome and survival in metastatic spinal cord compression Eur. Spine J. 2013 22 1383 1388 10.1007/s00586-012-2635-y 23275041 39. van Tol F.R. Choi D. Verkooijen H.M. Oner F.C. Verlaan J.J. Delayed presentation to a spine surgeon is the strongest predictor of poor postoperative outcome in patients surgically treated for symptomatic spinal metastases Spine J. 2019 19 1540 1547 10.1016/j.spinee.2019.04.011 31005624

PMC009xxxxxx/PMC9099618.txt

==== Front Int J Environ Res Public Health Int J Environ Res Public Health ijerph International Journal of Environmental Research and Public Health 1661-7827 1660-4601 MDPI 10.3390/ijerph19095476 ijerph-19-05476 Article Non-Farm Employment, Farmland Renting and Farming Ability: Evidence from China Li Jinning 1* Song Shunfeng 2* Sun Guanglin 3 Tchounwou Paul B. Academic Editor 1 Xingzhi College, Zhejiang Normal University, Jinhua 321004, China 2 Department of Economics, University of Nevada, Reno, NV 89557, USA 3 School of Finance, Nanjing University of Finance and Economics, Nanjing 210023, China; sunguanglin008@126.com * Correspondence: lijinning529077@163.com (J.L.); song@unr.edu (S.S.) 30 4 2022 5 2022 19 9 547622 3 2022 28 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). In the process of China’s urbanization, non-farm employment and farmland rental activity are closely correlated. Using data from a survey on rural households in three Chinese provinces, this article examines the relationship between farmland renting activity and non-farm employment with simultaneous equations that consider the farming ability of farmers. Our results are fourfold. First, farmland renting-out promotes non-farm employment, while farmland renting-in reduces non-farm employment. Second, non-farm employment encourages farmland renting-out and decreases farmland renting-in. Third, farming ability increases farmland renting-in but decreases non-farm employment. Fourth, non-farm employment decreases the farming ability of farmers. Based on our empirical findings, we would suggest that the Chinese government further reforms its land system in rural areas, which could better facilitate land-use-right transfer and promote farmland rental market. non-farm employment farmland renting-out farmland renting-in farming ability National Natural Science Foundation of China72003086 National Natural Science Foundation of China72002204 Natural Science Foundation of Zhejiang ProvinceLY20G020012 This research was funded by National Natural Science Foundation of China(72003086), National Natural Science Foundation of China(72002204), Natural Science Foundation of Zhejiang Province (LY20G020012). ==== Body pmc1. Introduction As China has been experiencing rapid urbanization over the past decades, urbanization rates increased from 25.84% in 1990, to 36.22% in 2000, 49.95% in 2010, and 63.89% in 2020 (China National Bureau of Statistics). The urbanization process involves both people and land transfers from the rural sector to the urban sector. The former is evident by the mass peasant migration, such as a rural-migrant population of about 290.77 million in 2019. The latter includes transferring land ownership from collectives to the state and reclassifying suburban rural areas into urban areas. Many studies have investigated the push and pull factors of rural-urban migration [1,2,3], the contributions to urbanization made by migrant workers [2], challenges faced by migrant workers [3], and land issues related to compensation on land ownership transfers and settlements paid to the farmers who lost their land due to urbanization [3,4]. Additionally, some researchers paid attention to the pressure on farmland loss, inefficient land-use patterns, food security, food sovereignty and de-agrarianization caused by urban expansion [5,6,7,8], and readily support the preservation of farmland in urbanization [9,10]. Land rental activities in rural areas are an important issue that is understudied. When farmers move to cities, they leave their land behind in their villages. If a migrant worker’s family lacks sufficient labor to farm its own land, it either leaves the land vacant or rents it out to other farmers. Rural households that have surplus labor or want to enjoy scale effects rent more land for larger-scale of production. A farmland rental market emerges in rural areas, and this is the research focus of our paper. Several interesting questions arise with the land rental market in rural areas. What affects a rural family’s decision to rent out its land? Would the rental activities further encourage farmers to migrate to urban areas and seek non-farm jobs? Recent research shows farming ability has an important influence on farmland rental activity and non-farm employment [11,12,13,14,15]. Do simultaneous relationships exist among non-farm employment, farmland renting, and farming ability? Do the market behaviors differ between the renting-out and the renting-in groups? This paper attempts to answer these questions and aims to enrich the literature on rural migration and land market. There are great differences in farmland transfer mechanisms between rural-to-urban and rural-to-rural. The scope of this research is limited to rural-to-rural land transfers. Since the farmland is owned by the village in China, in this paper, farmland transfers refer to the right-to-use only, with the land ownership remaining to the village. This article uses survey data collected from rural areas in Zhejiang, Henan, and Shaanxi provinces in 2017. It employs a simultaneous equation system to examine the determinants of non-farm employment, farmland renting, and farming ability. Our empirical results show that: (1) farmland renting-out promotes non-farm employment while farmland renting-in reduces non-farm employment; (2) non-farm employment encourages farmers to rent out farmland and decreases the chances for farmers to rent in farmland; (3) farming ability decreases non-farm employment, while increases farmland renting-in; (4) non-farm employment decreases farming ability. The rest of this paper is organized as follows. In Section 2, we discuss the research background, conduct a literature review, and propose hypotheses. Section 3 develops the simultaneous equations, describes the data, and defines the variables. Section 4 presents the empirical results, performs robustness checks, and discusses identification issues. Section 5 provides conclusions and policy implications. 2. Background, Literature Review, and Hypotheses 2.1. Background Over the past decades, a mass of farmers has moved to urban areas. On the one hand, China’s rapid urbanization and industrialization call for rural-to-urban migration of labor [16]. Pulling factors in urban areas (such as higher wages, better education, and more opportunities) also attract rural laborers to flow into non-farm sectors. On the other hand, pushing factors in rural areas (such as lack of farmland, low agricultural productivity, high tax burden before 2004, and poor quality of life) encourage farmers to move into cities [2]. Both the pulling and pushing factors have caused millions of migrants to move or drift into non-farm sectors, often leaving farmland behind in their villages. However, farmers are mandated to work on their farmland allocated by their villages. Otherwise, according to the Land Management Law of China, village committees could retrieve farmland from rural households if the land stays vacant over two consecutive years [17]. Therefore, farmers face choices in their decision-making. They could seek non-farm jobs in urban areas, continue to work on their farmland, rent out or rent in farmland, or let their farmland vacant and face the risk of being retrieved. Decisions on these choices are not independent, due to the fact that seeking jobs in urban areas often implies that they may not have sufficient labor to work on farmland and thus have to either rent out farmland or let the farmland vacant. Likewise, staying in rural areas may suggest that farmers have surplus labor; it would be more productive if they could rent in some farmland to expand the scale of production. Also, in farmers’ decision-making, farming ability could be a factor, which itself could diminish with time that farmers have moved to urban areas and over generations. The above discussions raise the following questions: Will non-farm employment necessarily lead to farmland rent-out or rent-in? Does farmland rent-out or rent-in have an impact on non-farm employment? What are the relationships among non-farm employment, farmland renting, and farming ability? Answering these questions will shed important insights on better understanding the rural farmland rental market, migration behavior, farmers’ non-farm employment, poverty reduction in rural areas, and farmland production efficiency [18,19]. 2.2. Literature Review Our literature review is two-fold. One is about the relationship between non-farm employment and farmland renting activities. Some researchers suggested that non-farm employment shapes farmland renting activities [20,21]. From the perspective of non-farm employment’s stability, Du and Li [22] showed that a permanent non-farm employment of a farmer has a positive impact on farmland transfer, while a temporary non-farm employment of a farmer is not conducive to farmland transfer. From the perspective of household resource allocation, Salvioni [23] showed that the opportunities of non- farm job may affect the decisions of peasant households in household resource allocation. In other words, in the lack of non- farm labor chances, pluriactivity is not available to solve the problem of low farming incomes, when farmers are more dependent on income from farming. This will make farmers more inclined to expand by renting in land from others. The conclusion was supported by Vranken [24] and Akter [25], who suggested farmland is usually rented out to farmers not involving in non-farm jobs, and the lack of non-farm labor opportunities encourages a farming household to rent in more land or rent out less land, so opportunities of non-farm employment are identified as critical factors affecting land rental activities of rural households especially in transition countries. As for the measurement of non-farm employment, we notice that previous studies measured non-farm employment differently. For example, Kung [20] and De Janvry [26] used the share of non-farm income, Che [27] used the number of non-farm workers in a household, and Willmore et al. [28] used the share of a household’s non-farm labor. Other studies suggested farmland renting activities affect non-farm employment. The exiting literature shows renting out farmland promotes non-farm employment of farmers [29]. It is especially true in government-oriented transfers of farmland [30]. Government-oriented transfers of farmland often happen in China, when farmers losing farmland have to seek for non-farm jobs to raise their family income, but not necessarily in other countries, so a large amount of relevant literature is concentrated in China. The other fold is about the role played by farming ability in farmland rental activity and non-farm employment. In terms of the relationship between farming ability and farmland renting activity, a number of previous studies concluded that farmers with greater farming ability are more likely to rent in farmland than those with poor farming ability [11,12,13,14,15]. For instance, Akter [25] suggested that farmland is transferred to those with relatively smaller holding, but greater ability to make productive use of land. This conclusion suggests that farmland rental markets facilitate agricultural production efficiency, as more able farmers rent in farmland from those with less ability. As far as the relationship between farming ability and non-farm employment is concerned, farming households usually seek to combine non-farm and on-farm generating activities in order to optimize their total income and establish a suitable balance between the amount of labor needed for farming activities and the labor required for non-farming activities [31]. In this way, farmers with great farming abilities can be allocated for farming jobs in peasant households instead of non-farm jobs. Besides that, Mishra and Goodwin [32] suggested that more farming experience corresponds to less non-farm employment, which likely reflects the fact that farming experience builds farming-specific human capital. As a result, farming ability and farming income are raised, which makes them less likely to be employed non-farm. The main findings from previous studies are (a) non-farm employment encourages farmland rent-out; (b) farmland rent-out promotes farmers’ non-farm employment; (c) farming ability plays an important role in non-farm employment and farmland rental activity. However, previous studies only focused on one-way causal relationships, although some studies tried to use instrumental variables to correct the biased estimates caused by endogeneity [20,33]. Particularly, farming ability of farmers was ignored as an endogenous variable, which renders inconsistent and biased results. Therefore, this paper makes two contributions to the existing literature. One is to consider farming ability as an endogenous variable. Another is the use of a simultaneous equation system in examining the mutual relationships among non-farm employment, farmland renting activities, and farming ability. In other words, we take the decisions of non-farm employment, farmland rental activity, and farming ability as a whole to estimate the mutual effects, rather than use separate equations to examine one-way causal relationships. Doing so, we can obtain results that are more robust, reliable, and consistent. 2.3. Hypotheses Development Historically speaking, China’s household contract responsibility system in rural areas has played a huge role in economic development. It was the first step of China’s economic reforms, which not only allowed farmers to decide what to grow and thus improved production efficiency, but also freed up rural labors which can better meet the increasing demand for labor in China’s industrialization and urbanization. However, such system created fragmented farmland for farmers and made rural households responsible for their land allocated by villages. On the one hand, the small-scale agricultural cultivation can create difficulties for both individual farms and farming communities [34], for example, it impedes the introduction of new technologies, farming machines, and novel production models, limiting the labor and machine efficiency [35,36,37,38,39]. On the other hand, China’s farmers have a deep local attachment to their farmland, which makes them want to stay home, which could cause an income loss and a labor shortage in urban sectors. A possible solution is to develop a farmland rental market in rural areas. In the market, some rural households can rent-out their farmland, so they can free up more time and engage in part- or full-time non-farm employment to increase income. Other farmers can rent-in farmland to increase their production scale, especially for farmers with greater farming ability [11,12,13,14,15]. A number of previous studies have proved that farmland transfer is an important way to improve farmland production efficiency [10,15,19,39]. With the development of farmland rental market, rural resources could be better used, since renting out farmland frees up some farmers for non-farm employment while renting in farmland allows other farmers to increase production scale, especially those who have greater farming ability. Therefore, this study proposes the following hypothesis: Hypothesis 1 (H1). Renting out farmland promotes the non-farm employment of farmers while renting in farmland decreases the non-farm employment of farmers. China’s industrialization and urbanization in the past decades provided a vast number of opportunities for farmers to engage in non-farm employment. The higher wages in urban areas also attracted many farmers to migrate into cities [2,3]. Among the migrant workers, most are young and middle-aged farmers who are the main labor force for farming activities [40]. One consequence is that some rural households became labor shortage. For these households, as an “economic man”, renting out part or all of their farmland could be a one-stone-two-birds choice under the current system and policy. In addition, the past commentators have indicated that taking up off-farm work would result in a decline in the technical efficiency of farming production [31], which is enable farmers involving in non-farm jobs to rent out their farmland for rent income. On the one hand, renting out farmland would allow farmers to collect some symbolic rent and receive government agricultural subsidies. On the other hand, renting out farmland helps farmers prevent their farmland from being confiscated by their villages. As we discussed earlier, under the current Chinese law, villages have the right to reclaim farmland if it has been vacant for more than two years. Based on these two reasons, farmers who engaged in non-farm employment would be more likely to rent their farmland out. For the same reasons, the lack of farming labor caused by non-farm employment limits the room on renting in farmland. Accordingly, this study proposes the following hypothesis: Hypothesis 2 (H2). The non-farm employment of farmers promotes the rent-out of farmland but reduces the rent-in of farmland. The existing literature shows that rural households with greater farming ability are more likely to rent in farmland [11,12,14,15,41]. A greater farming ability suggests a higher marginal product of labor. It has two implications for farmer’s decision-making. One is a direct impact. A higher marginal product suggests a bigger return from farming on land, causing a stronger demand for farmland. This, in turns, decreases farmland rent-out but increases farmland rent-in. The other is an indirect impact. A higher marginal agricultural product suggests a higher opportunity cost of taking non-farm jobs, giving farmers less incentive to engage in non-farm employment. Therefore, this study proposes the following two hypotheses: Hypothesis 3 (H3). A household’s farming ability decreases farmland rent-out but increases farmland rent-in. Hypothesis 4 (H4). A household’s farming ability decreases non-farm employment. Off-farm employment may accelerate exits from farming [42]. Thus, farmers’ farming ability declines with the length of their non-farm employment. Physically, farm and non-farm jobs could demand quite differently. After a long time of non-farm employment, a farmer could become incompetent doing farming work any longer. Technologically, farming changes over time, including what machines to run, what fertilizers and pesticides to use, and even what and when to produce. If farmers have worked in urban sectors for a long time, they could become unfamiliar with the current cultivating skills and their farming ability declines. Based on the above analysis, the following hypothesis is proposed: Hypothesis 5 (H5). Non- farm employment decreases a household’s farming ability. Figure 1 depicts the possible causal relationships between non-farm employment, farmland renting, and farming ability. We aim to provide statistical evidence on them in our analysis below. 3. Simultaneous Equation Model, Data, and Variables 3.1. Simultaneous Equation Model Given the above discussions on non-farm employment, farmland renting, and farming ability, simultaneous equation models are suitable for our empirical analysis. We specify the following simultaneous equations. NONFARM = a10 + a11RENT + a12ABILITY + a13X1 + ε1(1) RENT = a20 + a21NONFARM + a22ABILITY + a23X2 + ε2(2) ABILITY = a30 + a31NONFARM+a32X3 + ε3(3) where NONFARM denotes non-farm employment; RENT denotes farmland renting activities, that refers to either rent-out or rent-in depending on sample groups. Farmland renting activities is measured by the ratio of rent-out or rent-in land amount to the household’s initial farmland; ABILITY represents farming ability; vectors Xj (j = 1, 2, 3) include all household-level characteristics, village-level characteristics, and policy-level characteristics; aij (i = 1, 2, 3 and j = 0, 1, 2, 3) are parameters to be estimated; and εj (j = 1, 2, 3) are error terms. 3.2. Data Data used in this paper come from the survey on rural households in Zhejiang, Henan, and Shaanxi provinces. The survey was conducted and administrated by the Rural Research Center at Zhejiang Normal University during the 2017 Spring Festival holiday. Since migrant farmers largely return to their hometowns and stay at home during this major Chinese holiday, the survey was able to obtain comprehensive and accurate information on rural households, with or without migrant farmers. As the largest developing country in the world, the levels of economic development and urbanization are quite uneven among different regions in China. We expect that non-farm employment, farmland rental activities, and farmland resources per capita also vary across regions. The survey selected Zhejiang Province in the eastern region, Henan Province in the central region, and Shaanxi Province in the western region. We understand that they may not be able to fully represent China’s three main regions. But given the resource constraint, the survey was unable to include more provinces. To better ensure the representativeness of the sample, in each sampled province, the survey used a stratified random sampling procedure. Specifically, the survey randomly selected six counties from each province, five townships from each county, two villages from each township, and 4–7 households from each village depending on the village size. When sampling, the survey tried to keep a balance between villages that are suburbs of cities and villages that are far from cities. The respondents of the survey include three groups: households who were renting in farmland from other rural families, households who were renting out their farmland, and households who did neither. Questions asked in the survey were about non-farm employment, farmland rental activities, farming ability, household situation, household head situation, village situation, and policy on farmland. Most information was collected at the household level; some was collected at the village level, such as the village location and the average income of households in the village. After dropping the observations with missing values on the key variables or obvious errors, we have 1001 observations from 180 different villages. The data on per capita GDP at prefecture-level city are derived from the China Statistical Yearbook. Figure 2 shows that the sample is quite evenly distributed across the three provinces, with 36.06%, 29.67%, and 34.27%, respectively, for Zhejiang, Shaanxi, and Henan. Figure 3 shows that about half of the households (47.85%) in the survey did not engage in farmland renting activities, one third of rural households (33.57%) rented out their farmland, and less than 20% of rural households (18.58%) rented in farmland from other farmers. Our data suggest that farmland renting activities are not uncommon in rural areas, although China’s farmland rental market has a big room for future development [43]. Figure 4 shows the sample shares of households based on the number of migrant workers for non-farm jobs. We observed that about two third of households have migrant workers who engage in non-farm employment, indicating that non-farm jobs are common for rural households. We also observed that more households have 2 or more migrant workers than those with single migrant worker, suggesting that rural households often migrant jointly. 3.3. Variables In this article, the core variables include non-farm employment, farmland renting, and farming ability, which are all endogenous. Non-farm employment is measured by the share of non-farm income as used in the literature [20,44]. For robustness tests, the study also uses the number of non-farm workers in a household [27,45] to replace the share of non-farm income. Farmland renting is measured by the farmland renting ratio, which is the proportion of farmland that is rented relative to the total amount of farmland initially endowed to the household by the village [20]. Also, we replace it with farmland renting incidence late for robustness tests. The farming ability of the household is measured with the Likert scale by farmer’s self-evaluation according to their farming experiences. Exogenous variables are those about household-level characteristics, village-level characteristics, and policy-level characteristics. Household-level characteristics include household head’s situation (physical condition, education level, cognitive level of farmland tenure, age, spouse, the training of farming skills), per capita farmland area, agricultural fixed assets, average farming years of household members, and dependency ratio. We consider agricultural fixed assets since the residual value of the used specialized agricultural fixed assets is low [46], which will increase the opportunity cost of giving up farming and engaging in non-farm employment. To measure the dependency ratio, we divided the number of non-working age members by the number of working age members in a household. Following the international practice and considering that most farmers still work on farms until 64, we define working age as the ages of 15-64. To measure dependency ratio, we divided the number of non-working age members by the number of working age members in a household. Village-level characteristics include the average income of households in the village, the location of the village, and the stability of farmland tenure, which is measured by the frequency of farmland adjustment by its village. Policy-level characteristics include farmland subsidies, new farmer insurance, and farmland title. It is worth mentioning that some variables are ordinal variables such as head physical condition, education level, cognitive level of farmland tenure, and the stability of farmland tenure; some are nominal (dummy) variables, such as spouse, farmland title, farmland subsidy, and new farmer insurance. In the article, the rent of farmland is not included as a variable since rent is not a critical factor in China’s farmland rental market. According to the Land Management Law of China, village committees could retrieve farmland from rural households if the land stays vacant over two consecutive years [17]. Therefore, rural households often rent out their farmland to their neighbors, relatives, or friends at a very low price or even for free to avoid from confiscating their farmland by their villages or disqualifying them for a farmland subsidy from the government. Table 1 presents descriptive statistics about the variables used in the study for both rent-out and rent-in groups. RENT denotes farmland renting ratio, NONFARM represents non-farm income share, and ABILITY means farming ability. We observe some significant differences in NONFARM, RENT, and ABILITY between the both groups. First, the mean non-farm income share is much higher for the renting-out group (81.98%) than that for the renting-in group (55.55%). This correlation between non-farm income shares and farmland rental activities is not a surprise. The mean of farmland renting ratio is also higher for the rent-in group (2.43) than that for the rent-out group (0.76), suggesting that farmers want to enjoy scale effect in their production when they rent in farmland. As for farming ability (ABILITY), the mean value is greater for the rent-in group than that for the rent-out group. It shows that farmers in the rent-in group are of better farming ability than those in the rent-out group. The mean value of SKILL is higher for the rent-in group (3.67) than that for the rent-out group (3.01) since the farmers with farming training have greater farming ability, thus are more likely to rent in farmland. Among the explanatory and control variables, most are similar between the rent-out and rent-in groups. As expected, households in the rent-in group have more agricultural fixed assets and per capita farmland area. 4. Empirical Analyses In this section, we perform empirical analyses on factors that affect non-farm income share, farmland renting ratio, and farming ability. In our analyses, we separate the rent-out group from the rent-in group. We estimated Equations (1)–(3) individually and simultaneously. The estimated coefficients from OLS are generally smaller than those from multi-equation system since OLS ignores the relationships among the equations. For the multi-equation system, if any endogenous independent variables are included in the equation system, the two-stage least squares’ (2SLS) estimation results of each equation are consistent. However, they are not the most effective since 2SLS ignores the possible correlation among the disturbances of equations. Both OLS and 2SLS belong to the single equation estimation method, which separately estimates each equation in the simultaneous equations. Our main estimation method is three-stage least squares (3SLS), which treats all of the equations as a whole. In the first stage, the reduced form of simultaneous equation system is estimated. Then, the 2SLS estimates of all equations in the simultaneous equation system are obtained by the fitting of all endogenous variables. Once the 2SLS parameters are calculated, the residuals of each equation can be used to estimate the variances and covariances between equations. In the third stage, the parameter estimators of the generalized least square method are obtained, so the most effective estimation results could be obtained. Therefore, the 3SLS can better address the problem of endogeneity and correlations among error terms in estimating. In the following discussions, this paper focuses on the estimation results from 3SLS, although it reports the estimation results from OLS and 2SLS for references. To save space, we only present and discuss the results on NONFARM, RENT, and ABILITY. The results on other explanatory and control variables are available upon request. 4.1. Regression Results and Discussion 4.1.1. Determinants of Non-Farm Income Share Panel A in Table 2 presents the regression results on the share of non-farm income (NONFARM) based on Equation (1), for the rent-out and rent-in groups, respectively. For the rent-out group, the 3SLS coefficient of RENT is statistically positive at the 1% significance level. This indicates that farmland rent-out ratio increases the non-farm income share. Since renting out farmland frees up time for non-farm employment and thus promotes non-farm employment. In some circumstances, to some extent farmland is compulsively rented out in China, especially in the government-oriented transfer model, when farmers have to seek non-farm jobs to increase their family income. Therefore, renting farmland out is regarded as an important aspect affecting non-farm employment. For the rent-in group, the 3SLS coefficient of RENT is statistically negative at the 5% significance level. This result shows farmland rent-in ratio lifts the share of non-farm employment. We interpret it as evidence that renting in farmland indeed increases the scale of agricultural production, which, in turn, increases the demand for farming labor. It further leads to the lack of migrant workers available for non-farm employment in households and decreases the share of non-farm income. In addition, farmers renting in farmland may obtain scale returns, which lifts the opportunity costs of non-farm employment. Therefore, they may give up some chances of non-farm employment, then both the probability and share of non-farm employment may fall. These results support our hypothesis H1: Renting out farmland promotes the non-farm employment of farmers while renting in farmland decreases the non-farm employment of farmers. Panel A in Table 2 shows that the coefficients of ABILITY are significantly negative for both rent-out and rent-in groups, suggesting that farming ability decreases the share of non-farm income. In detail, farmers with great ability can obtain a more marginal return from farming than others, thus compared to others, those farmers are more likely to engage in farming jobs and less likely to involve in non-farm jobs. This finding verifies our hypothesis H4: A household’s farming ability decreases non-farm employment. 4.1.2. Determinants of Farmland Renting Ratio Panel B in Table 2 presents the regression results on farmland renting ratio based on Equation (2). For the rent-out group, we observed a positive and significant 3SLS coefficient of NONFARM, showing that the share of non-farm income promotes the rent-out farmland activities. A higher share of non-farm income often suggests a higher non-farm employment share for a rural family. In turn, a higher non-farm employment share could suggest a lack of farming labor, causing the household to rent out more farmland. Moreover, vacant farmland for over two consecutive years will be retrieved by the village in China, so farmers without enough farming labor have to rent out all or part of their farmland. So non-farm employment promotes renting out farmland, which is consistent with previous literature [47]. For the rent-in group, in contrast, we observed a statistically negative 3SLS coefficient of NONFARM, showing that non-farm employment (represented here by the share of non-farm income) decreases the rent-in ratio of farmland, probably since non-farm employment reduces farmers’ capacity to rent in farmland for a larger scale of agricultural production. The higher the share of non-farm income, the more time or labor the households spend on non-farm jobs. Correspondingly, less time will be spent on farming activities, so they are more incentive to rent out their farmland and less incentive to rent in farmland from others. These results support our hypothesis H2: The non-farm employment of farmers promotes the rent-out of farmland and reduces the rent-in of farmland. Panel B in Table 2 shows a negative but insignificant result of ABILITY on RENT for the rent-out group, suggesting that farming ability has little impact on farmers’ rent-out activities. This weak relationship between farming ability and rent-out farmland activity could suggest that rural households rent out farmland, not due to poor farming ability but due to other reasons, e.g., high risks and low income from farming, especially lacking the access to non-farm employment (hypothesis 2). The conclusion is supported by much existing literature [22,47]. In short, some farmers are reluctant to cultivate their farmland for various reasons but are under pressure of being retrieved farmland by villages, so they select to rent out their farmland. if the land stays vacant over two consecutive years. For the rent-in group, however, the coefficient of ABILITY is significantly positive, which shows that farming ability promotes the rent-in ratio of farmland. This finding is not surprising, since higher farming ability suggests better labor productivity, which increases the marginal product of farmland, calling for more land input in production. These results partly support our hypothesis H3: A household’s farming ability decreases the rent-out of farmland but promotes the rent-in of farmland. 4.1.3. Determinants of Farming Ability Panel C in Table 2 shows the regression results on farming ability based on Equation (3). For both rent-out and rent-in groups, we found that the coefficients of NONFARM are statistically negative at the 1% significance level, indicating that the share of non-farm income decreases farming ability. In reality, farmers with non-farm experience tend to select renting out farmland and non-farm employment in future [32,48], which pushes them further away from farming activities. This makes them worse at farming. This empirical finding confirms our hypothesis H5: Non-farm employment decreases a household’s farming ability. Panel A, B, and C in Table 2 have the same sample size. 4.2. Robustness Checks The statistical results for the relationships among non-farm employment, farmland renting, and farming ability are intuitively plausible in light of the literature on agricultural economics. But how robust are they across alternative specifications of the variables? To the best of our knowledge, non-farm employment and farmland renting could be measured in different ways as mentioned earlier, and we choose the following two measurements for our robustness tests. 4.2.1. The Number of Migrant Workers In addition to the share of non-farm income, non-farm employment could be measured by the number of migrant workers in a household [27,45]. We regard the number of migrant workers as the measurement of non-farm employment in the robustness test, and the symbol is MIRATE, keeping other variables unchanged. The robustness check results are shown in Table 3. Panel A in Table 3 presents the results on the number of migrant workers in a rural household for both rent-out and rent-in groups. The results confirm the findings we observed in Table 2 Panel A. For the rent-out group, the 3SLS estimated coefficient of RENT is significantly positive, indicating that the rent-out ratio of farmland frees up rural labors and thus increases the number of migrant workers. For the rent-in group, the 3SLS estimated coefficient of RENT is significantly negative, suggesting that renting in farmland demands more rural labor and thus decreases the number of migrant workers. As for ABILITY, the estimated coefficients are significantly negative for both groups, confirming that farming ability decreases the number of migrant workers. These results strongly support the finding that the renting out of farmland promotes non-farm employment, renting-in of farmland decreases non-farm employment, and farming ability decreases non-farm employment. Panel B in Table 3 presents the results on the farmland renting ratio. For the rent-out group, the results show significant and positive coefficients of MIRATE, indicating that the number of migrant workers increases the rent-out ratio of farmland. For the rent-in group, the results show that significant but negative coefficients of MIRATE, suggesting that the number of migrant workers reduces the rent-in ratio of farmland. The results on ABILITY show that farming ability has little impact on renting-out activities but enables farmers to rent-in more farmland in their agricultural production. These results are consistent with our earlier findings that non-farm employment increases farmland rent-out of but decreases farmland rent-in, and farming ability increases farmland rent-in. Panel C in Table 3 shows negative but significant results for all models and for both rent-out and rent-in groups. These results further provide evidence that non-farm employment decreases farming ability. Panel A, B, and C in Table 3 have the same sample size. 4.2.2. Farmland Renting Incidence Following Deaton et al. [49], we measure farmland renting activities by farmland renting incidence (INCIRENT), with INCIRENT taking a value of 1 if a rural household rents out or rents in farmland, or it takes a value of 0 otherwise. Therefore, in our regression analysis below, sample points which show neither rent-out nor rent-in are double used as the reference group, respectively for the rent-out and rent-in groups. The results of our robustness checks are presented in Table 4. Panel A shows that INCIRENT has a positive estimated coefficient for the rent-out group, while it has a negative estimated coefficient for the rent-in group, both are statistically significant. Therefore, we conclude that farmland renting incidence helps to raise the share of non-farm income, but it makes a rural household become less dependent on non-farm income if it rents in farmland from other villagers. We won’t interpret these results undesirable, since renting out farmland could free up rural labor for more non-farm employment and earn a higher total income, while renting in farmland could expand the scale of agricultural production and raise rural household’s overall income. The coefficients of ABILITY are significantly negative for both groups, which shows that farming ability always decreases the share of non-farm income. Panel A confirms that our previous results on the determinants of non-farm income share are robust. The 3SLS results in Panel B of Table 4 show that a rural household with a higher share of non-farm income becomes more (less) likely to rent out (in) farmland. We consider this finding expected, since a higher share of non-farm income indicates a larger share of non-farmland employment among a family’s labor. In turns, when a rural household lacks of farming labor, due to non-farm employment, a natural choice is to rent out its farmland. Renting in farmland is the other side of the same story. The estimated coefficients of ABILITY are significant and positive for the rent-in group but not significant for the rent-out group. Therefore, farming ability only helps farmers when they decide to rent in farmland; it does not matter for farmers who choose to rent out their farmland. Panel B confirms that our previous results on the determinants of farmland renting decision are robust. Panel C in Table 4 shows the estimation results of farming ability. For both rent-out and rent-in groups, based on the significantly negative estimated coefficients of NONFARM, the results indicate that non-farm employment decreases farming ability, confirming our earlier finding. Panel A, B, and C in Table 4 have the same sample size. Panel C in Table 4 shows the estimation results of farming ability. For both rent-out and rent-in groups, based on the significantly negative estimated coefficients of NONFARM, the results indicate that non-farm employment decreases farming ability, confirming our earlier finding. Panel A, B, and C in Table 4 have the same sample size. 5. Conclusions Over the past decades, a mass of farmers has moved to urban areas and become migrant workers in China’s rapid process of urbanization and industrialization. However, to some extent, farmers do not want to part with their farmland allocated by their villages, which also happened in other countries [47]. Otherwise, according to the Land Management Law of China, village committees could retrieve farmland from rural households if the land stays vacant over two consecutive years. Therefore, farmers face choices in their decision-making. They could seek non-farm jobs in urban areas, continue to work on their farmland, rent out or rent in farmland, or let their farmland vacant and face the risk of being retrieved. Decisions on these choices are not independent, since seeking jobs in urban areas often implies that they may not have sufficient labor to work on farmland and thus have to either rent out farmland or let the farmland vacant. Likewise, staying in rural areas may suggest that farmers have surplus labor; it would be more productive if they could rent in some farmland to expand the production scale. Also, in farmers’ decision-making, farming ability could be a factor, which itself could diminish with time that farmers have moved to urban areas and over generations. Using survey data from Zhejiang, Henan, and Shaanxi provinces in 2017, this paper has investigated the simultaneous relationships among non-farm employment, farmland renting, and farming ability. We obtained three main conclusions. First, farmland rent-out promotes non-farm employment while farmland rent-in decreases non-farm employment. Farming ability decreases non-farm employment. Second, non-farm employment encourages farmland rent-out whereas it decreases farmland rent-in. Farming ability promotes farmland rent-in but it has no effect on farmland rent-out. Third, non-farm employment decreases farming ability. Policy Implications Based on our empirical results, we would propose the following two policy implications. First, the current fragmented land allocation system in rural China results in a small per capita farmland and causes low agricultural production efficiency. An effective farmland rental market could provide a win-win solution for all farmers. On the one hand, renting out farmland would free up more rural laborers so that they can work in the urban sector. On the other hand, renting in farmland would enable farmers to expand their production scale, apply new technology, and improve production efficiency. Therefore, it is important for the Chinese government to promote farmland rental markets in rural areas. Second, under the current system, agricultural land is owned collectively by villages. Farmers have the right to use, but they are not allowed to sell any farmland individually. We would propose that China could treat rural households as “shareholders” of their collectively-owned land and allow them to sell part of their shares at least back to their villages. Such a reform could help migrant workers better settle down in cities, as rural migrants could use the money to buy homes or cover the down payment when they buy houses in urban areas. In reality, many migrant workers have worked in cities for many years, and they will continue to live and work in cities. Deregulating the rural land system and improving the farmland market could help China narrow the rural-urban divide and promote a common prosperity development. Author Contributions Conceptualization, J.L.; methodology, J.L. and G.S.; software, J.L. and G.S.; validation, J.L. and S.S.; formal analysis, J.L.; investigation, J.L.; resources, J.L.; data curation, J.L.; writing—original draft preparation, J.L.; writing—review and editing, J.L. and S.S.; visualization, J.L. and S.S.; supervision, J.L. and S.S.; project administration, J.L. and S.S.; funding acquisition, J.L. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement The study did not report any data. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Possible causal relationships among non-farm employment, farmland renting, and farming ability. Figure 2 Share of samples, by region. Figure 3 Share of samples, by renting activity. Figure 4 Share of samples, by the number of migrant workers in a household. ijerph-19-05476-t001_Table 1 Table 1 Descriptive statistics on the variables. Symbols Variables Measurements(Unit) Rent-Out Group Rent-In Group Mean SD Mean SD NONFARM Non-farm employment Non-farm income share(non-agricultural income/total household’s income) (%) 81.98 28.51 55.55 33.4 RENT Farmland renting Farmland renting ratio(rental farmland /total household’s initial farmland) (%) 0.76 0.28 2.43 5.35 ABILITY Farming ability Likert five scale 3.01 1 3.67 0.86 PHY Head physical condition Likert five scale 3.74 0.64 3.83 0.96 EDU Head educational level Illiterate = 1, elementary = 2 school, middle school = 3, high school = 4, college or above = 5 2.69 0.83 2.75 0.95 FYEAR Years of farming (year) Average farming years of household members 15.20 14.40 14.50 11.68 SPOUSE Spouse yes = 1, no = 0 0.96 0.21 0.96 0.19 DASSET Agricultural fixed asset Agricultural fixed asset value per household (10,000 yuan) 0.92 2.66 4.08 9.72 AGE Head age 54.47 10.77 52.42 9.98 SKILL Farm skills training Length of farming skills training of household head (months) 0.28 2.11 0.70 1.72 DERATE Dependency ratio The number of non-working age members/the number of working age members in a household (%) 0.56 0.65 0.51 0.49 ACRPER Per capital farmland acres Total farmland acres/the number of family members (mu) 0.86 0.73 2.65 3.93 COGNI Cognitive level of farmers about farmland tenure How many rights a farmer think he has to his(her) farmland 1.95 0.95 1.87 0.86 STAB Stability of farmland tenure Adjusting frequency of farmland by its village 2.22 0.82 2.13 0.91 VPOSI Location of the village The distance of the village from the nearest town (km) 1.57 5.44 1.93 7.54 VINCO Village income level Per household income in village (10,000 yuan) 2.74 1.85 3.21 3.17 SUBSIDY Farmland subsidy policy yes = 1, no = 0 0.650 0.480 0.750 0.43 INSURE New farmer insurance policy yes = 1, no = 0 0.880 0.330 0.700 0.46 TITLE Farmland title yes = 1, no = 0 0.710 0.460 0.660 0.48 GDP Per capita GDP (10,000 yuan) 4.51 0.57 4.97 0.66 ijerph-19-05476-t002_Table 2 Table 2 Determinants of non-farm income share, farmland renting ratio, and farming ability. Variables Rent-Out Group Rent-In Group (1) OLS (2) 2SLS (3) 3SLS (4) OLS (5) 2SLS (6) 3SLS Panel A: Determinants of non-farm income share RENT 0.010 *** 0.054 *** 0.404 *** −0.439 ** −1.718 * −1.846 ** (0.002) (0.017) (0.014) (0.177) (0.954) (0.782) ABILITY −0.036 *** −0.052 −0.464 *** −0.539 *** −0.412 * −1.114 *** (0.064) (0.128) (0.111) (0.192) (0.238) (0.192) control yes yes yes yes yes yes Panel B: Determinants of farmland renting ratio NONFARM 4.224 *** 7.150 9.950 *** −0.089 *** −0.142 −0.337 *** (1.288) (7.153) (3.159) (0.034) (0.112) (0.078) ABILITY −4.316 ** −3.280 −0.614 0.012 0.022 * 0.093 ** (2.068) (3.249) (2.229) (0.060) (0.075) (0.065) control yes yes yes yes yes yes Panel C: Determinants of farming ability NONFARM −0.154 *** −0.789 *** −0.907 *** −0.097 ** −0.818 *** −0.816 *** (0.042) (0.171) (0.148) (0.043) (0.164) (0.156) control yes yes yes yes yes yes N 336 336 336 186 186 186 *, **, *** denote significance at the 10%, 5% and 1% level respectively. ijerph-19-05476-t003_Table 3 Table 3 Robustness check results of simultaneous equations: with the number of non-farm workers. Variables Rent-Out Group Rent-In Group (1) OLS (2) 2SLS (3) 3SLS (4) OLS (5) 2SLS (6) 3SLS Panel A: Determinants of the number of non-farm workers RENT 0.270 * 0.274 * 0.542 *** 0.012 −0.479 * −0.729 *** (0.139) (0.143) (0.069) (0.069) (0.287) (0.143) ABILITY −0.003 * −0.075 −0.104 ** −0.167 * −0.532 * −1.315 *** (0.067) (0.169) (0.157) (0.102) (0.427) (0.279) control yes yes yes yes yes yes Panel B: Determinants of farmland renting ratio (1) OLS (2) 2SLS (3) 3SLS (4) OLS (5) 2SLS (6) 3SLS MIRATE 0.031 1.584 ** 1.690 *** 0.031 −1.018 −1.309 *** (0.022) (0.623) (0.469) (0.085) (0.787) (0.329) ABILITY −0.042 −0.226 0.111 0.266 ** 0.783 1.502 *** (0.027) (0.283) (0.257) (0.113) (0.658) (0.438) control yes yes yes yes yes yes Panel C: Determinants of farming ability MIRATE −0.146 *** −0.479 *** −0.683 *** −0.088 −0.390 −0.614 ** (0.042) (0.163) (0.143) (0.057) (0.323) (0.261) control yes yes yes yes yes yes N 336 336 336 186 186 186 *, **, *** denote significance at the 10%, 5% and 1% level respectively. ijerph-19-05476-t004_Table 4 Table 4 Robustness check results of simultaneous equations (with farmland renting incidence). Rent-Out Group Rent-In Group (1) OLS (2) 2SLS (3) 3SLS (4) OLS (5) 2SLS (6) 3SLS Panel A: Determinants of non-farm income share RENTINCI 0.102 ** 0.647 ** 0.898 *** −0.386 *** −2.590 *** −3.011 *** (0.046) (0.262) (0.198) (0.128) (0.670) (0.566) ABILITY −0.234 *** −0.806 *** −0.884 *** −0.306 *** −0.976 *** −1.206 *** (0.049) (0.170) (0.162) (0.063) (0.294) (0.242) control yes yes yes yes yes yes Panel B: Determinants of farmland renting incidence NONFARM 0.046 * 0.689 1.002 *** −0.020 * −0.121 * −0.257 *** (0.027) (0.505) (0.241) (0.012) (0.068) (0.041) ABILITY −0.120 *** 0.634 0.890 0.080 *** 0.154 0.309 *** (0.037) (0.443) (0.252) (0.019) (0.112) (0.091) control yes yes yes yes yes yes Panel C: Determinants of farming ability NONFARM −0.080 *** −0.050 −0.090 ** −0.099 *** −0.174 ** −0.187 ** (0.024) (0.079) (0.075) (0.024) (0.080) (0.073) control yes yes yes yes yes yes N 815 815 815 665 665 665 *, **, *** denote significance at the 10%, 5% and 1% level respectively. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Williamson J.G. Migration and urbanization Handb. Dev. Econ. 1988 1 425 465 2. Zhang K.H. Song S. Rural–Urban Migration and Urbanization in China: Evidence from Time-Series and Cross-Section Analyses China Econ. Rev. 2003 14 386 400 10.1016/j.chieco.2003.09.018 3. Lu Z. Song S. Rural–Urban migration and wage determination: The case of Tianjin, China China Econ. Rev. 2006 17 337 345 10.1016/j.chieco.2006.04.007 4. Zhou B. Analysis on the rights and interests of land-lost farmers from the perspective of social exclusion Issues Agri-Cult. Econ. 2015 36 59 111 59–65, 111 (In Chinese) 5. Seto K.C. Ramankutty N. Hidden linkages between urbanization and food systems Science 2016 352 943 945 10.1126/science.aaf7439 27199419 6. Heimlich R.E. Vesterby M. Krupa K.S. Urbanizing Farmland: Dynamics of Land Use Change in Fast-Growth Counties: Concern Over Farmland Loss United States Department of Agriculture Washington, DC, USA 1991 No. 1474-2021-728 7. Agergaard J. Fole N. Gough K. Rural Urban Dynamics Routledge New York, NY, USA 2010 8. García-Sempere A. Hidalgo M. Morales H. Ferguson B.G. Nazar-Beutelspacher A. Rosset P. Urban transition toward food sovereignty Globalizations 2018 15 390 406 10.1080/14747731.2018.1424285 9. Kuminoff N.V. Sokolow A.D. Sumner D.A. Farmland Conversion: Perceptions and Realities (No. 16) University of California, Agricultural Issues Center Los Angeles, CA, USA 2001 10. Ruel M.T. Garrett J. Yosef S. Olivier M. Urbanization, food security and nutrition Nutrition and Health in a Developing World Springer Berlin/Heidelberg, Germany 2017 705 735 11. Chamberlin J. Ricker-Gilbert J. Participation in Rural Land Rental Markets in Sub-Saharan Africa: Who Benefits and by How Much? Evidence from Malawi and Zambia Am. J. Agric. Econ. 2016 98 1507 1528 10.1093/ajae/aaw021 12. Otsuka K. Place F. Holden S. Land Markets in Uganda: What Is Their Impact and Who Benefits? The Emergence of Land Markets in Africa Routledge London, UK 2009 144 169 13. Holden S. Keijiro O. Frank P. The Emergence of Land Markets in Africa Routledge London, UK 2010 14. Jin S. Deininger K. Land rental markets in the process of rural structural transformation: Productivity and equity impacts from China J. Comp. Econ. 2009 37 629 646 10.1016/j.jce.2009.04.005 15. Jin S. Jayne T.S. Land Rental Markets in Kenya: Implications for Efficiency, Equity, Household Income, and Poverty Land Econ. 2013 89 246 271 10.3368/le.89.2.246 16. Phimister E. Roberts D. The Effect of Off-farm Work on the Intensity of Agricultural Production Environ. Resour. Econ. 2006 34 493 515 10.1007/s10640-006-0012-1 17. Hong M. Zheng L. The Damage of Farmers’ Rights and Interests in theAgricultural Land Transfer J. Hebei Univ. Econ. Bus. 2019 40 14 21 (In Chinese) 18. Rozelle S. Guo L. Shen M. Hughart A. Giles J. Leaving China’s Farms: Survey Results of New Paths and Remaining Hurdles to Rural Migration China Q. 1999 158 367 393 10.1017/S0305741000005816 19. Kimura S. Otsuka K. Sonobe T. Rozelle S. Efficiency of Land Allocation through Tenancy Markets: Evidence from China Econ. Dev. Cult. Chang. 2011 59 485 510 10.1086/649639 20. Kung J.K. Off-Farm Labor Markets and the Emergence of Land Rental Markets in Rural China J. Comp. Econ. 2002 30 395 414 10.1006/jcec.2002.1780 21. Muhammad T. Khan H. Qahar A. Factors determining off-farm employment on small farms in district Nowshehra of North West Pakistan Sarhad J. Agric. 2012 28 333 343 22. Du G. Li R. Form of rural population non-agriculturalization and its impact on agricultural land transfer Agric. Econ. Manag. 2014 28 48 54 (In Chinese) 23. Salvioni C. Papadopoulou E. Santos M.D. Small farm survival in Greece, Italy and Portugal EuroChoices 2014 13 52 57 10.1111/1746-692X.12053 24. Vranken L. Swinnen J. Land rental markets in transition: Theory and evidence from Hungary World Dev. 2006 34 481 500 10.1016/j.worlddev.2005.07.017 25. Akter S. Farrington J. Deshingkar P. Sharma P. Rao L. Land Rental Markets in India: Efficiency and Equity Considerations International Association of Agricultural Economists Queensland, Australia 2006 No. 1004-2016-78613 26. De Janvry A. Sadoulet E. Income strategies among rural households in Mexico: The role of off-farm activities World Dev. 2001 29 467 480 10.1016/S0305-750X(00)00113-3 27. Che Y. Off-Farm employments and land rental behavior: Evidence from rural China China Agric. Econ. Rev. 2016 8 37 54 10.1108/CAER-09-2014-0086 28. Willmore L. Cao G. Xin L. Determinants of Off-Farm Work and Temporary Migration in China Popul. Environ. 2012 33 161 185 10.1007/s11111-011-0135-3 29. Tao R. Xu Z. Urbanization, Rural Land System and Social Security for Migrants in China J. Dev. Stud. 2007 43 1301 1320 10.1080/00220380701526659 30. Wu A.D. The Impacts of Land Transfer Controlled by Government on Household Off-Farm Employment Northwest A&F University Yangling, China 2018 (In Chinese) 31. Kinsella J. Wilson S. De Jong F. Renting H. Pluriactivity as a livelihood strategy in Irish farm households and its role in rural development Sociol. Rural. 2000 40 481 496 10.1111/1467-9523.00162 32. Mishra A.K. Goodwin B.K. Farm income variability and the supply of off-farm labor Am. J. Agric. Econ. 1997 79 880 887 10.2307/1244429 33. Hu X. Hong W. Which is the cause, labor migration and land rental? J. Huazhong Agric. Univ. 2019 139 137 169 137–145, 169 (In Chinese) 34. Forbord M. Bjørkhaug H. Burton R.J. Drivers of change in Norwegian agricultural land control and the emergence of rental farming J. Rural Stud. 2014 33 9 19 10.1016/j.jrurstud.2013.10.009 35. Tan S.H. Qu F.T. Heerink N. Causes of land fragmentation and driving forces China Rural Surv. 2003 6 24 30 (In Chinese) 36. Todorova S.A. Lulcheva D. Economic and social effects of land gragmentation on Bulgarian agriculture J. Cent. Eur. Agric. 2006 6 555 562 37. Kawasaki K. The impact of land fragmentation on rice production cost and input use Jpn. J. Rural Econ. 2011 13 1 14 10.18480/jjre.13.1 38. Manjunatha A.V. Anik A.R. Speelman S. Nuppenau E. Impact of Land Fragmentation, Farm Size, Land Ownership and Crop Diversity on Profit and Efficiency of Irrigated Farms in India Land Use Policy 2013 31 397 405 10.1016/j.landusepol.2012.08.005 39. Deininger K. Jin S. Zude X. Rozelle S. Implementing China’s New Land Law: Evidence and Policy Lessons. World Bank Policy Research Working Paper 2004 Available online: http://citeseerx.ist.psu.edu/viewdoc/download;jsessionid=A4FC5BFAB6DEEB5419CBAF356E9D4F25?doi=10.1.1.522.7556&rep=rep1&type=pdf (accessed on 5 June 2020) 40. Wang X. Huang J. Rozelle S. Off-Farm employment and agricultural specialization in China China Econ. Rev. 2017 42 155 165 10.1016/j.chieco.2016.09.004 41. Deininger K. Jin S. The Potential of Land Rental Markets in the Process of Economic Development: Evidence from China J. Dev. Econ. 2005 78 241 270 10.1016/j.jdeveco.2004.08.002 42. Mishra A.K. Fannin J.M. Joo H. Off-Farm work, intensity of government payments, and farm exits: Evidence from a ational survey in the United States Can. J. Agric. Econ./Rev. Can. D’agroeconomie 2014 62 283 306 10.1111/cjag.12027 43. Liu S. Deepening land reform faces five difficulties China Bus. News 2014 3 48 44. Qian L. Hong M. Non-Agricultural employment, farmland transferring and changes of farming production efficiency: Based on the empirical study of CFPS China Rural Econ. 2016 12 2 16 (In Chinese) 45. Su W. Eriksson T. Zhang L. Bai Y. Off-Farm employment and time allocation in on-farm work in rural China from gender perspective China Econ. Rev. 2016 41 34 45 10.1016/j.chieco.2016.08.006 46. Williamson O.E. The Economic Institutions of Capitalism Free Press New York, NY, USA 1985 47. Falkowski J. Key Determinants of Land Rentals in Poland European Association of Agricultural Economists Copenhagen, Denmark 2005 No. 724-2016-49197 48. Davis J.R. Bezemer D.J. The Development of the Rural Non-Farm Economy in Developing Countries and Transition Economies: Key Emerging and Conceptual Issues Natural Resources Institute Chatham, UK 2004 49. Deaton B.J. Lawley C. Nadella K. Renters, Landlords, and Farmland Stewardship Agric. Econ. 2018 49 521 531 10.1111/agec.12433

PMC009xxxxxx/PMC9099619.txt

==== Front Nutrients Nutrients nutrients Nutrients 2072-6643 MDPI 10.3390/nu14091760 nutrients-14-01760 Article Vitamin D and Chronic Diseases among First-Generation Immigrants: A Large-Scale Study Using Canadian Health Measures Survey (CHMS) Data Yousef Said 12* Colman Ian 1 Papadimitropoulos Manny 34 Manuel Douglas 567 https://orcid.org/0000-0002-7614-0072 Hossain Alomgir 16 https://orcid.org/0000-0002-7970-2616 Faris MoezAlIslam 8 Wells George A. 12 Barker Tyler Academic Editor 1 Faculty of Medicine, School of Epidemiology and Public Health, University of Ottawa, Ottawa, ON K1G 5Z3, Canada; icolman@uottawa.ca (I.C.); alhossain@ottawaheart.ca (A.H.); gawells@ottawaheart.ca (G.A.W.) 2 Cardiovascular Research Methods Centre, University of Ottawa Heart Institute, Ottawa, ON K1Y 4W7, Canada 3 Eli Lilly Canada Inc., Toronto, ON M5X 1B1, Canada; papadimitropoulos_manny@lilly.com 4 Faculty of Pharmacy, University of Toronto, Toronto, ON M5S 3M2, Canada 5 Ottawa Hospital Research Institute, Ottawa, ON K1Y 4E9, Canada; dmanuel@ohri.ca 6 Institute for Clinical Evaluative Sciences, Ottawa, ON K1Y 4E9, Canada 7 Department of Family Medicine, University of Ottawa, Ottawa, ON K1H 8M5, Canada 8 Department of Clinical Nutrition and Dietetics, Research Institute of Medical and Health Sciences (RIMHS), College of Health Sciences, University of Sharjah, Sharjah 27272, United Arab Emirates; mfaris@sharjah.ac.ae * Correspondence: sabde029@uottawa.ca 22 4 2022 5 2022 14 9 176016 3 2022 18 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Background: Nearly 22% of the Canadian population are first-generation immigrants. We investigated immigrants’ health status and health deterioration over time in terms of the prevalence of chronic diseases (CDs) and their relationship to vitD status. Methods: We used cycles three (2012–2013) and four (2014–2015) of the Canadian Health Measures Survey. These data contained unique health information and direct physical/blood measures, including serum 25-hydroxyvitamin D (S-25(OH)D). Indicators of health status and deterioration were the prevalence of CDs diagnosed by healthcare professionals, self-reported general and mental health, and CD-related biomarkers. Results: The data (n = 11,579) included immigrants from more than 153 countries. Immigrants were healthier than non-immigrants for most health status measures. The prevalence of CDs was higher among those who migrated to Canada aged ≥ 18 years. A longer time in Canada after immigration was associated with a higher risk for CDs. The mean S-25(OH)D was lower among immigrants, higher among patients with CDs, and inversely associated with glycated hemoglobin, total cholesterol/high-density lipoprotein ratio, immunoglobulin E, serum ferritin, and blood hemoglobin. After adjusting for covariates, no association was found between S-25(OH)D and the prevalence of CDs. Conclusions: Lower levels of accumulated S-25(OH)D among immigrants may impact their health profile in terms of CD-related biomarkers, which partially explains immigrants’ health deterioration over time. We recommend further longitudinal research to investigate immigrants’ vitD and health deterioration. immigrants’ health deterioration vitamin D serum 25-hydroxyvitamin D chronic diseases ==== Body pmc1. Introduction Immigrants are a fundamental part of the Canadian population and policy framework [1], with nearly 22% of the Canadian population being first-generation immigrants [2]. Statistics Canada estimated an increase of approximately 1 million foreign-born residents every 3 years by 2035/2036 [3]. Therefore, the health of immigrants will impact Canada’s overall future healthcare system. Accumulated evidence suggests that recent immigrants have better health than people of the host country, including long-term immigrants who had previously migrated to the host country. This phenomenon is called the “healthy immigrant effect” or “healthy immigrant advantage” [4,5,6]. However, immigrants’ physical and mental health declines after arrival in Canada, with health changes often occurring within 5–10 years [4,5]. Pre-migration factors, such as rigorous selection criteria that favor immigrants that are healthier, wealthier, and better educated, along with post-immigration factors, such as environmental and lifestyle changes, may explain this health decline after migration [4,7,8,9,10]. In addition, the healthy immigrant effect varies through the course of life [6]. Moreover, more immigrants have deficient and insufficient vitamin D (vitD) levels compared with native-born populations [2,11,12]. However, vitD status varies among immigrants because of ethnic variations, skin pigmentation, resettlement changes in diet, physical activity, and sun exposure [2,11,13,14]. The concentration of serum 25-hydroxyvitamin D (S-25(OH)D) represents the combined contributions of the cutaneous synthesis and dietary intake of vitD and is considered the most important clinical marker for the overall vitD level [12,15,16]. A lower concentration of S-25(OH)D was found to be a risk factor for CDs [17,18,19,20] and could predict all-cause mortality [17,18,21]. Moreover, Jablonski et al. expected the diseases associated with vitD deficiency to be a major source of morbidity and mortality in the 21st century [22]. Most recent guidelines have emphasized the importance of vitD for skeletal health, although evidence supporting its benefits for non-skeletal health outcomes is either weak or conflicting [23,24]. There is a range of recommended vitD thresholds that are used to study health effects. Mean S-25(OH)D (nmol/L) values or ranges at various thresholds (e.g., deficiency: 25–30 nmol/L; insufficiency: 25–49 nmol/L; and sufficiency: 50–75 nmol/L) are commonly used in studies to describe vitD status [12]. An insufficient vitD level (<50 nmol/L) is more frequently used to describe hypovitaminosis D [25]. vitD deficiency has been linked to physical and mental illnesses [26,27]. Recent research supported the hypothesis that vitD deficiency was related to the incidence, severity [28,29], and mortality attributed to COVID-19 [29,30]. These studies concluded that those at the highest risk for severe COVID-19 matched those at the highest risk for severe vitD deficiency, including older people, darker-skinned ethnic groups, and obese people [28,30,31]. Moreover, research has highlighted the importance of further studies investigating immigrants’ health deterioration in the context of vitD [2,32], lifelong adversity [33], and changes in lifestyle and dietary intake [5,32]. However, leading causes of death, such as CDs (cardiovascular disease, cancer, chronic lung disease, and diabetes), are used as reliable measures for public health [34]. The main objective of this study was to investigate immigrants’ health status and possible health deterioration in relation to serum vitD levels. Moreover, we used CDs and CD-related biomarkers as proxy indicators to investigate risk factors for health deterioration in the context of vitD status. In addition, we compared the health status (in relation to the vitD status) of immigrants from different ethnic groups and origins with non-immigrants. 2. Materials and Methods 2.1. Study Design and Participants We previously published the main characteristics, mean S-25(OH)D levels, and prevalence and leading determinants of vitD deficiency/insufficiency among immigrants and non-immigrants in Canada [2]. The present study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement in the planning, implementation, and reporting [35]. Health and vitD status among first-generation immigrants (foreign-born) were ascertained using Canadian Health Measures Survey (CHMS) data and compared with those in the non-immigrant (native-born) population. The CHMS represents the first national data on vitD, includes the Canadian immigrant population, addresses gaps in existing national information, and contains unique health information, including direct physical and blood measures [36,37]. The sample size for each CHMS cycle was carefully selected to provide a reliable and representative estimate at the national level for sex and age groups. The survey covered approximately 96% of the Canadian population. A dwelling stratification stage was applied, followed by a roster list of all people living in each household, from which individuals aged 3–79 years were randomly selected [36,37]. The CHMS sample population weight was adjusted for age and sex across Canada’s five standard geographic regions. All participants provided written informed consent, and the CHMS was approved by the Health Canada Research Ethics Board [36,37]. 2.2. Measures The S-25(OH)D is expressed in nanomoles per liter (nmol/L) and is measured using chemiluminescence immunoassay technology (DiaSorin®, Ltd., Stillwater, MN, USA). The analytical detection limit for S-25(OH)D was 10–375 nmol/L. We used S-25(OH)D cut-off points to identify vitD status as either sufficient (≥50 nmol/L) or insufficient (<50 nmol/L), as defined by the Institute of Medicine (IOM) and other experts [12,38,39]. We used the presence of CDs as the outcome-dependent factor for health deterioration. CDs included 24 chronic diseases and conditions (e.g., type 1 and 2 diabetes, heart disease (not limited to ischemic heart disease), and mood disorder). Respondents were asked “Do you have a long-term condition that was diagnosed by a healthcare professional and expected to last or has already lasted 6 months or more?” (yes/no). We used CDs (e.g., diabetes, heart disease, cancer, and asthma) for which a date of diagnosis was listed. We combined CDs that had similar inflammatory characteristics (including, but not limited to chronic obstructive pulmonary disease, bronchitis, and emphysema) as one variable called “inflammatory lung diseases.” Inflammatory lung diseases were measured for participants aged > 29 years, but no date of diagnosis was available. CDs were considered present if at least one of the five listed health conditions (diabetes, heart disease, cancer, asthma, and inflammatory lung diseases) received a “yes” response as a separate variable called “CDs”. The date of diagnosis for the CDs was not available. Our analysis examined: five health conditions (diabetes, heart disease, cancer, asthma, and inflammatory lung diseases), CDs (at least one), self-rated general health, and self-rated mental health. We also considered seven biomarkers associated with CDs: glycated hemoglobin (HbA1c), the total cholesterol/high-density lipoprotein cholesterol ratio (TC/HDL ratio), high-sensitivity C-reactive protein (hs-CRP), immunoglobulin E (IgE), ferritin, hemoglobin (Hb), and S-25(OH)D (the serum concentration of vitD). In addition, 16 sociodemographic and health behavior variables were examined as independent factors: immigration status, sex, age, income, education level, body mass index (BMI; kg/m2), smoking status, alcohol consumption, age at immigration, length of time in Canada, physical activity, region, country of birth, ethnicity, dairy intake, and vitD supplement use. The CHMS was a cross-sectional survey. Data for health measures, including several temporal attributes such as date of immigration and mobile clinic visit dates (e.g., blood sampling including S-25(OH)D), were used to create proxy health indicators for health deterioration. To examine longitudinal changes in health, we created a new variable to compare “diagnosis before immigration” versus “diagnosis after immigration”. We also created a new variable “time after diagnosis” to investigate the relationship between S-25(OH)D and the presence of CDs. We used two categories (recent vs. long-term) for diagnosis at two time-points (1 year and 5 years). We used the lowest cut-off value to represent the most recent S-25(OH)D measures at the date of diagnosis (e.g., ≤1 year and ≤5 years) and the highest to represent the long-term measures for S-25(OH)D at the date of diagnosis (e.g., >1 year and >5 years). A Likert scale was used for the two self-reported questions that reflected the quality of general and mental health. Participants aged over 12 years rated these questions, with the ratings for each question dichotomized in two categories: “fair/poor” versus “good” (excellent/very good/good). Detailed information about the CHMS data, methods, and merging of the two cycles can be found in our previous work [2] and on the Statistics Canada website (http://www.statcan.gc.ca, accessed 14 February 2021) [40,41]. The method of measurement and the analytical detection limit for each biomarker are available in Appendix A, Table A1. 2.3. Statistical Analysis Following the recommendations of Statistics Canada, all analyses used population weights that reflected the probability a respondent was included in the survey. Descriptive statistics were used to investigate immigrants’ demographic, clinical, and behavioral characteristics compared with non-immigrants. We used the mean and standard error (SE) for continuous variables. Data were stratified based on the study outcomes (CDs) and variables of interest (e.g., immigration status). To account for the unequal probability of selection and present an accurate estimate of the Canadian population, we used the survey command, recommended sample weight, and degrees of freedom in the analyses. Therefore, all results were weighted values. In addition to the two S-25(OH)D cutoff points (sufficient ≥ 50 nmol/L and insufficient < 50 nmol/L), we used the continuous values for S-25(OH)D and other biomarkers. We used univariate analyses to identify independent covariates. Multivariable logistic regression models (odds ratio (OR) and 95% confidence interval (CI)) were then used to evaluate the association between CDs (indication of health deterioration) and vitD status for immigrants compared with non-immigrants. The final model was adjusted for immigration status, age, sex, household income, education, ethnicity, BMI, physical activity, smoking behavior, alcohol consumption, vitD supplement use, HbA1c, the TC/HDL ratio, Hb, ferritin, IgE, and hs-CRP. Statistical significance was set at p ≤ 0.05. All analyses were performed with SPSS version 26.0 (IBM Corp., Armonk, NY, USA) and Stata version 16.0 (StataCorp, College Station, TX, USA). 3. Results We examined data for 11,579 CHMS participants, including immigrants from more than 153 countries. The main characteristics of the study sample have previously been published [2]. The prevalence rates of CDs based on the sociodemographic and immigration characteristics of all study participants are presented in Table 1. Immigrants had a lower prevalence of CDs than non-immigrants. Those who had migrated to Canada at age ≥ 18 years had a higher prevalence of CDs than those who migrated at under 18 years of age. At two different time points after the migration (5 and 10 years), the prevalence was higher among those who had stayed in Canada for ≥5 years and ≥10 years than among those who had stayed less than 5/10 years. The prevalence of CDs was higher among immigrants with older age, lower household income, underweight, overweight, and obesity compared with their younger, higher income, and average body weight counterparts. Participants who did not meet the Canadian physical activity recommendations had a higher prevalence of CDs. A lower prevalence of CDs was observed among Chinese, Blacks, Filipinos, Arabs, and the “all other ethnicities” group compared with the white group (Table A2). The prevalence of CDs based on region and place of birth is presented in Table A3. Those born in South/Central America, the Caribbean, Africa, and Asia had a lower prevalence of CDs than those born in Canada and North America. A detailed analysis based on the country of birth showed that, compared with native-born Canadians, more participants born in Italy had CDs and fewer of those born in the Philippines had CDs. Compared with non-immigrants, immigrants had a lower prevalence of heart disease, cancer, asthma, inflammatory lung diseases, and CDs but a higher prevalence of vitD insufficiency (52.82 vs. 31.75, p < 0.001). Immigrants also had lower mean concentration levels of S-25(OH)D, serum hs-CRP, and Hb and higher mean levels of HbA1c, the TC/HDL ratio, and ferritin than non-immigrants (Table 2). Immigrants with diabetes, heart disease, and poor/fair self-reported general and mental health had lower S-25(OH)D levels than non-immigrants. More immigrants with diabetes, heart disease, cancer, asthma, inflammatory lung diseases, CDs, and poor/fair general or mental health had insufficient vitD levels compared with non-immigrants. However, there was no difference in vitD supplement use for any health indicators between immigrants and non-immigrants (Table 3). There were no differences in S-25(OH)D levels between patients diagnosed before and after immigration for these health indicators (Table 4). There were no dates of diagnosis for inflammatory lung diseases, CDs, and self-reported general and mental health. The results shown in Appendix A and Table A4, Table A5, Table A6 and Table A7 represent all study participants and are not stratified by immigration status or CDs. No differences were observed in mean S-25(OH)D levels by recent or long-term diagnosis (Table A4). The data were not analyzed for inflammatory lung diseases, CDs, and self-reported general and mental health because the date of diagnosis was not given. The mean values for S-25(OH)D, vitD status (sufficiency/insufficiency), and the use of vitD supplements for each health indicator, including CDs, are presented in Table A5. Patients diagnosed with diabetes had insufficient vitD compared with those without diabetes. Patients with heart disease, CDs, and who self-reported poor/fair general health had higher S-25(OH)D levels than their counterparts without heart disease or CDs and with good general health. Patients with cancer tended to use more supplements than those without cancer. In addition, the mean concentration levels of S-25(OH)D, HbA1c, and IgE were higher among patients with CDs than among those without CDs (Table A6). The linear regression analysis showed that when the concentration levels of S-25(OH)D increased by 1 nmol/L, HbA1c, the TC/HDL ratio, IgE, ferritin, and Hb decreased (Table A7). In the multivariable logistic regression analysis, S-25(OH)D was associated with the prevalence of CDs (unadjusted OR: 1.006; 95%CI: 1.00, 1.01; p = 0.007) (result not shown). After adjusting for covariates (Table 5), the final model indicated there was no association between S-25(OH)D and CDs (OR: 1.007; 95%CI: 1.00, 1.01; p = 0.070). Immigrants were healthier (OR for CDs decreased by 60%) than non-immigrants. We also found that when age increased by 1 year, the OR for CDs increased by 2%. The OR increased by 71% among obese participants, by 174% when the HbA1c increased by 1%, and by 0.001% when IgE increased by 1 IU/mL. 4. Discussion This study used data for a large population of native-born Canadians and immigrants from 153 countries. Immigrants had a lower prevalence of CDs than native-born Canadians, including heart disease, cancer, asthma, and inflammatory lung diseases. Immigrants who came to Canada aged ≥ 18 years had more CDs than those younger than 18 years. There was no difference in the prevalence of CDs before immigration compared with post immigration. However, using two different time points (5 and 10 years), we found that the longer immigrants had lived in Canada, the higher the prevalence of CDs, which may be considered an early sign of health deterioration. Although the CHMS used a cross-sectional design, we used proxy indicators to verify the existence of the healthy immigrant effect (immigrants’ health advantage in Canada). Our results were consistent with those of Vang et al. (2015) who conducted a systematic review involving 77 studies of migration and health in Canada. That study found the healthy immigrant effect was more noticeable among recent immigrants and gradually disappeared among immigrants that were well-established in the host country. Moreover, the advantage of the healthy immigrant effect was also found in terms of protecting immigrants against CDs, including cancer, diabetes, heart disease, asthma, and obesity [6]. Those authors indicated there was a significant deviation for morbidity over time in the healthy immigrant effect, and there were no identifiable underlying reasons for health deterioration [6]. Moreover, Paszat et al. (2017) found that immigration was a determinant for developing colorectal cancer after the first 10 years of arrival in Canada [42]. Another study reported that Canadian immigrants demonstrated lower cancer-specific mortality, but this benefit diminished over time. After adjusting for age, each year following arrival was associated with increased mortality [43]. We hypothesized that lower S-25(OH)D levels, along with environmental and behavioral changes, may explain health deterioration. However, the relationship between vitD and CDs and the acculturation process may be affected by many factors, such as whether immigrants adopted healthy or unhealthy behaviors after immigration compared with before immigration. For example, in previous immigration, acculturation, and vitD studies, it was reported that immigrants less frequently consumed vitD-rich foods, engaged in less physical activity, and were less exposed to sun than non-immigrants [2,6]. Other studies found the length of residency since immigration was a crucial indicator of lifestyle acculturation. Higher acculturation levels were associated with significantly higher S-25(OH)D [2,5,44]. This study found that, compared with non-immigrants, immigrants had lower mean S-25(OH)D levels and more vitD insufficiency. Similarly, immigrant patients with CDs (e.g., diabetes, heart diseases, cancer, asthma, and inflammatory lung diseases) and poor/fair self-reported general and mental health had higher insufficiency levels than non-immigrants. In a systematic review and meta-analysis of 95 studies (880,128 participants), Chowdhury et al. (2014) reported an inverse association between circulating S-25(OH)D and the risk for death due to cardiovascular disease, cancer, and other mortality causes. Moreover, vitD3 supplementation was inversely associated with lower overall mortality among older adults [17]. Several other studies, including reviews, found that inadequate S-25(OH)D concentration was associated with an increased risk of CDs [26,27]. Despite the high level of evidence supporting these findings, there remains some controversy regarding the causative nature and efficacy of vitD. For example, there was a two-way correlation between vitD deficiency and disease outcomes such as infectious diseases, rheumatoid arthritis, and obesity [45]. Our previous findings indicated that immigrants had higher S-25(OH)D levels the longer they lived in Canada [2]. To gain better insights about the relationship between the duration of diagnoses and the prevalence of CDs and S-25(OH)D levels, we hypothesized that those who were recently (≤1 year or ≤5 years) diagnosed would have lower S-25(OH)D levels than those with a long duration of diagnosis (>1 year or >5 years). A previous study reported that S-25(OH)D may be sensitive to changes in health status [19]. However, our analysis revealed no difference in S-25(OH)D between recently diagnosed patients and their counterparts with a longer time since diagnosis. Another study recommended investigating pre-and post-migration experiences to better understand the healthy immigrant effect and health changes over time in the host country [6]. Therefore, we evaluated the prevalence of CDs and pre-and post-immigration vitD by the date of diagnosis and date of immigration. The results showed no differences in mean S-25(OH)D among immigrants diagnosed before immigration compared with those diagnosed after immigration for any of the studied health indicators (diabetes, heart disease, cancer, and asthma). We assumed that having lower S-25(OH)D levels (deficient or insufficient vitD) for a long time may increase the risk of a CD diagnosis. Therefore, we investigated the relationships between the length of time since immigration and vitD status and CDs. Our results showed that the longer the time after immigration, the higher the prevalence of CDs among immigrants. Our univariate analyses showed that the mean concentrations of CD-related biomarkers (HbA1c, hs-CRP, and IgE) were higher among patients with CDs compared with those without CDs. We did not expect to find the mean concentration of S-25(OH)D was higher among patients with CDs compared with their non-CD counterparts: mean (SE) 63.118 (1.787) versus 59.42 (1.76) (95%CI: −6.27, −1.13; p = 0.007). However, the findings related to immigration status and the use of vitD supplements along with other covariates may change the direction of the relationship between CDs and S-25(OH)D. For example, non-immigrants had more CDs and higher S-25(OH)D, and patients with CDs used more vitD supplementation than those without CDs. After adjusting for immigration status, vitD supplementation, and other covariates, S-25(OH)D was not associated with CDs, although HbA1c and IgE remained associated with a higher prevalence of CDs. Our analysis showed a negative association between S-25(OH)D and HbA1c, the TC/HDL ratio, IgE, ferritin, and Hb levels. Using the same data for Canadian adults, another study found the same inverse association between S-25(OH)D and elevated ferritin [46]. In addition, vitD supplementation was found to improve S-25(OH)D levels among patients with diabetes with vitD deficiency [47]. Our investigation of the concentrations of these biomarkers among immigrants compared with non-immigrants showed that immigrants had higher concentrations of HbA1c, and ferritin, a higher TC/HDL ratio, and lower levels of hs-CRP and Hb than non-immigrants. These deviations in CD-related biomarkers may be a preliminary indicator for the decline in the health status of immigrants in the long term. Our univariate analyses revealed that immigrants with insufficient vitD were more likely to self-report poor/fair general and mental health than non-immigrants. Previous studies hypothesized that low vitD concentration was associated with depression [48,49]. Another study found that anxiety, depression, and health-related quality of life were not associated with S-25(OH)D levels among the immigrant population [50]. Our analysis revealed that the prevalence of CDs was higher among those with older age, lower household income, and obesity. White ethnic groups had a higher prevalence of CDs than non-white groups, and CDs varied among immigrants from different ethnic groups and countries and regions of birth. Our finding concerning the health advantage of being immigrants (i.e., fewer CDs) was observed among Chinese, Black, Filipino, and Arab ethnicities as well as those born in South/Central America and the Caribbean, Africa, Asia, and the Philippines. In contrast, this effect was less consistent (i.e., more CDs) for those born in Italy. However, the difference in health status among immigrants may reflect variations in acculturation. The effect of origin in terms of health, life expectancy, and mortality patterns among immigrant populations varies across different ethnic groups [51]. There was no association between S-25(OH)D and CDs in our adjusted regression model. However, in addition to the environmental and lifestyle changes after immigration, we assumed the cumulative impact of immigrants’ lower levels of S-25(OH)D and the deviation in CD-related biomarkers over time played a crucial role in the subsequent health deterioration among immigrants in terms of the investigated proxy indicators. Furthermore, we assumed that improving S-25(OH)D over time, as found in our previous analysis [2], did not enhance the deviated biomarkers or cure CDs once they had developed. In a randomized control trial, vitD supplementation for patients with type 2 diabetes and asymptomatic vitD deficiency did not improve HbA1c levels [47]. However, considering the consistency of results related to higher deficiency levels of S-25(OH)D among immigrants and the relatively new findings relating to vitD and health deterioration, we cannot quantify the degree and time of deterioration nor guarantee such an association. Therefore, we recommend further research on this topic using longitudinal designs. Immigrants can be followed over their residency time in the host country to explain the patterns, developments, and direction of health deterioration. Cross-sectional surveys such as the CHMS are not usually used to examine changes in health over time. The bidirectional association between serum S-25(OH)D and CDs is complex, and it is challenging to address residual confounding using an observational study design. Longitudinal observation and interventional trials are warranted to further understand the relationship between S-25(OH)D and immigrants’ health deterioration. A limitation of using secondary data is that some of the chronic diseases could not be included in the analysis because the date of diagnosis was not provided. The CHMS data did not differentiate immigrants by specific immigration class or type (e.g., refugees, family class, economic migrants, or asylum seekers). Moreover, native-born Canadians formed a single group, despite some being second- or third-generation immigrants; this resulted in a high level of heterogeneity in the reference population. Previous immigration studies recommended assessing vitD status and its determinants among subgroups living in the same country [52]. To compare the same generation of immigrants rather than aggregated generations [53], it will be necessary to gather evidence and formulate recommendations specific to sub-populations that may differ from the overall immigrant population [32]. 5. Conclusions This study confirmed that the current health status of immigrants is good (healthy), but over time they may experience more CDs than their non-immigrant counterparts. S-25(OH)D is associated with environmental and behavioral factors and acts as a biological measure but is not associated with immigrants’ current health status. However, in addition to the adoption of unhealthy diets and lifestyles, lower levels of accumulated S-25(OH)D may impact the health profile of immigrants in terms of CD-related biomarkers. This may partially explain immigrants’ health deterioration over time. Given the emerging research interest in immigrants’ health deterioration, we recommend further longitudinal research to investigate the relationship between vitD and health deterioration, accounting for the dietary and lifestyle acculturation process. Summary of Key Issues and Findings Extensive research evidence indicates that Canadian immigrants tend to be healthier at the time of immigration than well-established immigrants and non-immigrants. This “healthy immigrant” effect lessens over time for immigrants living in Canada. Nevertheless, immigrants are at higher risk for vitD deficiency compared with non-immigrants, with the risk highest in younger and more recent immigrants. The longer an immigrant lived in Canada, the better their serum vitD levels; ethnicity was also a significant indicator of vitD deficiency. This study highlighted that immigrants tend to have better health than their counterparts in terms of the prevalence of chronic diseases (CDs) such as heart disease, cancer, asthma, and inflammatory lung diseases, but there is no difference in self-rated general and mental health. In contrast, CD-related biomarkers (e.g., glycated hemoglobin and total cholesterol/high-density lipoprotein cholesterol ratio) and serum vitD differed in favor of non-immigrants. There was no difference in the pre-migration prevalence of CDs compared with post-immigration; however, the longer immigrants lived in Canada, the higher the prevalence of CDs. The adjusted regression model showed serum vitD was not associated with CDs. Our findings suggest the cumulative impact of lower serum vitD among im-migrants over time, changing biomarkers, and environmental and lifestyle changes after immigration may play crucial roles in subsequent health deterioration among immigrants. Further longitudinal research is needed to clarify immigrants’ health deterioration and the complicated and controversial bidirectional association between serum vitD and CDs. For instance, immigrants may be followed over their residency in their host country to explore the patterns, developments, and direction of health deterioration. Acknowledgments The authors acknowledge the Carleton, Ottawa, Outaouais Local Research Data Centre team for assistance with access to the CHMS data. The CHMS data were obtained and supported by funding to the Canadian Research Data Network from the Social Sciences and Humanities Research Council, the Canadian Institute for Health Research, the Canadian Foundation for Innovation, and Statistics Canada. Although the research and analysis were based on data from Statistics Canada, the opinions expressed do not represent the views of Statistics Canada. Author Contributions S.Y. and G.A.W. conceived this study with close supervision from G.A.W.; S.Y. managed the data, verified the analytical methods, and prepared the tables and figures; A.H. verified the underlying data at the research data center; S.Y. interpreted the results and prepared the final draft; All authors, including I.C., M.P., D.M. and M.F., contributed to the design and analysis of the research and the substantive review of the manuscript. All authors have read and agreed to the published version of the manuscript. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Institutional Review Board Statement All components of the national survey were reviewed and authorized annually by the Health Canada/Public Health Agency of Canada Research Ethics Board (REB# 2005-0025). The Health Canada Research Ethics Board approved the CHMS. Informed Consent Statement Statistics Canada collected written consent from all participants for a personal interview, physical measures, and biospecimen collection. Data Availability Statement The data described in the manuscript, codebook, and analytic codes are not publicly available because the data are confidential national data hosted by Statistics Canada. Conflicts of Interest The authors declare no conflict of interest. Appendix A Supporting Information Including (Table A1, Table A2, Table A3, Table A4, Table A5, Table A6 and Table A7). nutrients-14-01760-t0A1_Table A1 Table A1 Biomarker details. Biomarkers Symbol Unit Method of Measurements/Name of the Machine and Related Information Analytical Detection Limit CHMS Reference Laboratory Precision Target Serum 25-hydroxyvitamin D [2,54] S-25(OH)D nmol/L Chemiluminescent immunoassay technology (DiaSorin, Ltd., Stillwater, MN, USA) 10–375 <20 nmol/L = 15% 20–100 nmol/L = 10% >100 nmol/L = 12% Glycated hemoglobin HbA1c % Immunoturbidimetric end-point spectrophotometry (anti-HbA1c antibody, polyhapten) on the Ortho Clinical Diagnostics Vitros 5, 1FS analyzer. %A1c is a derived test calculated from the quantitative measurements of hemoglobin and hemoglobin A1c Hb: 6.0–22.0 g/dL(component) HbA1c: 0.2–2.5 g/dL (component) %A1c: 4–14% NGSP 5% Total cholesterol/HDL cholesterol ratio TC/HDL ratio Derived calculation N/A N/A High-sensitivity C-reactive protein hs-CRP mg/L Immunoturbidimetric two-point rate reflectance spectrophotometry (monoclonal antibodies) on the Ortho Clinical Diagnostics Vitros 5, 1FS analyzer 0.10–15.0 5% Total immunoglobulin E IgE IU/mL Two-site sandwich immunoassay using direct chemiluminescent technology on the Siemens ADVIA Centaur XP analyzer 1.5–3300 10% Ferritin Ferritin µg/L Two-site sandwich immunoassay using direct chemiluminescent technology on the Siemens ADVIA Centaur XP analyzer 0.5–1650 10% Hemoglobin Hb g/L Hemoglobin is a non-cyanide-based method read photometrically at 555 nm on the POCH-i100 Sysmex hematology analyzer 0–25.0 1.5% nutrients-14-01760-t0A2_Table A2 Table A2 Weighted prevalence of chronic diseases by ethnicity. No CDs (78.3%), % CDs (21.7%), % p-Value White † 76.72 23.28 - Aboriginal 74.47 25.53 0.446 South Asian 80.85 19.15 0.309 Chinese 86.08 13.92 0.014 Black 89.33 10.67 0.002 Filipino 89.77 10.23 0.011 Latin American 86.48 13.52 0.155 Arab 88.92 11.08 0.049 Southeast Asian 83.78 16.22 0.137 West Asian 90.89 9.11 0.196 Japanese 60.2 39.8 0.141 Multiple ethnicities 72.13 27.87 0.250 Other ethnicities 84.05 15.95 <0.001 CDs (chronic diseases): at least one of the listed CDs (diabetes, heart disease, cancer, asthma, and inflammatory lung diseases). † Reference value; p ≤ 0.05. Data for Koreans were insufficient for analysis. nutrients-14-01760-t0A3_Table A3 Table A3 Weighted prevalence of chronic diseases by country and region of birth. No CDs (78.3%), % CDs (21.7%), % p-Value Region of birth Canada and North America † 76.53 23.47 - South/Central America and the Caribbean 86.75 13.25 0.011 Europe 78.96 21.04 0.314 Africa 88.74 11.26 0.008 Asia 85.94 14.06 0.006 Country of birth Canada † 76.39 23.61 - China 80.81 19.19 0.575 USA 84.31 15.69 0.224 France 89.7 10.3 0.066 Jamaica 81.21 18.79 0.648 UK 72.97 27.03 0.397 Mexico 81.54 18.46 0.527 Netherlands 60.97 39.03 0.155 India 82.51 17.49 0.115 Philippines 94.64 5.36 <0.001 Hong Kong 77.8 22.2 0.907 Germany 72.55 27.45 0.625 Italy 53.56 46.44 0.004 Iran 84.93 15.07 0.537 Lebanon 49.67 50.33 0.126 Others 87.37 12.63 <0.001 CDs (chronic diseases): at least one of the listed CDs (diabetes, heart disease, cancer, asthma, and inflammatory lung diseases). † Reference value; p ≤ 0.05. The data for Algeria, Pakistan, Romania, Colombia, and Morocco were insufficient for analysis. nutrients-14-01760-t0A4_Table A4 Table A4 Weighted mean 25(OH)D for each chronic disease based on the date of diagnosis: recent (e.g., ≤1 year) versus long-term (e.g., >1 year). S-25(OH)D, (nmol/L) Diagnosis ≤ 1 year † Diagnosis > 1 year Diagnosis ≤ 5 years † Diagnosis > 5 years % Mean (SE) % Mean (SE) 95% CI p-Value % Mean (SE) % Mean (SE) (95% CI) p-Value Diabetes 9.47 45.63 (10.28) 90.53 62.58 (3.16) −38.21, 4.31 0.113 61.8 64.80 6.41) 38.92 58.52 (3.15) −8.07, 20.62 0.374 Heart disease 7.17 58.00 (5.97) 92.83 65.80 (1.98) −20.58, 4.99 0.220 69.32 63.19 (2.58) 30.68 66.14 (2.04) −7.78, 1.88 0.218 Cancer 11.78 75.47 (3.82) 88.22 69.55 (2.68) −2.93, 14.76 0.179 59.76 66.89 (3.82) 40.24 72.44 (3.34) −16.57, 5.47 0.308 Asthma 3.56 61.17 (4.50) 96.44 57.82 (2.34) −5.19, 11.89 0.425 83.21 61.12 (2.78) 16.79 57.31 (2.43) −1.19, 8.81 0.129 † Reference value; p ≤ 0.05; SE, standard error; CI, confidence interval. nutrients-14-01760-t0A5_Table A5 Table A5 Weighted mean and status of 25(OH)D for each chronic disease and the use of vitD supplements. 25(OH)D, (nmol/L) S-25(OH)D Status VitD Supplements Mean (SE) 95% CI p-Value ≥50 nmol/L, % <50 nmol/L, % p-Value No, % Yes, % p-Value Diabetes No † 60.23 (1.66) 64.14 35.86 94.84 5.16 Yes 60.32 (3.15) −5.08, 4.89 0.968 55.11 44.89 0.043 92.6 7.54 0.142 Heart disease No † 60.04 (1.72) 63.24 36.76 94.92 5.08 Yes 65.21 (1.94) −9.42, −0.91 0.020 75.78 24.22 0.005 88.56 11.44 0.102 Cancer No † 59.51 (1.78) 62.84 37.16 95.04 4.96 Yes 70.23 (2.46) −16.44, −4.99 0.001 74.76 25.24 0.068 89.06 10.94 0.025 Asthma No † 60.40 (1.69) 64.22 35.78 94.51 5.49 Yes 58.57 (2.20) −1.02, 4.68 0.197 58.68 41.32 0.095 96.80 3.20 0.104 Inflammatory lung diseases a No † 60.90 (1.70) 64.11 35.89 92.80 7.20 Yes 61.74 (2.38) −5.03, 3.36 0.685 61.92 38.08 0.538 84.99 15.01 0.161 CDs b No CDs † 59.41(1.76) 63.41 36.59 95.11 4.89 CDs 63.11(1.79) −6.27, −1.13 0.007 64.59 35.41 0.5945 93.19 6.81 0.119 Self-rated general health Good † 56.35 (2.01) 64.4 35.6 94.81 51.9 Poor/Fair 60.66 (1.73) 1.10, 7.52 0.011 56.98 43.02 0.023 94.00 6.00 0.479 Self-rated mental health Good † 57.33 (2.24) 62.89 37.11 94.45 5.55 Poor/Fair 59.10 (1.70) −1.11, 6.44 0.157 56.51 43.49 0.197 91.61 8.39 0.401 a Inflammatory lung diseases: bronchitis + chronic obstructive pulmonary disease + emphysema; (age > 29 years). b CDs (chronic diseases): at least one of the listed CDs (diabetes, heart disease, cancer, asthma, and inflammatory lung diseases). † Reference value; p ≤ 0.05; SE, standard error; CI, confidence interval. Good = Excellent/Very Good/Good; <50 = Insufficient; ≥50 = Sufficient. nutrients-14-01760-t0A6_Table A6 Table A6 Weighted means for biomarkers by the presence of chronic diseases. Chronic Diseases (CDs) No CDs † Mean (SE) CDs Mean (SE) 95%CI p-Value Serum vitD S-25(OH)D, nmol/L 59.42 (1.76) 63.118 (1.787) −6.27, −1.13 0.007 Glucose homeostasis HbA1c, % 5.30 (0.2) 5.82 (0.05) −0.62, −0.41 <0.001 Lipid profile TC/HDL ratio 3.65 (0.03) 3.72 (0.07) −0.20, 0.07 0.309 Immune system and inflammation hs-CRP, mg/L 2.22 (0.08) 2.91 (0.14) −1.03, −0.35 <0.001 IgE, IU/mL 95.48 (5.85) 181.74 (20.96) −134.29, −38.23 0.001 Hematology Ferritin, µg/L 106.02 (2.57) 114.90 (5.02) −20.95, 3.19 0.141 Hb, g/L 139.84 (042) 140.22 (0.74) −1.86, 1.12 0.609 † Reference value; p ≤ 0.05; CDs, chronic diseases; SE, standard error; CI, confidence interval. nutrients-14-01760-t0A7_Table A7 Table A7 Linear regression of S-25(OH)D and chronic-disease-related biomarkers. S-25(OH)D and CD-Related Biomarkers Mean (SE) β (SE) 95%CI p-Value Serum vitD S-25(OH)D, nmol/L 60.22 (1.69) - - - Glucose homeostasis HbA1c, % 5.42 (0.03) −1.79 (0.79) −3.43, −0.142 0.035 Lipid profile TC/HDL ratio 3.67 (0.03) −4.09 (0.39) −4.90, −3.28 <0.001 Immune system and inflammation hs-CRP, mg/L 2.38 (0.07) −0.50 (0.23) −0.94, 0.02 0.058 IgE, IU/mL 114.6 (5.55) −0.01 (0.00) −0.01, −0.00 0.007 Hematology Ferritin, µg/L 108.05 (2.18) −0.02 (0.00) −0.03, −0.01 0.003 Hb, g/L 139.93 (0.41) −0.10 (0.04) −0.19, −0.02 0.016 p ≤ 0.05; SE, standard error; CI, confidence interval; CD, chronic disease. nutrients-14-01760-t001_Table 1 Table 1 Weighted prevalence of chronic diseases by sociodemographic and behavioral characteristics and immigration status. No CDs † (78.3%), % CDs (21.7%), % p-Value Immigration status Non-immigrant † 76.80 23.20 <0.001 Immigrant 83.36 16.64 Age at immigration, years <18 years † 91.69 8.31 <0.001 ≥18 74.75 25.25 Years after immigration ≤5 years † 91.72 8.28 0.004 >5 80.98 19.02 Years after immigration ≤10 years † 92.84 7.16 <0.001 >10 77.34 22.66 Sex Male † 78.67 21.33 0.641 Female 77.88 22.12 Age group, years <5 92.92 7.08 <0.001 5–11 89.83 10.17 12–17 85.14 14.86 18–64 † 80.01 19.99 >64 52.7 47.3 Household income (CAD) <50,000 72.34 27.66 <0.001 50,000–100,000 † 80.35 19.65 >100,000 83.09 16.91 BMI (kg/m2) Underweight 75.95 24.05 <0.001 Normal weight † 84.72 15.28 Overweight 78.64 21.36 Obese 67.56 32.44 Physical activity Yes† 84.36 15.64 <0.001 No 74.58 25.42 Education >Secondary school † 78.72 21.28 0.660 ≤Secondary school 77.87 22.13 Smoking status No/former † 77.74 22.26 0.137 Current smoker 76.27 23.73 Alcohol status No/former † 74.15 25.85 0.155 Current drinker 77.6 22.4 † Reference value; BMI, body mass index; CAD, Canadian dollars; p ≤ 0.05. nutrients-14-01760-t002_Table 2 Table 2 Weighted prevalence of chronic diseases, chronic-disease-related biomarkers, and self-rated general and mental health by immigration status. Non-Immigrants † (78.1%), % Immigrants (21.9%), % All Participants, (100%), % p-Value Diabetes (Yes) 4.82 6.59 5.21 0.095 Heart disease (Yes) 3.55 2.37 3.29 0.029 Cancer (Yes) 6.41 4.65 6.02 0.041 Asthma (Yes) 11.46 5.06 10.05 <0.001 Inflammatory lung diseases a (Yes) 3.86 1.95 3.34 0.028 CDs b (Yes) 23.2 16.64 21.72 <0.001 Self-rated general health (Poor/fair) 9.67 11.67 10.11 0.121 Self-rated mental health (Poor/fair) 8.31 6.76 7.95 0.334 S-25(OH)D, nmol/L (<50) 31.75 52.82 36.34 <0.001 CD-Related Biomarkers Non-Immigrants † Mean (SE) Immigrants Mean (SE) 95% CI p-Value Serum vitD S-25(OH)D (nmol/L) 62.72 (1.73) 51.23 (1.41) 8.37, 14.62 <0.001 Glucose homeostasis HbA1c (%) 5.39 (0.03) 5.52 (0.04) −0.19, −0.07 <0.001 Lipid profile TC/HDL ratio 3.63 (0.035) 3.79 (0.07) −0.30, −0.02 0.030 Immune system and inflammation hs-CRP (mg/L) 2.46 (0.07) 2.12 (0.13) 0.06, 0.6 0.017 IgE (IU/mL) 118.37 (6.46) 102.75 (6.59) −1.14, 32.38 0.066 Hematology Ferritin (µg/L) 103.07 (2.42) 123.07 (6.20) −34.45, −5.54 0.009 Hb (g/L) 140.54 (0.43) 137.73 (0.70) 1.22, 4.42 0.001 a Inflammatory lung diseases: bronchitis + chronic obstructive pulmonary disease + emphysema; (age > 29 years); b CDs (chronic diseases): at least one of the listed CDs (diabetes, heart disease, cancer, asthma, and inflammatory lung diseases). † Reference value; p ≤ 0.05; SE, standard error; CI, confidence interval. nutrients-14-01760-t003_Table 3 Table 3 Weighted mean 25(OH)D (nmol/L), prevalence of vitD insufficiency (<50 nmol/L), and vitD supplement use by immigration status. S-25(OH)D, (nmol/L) <50 (nmol/L) vitD Supplement Use Non-Immigrant † Mean (SE) Immigrant Mean (SE) 95% CI p-Value Non-Immigrant † % Immigrant % p-Value Non-Immigrant † % Immigrant % p-Value Diabetes 65.30 (4.05) 47.56 (4.46) −31.35, −4.13 0.013 31.43 61.63 <0.001 5.28 6.31 0.642 Heart disease 66.63 (2.15) 57.87 (3.80) −17.42, −0.11 0.048 32.12 38.38 <0.001 c c Cancer 70.71 (2.86) 67.96 (5.71) −16.49, 10.98 0.681 29.08 32.28 <0.001 5.02 8.14 0.083 Asthma 59.13 (2.37) 54.23 (3.36) −12.97, 3.18 0.222 30.68 53.59 <0.001 c c Inflammatory lung diseases a 63.87 (1.80) 53.27 (5.80) −22.09, 0.88 0.069 30.76 3841 <0.001 c c CDs b 62.18 (1.80) 56.29 (2.49) −11.23, −0.53 0.033 31.35 47.38 <0.001 4.91 4.97 0.379 Self-rated general health (poor/fair) 58.69 (1.94) 48.68 (2.98) 3.61, 16.43 0.004 31.13 61.2 <0.001 5.28 4.37 0.807 Self-rated mental health (poor/fair) 59.83 (2.43) 47.55 (3.75) 4.12, 20.45 0.005 32.52 64.39 <0.001 5.67 3.67 0.271 a Inflammatory lung diseases: bronchitis + chronic obstructive pulmonary disease + emphysema; (age > 29 years). b CDs (chronic diseases): at least one of the listed CDs (diabetes, heart disease, cancer, asthma, and inflammatory lung diseases). c Data are not sufficient for analysis. † Reference value; p ≤ 0.05; SE, standard error; CI, confidence interval. nutrients-14-01760-t004_Table 4 Table 4 Weighted mean 25(OH)D (nmol/L) based on the time of diagnosis (before or after immigration). S-25(OH)D, (nmol/L) Before Immigration † Mean (SE) After Immigration Mean (SE) 95% CI p-Value Diabetes 41.80 (3.87) 48.48 (5.31) −20.59, 7.24 0.331 Heart disease 44.38 (8.01) 59.55 (4.346) −35.52, 5.18 0.136 Cancer 71.37 (16.93) 67.8 (5.95) −36.74, 43.80 0.858 Asthma 52.82 (5.88) 56.57 (4.63) −18.74, 11.25 0.609 † Reference value; p ≤ 0.05; SE, standard error; CI, confidence interval. nutrients-14-01760-t005_Table 5 Table 5 Multivariable logistic regression analysis for chronic diseases and serum 25(OH)D (contentious) in a predefined model. CDs OR (SE) 95% CI p-Value S-25(OH)D, nmol/L 1 nmol/L 1.007 (0.004) 1.00, 1.01 0.070 Immigration status Immigrants 0.40 (0.09) 0.25, 0.63 <0.001 Age, years 1 year 1.02 (0.008) 1.00, 1.04 0.024 Sex Female 0.86 (0.18) 0.56, 1.33 0.494 Household income, CAD (reference < 50,000) 50,000–100,000 0.61 (0.11) 0.42, 0.88 0.011 ≥100,000 0.56 (0.15) 0.32, 0.99 0.046 Ethnic group Non-white 1.00 (0.29) 0.54, 1.86 0.998 BMI, kg/m2 (reference normal weight) Obese 1.92 (0.47) 1.15, 3.20 0.014 Met physical activity recommendations No 0.80 (0.21) 0.46, 1.38 0.408 Education ≤Secondary school 0.97 (0.19) 0.69, 1.39 0.841 Smoking status Current smoker 1.45 (0.34) 0.89, 2.36 0.126 Alcohol Current drinker 0.64 (0.17) 0.37, 1.12 0.114 VitD supplement or analog use Yes 0.83 (0.25) 0.45, 1.54 0.540 HbA1c (%) 1% 2.63 (0.37) 1.96, 3.51 <0.001 IgE (IU/mL) 1 IU/mL 1.01 (0.00) 1.00, 1.001 0.001 TC/HDL ratio Ratio 0.98 (0.10) 0.80, 1.21 0.843 hs-CRP mg/L 1 mg/L 1.01 (0.04) 0.92, 1.10 0.889 Ferritin (µg/L) 1 µg/L 1.00 (0.00) 1.00, 1.001 0.380 Hb (g/L) 1 g/L 0.98 (0.01) 0.97, 1.00 0.060 Constant 0.005 (0.01) 0.000, 0.10 0.001 CDs (chronic diseases): at least one of the listed CDs (diabetes, heart disease, cancer, asthma, and inflammatory lung diseases). Adjusted for immigration status, age, sex, household income, education, ethnicity, BMI, physical activity, smoking behavior, alcohol consumption, vitD supplement use, HbA1c (glycated hemoglobin), TC/HDL-R (total cholesterol/high-density lipoprotein cholesterol ratio), Hb (hemoglobin), ferritin, IgE (immunoglobulin E), and hs-CRP (high-sensitivity C-reactive protein). Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Gushulak B.D. Williams L.S. National Immigration Health Policy Existing Policy, Changing Needs, and Future Directions Can. J. Public Health 2004 95 I27 I29 10.1007/BF03403662 15191129 2. Yousef S. Manuel D. Colman I. Papadimitropoulos M. Hossain A. Faris M. Wells G. Vitamin D Status among First-Generation Immigrants from Different Ethnic Groups and Origins: An Observational Study Using the Canadian Health Measures Survey Nutrients 2021 13 2702 10.3390/nu13082702 34444863 3. StatisticsCanada Population Projections for Canada, Provinces and Territories 2014 Available online: https://www150.statcan.gc.ca/n1/en/pub/91-520-x/91-520-x2010001-eng.pdf?st=LWFifucU (accessed on 20 November 2021) 4. Newbold B. The short-term health of Canada’s new immigrant arrivals: Evidence from LSIC Ethn. Health 2009 14 315 336 10.1080/13557850802609956 19263262 5. Sanou D. O’Reilly E. Ngnie-Teta I. Batal M. Mondain N. Andrew C. Newbold B.K. Bourgeault I. Acculturation and Nutritional Health of Immigrants in Canada: A Scoping Review J. Immigr. Minority Health 2014 16 24 34 10.1007/s10903-013-9823-7 23595263 6. Vang Z. Sigouin J. Flenon A. Gagnon A. The Healthy Immigrant Effect in Canada: A Systematic Review Popul. Change Lifecourse Strateg. Knowl. Clust. Discuss. Pap. Ser. Réseau Strat. Connaiss. Changements Popul. Parcour. Vie Doc. Trav. 2015 3 4 7. Autier P. Vitamin D status as a synthetic biomarker of health status Endocrine 2016 51 201 202 10.1007/s12020-015-0837-x 26718192 8. Chiu M. Austin P.C. Manuel D. Tu J.V. Cardiovascular Risk Factor Profiles of Recent Immigrants vs Long-term Residents of Ontario: A Multi-Ethnic Study Can. J. Cardiol. 2012 28 20 26 10.1016/j.cjca.2011.06.002 21827964 9. Kinnon D. Canadian Research on Immigration and Health Cat. No H21-149/1999E Health Canada Ottawa, ON, Canada 1999 10. Williams D.R. Mohammed S.A. Leavell J. Collins C. Race, socioeconomic status, and health: Complexities, ongoing challenges, and research opportunities Ann. N. Y. Acad. Sci. 2010 1186 69 101 10.1111/j.1749-6632.2009.05339.x 20201869 11. Martin C.A. Gowda U. Renzaho A.M.N. The prevalence of vitamin D deficiency among dark-skinned populations according to their stage of migration and region of birth: A meta-analysis Nutrition 2016 32 21 32 10.1016/j.nut.2015.07.007 26643747 12. Spiro A. Buttriss J.L. Vitamin D: An overview of vitamin D status and intake in Europe Nutr. Bull. 2014 39 322 350 10.1111/nbu.12108 25635171 13. Centers for Disease, Control Prevention Guidelines for Evaluation of the Nutritional Status and Growth in Refugee Children during the Domestic Medical Screening Examination Centers for Disease, Control Prevention, Division of Global Migration and Quarantine Atlanta, GA, USA 2013 14. Granlund L. Ramnemark A. Andersson C. Lindkvist M. Fhärm E. Norberg M. Prevalence of Vitamin D deficiency and its association with nutrition, travelling and clothing habits in an immigrant population in Northern Sweden Eur. J. Clin. Nutr. 2016 70 373 379 10.1038/ejcn.2015.176 26508457 15. De Boer I.H. Kestenbaum B. Shoben A.B. Michos E.D. Sarnak M.J. Siscovick D.S. 25-Hydroxyvitamin D levels inversely associate with risk for developing coronary artery calcification J. Am. Soc. Nephrol. 2009 20 1805 1812 10.1681/ASN.2008111157 19443637 16. Zmijewski M.A. Vitamin D and Human Health Int. J. Mol. Sci. 2019 20 145 10.3390/ijms20010145 30609781 17. Chowdhury R. Kunutsor S. Vitezova A. Oliver-Williams C. Chowdhury S. Kiefte-de-Jong J.C. Khan H. Baena C. Prabhakaran D. Hoshen B. Vitamin D and risk of cause specific death: Systematic review and meta-analysis of observational cohort and randomised intervention studies BMJ 2014 348 g1903 10.1136/bmj.g1903 24690623 18. Holick M.F. The Vitamin D Defi ciency Pandemic: A Forgotten Hormone Important for Introduction: Historical Perspective Public Health Rev. 2010 32 267 283 10.1007/BF03391602 19. McNeill G. Vyvyan J. Peace H. McKie L. Seymour G. Hendry J. MacPherson I. Predictors of micronutrient status in men and women over 75 years old living in the community Br. J. Nutr. 2002 88 555 561 10.1079/BJN2002706 12425736 20. Sun Y.-Q. Langhammer A. Skorpen F. Chen Y. Mai X.-M. Serum 25-hydroxyvitamin D level, chronic diseases and all-cause mortality in a population-based prospective cohort: The HUNT Study, Norway BMJ Open 2017 7 e017256 10.1136/bmjopen-2017-017256 21. Gaksch M. Jorde R. Grimnes G. Joakimsen R. Schirmer H. Wilsgaard T. Mathiesen E.B. Njølstad I. Løchen M.-L. März W. Vitamin D and mortality: Individual participant data meta-analysis of standardized 25-hydroxyvitamin D in 26,916 individuals from a European consortium PLoS ONE 2017 12 e0170791 10.1371/journal.pone.0170791 28207791 22. Jablonski N.G. Chaplin G. Human skin pigmentation, migration and disease susceptibility Philos. Trans. R. Soc. B Biol. Sci. 2012 367 785 792 10.1098/rstb.2011.0308 23. Heaney R.P. Holick M.F. Why the IOM recommendations for vitamin D are deficient J. Bone Miner. Res. 2011 26 455 457 10.1002/jbmr.328 21337617 24. Pham T. Vitamin D Status of Immigrant and Ethnic Minority Children Ages 2 to 5 y in Montréal Ph.D. Thesis McGill University Montreal, QC, Canada 2012 25. Mithal A. Wahl D.A. Bonjour J.P. Burckhardt P. Dawson-Hughes B. Eisman J.A. El-Hajj Fuleihan G. Josse R.G. Lips P. Morales-Torres J. Global vitamin D status and determinants of hypovitaminosis D Osteoporosis International Springer London, UK 2009 1807 1820 26. Peterlik M. Boonen S. Cross H.S. Lamberg-Allardt C. Vitamin D and calcium insufficiency-related chronic diseases: An emerging world-wide public health problem Int. J. Environ. Res. Public Health 2009 6 2585 2607 10.3390/ijerph6102585 20054456 27. Renzaho A.M.N. Bhsci J.A.H. Nowson C. Vitamin D, obesity, and obesity-related chronic disease among ethnic minorities: A systematic review Nutrition 2011 27 868 879 10.1016/j.nut.2010.12.014 21704500 28. Cristian Ilie P. Stefanescu S. Smith L. The role of vitamin D in the prevention of coronavirus disease 2019 infection and mortality Aging Clin. Exp. Res. 2020 32 1195 1198 10.1007/s40520-020-01570-8 32377965 29. Ghasemian R. Shamshirian A. Heydari K. Malekan M. Alizadeh-Navaei R. Ebrahimzadeh M.A. Warkiani M.E. Jafarpour H. Bazaz S.R. Shahmirzadi A.R. The Role of Vitamin D in The Age of COVID-19: A Systematic Review and Meta-Analysis MedRxiv 2021 10.1111/ijcp.14675 30. Mercola J. Grant W.B. Wagner C.L. Evidence Regarding Vitamin D and Risk of COVID-19 and Its Severity Nutrients 2020 12 3361 10.3390/nu12113361 31. Meltzer D.O. Best T.J. Zhang H. Vokes T. Arora V. Solway J. Association of Vitamin D Status and Other Clinical Characteristics with COVID-19 Test Results JAMA Netw. Open 2020 3 e2019722 10.1001/jamanetworkopen.2020.19722 32880651 32. Yousef S. Elliott J. Manuel D. Colman I. Papadimitropoulos M. Hossain A. LeClair N. Wells G.A. Study protocol: Worldwide comparison of vitamin D status of immigrants from different ethnic origins and native-born populations—A systematic review and meta-analysis Syst. Rev. 2019 8 211 10.1186/s13643-019-1123-4 31439035 33. Levinsky M. Schiff M. Lifetime cumulative adversity and physical health deterioration in old age: Evidence from a fourteen-year longitudinal study Soc. Sci. Med. 2021 289 114407 10.1016/j.socscimed.2021.114407 34555682 34. Thacker S.B. Stroup D.F. Carande-Kulis V. Marks J.S. Roy K. Gerberding J.L. Measuring the public’s health Public Health Rep. 2006 121 14 22 10.1177/003335490612100107 16416694 35. Von Elm E. Altman D.G. Egger M. Pocock S.J. Gøtzsche P.C. Vandenbroucke J.P. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: Guidelines for reporting observational studies Bull. World Health Organ. 2007 85 867 872 10.2471/BLT.07.045120 18038077 36. Day B. Langlois R. Tremblay M. Knoppers B.M. Canadian Health Measures Survey: Ethical, legal and social issues Health Rep./Stat. Can. Can. Cent. Health Inf. 2007 18 37 51 37. Ng E. Canadian Health Measures Survey: A Tool for Immigrant Health Research Health Report Statistics Canada Ottawa, ON, Canada 2015 38. Dawson-Hughes B. Vitamin D deficiency in Adults: Definition, Clinical Manifestations, and Treatment-UpToDate Update-2021 2021 Available online: https://www.uptodate.com/contents/vitamin-d-deficiency-in-adults-definition-clinical-manifestations-and-treatment?search=vitamin (accessed on 31 January 2022) 39. Ross A.C. Taylor C.L. Yaktine A.L. Del Valle H.B. Dietary Reference Intakes. Calcium Vitamin D Committee to Review Dietary Reference Intakes for Vitamin D and Calcium Food and Nutrition Board National Academies Press Washington, DC, USA 2011 40. StatisticsCanada Canadian Health Measures Survey, Cycle 3—Privacy Impact Assessment. 2012–2013 Statistics Canada Ottawa, ON, Canada 2013 41. StatisticsCanada Canadian Health Measures Survey, Cycle 4, 2014–2015—Privacy Impact Assessment Summary Statistics Canada Ottawa, ON, Canada 2015 42. Paszat L. Sutradhar R. Liu Y. Baxter N.N. Tinmouth J. Rabeneck L. Risk of colorectal cancer among immigrants to Ontario, Canada BMC Gastroenterol. 2017 17 85 10.1186/s12876-017-0642-5 28683721 43. Cheung M.C. Earle C.C. Fischer H.D. Camacho X. Liu N. Saskin R. Shah B.R. Austin P.C. Singh S. Impact of immigration status on cancer outcomes in Ontario, Canada J. Oncol. Pract. 2017 13 e602 e612 10.1200/JOP.2016.019497 28605254 44. Salant T. Diane S.L. Measuring culture: A critical review of acculturation and health in Asian immigrant populations Soc. Sci. Med. 2003 71 71 90 10.1016/S0277-9536(02)00300-3 45. Spedding S. Vitamin D and Human Health 2015th ed. MDPI Basel, Switzerland Beijing, China Wuhan, China 2015 46. Munasinghe L.L. Ekwaru J.P. Mastroeni M.F. Mastroeni S.S.B.S. Veugelers P.J. The association of serum 25-hydroxyvitamin D concentrations with elevated serum ferritin levels in normal weight, overweight and obese Canadians PLoS ONE 2019 14 e0213260 10.1371/journal.pone.0213260 30845159 47. Randhawa F.A. Mustafa S. Khan D.M. Hamid S. Effect of Vitamin D supplementation on reduction in levels of HbA1 in patients recently diagnosed with type 2 Diabetes Mellitus having asymptomatic Vitamin D deficiency Pak. J. Med. Sci. 2017 33 881 10.12669/pjms.334.12288 29067058 48. Anglin R.E. Samaan Z. Walter S.D. McDonald S.D. Vitamin D deficiency and depression in adults: Systematic review and meta-analysis Br. J. Psychiatry 2013 202 100 107 10.1192/bjp.bp.111.106666 23377209 49. Chu F. Ohinmaa A. Klarenbach S. Wong Z.-W. Veugelers P. Serum 25-Hydroxyvitamin D Concentrations and Indicators of Mental Health: An Analysis of the Canadian Health Measures Survey Nutrients 2017 9 1116 10.3390/nu9101116 50. Granlund L.E. Ramnemark A.K. Andersson C. Lindkvist M.C. Norberg M. Fhärm E.M. Vitamin D status was not associated with anxiety, depression, or health-related quality of life in Middle Eastern and African-born immigrants in Sweden Nutr. Res. 2020 75 109 118 10.1016/j.nutres.2020.02.006 32120223 51. Huh J. Prause J.A. Dooley C.D. The impact of nativity on chronic diseases, self-rated health and comorbidity status of Asian and Hispanic immigrants J. Immigr. Minority Health 2008 10 103 118 10.1007/s10903-007-9065-7 52. Hilger J. Friedel A. Herr R. Rausch T. Roos F. Wahl D.A. Pierroz D.D. Weber P. Hoffmann K. A systematic review of vitamin D status in populations worldwide Br. J. Nutr. 2014 111 23 45 10.1017/S0007114513001840 23930771 53. Zhu L. Depression risks and correlates among different generations of Chinese Americans: The effects of relationships with friends and relatives Soc. Sci. 2017 6 56 10.3390/socsci6020056 54. Munasinghe L.L. Yuan Y. Willows N.D. Faught E.L. Ekwaru J.P. Veugelers P.J. Vitamin D deficiency and sufficiency among Canadian children residing at high latitude following the revision of the RDA of vitamin D intake in 2010 Br. J. Nutr. 2017 117 457 465 10.1017/S0007114517000320 28245892

PMC009xxxxxx/PMC9099620.txt

==== Front Polymers (Basel) Polymers (Basel) polymers Polymers 2073-4360 MDPI 10.3390/polym14091700 polymers-14-01700 Article Green Corrosion Inhibition on Carbon-Fibre-Reinforced Aluminium Laminate in NaCl Using Aerva Lanata Flower Extract https://orcid.org/0000-0002-2679-473X Hynes Navasingh Rajesh Jesudoss 1* Vignesh Nagarajan Jawahar 1 https://orcid.org/0000-0002-8682-6078 Barile Claudia 2 https://orcid.org/0000-0001-7963-5671 Velu Pitchumani Shenbaga 3 https://orcid.org/0000-0002-9301-9776 Baskaran Thangagiri 4 Jappes Jebas Thangiah Winowlin 5 https://orcid.org/0000-0002-3282-2346 Al-Khashman Omar Ali 6 https://orcid.org/0000-0002-8158-1278 Brykov Michail 7 https://orcid.org/0000-0002-6976-0767 Ene Antoaneta 8* Ku Bon-Cheol Academic Editor 1 Department of Mechanical Engineering, Mepco Schlenk Engineering College, Sivakasi 626005, Virudhunagar, India; vigneshmech@mepcoeng.ac.in 2 Dipartimento di Meccanica, Matematica e Management, Politecnico di Bari, Viale Japigia 182, 70126 Bari, Italy; claudia.barile@poliba.it 3 Department of Mechanical Engineering, PSR Engineering College, Sivakasi 626140, Virudhunagar, India; velupitchumani@gmail.com 4 Department of Chemistry, Mepco Schlenk Engineering College, Sivakasi 626005, Virudhunagar, India; thangagiri@gmail.com 5 School of Automotive and Mechanical Engineering, Kalasalingam Academy of Research and Education, Krishnankoil 626126, Virudhunagar, India; winowlin@yahoo.com 6 Department of Environmental Engineering, Faculty of Engineering, Al-Hussein Bin Talal University, Ma’an P.O. Box 20, Jordan; omarkhashman@yahoo.com 7 Zaporizhzhya Polytechnic National University, 69600 Zaporizhzhya, Zaporiz′ka Oblast′, Ukraine; m@brykov.com 8 Department of Chemistry, Physics, and Environment, INPOLDE Research Center, Dunarea de Jos University of Galati, 800008 Galaţi, Romania * Correspondence: findhynes@yahoo.co.in (N.R.J.H.); antoaneta.ene@ugal.ro (A.E.) 21 4 2022 5 2022 14 9 170018 3 2022 12 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Aluminium-based fibre–metal laminates are lucrative candidates for aerospace manufacturers since they are lightweight and high-strength materials. The flower extract of aerva lanata was studied in order to prevent the effect of corrosion on the aluminium-based fibre–metal laminates (FMLs) in basic media. It is considered an eco-friendly corrosion inhibitor using natural sources. Its flower species belong to the Amaranthaceae family. The results of the Fourier-transform infrared spectroscopy (FTIR) show that this flower extract includes organic compounds such as aromatic links, heteroatoms, and oxygen, which can be used as an organic corrosion inhibitor in an acidic environment. The effectiveness of the aerva-lanata flower behaviour in acting as an inhibitor of the corrosion process of FMLs was studied in 3.5% NaCl solution. The inhibition efficiency was calculated within a range of concentration of the inhibitor at room temperature, using the weight-loss method, potentiodynamic polarization measurements and electrochemical-impedance spectroscopy (EIS). The results indicate a characterization of about 87.02% in the presence of 600 ppm of inhibitor. The Tafel curve in the polarization experiments shows an inhibition efficiency of 88%. The inhibition mechanism was the absorption on the FML surface, and its absorption was observed with the aid of the Langmuir adsorption isotherm. This complex protective film occupies a larger surface area on the surface of the FML. Hence, by restricting the surface of the metallic layer from the corrosive medium, the charge and ion switch at the FML surface is reduced, thereby increasing the corrosion resistance. Aerva lanata green corrosion inhibitor carbon-fibre-reinforced aluminium laminates electrochemical-impedance spectroscopy polarization measurements SEM study langmuir absorption technique ==== Body pmc1. Introduction Nowadays, the great part of the automotive and aerospace industries is searching for improvements to the use of lightweight components in order to increase the strength-to-density ratio, which implies the possibility of decreasing the weight of structures while simultaneously ensuring the high performance in terms of strength, stiffness, flexibility, corrosion resistance, wear resistance, etc. The most famous automobile companies are looking for hybrid materials that could assist in achieving the above-mentioned properties [1,2]. Friction stud welding [3], diffusion bonding [4], friction welding [5], friction drilling, and friction riveting are some of the processes employed for joining multi-material structures. Nevertheless, multi-material structures developed by combining fibres and metal laminates have promising properties and applications. An analysis of new works shows that sufficiently advanced protection methods, such as arc spraying with flux-cored wire [6], novel superhydrophobic coatings [7], and superdispersed polytetrafluoroethylene (SPTFE) and polyvinylidene fluoride (PVDF) coatings [8] can be used for various connected structural elements, including aluminium alloys, which makes them promising methods for the aerospace industry. Carbon-fibre-reinforced plastics (CFRPs) are characterised by the best specific strength that is at least two times higher than steel. The weight of a component could be decreased by up to half if a CFRP is used instead of steel. For vehicle applications, this means the possibility of making lighter vehicles that consume less fuel. As a result, they are considered advanced structural materials [9]. Currently, the design of Al-based fibre–metal laminates is restricted to the use of glass fibre (GF), since the risk of galvanic corrosion prohibits the use of carbon-fibre-reinforced plastics. Previous research has explored galvanic corrosion between fibre composites and a variety of metals. Tavakkolizadeh et al. [9] suggested that galvanic corrosion occurs when carbon fibres have been in electrical contact with steel in an electrolyte. Moreover, they have shown that the use of coating could reduce the rate of galvanic corrosion, but not eliminate it. Ireland et al. studied the effect of carbon nanotubes, including GF-reinforced epoxy, on galvanic corrosion with aluminium. They assessed that galvanic corrosion did occur between them, even if a polymer barrier existed between the two conductive materials. Plant extracts are highly available natural products that have corrosion-inhibition properties and are a renewable source and by nature, they are non-poisonous. The ample chemical constituents that are present in the plant extracts such as alkenes, polyphenols, and aromatics are capable of inhibiting the corrosion process in mild steel [10]. Most of the herbal products containing functional groups such as C–Cl, C–O, NH2, C–H, C=O, O–H and CHO are potential inhibitors. The above compounds become adsorbed and form a protective layer on the surface of the steel to restrict the formation of corrosion. Steven et al. [11] investigated the probability of galvanic corrosion of carbon fibre and aluminium in FML. In this examination, the authors studied the galvanic corrosion behaviour between the bulk metallic glass (BMG) and the CFRP. They assessed that the BMG showed less corrosion than the Al combined with CFRP. Mehdi Yari et al. [12] investigated the properties of carbon composites. The authors determined that when a considerable region of carbon composites is coupled to small metallic parts such as nuts, screws and clasps, the galvanic corrosion was a significant threat. Perez et al. [13] researched the galvanic corrosion between carbon steel and a hardened metal that was treated with 1M NaOH. The evaluation was carried out at two different conditions: with and without chloride. They determined that there is no significant threat or massive damage due to galvanic corrosion when carbon steel and hardened metal are electrically coupled in a strong, fortified structure. Akhil et al. [14] investigated the corrosion behaviour of mild steel in 0.5 M H2SO4 by using Saraka Ashoka. They proposed that the application of this extract containing epicatechin helps in limiting the corrosion rate of the mild steel. The excellent inhibition effect of mild metal in 0.5 M H2SO4 was evaluated at 100 mg/L by using the electrochemical and weight-loss measurements. Atomic force microscopy (AFM) and scanning electron microscopy (SEM) were also used for analysing the surface morphology. The electrochemical studies showed that there was a 95.48% inhibition effectivity at 100 mg/L inhibitor concentration. The following Table 1 describes the inhibition efficiencies exhibited by different plant extracts and the corresponding medium used during the process [15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30]. Jakubczak et al. [15] carried out their study on the interlaminar shear strength of the CARALLs with different aluminium-surface preparations and various fibre combinations because of the effect of thermal ageing. Their study revealed that galvanic corrosion is reduced when there is an insertion of a thin glass ply in between the metal sheets and carbon laminates. However, this insertion does not have any effect on the ILSS or the thermal fatigue of the laminates. Pan et al. performed a study on the influence of anodizing and annealing the aluminium sheets in an FML [16]. They found that the mechanical properties lowered due to the effect of annealing. Kim et al. (2010) carried out a systematic investigation on the adhesion strength of the CFRP/steel bond to determine the effect of the surface morphology of steel. They incorporated a micro-periodic line pattern on the steel surface for carrying out the investigation. Their findings show that the enhancement of strength is due to the transition from interfacial failure to cohesive failure. Reyes and Gupta (2009) included glass fibre-reinforced polypropylene in place of conventional thermosetting parts used in FMLs. They applied a zinc coating to the surface of the steel layer to achieve very good adhesion on the polymer–metal interface. Carbon-fibre-reinforced aluminium laminates were used for carrying out the low-velocity impact on the specimens. Experimental and numerical simulations were performed on the prepared samples, both qualitatively and quantitatively. It was found that matrix fractures, carbon-fibre cracking, and delamination were the important modes of damage [31]. Novel superhydrophobic coatings were applied to the surface of an aluminium alloy to prevent the effect of corrosion by using Al2O3/siloxane hybrids that were grown in situ on the surface of the aluminium. The experiment proved that the aluminium alloy AA 2024 has excellent corrosion resistance in NaCl, alkaline and other acidic environments [32]. Superdispersed polytetrafluoroethylene (SPTFE) and polyvinylidene fluoride (PVDF) were used as the coating medium for studying the anti-icing properties of the samples, which were oxidised by plasma electrolytic oxidation. The coatings having the combined PVDF–SPFTE layers with a ratio of 1:4 showed significant performance with hydrophobicity, ice-phobic and electrochemical characteristics compared to all the other sample combinations. The use of this PVDF–SPFTE coating also proved that there was a reduction in the corrosion current density in powers of order 5 when compared with uncoated aluminium alloy [32]. The cored-wire arc-spraying technique was used for spraying ultra-high-molecular-weight polyethene (UHMWPE) particles onto the surface of aluminium during the aluminium spray coatings. These particles act as sealants. The microstructure evaluation was carried out in the study. To study the effect of corrosion, neutral-salt spraying and electrochemical analysis were performed on the coated aluminium sample. The study revealed that the UHMWPE particles acted as sealants, helping to improve corrosion resistance [33]. The present paper aims to study the behaviour of aerva lanata in aluminium-based FMLs, in terms of corrosion and absorption. It is an Indian plant species and belongs to the family of Amaranthaceae. The extraction process was carried out on a soxhlet-extraction handle apparatus with a 3.5% NaCl solution, and its inhibition efficiencies on carbon-fibre-and-aluminium-metal-based FMLs were investigated. The performance and mechanism of inhibition was also discussed. The characteristics of FMLs were studied by employing Fourier-transform infrared spectroscopy (FTIR) and SEM analysis was carried out on the prepared surface. 2. Experimentation 2.1. Materials Preparation In the present study, carbon-fibre-and-aluminium-metal-based FMLs were studied. The fibre–metal laminates were made by stacking alternating layers of AA 6061 sheets and carbon fibre for a total of 5 layers (starting from AA 6061 and again ending at AA 6061 with carbon fibre in alternate layers). The dimensions of the sheets and fibres were cut to 150 mm × 250 mm, which was equal to the size of the die. The thickness or depth of the die used was about 5 mm. The thickness of the AA 6061 sheets was equal to 0.5 mm and that of the carbon fibre was 0.25 mm. The elemental composition of AA 6061 is given in the following Table 2. For preparing these FMLs, carbon fibre and aluminium sheets were bonded together by deposing a layer of reinforcing adhesive. They were joined together by using epoxy resin with the commercial name Araldite LY 556 bought from a retail manufacturer Excel Trading Corporation, Pune, India. The hardener used was Aradur HY 951 bought from a retail manufacturer Aerium Tech Private Limited, Mumbai, India. Both the epoxy resin and the hardener were mixed in the ratio of 10:1 parts by weight using the rule of mixture calculations. Then, a compressive force of about 10 bars was applied to the die using a compressive moulding machine available at Department of Mechanical Engineering, Mepco Schlenk Engineering College, Sivakasi, India. The following Figure 1 shows the prepared FML. After this, for carrying out the corrosion tests, the samples were cut from the FML in dimensions of 5 mm × 5 mm. The following Figure 1 shows the cross-sectional view of the carbon fibre/aluminium 6061 FML laminate sandwich. The preparation of the flower extract as a corrosion inhibitor is discussed herein. Initially, the Aerva-lanata flower was dried for 7 sunny days. Then the extraction process was carried out in a soxhlet-extraction mantle apparatus. A total of 10 g of aerva-lanata-flower powder was mixed with 170 mL of distilled water. The powdered sample was refluxed for four hours using distilled water at 80 °C. Then it was filtered to obtain the extract solution of 70 mL. Finally, the filtered solution was heated on a hot plate. The hot plate was maintained at 100 °C for more than 1 h. A quantity equal to 70 mL was boiled by a hot plate until 40–35 mL remained, then it was poured in petri dish. The petri dish was kept in an open atmosphere for three days. Finally, the Aerva-lanata-extract powder was collected. 2.2. Weight-Loss Measurements For evaluating the rate of corrosion in aqueous solutions, the method of immersion represents an easy technique. In the present work, the immersion corrosion tests on FMLs were carried out for a period of 5 days in a medium of 3.5% NaCl. For weight-loss estimation, the carbon-fibre-and-aluminium-metal-based FMLs were prepared according to ASTM G 31-72 standards. Before the FMLs were exposed to the environment, they were cleaned, dried and weighed, and then exposed to 3.5% NaCl. The results were examined at 25 °C in the presence and absence of inhibitor for an immersion period of 5 days. By using this method, the corrosion rate and the efficiency of the inhibitor was also determined. 2.3. Electrochemical Measurements Electrochemical measurements were carried out using the simple three-electrode cell system. It was a Cyclic Voltammetry Electrochemical Cell manufactured by Ossila Ltd., Sheffield, UK. It involved an easier method containing a mild steel electrode, a platinum electrode, and a saturated calomel electrode (SCE) as the working, counter and reference electrodes, respectively. During the tests, the working electrode was immersed in the 3.5% NaCl test solution for 1 h, to obtain a stabilised open-circuit potential (OCP). Electrochemical-impedance spectroscopy (EIS) was used for scanning from 100 kHz to 0.01 Hz frequency, with a sign-amplitude perturbation of 5 mV at OCP. From this value, the Tafel and Nquist graphs were drawn in order to determine the inhibition efficiency of the extracted Aerva-lanata sample on the FMLs. 2.4. FTIR Spectroscopy In order to understand the inhibition mechanism in a better way, the FTIR spectra of Aerva-lanata extract were examined. The FTIR measurements were carried out on a IRSpirit FTIR Spectrometer, Japan purchased from Toshvin Analytical Pvt. Ltd., Mumbai, India. The Aerva-lanata extract was reduced into powder form for FTIR characterization by means of a FTIR 8400s spectrophotometer with a wave number between 500–4000 cm−1. 2.5. Hardness Studies The hardness test was used to study the influence of the corrosive NaCl on the hardness of the FMLs. The hardness tests were carried out on a micro Vickers hardness tester purchased from Walter Uhl technische Mikroskopie Gmbh & Co. KG, Asslar, Germany. Tests were carried out on a micro Vickers hardness tester with an indentation load of 500 g for 10 s on the FMLs under three conditions. First, the FML was tested for hardness without being subjected to corrosive NaCl. Secondly, the FML specimen that had not been treated with the Aerva lanata extract surface coating was subjected to NaCl corrosion and then measured with Vickers hardness. Finally, the hardness value was also measured for the specimen that was first coated with the Aerva lanata extract and then subjected to corrosion. 2.6. SEM Imaging Scanning-electron-microscopic studies were carried out on the prepared samples. The SEM analysis was performed on a ZEISS GeminiSEM Field Emission Scanning Electron Microscope, made in Oberkochen, Germany. Images of the carbon-fibre-reinforced aluminium laminate, the bare material, and the flower-extract-coated specimen, both before and after the immersion test, were obtained. 3. Results and Discussion 3.1. FTIR Characterization Figure 2 reports the FTIR results. It shows that the Aerva lanata sample had the stretching vibration of O–H, causing the peak centre at 3522.16 cm−1 belonging to the amine functional group [28]. The stretching vibration of C–Cl caused the peak at 2362.34 cm−1, which belongs to carbolic-acid functional group. The stretching vibration of C–O caused the peak at 1556.44 cm−1, which belongs to useful alkene group. The stretching vibration of C–H caused the peak at 593.40 cm−1, which belongs to aromatic functional group. The FTIR results confirm the presence of anticorrosive properties of the Aerva lanata extract that prevent corrosion on the surface. Since the extracts are mainly composed of few low-molecular-weight compounds, FTIR analysis is a suitable method to identify them. Therefore, to examine the more prevalent compounds in the Aerva lanata extract, FTIR analysis was used. 3.2. Weight-Loss Measurements Table 3 shows the weight-loss values, inhibition efficiency (η %) and surface coverage (θ) for the carbon-fibre FML at different concentrations of Aerva-lanata extract. From the weight-loss values, corrosion rates (CR) were calculated by the following equation, CR = (ΔW × K)/ρ × A × t(1) where CR is the corrosion rate in mmpy, ΔW is the weight loss before/after the immersion test, K is a constant (for mmpy K value be 8.75 × 104), ρ is the specimen density (g/cm3), A is the exposed area (cm2) and t is the exposure time (h). The inhibition efficiency (η) was calculated as follows:(2) η=Wo−WiWo×100 where Wi and W0 are the weight of the specimen in the presence and absence of the inhibitor, respectively. Table 3 also shows the corrosion rate (mmpy) and the inhibition efficiency (η %) of the FML in 3.5% NaCl at different concentrations of Aerva-lanata extract. From the above results, it can be seen that the corrosion rate of the FML was reduced as the concentration of Aerva-lanata extract was increased. This is due to the phenomenon of the precipitation reaction caused by the adsorption of the active ingredients of the Aerva-lanata extract on the carbon-fibre-and-aluminium-metal surface. The average inhibition effectivity of the Aerva-lanata inhibitor on the FML was 87.032%. Table 4 indicates the weight of the specimens before and after coating. Figure 3 indicates the weight loss of FML due to the action of corrosion with respect to the number of days for each of the bare and coated FML in 3.5% NaCl. Figure 3 indicates that there was only a small reduction in the weight of the coated FML when compared to the uncoated bare FML. So, the Aerva lanata extract yields a great efficiency by providing a protective covering over the surface of the FML. Figure 4 indicates the corrosion rate of FML with respect to the number of days in which the FML was immersed in 3.5% NaCl. This Aerva lanata extract was capable of yielding an effectivity of about 87% on carbon-fibre-and-aluminium-metal-based FMLs in 3.5% NaCl. 3.3. Electrochemical Study 3.3.1. Polarization Measurements The effect of the concentration of the Aerva-lanata extract on the polarization behaviour of the FML in 3.5% NaCl were analysed and the Tafel plots were recorded for different inhibitor concentrations, as shown in Figure 5. The corrosion current densities were calculated by the intersection corresponding to the corrosion potential. When there was higher concentration, it resulted in a lower current density at 600 ppm [29]. (3) η=Io corr−Ii corrIo corr where Io corr and Ii corr represent the corrosion-density values with and without the inhibitor on the FML surface, respectively. Figure 5 clearly shows that the anodic metallic-dissolution and cathodic hydrogen-evolution reactions were inhibited when the concentration of aerva lanata was increased in the aggressive medium. The Tafel plots were plotted for the FML for two conditions viz. before and after adding the aerva-lanata-flower extract to the FML surface, which were controlled by means of charge transfer between the cathodic and anodic reaction mechanism. The combination of physisorption and chemisorption properties causes the active ingredients of the Aerva-lanata-flower extract to be absorbed strongly onto the FML surface. Corrosion on the surface was prevented by using this flower extract on the surface of the FML, and it can be proved by fact that the corrosion current-density value decreasing by increasing the Aerva-lanata inhibitor concentration. The result from Table 4 suggests that by increasing the concentration of the Aerva-lanata extract, there is decrease in the corrosion current density. The lowest corrosion current density of 2.335 × 10−3 A/cm2 was obtained at 600 ppm concentration at the rate of 88% on the carbon-fibre-and-aluminium-metal-based FML. From Figure 5 it can be seen that there was a positive shift, which is due to the corrosion resistance of the carbon-fibre-and-aluminium-metal-based FMLs with the Aerva-lanata extract concentration on the surface. It can be seen that the maximum positive shift happens at 600 ppm compared to the other inhibitor concentrations. Additionally, this 600 ppm concentrated solution yielded the highest efficiency of about 88%, which is shown in Table 5. 3.3.2. Electrochemical-Impedance Spectroscopy Electrochemical-impedance-spectroscopy (EIS) measurements were performed on the carbon-fibre-and-aluminium-metal-based fibre–metal laminates to study the impedance parameters in a 3.5% NaCl environment with different Aerva-lanata-flower-extract concentrations. Table 6 shows the results of the EIS process. Figure 6a shows the EIS Nyquist curves with different concentrations of Aerva-lanata extract. It can be seen that as the concentration of Aerva lanata extract increased, the values of charge-transfer resistance (Rct) also increased. A maximum value of 301.15 Ω cm2 was reached. This rise in charge-transfer resistance reduced the number of active sites created by the adsorption of chloride ions on the surface of the FML, which led to the protection. The equivalent circuit is shown in Figure 6b. In the Nyquist plot, it can be seen that a semicircle is formed in each curve, which is due to the load-transfer resistance, which stands for a particular time constant. Increasing the concentration of the Aerva lanata extract enlarged the capacitive-loop diameter from uncoated to coated at about 600 ppm, from which it is understood that an inhibition effect was being developed. The inhibition efficiency can be calculated in the impedance study using the following formula, (4) η=Rct−R°ctRct×100 where Rct and R°ct are the inhibitor and non-inhibitor load-transfer resistances, respectively. The efficiency of inhibition is enhanced by increasing the Aerva lanata extract concentration, reaching the highest value of 85.9 percent at 600 mg/L. Table 6 clearly shows that when the Rct value increases, the CPE value correspondingly increases. This is due to the formation of the protective film on the surface of the carbon-fibre-and-aluminium-metal laminates. This transition in Rct and CPE values was caused by the substitution of water molecules on the carbon-fibre-and-aluminium-metal-based FML surface by the adsorption of the Aerva lanata inhibitor. The inhibited solutions had higher values of n than those that were uninhibited due to a reduction in the surface heterogeneity as a result of the adsorption of the electrode electrolyte Aerva lanata inhibitor on the FML interface based on carbon fibre and aluminium metal. Inductive loops were present at low frequency for the EIS curves of a blank solution, which is due to the result of the surface relaxation of the adsorbed intermediate products. The inductive loops disappeared for the remaining concentrations. 3.4. Mechanical-Hardness Test The Vickers hardness test conducted on the carbon-fibre-and-aluminium-metal laminates that were immersed in the 3.5% NaCl corrosion test solution are shown in Figure 7. The raw specimen indicated a most astounding hardness strength of 522 VHN, whereas when it was subjected to corrosion its hardness value drastically decreased to 209 VHN. However, at the same time, when the corroded FML was treated with the Aerva lanata extract, it improved the hardness ability up to 299 VHN. Table 7 shows the Vickers hardness values for the FML at various conditions. 3.5. Surface Analysis Scanning Electron Microscope Figure 8 shows the carbon-fibre-and-aluminium-metal-based fibre–metal laminate surface morphology in 3.5% NaCl solution with and without the aerva-lanata-flower inhibitor. Figure 8a shows the bare carbon-fibre-and-aluminium-metal-based fibre–metal laminates that were subjected to polishing before undergoing SEM, whereas Figure 8b shows the SEM image of the bare carbon-fibre-and-aluminium-metal-based fibre–metal laminate after eight hours of immersion in 3.5% NaCl, and Figure 8c shows the SEM image of the carbon-fibre-and-aluminium-metal-based fibre–metal laminates that were immersed in 3.5% NaCl and also coated with the Aerva-lanata extract. The surface morphology of the carbon-fibre-and-aluminium-metal-based fibre–metal-laminate specimen was incredibly rough due to surface corrosion when the specimen was immersed in the 3.5% NaCl solution, but the specimen that was coated with the Aerva-lanata extract showed less surface morphology than Figure 8b. This is due to the formation of an extract-solution protective layer on the carbon-fibre-and-aluminium-metal-based fibre–metal-laminate surface. The smoother surface was obtained by the presence of the green corrosion inhibitor on the surface of the carbon-fibre-and-aluminium-metal-based fibre–metal laminates [29]. The effect of microstructure was investigated by comparing the corrosion behaviour of the devitrified (crystalline) to that of the non-crystalline BMG. The results were similar to those reported in a previous study by Peter et al., wherein the corrosion current density from crystalline BMG was slightly greater than the non-crystalline structure. Surface oxide layers can exhibit different potentials than the base metal, and thus affect corrosion behaviour. For example, the standard reversible electrode potential for titanium is negative, but in practice, the electrode potential of titanium in the galvanic series is positive because of the passive oxide layer on the surface. 4. Conclusions For the first time in a basic medium, the Aerva-lanata flower was successfully employed as a green corrosion inhibitor for carbon-fibre-and-aluminium-metal-based fibre–metal laminates. In fact, the flower of aerva lanata contains aromatic rings, heteroatoms, and oxygen, which makes it a suitable candidate for acting as an inhibitor in the basic medium. The inhibition efficiency reached a maximum of about 92 percent with electrochemical techniques in the presence of 600 ppm inhibitor. SEM analysis was used to analyse the surface morphology of the carbon-fibre-and-aluminium-metal-based fibre–metal-laminate microstructure in the absence and presence of the aerva-lanata-extract inhibitor. The Langmuir adsorption isotherm was detected as the inhibition mechanism. Due to the complex protective film formation between the inhibitor and the metal surface ions, there was a decrease in charge and ion transfer on the metal surface, which was recognised as the reason for the existence of the inhibition property. Acknowledgments The work of Antoaneta Ene was financed by Dunarea de Jos University of Galati, through the internal grant with contract no. RF 3621/2021. Author Contributions Conceptualization, N.R.J.H. and A.E.; methodology, N.J.V. and C.B.; validation, P.S.V.; formal analysis, A.E. and M.B.; investigation, N.R.J.H.; Writing—Original draft preparation, N.R.J.H.; Writing—Review and editing, T.B. and J.T.W.J.; visualization, O.A.A.-K.; supervision, N.R.J.H. and A.E.; project administration, A.E.; funding acquisition, A.E. All authors have read and agreed to the published version of the manuscript. Funding The research is partially funded by “Dunarea de Jos” University of Galati, Romania, grant no. RF 3621/2021. Data Availability Statement Data is contained within the article. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Cross-sectional view of the laminate sandwich. Figure 2 FTIR results of Aerva lanata. Figure 3 Weight Loss (g) Vs. No. of Days in NaCl. Figure 4 Corrosion rate (mmpy) vs. no. of days in NaCl. Figure 5 Tafel graph for FML in NaCl. Figure 6 (a) Nyquist plot for FML in NaCl; (b) Equivalent-circuit for calculation. Figure 7 Vickers hardness test on FML coated with Aerva-lanata extract. Figure 8 SEM images of carbon-fibre-reinforced aluminium laminate. (a) carbon-fibre-reinforced aluminium laminate; (b) carbon-fibre-reinforced aluminium laminate after immersion test; (c) View of corroded region in carbon-fibre-reinforced aluminium laminate after immersion test. polymers-14-01700-t001_Table 1 Table 1 Inhibition Efficiency Developed from Different Plant Extracts. Plant Name Inhibition Efficiency Material Used Medium Saraka Ashoka 95.48 carbon steel 0.5 M H2SO4 Chitosan 92% 300 mg/L on mild steel 1M sulfamic acid Glycyrrhiza glabra leaves 88% 800 ppm on mild steel 1 M HCl Sunflower-seed hull 98% 300 ppm on aluminium 1 M HCl Pyridazinium 84% 100 mg/L on mild steel 1 M HCl Pyrazolo-pyridines 97% 100 mg/L on carbon steel 1 M HCl Salvia officinalis 96% 2500 mg/L on stainless steel HCl Osmanthus fragran 94% 340 mg/L on carbon steel HCl Musa paradisica 90% 300 mg/L on carbon steel HCl Mangrove tannins 89% 6000 mg/L on metal Acidic medium Jasminumnudiflorum 92% 1000 mg/L on aluminium HCl Lawsonia inermis 92% 1200 mg/L on moderate steel 1 m HCl Dendrocalamus brandisii 90% 1000 mg/L on aluminium HCl, H3PO4 Aqueous coffee grounds 83% 400 mg/L on carbon steel 1 M HCl Phyllanthus amarus 81% 4000 mg/L on mild steel Acidic media Black radish 92% 1000 mg/L on carbon steel - Ginkgo 80% 100 mg/L on carbon steel HCl and H2SO4 polymers-14-01700-t002_Table 2 Table 2 Weight-Loss-Method Calculation of FML in 3.5% NaCl. Elemental Composition % of Element Manganese (Mn) 0.0–0.15 Iron (Fe) 0.0–0.70 Magnesium (Mg) 0.80–1.20 Silicon (Si) 0.40–0.80 Copper (Cu) 0.15–0.40 Zinc (Zn) 0.0–0.25 Titanium (Ti) 0.0–0.15 Chromium (Cr) 0.04–0.35 Other (Each) 0.0–0.05 Others (Total) 0.0–0.15 Aluminium (Al) Balance polymers-14-01700-t003_Table 3 Table 3 Weight-Loss-Method Calculation of FML in 3.5% NaCl. Days Weight of the Bare FML (g) Corrosion Rate (mmpy) Weight of the Coated FML (g) Corrosion Rate (mmpy) Efficiency (%) Initial 1.5394 - 1.8249 - - 1 1.5250 3.5358 1.8230 0.4665 86.8 2 1.5079 4.1982 1.8210 0.4912 88.29 3 1.4924 3.8065 1.8189 0.5056 86.71 4 1.4768 3.8312 1.8168 0.5147 86.56 5 1.4608 3.9287 1.8147 0.5184 86.8 Average 87.032 polymers-14-01700-t004_Table 4 Table 4 Weight of the Specimen before and after Coating. Concentration (ppm) Before Coating (g) After Coating (g) Bare 1.5394 - 600 1.5346 1.8890 polymers-14-01700-t005_Table 5 Table 5 Polarization parameters for FML in NaCl. Specimen Ecorr (V) Icorr (A cm−2) Efficiency (η %) Brae FML −0.79 0.001959 - Coated FML −0.77 0.0002335 88 polymers-14-01700-t006_Table 6 Table 6 EIS parameters for FML in NaCl. Specimen Rct (Ω cm2) CPE (μF cm−2) n Efficiency (η %) Brae FML (0) 300.23 1.09 × 10−3 0.0715 - Coated FML (150) 301.15 1.72 × 10−3 0.0785 85.9 polymers-14-01700-t007_Table 7 Table 7 Vickers hardness values for the FML at various conditions. Specimen Applied Force (P) in kg Vickers Hardness Number (VHN) VHN = (1824 × P)/d2 Average Vickers Hardness Number Standard Deviation Trial 1 Trial 2 Trial 3 Coated FML after immersion 30 293 307 297 299 5.89 Bare FML after immersion 30 211 215 201 209 5.89 Untreated FML 30 530 518 518 522 5.66 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Sinmazcelik T. Avcu E. Bora M.O. Çoban O. A review: Fibre metal laminates, background, bonding types and applied test methods Mater. Des. 2011 32 3671 3685 10.1016/j.matdes.2011.03.011 2. Sankaranarayanan R. Rajesh Jesudoss Hynes N. Prospects of Joining Multi-Material Structures AIP Conf. Proc. 2017 1953 130021 130025 3. Hynes N.R.J. Velu P.S. Simulation of friction welding of alumina and steel with aluminum interlayer Int. J. Adv. Manuf. Technol. 2017 93 121 127 10.1007/s00170-015-7874-8 4. Velu P.S. Hynes N.R.J. Numerical Analysis of Friction Welded Titanium Joints J. Achiev. Mater. Manuf. 2016 76 26 29 5. Kumar R. Hynes N.R.J. Finite-element simulation and validation of material flow in thermal drilling process J. Braz. Soc. Mech. Sci. Eng. 2018 40 162 172 10.1007/s40430-018-1091-y 6. Wang X. Zhao H. Wu S. Suo X. Wei X. Li H. Aluminum-polyethylene composite coatings with self-sealing induced anti-corrosion performances J. Mater. Process. Technol. 2020 282 116642 10.1016/j.jmatprotec.2020.116642 7. Jakubczak P. Bieniaś J. Dadej K. Experimental and numerical investigation into the impact resistance of aluminium car-bon laminates Compos. Struct. 2020 244 112319 10.1016/j.compstruct.2020.112319 8. Li F. Zhao M. Zhan Y. Wu C. Zhang Y. Jiang X. Sun Z. Facile fabrication of novel superhydrophobic Al2O3/polysiloxane hybrids coatings for aluminum alloy corrosion protection Colloids Surf. A Physicochem. Eng. Asp. 2022 640 128444 10.1016/j.colsurfa.2022.128444 9. Tavakkolizadeh M. Saadatmanesh H. Repair of Damaged Steel-Concrete Composite Girders Using Carbon fibre—Reinforced Polymer Sheets Compos. Struct. 2003 7 311 322 10. Krishnegowda P. Venkatesha V. Krishnegowda P. Shivayogiraju S. Acalypha torta leaf extract as green corrosion inhibitor for mild steel in hydrochloric acid solution Ind. Eng. Chem. Res. 2013 52 722 728 10.1021/ie3018862 11. Hamill L. Hofmann D.C. Nutt S. Galvanic Corrosion and Mechanical Behaviour of Fibre Metal Laminates of Metallic Glass and Carbon Fibre Composites Adv. Eng. Mater. 2018 2 1700711 10.1002/adem.201700711 12. Qu D. Qian S.Y. Baldock B. Effects of galvanic coupling between carbon steel and stainless-steel reinforcements NACE North. Area Conf. Proc. 2006 45 475 483 13. D’Urso G. Giardini C. Lorenzi S. Cabrini M. Pastore T. The Effects of process parameters on mechanical properties and corrosion behavior in friction stir welding of aluminum alloys Procedia Eng. 2017 183 270 276 10.1016/j.proeng.2017.04.038 14. Gupta N.K. Joshi P.G. Srivastava V. Quraishi M.A. A macromolecule as green corrosion inhibitor for mild steel in sulfamic acid useful for sugar industry Int. J. Biol. Macromol. 2018 106 704 711 10.1016/j.ijbiomac.2017.08.064 28818723 15. Alibakhshi E. Ramezanzadeh M. Bahlakeh G. Ramezanzadeh B. Glycyrrhiza glabra leaves extract as a green corrosion inhibitor for mild steel in 1 M hydrochloric acid solution J. Mol. Liq. 2018 255 185 198 10.1016/j.molliq.2018.01.144 16. Hassannejad H. Nouri A. Sunflower seed hull extract as a novel green corrosion inhibitor for aluminium in HCl solution J. Mol. Liq. 2018 254 377 382 10.1016/j.molliq.2018.01.142 17. El-Hajjaji F. Messali M. Aljuhani A. Aouad M.R. Hammouti B. Belghiti M.E. Chauhan D.S. Quraishi M.A. Pyridazinium-based ionic liquids as novel and green corrosion inhibitors of carbon steel in acidmedium J. Mol. Liq. 2018 249 997 1008 10.1016/j.molliq.2017.11.111 18. Dohare P. Quraishi M.A. Lgaz H. Salghi R. Ultrasound induced green synthesis of pyrazolo-pyridines as novel corrosion inhibitors useful for industrial pickling process: Experimental and theoretical approach Results Phys. 2019 13 102344 10.1016/j.rinp.2019.102344 19. Soltani N. Tavakkoli N. Khayatkashani M. Jalali M.R. Mosavizade A. Green approach to corrosion inhibition of 304 stainless steel in hydrochloric acid solution by the extract of Salvia officinalis leaves Corros. Sci. 2012 62 122 135 10.1016/j.corsci.2012.05.003 20. Li L. Zhang X. Lei J. He J. Zhang S. Pan F. Adsorption and corrosion inhibition of Osmanthus fragran leaves extract on carbon steel Corros. Sci. 2012 63 82 90 10.1016/j.corsci.2012.05.026 21. Ji G. Anjum S. Sundaram S. Prakash R. Musa paradisica peel extract as green corrosion inhibitor for mild steel in HCl solution Corros. Sci. 2015 90 107 117 10.1016/j.corsci.2014.10.002 22. Rahim A. Rocca E. Steinmetz J. Kassim M. Adnan R. Ibrahim M.S. Mangrove tannins and their flavanoid monomers as alternative steel corrosion inhibitors in acidic medium Corros. Sci. 2007 49 402 417 10.1016/j.corsci.2006.04.013 23. Deng S. Li X. Inhibition by Jasminum nudiflorum Lindl leaves extract of the corrosion of aluminium in HCl solution Corros. Sci. 2012 64 253 262 10.1016/j.corsci.2012.07.017 24. Ostovari A. Hoseinieh S.M. Peikari M. Shadizadeh S.R. Hashemi S.J. Corrosion inhibition of mild steel in 1 M HCl solution by henna extract: A comparative study of the inhibition by henna and its constituents (Lawsone, Gallic acid, alpha-d-Glucose and Tannic acid) Corros. Sci. 2009 51 1935 1949 10.1016/j.corsci.2009.05.024 25. Li X. Deng S. Inhibition effect of Dendrocalamus brandisii leaves extract on aluminum in HCl, H3PO4 solutions Corros. Sci. 2012 65 299 308 10.1016/j.corsci.2012.08.033 26. Torres V. Inhibitory action of aqueous coffee ground extracts on the corrosion of carbon steel in HCl solution Corros. Sci. 2011 53 2385 2392 10.1016/j.corsci.2011.03.021 27. Okafor P.C. Ikpi M.E. Uwah I.E. Ebenso E.E. Ekpe U.J. Umoren S.A. Inhibitory action of Phyllanthus amarus extracts on the corrosion of mild steel in acidic media Corros. Sci. 2008 50 2310 2317 10.1016/j.corsci.2008.05.009 28. Radojčić I. Berković K. Kovač S. Vorkapić-Furač J. Natural honey and black radish juice as tin corrosion inhibitors Corros. Sci. 2008 50 1498 1504 10.1016/j.corsci.2008.01.013 29. Deng S. Li X. Inhibition by Ginkgo leaves extract of the corrosion of steel in HCl and H2SO4 solutions Corros. Sci. 2012 55 407 415 10.1016/j.corsci.2011.11.005 30. Goyal M. Pareek A. Nagori B.P. Sasmal D. Aerva Lanata: A Review on Phytochemistry And Pharmacological Aspects Pharm. Rev. 2011 5 195 198 10.4103/0973-7847.91120 31. Hynes N.R.J. Selvaraj R.M. Mohamed T. Mukesh A.M. Olfa K. Nikolova P.M. Aerva lanata flowers extract as green corrosion inhibitor of low-carbon steel in HCl solution: An in vitro study Chem. Pap. 2020 75 1165 1174 10.1007/s11696-020-01361-5 32. Jafari Y. Nooshabadi M.S. Ghoreishi S.M. Electropolymerized coatings of poly(oanisidine) and poly(o-anisidine)-TiO2 nanocomposite on aluminium alloy 3004 by using the galvanostatic method and their corrosion protection performance Poly. Adv. Tech. 2014 25 279 287 10.1002/pat.3233 33. Egorkin V.S. Mashtalyar D.V. Gnedenkov A.S. Filonina V.S. Vyaliy I.E. Nadaraia K.V. Imshinetskiy I.M. Belov E.A. Izotov N.V. Sinebryukhov S.L. Icephobic Performance of Combined Fluorine-Containing Composite Layers on Al-Mg-Mn–Si Alloy Surface Polymers 2021 13 3827 10.3390/polym13213827 34771383

PMC009xxxxxx/PMC9099621.txt

==== Front Int J Environ Res Public Health Int J Environ Res Public Health ijerph International Journal of Environmental Research and Public Health 1661-7827 1660-4601 MDPI 10.3390/ijerph19094961 ijerph-19-04961 Article Influence of Matrix Type on Marginal Gap Formation of Deep Class II Bulk-Fill Composite Restorations https://orcid.org/0000-0002-1968-8453 Hahn Britta * Haubitz Imme Krug Ralf Krastl Gabriel https://orcid.org/0000-0003-0887-9550 Soliman Sebastian Mertens Christian Academic Editor Tchounwou Paul B. Academic Editor Department of Conservative Dentistry and Periodontology, University Hospital Würzburg, Pleicherwall 2, 97070 Würzburg, Germany; imme.haubitz@gmx.de (I.H.); krug_r@ukw.de (R.K.); krastl_g@ukw.de (G.K.); soliman_s@ukw.de (S.S.) * Correspondence: hahn_b1@ukw.de 19 4 2022 5 2022 19 9 496118 3 2022 15 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Background: To test the hypothesis that transparent matrices result in more continuous margins of bulk-fill composite (BFC) restorations than metal matrices. Methods: Forty standardized MOD cavities in human molars with cervical margins in enamel and dentin were created and randomly assigned to four restorative treatment protocols: conventional nanohybrid composite (NANO) restoration (Tetric EvoCeram, Ivoclar Vivadent, Schaan, Liechtenstein) with a metal matrix (NANO-METAL) versus transparent matrix (NANO-TRANS), and bulk-fill composite restoration (Tetric EvoCeram Bulk Fill, Ivoclar Vivadent, Schaan, Liechtenstein) with a metal matrix (BFC-METAL) versus transparent matrix (BFC-TRANS). After artificial aging (2500 thermal cycles), marginal quality was evaluated by scanning electron microscopy using the replica technique. Statistical analyses were performed using the Mann–Whitney U-test and Wilcoxon test. The level of significance was p < 0.05. Results: Metal matrices yielded significantly (p = 0.0011) more continuous margins (46.211%) than transparent matrices (27.073%). Differences in continuous margins between NANO (34.482%) and BFC (38.802%) were not significant (p = 0.56). Matrix type did not influence marginal gap formation in BFC (p = 0.27) but did in NANO restorations (p = 0.001). Conclusion: Metal matrices positively influence the marginal quality of class II composite restorations, especially in deep cavity areas. The bulk-fill composite seems to be less sensitive to the influence of factors such as light polymerization and matrix type. transparent matrix metal matrix bulk-fill technique centripetal technique marginal gap formation class II restoration SEM ==== Body pmc1. Introduction The impacts of various oral health conditions on oral health-related quality of life (OHRQoL) have been extensively studied in the literature [1]. It is well documented that higher DMFS (Decayed Missed Filled Surfaces) scores are associated with a significantly greater impact on self-reported OHRQoL than lower DMFS scores [2]. Thus, modern restorative dentistry should focus on prevention and high-quality, long-lasting restorations in order to slow down the “restorative death spiral”. In recent decades, considerable developments have been made in dental resin composites [3]. Bulk-fill composite (BFC) materials, in particular, have gained considerable clinical acceptance [4,5], because they enable the placement of thicker composite layers (~4 mm) with a sufficient depth of cure and less polymerization shrinkage stress [4,6,7,8]. A higher depth of cure has been achieved by using higher-translucency composite materials to improve light transmission or by adding optimized highly reactive photo-initiators such as a dibenzoyl germanium derivative (e.g., Ivocerin® in Tetric EvoCeram® Bulk Fill; Ivoclar Vivadent, Schaan, Liechtenstein), besides the conventional camphorquinone [9,10,11,12]. Bulk filling simplifies the restorative process, saves time, and reduces the risk of technical errors, such as the formation of voids between layers [12,13]. In view of their properties, it can be concluded that bulk-fill materials can be recommended for large and deep cavities [14,15]. In clinical practice, dentists are often confronted with cavities significantly deeper than 4 mm, which are especially demanding with regard to light polymerization. Furthermore, such deep defects are difficult to seal with a matrix and moisture control remains a major challenge. It has been shown that pre-contoured matrices are beneficial for creating proximal contacts [16,17,18,19], especially when combined with separation rings, and for reducing overhangs [20,21]. Flat matrix bands also produced satisfactory results in other studies [22,23,24,25,26]. When restoring deep cavities with margins below the cementoenamel junction (CEJ), rigid metallic matrices may facilitate matrix placement and adaptation [20,26,27,28,29,30]. However, light polymerization may be compromised if the light guide tip is partially covered when using a metal matrix [7,31]. On the other hand, an older study showed that metal matrices with a reflective surface can focus the light cervically within the cavity and thus achieve a higher depth of cure than transparent matrices [26]. Optimal positioning and angulation of the light guide tip is the key to ensuring light transmission to each area of the composite layer [30,32]. Accordingly, use of the three-sited light curing technique after metal matrix removal has been recommended to ensure a sufficient depth of cure [33,34]. Nevertheless, even this polymerization technique does not prevent the attenuation of light intensity during the penetration of dental hard tissue, so the extension of curing time also seems necessary [35,36]. Countless matrix systems are available on the market, including flat or pre-contoured bands, retainer-fixed circumferential systems, and sectional matrices, and most feature either metal or plastic matrices [17,18,20,21,22,37,38,39]. A recent survey by Schaalan [22] revealed that Egyptian dentists prefer sectional matrix systems over circumferential matrix systems, but the author did not mention whether there was a difference between plastic and metal matrices [22]. In a clinical trial by Demarco et al. [27], however, the clinical performance of composite restorations did not depend on whether a transparent plastic or metallic matrix was used, but rather was more strongly influenced by deterioration of the adhesive bond and composite material—a conventional micro-hybrid composite (Filtek P60, 3M ESPE, St. Paul, MN, USA) in this case. However, there are no studies investigating this question for bulk-fill materials, which are usually placed using the bulk-fill technique. It has been shown that restoring deep cavities leads to large volumes of composite material if filled in bulk, and that the larger the volume of composite material, the greater the marginal gap formation [40]. The current literature lacks information on the extent to which the type of matrix (transparent or metal) might influence marginal gap formation in deep class II bulk-fill composite restorations. Such data would be useful, since metal matrices are easier to place but can impair light polymerization, as described above. Therefore, this in vitro study aims to test the hypothesis that transparent matrices result in more continuous margins of bulk-fill composite restorations than metal matrices. 2. Materials and Methods Ethical approval for the use of extracted human teeth for material testing of dental restorations was obtained from the local Ethics Committee (approval number: AZ 15/15). Forty freshly extracted, caries-free human molars of nearly equal size were stored in 0.1% chloramine T solution until further processing. All mesio-occlusal-distal (MOD) cavities were prepared and filled within seven consecutive days. The specimens were randomly assigned to four treatment groups of ten specimens each featuring two types of restorative materials and techniques—conventional nanohybrid composite (Tetric EvoCeram, Ivoclar Vivadent, Schaan, Liechtenstein) for centripetal layering versus bulk-fill composite (Tetric EvoCeram Bulk Fill, Ivoclar Vivadent, Schaan, Liechtenstein) for bulk-filling—and two types of matrix systems—metal (METAL) matrices versus transparent plastic matrix bands (TRANS) secured in a Tofflemire retainer, respectively. Self-curing resin (Paladur, Heraeus Kulzer, Hanau, Germany) was used to embed the teeth by means of a Teflon mold with the occlusal surfaces parallel to the ground. Box-shaped MOD cavities (occlusal box: 2.0 mm deep, 3.5 mm wide) were prepared using hand-held cylindrical 1.2 mm diamond burs (grain size 80–100 µm and 40 µm; Komet, Lemgo, Germany) in a high-speed contra-angle handpiece (INTRAmatic Lux 3 25 LH, KaVo, Biberach, Germany). Interproximal boxes were prepared using the same instrument to a buccolingual width of 3.5 mm. The cervical margin of the mesial box was located 1.5 mm above the cementoenamel-junction, but not deeper than 4.0 mm from the occlusal surface, and that of the distal box was located 1.5 mm beyond the CEJ, but not deeper than 7.0 mm from the occlusal surface. The enamel parts of the interproximal boxes were converted into a bevel design using a sonic preparation system (SONICflex LUX 2000 L, KaVo, Biberach, Germany) with a standardized oscillating diamond tip (SONICsys Approx, No. 36, KaVo, Biberach, Germany), which was completely immersed into the tooth. The beveled design in the enamel was finished with an oscillating Bevelshape file (No. 01, Intensiv, Montagnola, Switzerland) in a contra-angle handpiece (INTRAmatic Lux 2 20 KN, KaVo, Biberach, Germany) with an oscillating head (Intra EVA Head L6 R, KaVo, Biberach, Germany). The bevel width was 1 mm. The box-shaped design in dentin was finished with an oscillating Cavishape file (CS 140, Intensiv, Grancia, Switzerland). All preparation instruments were replaced with new instruments after ten completed cavity preparations. Cavity design is shown in Figure 1. Cavity dimensions were continuously monitored during preparation by means of loupes (2.5× magnification, Zeiss, Oberkochen, Germany) and a periodontal probe. As shown in Figure 2, the test teeth were mounted between artificial tooth models to simulate physiological interproximal relations. The mounted specimens were restored using either metal matrix bands (399 C, Kerr, Bioggio, Switzerland) or transparent matrix bands (DEL, Dental Exports London, Feltham, UK), respectively, secured in a Tofflemire retainer (Omnident, Rodgau Nieder-Roden, Germany). Each matrix band was secured interdentally–cervically with wooden wedges (Hawe Sycamore Interdental Wedges, Kerr; Orange, CA, USA), and laterally, at the vertical cavity margins, with separation rings (Composi-Tight 3D 400 Thin Tine G/Ring, Garrison Dental Solutions, Spring Lake, MI, USA). The contact area was burnished with a hand instrument (PFI19, Hu-Friedy, Frankfurt, Germany) so that no visual space was left between the matrix and the adjacent tooth. Enamel and dentin were etched (30 and 15 s, respectively) with 37% phosphoric acid gel (Omni-Etch, Omnident, Rodgau, Germany) and then rinsed with water spray for 20 s. A two-step etch-and-rinse bonding agent (OptiBond FL, Kerr Italia S.r.l., Scafati, Italy) was applied and processed according to the manufacturer’s instructions. Bonding agent was polymerized from the occlusal direction, and each proximal box was light-cured for 20 s. Cavities were filled with conventional nano-hybrid composite (Tetric EvoCeram, Ivoclar Vivadent, Schaan, Liechtenstein) using a centripetal layering technique or with bulk-fill composite (Tetric EvoCeram Bulk Fill, Ivoclar Vivadent, Schaan Liechtenstein) using a bulk-fill technique (see Table 1). The centripetal layering technique involves initial restoration of the absent proximal wall, thus transforming the class II cavity into a class I cavity. Each increment of composite was light-cured for 20 s with a mono-wave LED light curing device (Elipar Freelight 2, 3 M ESPE, Seefeld, Germany) at 1020 mW/cm2, verified with a radiometer (Bluephase Meter II, Ivoclar Vivadent, Schaan, Liechtenstein). With the bulk-fill technique, intermediate light-curing was performed once after filling the proximal boxes and modeling the proximal wall, as otherwise, the maximum increment thickness of 4 mm would have been significantly exceeded. With the centripetal technique, on the other hand, intermediate light-curing was performed after each individual increment. After removal of the matrix band, restorations were post-cured for a further 20 s from the buccal and lingual side, respectively, with the specimen teeth still secured within the artificial tooth model. An overview of the experimental groups and restorative techniques is given in Figure 3. The test teeth were then taken off the artificial tooth model for hand-held finishing. Composite overhangs were removed with a scalpel (No. 15, Braun, Aesculap AG, Tuttlingen, Germany), and the restorations were finished with a brown rubber polisher (Komet, Lemgo, Germany) at 10,000 rpm with water spray cooling to allow SEM analysis of the restoration margins. All restorations and measurements were performed by one calibrated operator (B.H.) after the samples were blinded by an independent observer (S.S.). For artificial aging, the specimens were stored in physiological saline solution in an incubator (Memmert, Schwabach, Germany) for seven days at 37 °C followed by thermal cycling (MT & UKT 600, Lauda, Lauda Königshofen, Germany). The specimens were subjected to 2500 cycles of alternating cold and hot water treatment (5 °C and 55 °C) following another seven days of storage in physiological saline solution. The specimen teeth (n = 40) were replicated with epoxy resin (Araldite, Ciba-Geigy, Basel, Switzerland) for analysis by scanning electron microscopy (SEM). The mesial and distal surfaces of each specimen were cast with silicone, yielding a total of n = 80 replicas. These were subsequently sputter-coated with gold in a sputter coater (EMITECH K550 Emitech, Taunusstein, Germany). Marginal quality was assessed by measuring the percentage of continuous margins and marginal gaps, respectively, using a scanning electron microscope (DSM 940, Zeiss, Oberkochen, Germany) with 100× to 1000× magnification and calibrated measuring software (RaEm©; programmer: Peter Müller, 97267 Himmelstadt, Germany). The results were expressed as a percentage of the respective quality outcome variables along the total margin length for each test group. The two different marginal qualities (continuous margin vs. marginal gap) are illustrated in Figure 4. For clarity, only the proportion of continuous margins [%] is depicted in the results section. Therefore, the proportion of marginal gaps is 100% minus the proportion of continuous margins. All statistical analyses were performed using the WinMEDAS statistical software package (Version 8/20, C. Grund, Würzburg, Germany). Since there was no Gaussian normal distribution of the measured values, rank tests were used. The Wilcoxon test (p-values depicted as Pw) was used for comparison between two measurements of dependent samples, i.e., to test for differences between enamel and dentin margins. The Mann–Whitney U-test (p-values depicted as Pu) was used for independent samples to compare measurements between the two composite materials or the two matrix systems, respectively. In case of statistically significant differences, Cohen’s effect size (ES dCohen) was calculated. Cohen’s effect size shows how strongly a parameter affects the outcome and reflects its clinical relevance. The effect sizes were classified as small (ES dCohen < 0.5), medium (ES dCohen = 0.5–0.8) or large (ES dCohen > 0.8). To compare the test results quantitatively, p-values were calculated. The significance level was set at p < 0.05. p-values were marked with asterisks to denote the significance level as follows: * p < 0.05, ** p < 0.01, *** p < 0.001. 3. Results SEM Analysis Figure 5 shows the proportions of continuous margins [%] in enamel and dentin in all groups. The percentage of continuous margins was significantly higher in cavity segments located in enamel than in dentin in all four test groups (Table 2 and Figure 5; Pw = 0.00005 ***). Metal matrices yielded significantly more continuous margins than transparent matrices (Pu = 0.0011 **; Table 3, line 3) with a large effect size in dentin (ES dCohen = 0.87; Table 3, line 2) and a medium effect size in enamel (ES dCohen = 0.77; Table 3, line 1). This result was mainly observed in the groups with the conventional nano-hybrid composite, as reflected by the statistically significant difference and large effect size (ES dCohen = 2.27) between the NANO-METAL and NANO-TRANS groups (Pu = 0.0010 **) (Table 4 and Figure 5). However, the bulk-fill groups (BFC-METAL and BFC-TRANS) had no statistically significant difference between the two matrix types (Pu = 0.27). Bulk-fill composite combined with the bulk-fill technique resulted in significantly more continuous margins within dentin (Table 3; Pu = 0.031 *, medium effect size dCohen = 0.58). On the other hand, the quality of margins located within enamel did not differ significantly between the two composite materials or restorative techniques (Pu = 0.87) (Table 3). 4. Discussion The aim of this study was to test the hypothesis that transparent matrices result in more continuous margins of bulk-fill composite restorations than metal matrices. The hypothesis was rejected, as no statistically significant difference in marginal quality between the two matrix systems could be detected. These findings are in agreement with those of other (laboratory and clinical) studies comparing transparent and metal matrices [27,34,41,42,43,44]. However, in the present study, the conventional nano-hybrid composite (Tetric EvoCeram) achieved significantly better marginal quality when applied using a metal matrix. This finding is in accordance with that of three older trials [26,30,35]. One explanation for this could be that the access cavity to the proximal box was smaller than the size of the light guide tip and thus blocked some of the polymerization light when the metal matrix was used [7,31]. This may have reduced the shrinkage stress of Tetric EvoCeram [45,46,47,48], resulting in fewer marginal gaps [35,49,50,51,52]. Whether this resulted in a lower depth of cure (DC) remains unclear as curing depth was not assessed in the present study. However, the three-sited light-curing technique was performed to achieve the best possible polymerization. Nevertheless, the data of Alshaafi et al. [7] and Price et al. [31] suggest that the depth of cure decreases if the tip of the light guide is partially covered, as might be the case when using a metal matrix. In the case of Tetric EvoCeram Bulk Fill, this effect might be less strong because its more efficient photo initiator makes polymerization of the material less susceptible to reduced radiant exposure while maintaining its physical properties and a sufficient depth of cure [11,15,53,54]. Therefore, we conclude that matrix type does not have such a strong influence on marginal gap formation with this bulk-fill composite. Another explanation for the metal matrix resulting in higher proportions of perfect margins, especially with the conventional nano-hybrid composite (Tetric EvoCeram), might be that its reflective surface may have concentrated the polymerization light within the cavity, thus achieving a better depth of cure in deeper areas of the restoration [26]. With a transparent matrix, on the other hand, more light can exit the tooth and, therefore, less light reaches the deeper areas of the proximal boxes, resulting in poorer curing and poorer marginal quality. This assertion cannot be proven by measurements of the present study and may be subject to future studies. However, the findings by Kays et al. [26] suggest such an effect. Although we performed three-sited light curing after matrix removal to compensate for this, it must be assumed that the adjacent teeth of the artificial dental model and the hard tissue of the sample tooth itself attenuate light intensity when curing the buccal and lingual surfaces [35,55]. Conversely, the bulk-fill material could still be better polymerized than the conventional nano-hybrid composite due to its more efficient photo initiator. Nevertheless, there was a detectable, albeit not statistically significant tendency towards metal matrices resulting in better marginal quality in deeper areas of bulk-fill composite restorations (Figure 5). Finally, this study is also subject to some methodological limitations, which must be discussed. First, artificial aging was achieved by performing 2500 cycles of thermocycling (5–55 °C), which is a rather short treatment period. Furthermore, the specimens were not loaded in a chewing simulator. However, a clear effect of the artificial aging protocol can be seen when looking at the proportions of continuous restoration margins and marginal gaps. This is supported by data from Frankenberger and Tay [56] and Peutzfeldt et al. [57], who observed marginal gap formation using either the same artificial aging protocol [56] or one with even fewer thermal cycles [57]. Nevertheless, it cannot be excluded that Tetric EvoCeram might have performed worse with the metal matrix due to a lower depth of cure, if more thermocycles or mechanical loading had been applied. However, in view of the large effect sizes (ES dCohen) found in the present study, it is likely that a longer artificial aging period would have affected marginal gap formation in all other test groups as well, and that the relations between the test groups would have remained the same. Another limitation of this study is that two materials from the same manufacturer were used. On the other hand, the two materials can be compared well with each other, as they are similar in terms of filler geometry and organic matrix. The results of the present study show that it might be worthwhile to conduct further studies on this research question with other materials. Furthermore, flat matrix tapes were used in the present study because this was the easiest way to seal the cavity in this specific artificial dental model. Although these bands were used in other studies [22,23,24,25,58], there is consensus in literature that pre-contoured matrices (sectional or circumferential) are superior in clinical situations, especially for creating interproximal contacts and profiles [16,17,18,19,22,59]. In preliminary tests of various matrix systems (pre-contoured, sectional and circumferential), we ultimately selected the flat matrix bands as the preferred matrix system for reasons of practicality, i.e., because the focus of the present study was marginal gap formation rather than proximal contact tightness. 5. Conclusions Taking into account the limitations of this study, it can be concluded that metal matrices have a positive influence on the marginal quality of deep class II composite restorations, and that this effect is more pronounced with conventional composite than with bulk-fill composite. Moreover, our findings indicate that bulk-fill composite achieves better marginal quality in deep cavity areas, and that its marginal quality is less sensitive to influence from factors such as light polymerization and the matrix system. Author Contributions Conceptualization, S.S.; Methodology, B.H. and S.S.; Investigation, B.H.; Statistical analysis, B.H. and I.H.; Writing—original draft preparation, S.S. and B.H.; Writing—review and editing, G.K. and R.K.; Project administration, S.S. and G.K. All authors have read and agreed to the published version of the manuscript. Funding This research received no external funding. Institutional Review Board Statement The study was conducted according to the guidelines of the Declaration of Helsinki, and the use of human teeth was approved by the local Ethics Committee of the University of Würzburg (approval no. 15/15) from 9 February 2015. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Cavity design and cavity dimensions (arrows); E = proximal box located within enamel; O = occlusal cavity; D = proximal box cervically located in dentin; CEJ = cementoenamel junction. Figure 2 Artificial dental model with mounted specimen tooth, metal matrix secured in a Tofflemire holder, wooden wedges, and separation rings. Figure 3 Experimental setup with the four experimental groups; red digits represent the order and number of composite increments. NANO = Tetric EvoCeram; BFC = Tetric EvoCeram Bulk Fill; METAL = metal matrix; TRANS = transparent matrix; CEJ = cementoenamel junction. Figure 4 Representative SEM images of the marginal quality outcomes: (a) continuous margin and (b) marginal gap. Figure 5 Mean percentages with standard deviation of continuous margins in enamel (E) and dentin (D) in all groups; NANO = Tetric EvoCeram; BFC = Tetric EvoCeram Bulk Fill; METAL = metal matrix band, TRANS = transparent matrix band. ijerph-19-04961-t001_Table 1 Table 1 Material compositions and physical properties. Tetric EvoCeram a Tetric EvoCeram Bulk Fill b Organical matrix [wt%] Bis-GMA Bis-EMA UDMA 16.8 Bis-GMA Bis-EMA UDMA 19.7 Fillers [wt%] Aluminoborosilicate glass, Ytterbiumtriflourid, Mixed oxides 48.5 Aluminoborosilicate glass, Ytterbiumtriflourid, Mixed oxides 62.5 Prepolymers 34.0 Prepolymers 17.0 Additives <0.8 Additives <1.0 Phototinitiators Lucirin®-TPO Camphorquinone Ivocerin® Lucirin®-TPO Camphorquinone Flexural strength [MPa] 120 120 Flexural modulus [MPa] 10,000 10,000 Water absorption [μg/mm3], 7d 21.2 24.8 Water solubility [μg/mm3], 7d <1.0 <1.0 Radio opacity [% Al] 400 (except for Bleach) 260 200 (Bleach I) 300 (Bleach L, M, XL) Depth of cure [mm] >1.5 4 Translucency [%] 6.5–20.0 14.0–16.0 Vickers hardness HV 0.5/30 [MPa] 580 620 Abbreviations: Bis-GMA, bisphenolglycidyl methacrylate; Bis-EMA, bisphenolglycidyl ethyl-methacrylate; UDMA, urethane dimethacrylate; TPO, Diphenyl (2,4,6-trimethylbenzoyl)-phosphine oxide. Materials compositition according to manufacturer’s scientific documentation from a February and b October 2011. ijerph-19-04961-t002_Table 2 Table 2 Percentages of continuous margins in enamel and dentin (n = 40). Continuous Margins [%] Margin Location Mean SD Median 68%-CI Pw dCohen ES s-m-l Lower Upper CI CI Enamel 46.125 25.962 45.194 20.589 70.414 0.00005 *** 0.78 m Dentin 22.577 24.349 15.674 0 42.839 Total 36.642 20.412 33.366 20.685 58.413 - - - Pw from Wilcoxon test, *** p < 0.001; ES, effect size dCohen; s, small effect (dCohen < 0.5); m, medium effect (dCohen = 0.5–0.8); l, large effect (dCohen > 0.8); CI = confidence interval; NANO = Tetric EvoCeram; BFC= Tetric EvoCeram Bulk Fill; METAL = metal matrix; TRANS = transparent matrix; CT = centripetal technique; BFT = bulk-fill technique. ijerph-19-04961-t003_Table 3 Table 3 Pairwise comparisons of the four test groups (n = 20 each) according to the parameter matrix type and composite material (filling technique) in enamel and dentin; continuous margins [%] (n = 20 per group). Continuous Margins [%] Margin Location Matrix Mean SD Median 68%-CI Pu dCohen ES s-m-l Lower Upper CI CI Enamel METAL 55.491 22.477 54.067 27.141 75.203 0.013 * 0.77 m TRANS 36.759 26.337 37.836 12.359 50.353 Dentin METAL 32.349 25.155 25.902 8.135 50.120 0.0038 ** 0.87 l TRANS 12.804 19.573 5.981 0.000 32.134 Total (enamel + dentin) METAL 46.211 14.912 47.799 33.028 63.291 0.0011 ** 1.052 l TRANS 27.073 20.978 27.553 9.899 33.477 Margin Location Composite (Filling Technique) Mean SD Median 68%-CI Pu dCohen ES s-m-l Lower Upper CI CI Enamel NANO (CT) 46.231 29.005 47.021 16.359 71.126 0.87 - - BFC (BFT) 46.019 23.286 43.859 26.790 66.368 Dentin NANO (CT) 15.716 20.832 6.636 0.000 32.192 0.031 * 0.58 m BFC (BFT) 29.437 26.151 23.865 7.198 44.914 Total (enamel + dentin) NANO (CT) 34.482 21.894 31.806 10.026 59.518 0.56 - - BFC (BFT) 38.802 19.132 33.366 24.686 53.162 Pu from Mann–Whitney U-test, * p < 0.05, ** p < 0.01; ES, effect size dCohen; s, small effect (dCohen < 0.5); m, medium effect (dCohen = 0.5–0.8); l, large effect (dCohen > 0.8); CI = confidence interval; NANO = nanohybrid composite (Tetric EvoCeram); BFC = bulk-fill composite (Tetric EvoCeram Bulk Fill); METAL = metal matrix; TRANS = transparent matrix; CT = centripetal technique; BFT = bulk-fill technique. ijerph-19-04961-t004_Table 4 Table 4 Percentage of continuous margins [%] by margin location (enamel or dentin), composite material and matrix type (n = 10 per group). Groups Continuous Margins [%] Margin Location Composite–Matrix Mean SD Median 68%-CI Pu dCohen ES s-m-l Lower Upper CI CI Enamel NANO–METAL 65.935 22.724 66.863 48.644 86.094 0.0017 ** 1.844 l NANO–TRANS 26.526 19.921 29.178 2.768 46.963 Dentin NANO–METAL 26.644 23.438 22.863 3.438 47.188 0.021 * 1.212 l NANO–TRANS 4.789 10.072 0.000 0.000 6.971 Total (E + D) NANO–METAL 50.857 15.947 51.389 36.907 66.868 0.0010 ** 2.270 l NANO–TRANS 18.107 12.720 22.809 2.451 29.456 Enamel BFC–METAL 45.047 17.545 46.335 26.909 61.575 0.91 - - BFC–TRANS 46.991 28.894 43.812 23.423 74.805 Dentin BFC–METAL 38.054 26.723 36.196 15.584 52.186 0.064 - - BFC–TRANS 20.819 23.760 10.797 4.073 34.958 Total (E + D) BFC–METAL 41.564 12.929 41.039 29.364 51.575 0.27 - - BFC–TRANS 36.040 24.261 28.686 16.419 56.898 Pu from Mann–Whitney U-test, * p < 0.05, ** p < 0.01, ES, effect size dCohen; s, small effect (dCohen < 0.5); m, medium effect (dCohen = 0.5–0.8); l, large effect (dCohen > 0.8); CI = confidence interval; NANO = Tetric EvoCeram applied in centripetal technique; BFC = Tetric EvoCeram Bulk Fill applied in bulk-fill technique; METAL = metal matrix; TRANS = transparent matrix; E = enamel; D = dentin. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Schierz O. Baba K. Fueki K. Functional oral health-related quality of life impact: A systematic review in populations with tooth loss J. Oral Rehabil. 2021 48 256 270 10.1111/joor.12984 32333415 2. Carvalho J.C. Mestrinho H.D. Stevens S. van Wijk A.J. Do oral health conditions adversely impact young adults? Caries Res. 2015 49 266 274 10.1159/000375377 25832802 3. Van Ende A. De Munck J. Lise D.P. Van Meerbeek B. Bulk-Fill Composites: A Review of the Current Literature J. Adhes. Dent. 2017 19 95 109 10.3290/j.jad.a38141 28443833 4. Cidreira Boaro L.C. Pereira Lopes D. de Souza A.S.C. Lie Nakano E. Ayala Perez M.D. Pfeifer C.S. Goncalves F. Clinical performance and chemical-physical properties of bulk fill composites resin—A systematic review and meta-analysis Dent. Mater. 2019 35 e249 e264 10.1016/j.dental.2019.07.007 31421957 5. Heck K. Manhart J. Hickel R. Diegritz C. Clinical evaluation of the bulk fill composite QuiXfil in molar class I and II cavities: 10-year results of a RCT Dent. Mater. 2018 34 e138 e147 10.1016/j.dental.2018.03.023 29636239 6. AlQahtani M.Q. Michaud P.L. Sullivan B. Labrie D. AlShaafi M.M. Price R.B. Effect of High Irradiance on Depth of Cure of a Conventional and a Bulk Fill Resin-based Composite Oper. Dent. 2015 40 662 672 10.2341/14-244-L 26237638 7. AlShaafi M.M. AlQussier A. AlQahtani M.Q. Price R.B. Effect of Mold Type and Diameter on the Depth of Cure of Three Resin-Based Composites Oper. Dent. 2018 43 520 529 10.2341/17-122-L 29570024 8. El-Damanhoury H. Platt J. Polymerization shrinkage stress kinetics and related properties of bulk-fill resin composites Oper. Dent. 2014 39 374 382 10.2341/13-017-L 23865582 9. Catel Y. Angermann J. Fassler P. Fischer U. Schnur T. Moszner N. High refractive index monofunctional monomers as promising diluents for dental composites Dent. Mater. 2021 37 351 358 10.1016/j.dental.2020.11.029 33357987 10. Kowalska A. Sokolowski J. Bociong K. The Photoinitiators Used in Resin Based Dental Composite-A Review and Future Perspectives Polymers 2021 13 470 10.3390/polym13030470 33540697 11. Moszner N. Fischer U.K. Ganster B. Liska R. Rheinberger V. Benzoyl germanium derivatives as novel visible light photoinitiators for dental materials Dent. Mater. 2008 24 901 907 10.1016/j.dental.2007.11.004 18155290 12. Van Ende A. De Munck J. Van Landuyt K.L. Poitevin A. Peumans M. Van Meerbeek B. Bulk-filling of high C-factor posterior cavities: Effect on adhesion to cavity-bottom dentin Dent. Mater. Off. Publ. Acad. Dent. Mater. 2013 29 269 277 10.1016/j.dental.2012.11.002 23228335 13. Benetti A.R. Havndrup-Pedersen C. Honore D. Pedersen M.K. Pallesen U. Bulk-fill resin composites: Polymerization contraction, depth of cure, and gap formation Oper. Dent. 2015 40 190 200 10.2341/13-324-L 25216940 14. Loguercio A.D. Rezende M. Gutierrez M.F. Costa T.F. Armas-Vega A. Reis A. Randomized 36-month follow-up of posterior bulk-filled resin composite restorations J. Dent. 2019 85 93 102 10.1016/j.jdent.2019.05.018 31100332 15. Bucuta S. Ilie N. Light transmittance and micro-mechanical properties of bulk fill vs. conventional resin based composites Clin. Oral Investig. 2014 18 1991 2000 10.1007/s00784-013-1177-y 16. Scholtanus J.D. Ozcan M. Clinical longevity of extensive direct composite restorations in amalgam replacement: Up to 3.5 years follow-up J. Dent. 2014 42 1404 1410 10.1016/j.jdent.2014.06.008 24994619 17. Loomans B.A. Opdam N.J. Roeters F.J. Bronkhorst E.M. Burgersdijk R.C. Dorfer C.E. A randomized clinical trial on proximal contacts of posterior composites J. Dent. 2006 34 292 297 10.1016/j.jdent.2005.07.008 16157438 18. Kampouropoulos D. Paximada C. Loukidis M. Kakaboura A. The influence of matrix type on the proximal contact in Class II resin composite restorations Oper. Dent. 2010 35 454 462 10.2341/09-272-L 20672731 19. Durr E.S. Ahmad M.Z. Gaikwad R.N. Arjumand B. Comparison of two different matrix band systems in restoring two surface cavities in posterior teeth done by senior undergraduate students at Qassim University, Saudi Arabia: A randomized controlled clinical trial Indian J. Dent. Res. 2018 29 459 464 10.4103/ijdr.IJDR_26_17 30127197 20. Loomans B.A. Opdam N.J. Roeters F.J. Bronkhorst E.M. Huysmans M.C. Restoration techniques and marginal overhang in Class II composite resin restorations J. Dent. 2009 37 712 717 10.1016/j.jdent.2009.05.025 19524348 21. Loomans B.A. Opdam N.J. Roeters F.J. Huysmans M.C. Proximal marginal overhang of composite restorations in relation to placement technique of separation rings Oper. Dent. 2012 37 21 27 10.2341/10-286-L 21942287 22. Shaalan O.O. Evaluation of Matrix Band Systems for Posterior Proximal Restorations among Egyptian Dentists: A Cross-Sectional Survey Acta Stomatol. Croat. 2020 54 392 400 10.15644/asc54/4/6 33642603 23. Lindberg A. van Dijken J.W. Lindberg M. Nine-year evaluation of a polyacid-modified resin composite/resin composite open sandwich technique in Class II cavities J. Dent. 2007 35 124 129 10.1016/j.jdent.2006.06.003 16956709 24. Coelho-De-Souza F.H. Camargo J.C. Beskow T. Balestrin M.D. Klein-Junior C.A. Demarco F.F. A randomized double-blind clinical trial of posterior composite restorations with or without bevel: 1-year follow-up J. Appl. Oral. Sci. 2012 20 174 179 10.1590/S1678-77572012000200009 22666833 25. Coelho-de-Souza F.H. Klein-Junior C.A. Camargo J.C. Beskow T. Balestrin M.D. Demarco F.F. Double-blind randomized clinical trial of posterior composite restorations with or without bevel: 6-month follow-up J. Contemp. Dent. Pract. 2010 11 1 8 10.5005/jcdp-11-2-1 26. Kays B.T. Sneed W.D. Nuckles D.B. Microhardness of Class II composite resin restorations with different matrices and light positions J. Prosthet. Dent. 1991 65 487 490 10.1016/0022-3913(91)90285-5 2066884 27. Demarco F.F. Pereira-Cenci T. de Almeida Andre D. de Sousa Barbosa R.P. Piva E. Cenci M.S. Effects of metallic or translucent matrices for Class II composite restorations: 4-year clinical follow-up findings Clin. Oral Investig. 2011 15 39 47 10.1007/s00784-009-0362-5 20049495 28. Mullejans R. Badawi M.O. Raab W.H. Lang H. An in vitro comparison of metal and transparent matrices used for bonded class II resin composite restorations Oper. Dent. 2003 28 122 126 12670066 29. Chan D.C. Modified matrix adaptation for sub-gingival Class II amalgam restorations Oper. Dent. 2003 28 469 472 12877435 30. Dietrich T. Losche A.C. Losche G.M. Roulet J.F. Marginal adaptation of direct composite and sandwich restorations in Class II cavities with cervical margins in dentine J. Dent. 1999 27 119 128 10.1016/S0300-5712(98)00035-9 10071469 31. Price R.B. Rueggeberg F.A. Harlow J. Sullivan B. Effect of mold type, diameter, and uncured composite removal method on depth of cure Clin. Oral Investig. 2016 20 1699 1707 10.1007/s00784-015-1672-4 26631060 32. Goodchild J.H. Class II composite placement is difficult! Solutions to help overcome the clinical challenges Dent. Today 2013 32 116 117 33. Lutz F. Krejci I. Luescher B. Oldenburg T.R. Improved proximal margin adaptation of Class II composite resin restorations by use of light-reflecting wedges Quintessence Int. 1986 17 659 664 3466196 34. Nguyen D.P. Motyka N.C. Meyers E.J. Vandewalle K.S. Depth of cure of proximal composite resin restorations using a new perforated metal matrix Gen. Dent. 2018 66 68 74 29714704 35. Lösche G.M. Marginal adaptation of Class II composite fillings: Guided polymerization vs reduced light intensity J. Adhes. Dent. 1999 1 31 39 11725684 36. Lima R.B.W. Troconis C.C.M. Moreno M.B.P. Murillo-Gomez F. De Goes M.F. Depth of cure of bulk fill resin composites: A systematic review J. Esthet. Restor. Dent. 2018 30 492 501 10.1111/jerd.12394 30375146 37. Cerdan F. Ceballos L. Fuentes M.V. Quality of approximal surfaces of posterior restorations in primary molars J. Oral Sci. 2021 63 347 351 10.2334/josnusd.21-0264 34511588 38. Loomans B.A. Opdam N.J. Roeters J.F. Bronkhorst E.M. Plasschaert A.J. Influence of composite resin consistency and placement technique on proximal contact tightness of Class II restorations J. Adhes. Dent. 2006 8 305 310 17080878 39. Saber M.H. El-Badrawy W. Loomans B.A. Ahmed D.R. Dorfer C.E. El Zohairy A. Creating tight proximal contacts for MOD resin composite restorations Oper. Dent. 2011 36 304 310 10.2341/10-210-L 21740239 40. Souza-Junior E.J. de Souza-Regis M.R. Alonso R.C. de Freitas A.P. Sinhoreti M.A. Cunha L.G. Effect of the curing method and composite volume on marginal and internal adaptation of composite restoratives Oper. Dent. 2011 36 231 238 10.2341/10-107-L 21702671 41. Szep S. Frank H. Kenzel B. Gerhardt T. Heidemann D. Comparative study of composite resin placement: Centripetal buildup versus incremental technique Pract. Proced. Aesthet. Dent. 2001 13 243 252 243–250; discussion 252 11360771 42. Hofmann N. Hunecke A. Influence of curing methods and matrix type on the marginal seal of class II resin-based composite restorations in vitro Oper. Dent. 2006 31 97 105 10.2341/04-194 16536200 43. Cenci M.S. Demarco F.F. Pereira C.L. Lund R.G. de Carvalho R.M. One-year comparison of metallic and translucent matrices in Class II composite resin restorations Am. J. Dent. 2007 20 41 45 17380807 44. Demarco F.F. Cenci M.S. Lima F.G. Donassollo T.A. Andre Dde A. Leida F.L. Class II composite restorations with metallic and translucent matrices: 2-year follow-up findings J. Dent. 2007 35 231 237 10.1016/j.jdent.2006.07.011 17034926 45. Mehl A. Hickel R. Kunzelmann K.H. Physical properties and gap formation of light-cured composites with and without ‘softstart-polymerization’ J. Dent. 1997 25 321 330 10.1016/S0300-5712(96)00044-9 9175364 46. Lee C.H. Ferracane J. Lee I.B. Effect of pulse width modulation-controlled LED light on the polymerization of dental composites Dent. Mater. 2018 34 1836 1845 10.1016/j.dental.2018.10.003 30509382 47. Piccioni M.A. Baratto-Filho F. Kuga M.C. Morais E.C. Campos E.A. Cuspal movement related to different polymerization protocols J. Contemp. Dent. Pract. 2014 15 26 28 10.5005/jp-journals-10024-1482 24939260 48. Uno S. Asmussen E. Marginal adaptation of a restorative resin polymerized at reduced rate Scand. J. Dent. Res. 1991 99 440 444 10.1111/j.1600-0722.1991.tb01052.x 1836644 49. Yoshikawa T. Burrow M.F. Tagami J. A light curing method for improving marginal sealing and cavity wall adaptation of resin composite restorations Dent. Mater. 2001 17 359 366 10.1016/S0109-5641(00)00095-6 11356214 50. Yoshikawa T. Nakaoki Y. Takada T. Burrow M.F. Tagami J. Effect of light-curing method and irradiation time on marginal sealing and cavity wall adaptation of resin composite restorations Am. J. Dent. 2003 16 63A 67A 51. Feilzer A.J. Dooren L.H. de Gee A.J. Davidson C.L. Influence of light intensity on polymerization shrinkage and integrity of restoration-cavity interface Eur. J. Oral Sci. 1995 103 322 326 10.1111/j.1600-0722.1995.tb00033.x 8521124 52. Unterbrink G.L. Muessner R. Influence of light intensity on two restorative systems J. Dent. 1995 23 183 189 10.1016/0300-5712(95)93577-O 7782531 53. Ganster B. Fischer U.K. Moszner N. Liska R. New Photocleavable Structures. Diacylgermane-Based Photoinitiators for Visible Light Curing Macromolecules 2008 41 2394 2400 10.1021/ma702418q 54. Ilie N. Bucuta S. Draenert M. Bulk-fill resin-based composites: An in vitro assessment of their mechanical performance Oper. Dent. 2013 38 618 625 10.2341/12-395-L 23570302 55. Sartori N. Knezevic A. Peruchi L.D. Phark J.H. Duarte S. Jr. Effects of Light Attenuation through Dental Tissues on Cure Depth of Composite Resins Acta Stomatol. Croat. 2019 53 95 105 10.15644/asc53/2/1 31341317 56. Frankenberger R. Tay F.R. Self-etch vs etch-and-rinse adhesives: Effect of thermo-mechanical fatigue loading on marginal quality of bonded resin composite restorations Dent. Mater. 2005 21 397 412 10.1016/j.dental.2004.07.005 15826696 57. Peutzfeldt A. Muhlebach S. Lussi A. Flury S. Marginal Gap Formation in Approximal “Bulk Fill” Resin Composite Restorations after Artificial Ageing Oper. Dent. 2018 43 180 189 10.2341/17-068-L 29148914 58. Alves dos Santos M.P. Luiz R.R. Maia L.C. Randomised trial of resin-based restorations in Class I and Class II beveled preparations in primary molars: 48-month results J. Dent. 2010 38 451 459 10.1016/j.jdent.2010.02.004 20188783 59. Loomans B.A. Roeters F.J. Opdam N.J. Kuijs R.H. The effect of proximal contour on marginal ridge fracture of Class II composite resin restorations J. Dent. 2008 36 828 832 10.1016/j.jdent.2008.06.001 18621458

PMC009xxxxxx/PMC9099622.txt

==== Front Int J Mol Sci Int J Mol Sci ijms International Journal of Molecular Sciences 1422-0067 MDPI 10.3390/ijms23094481 ijms-23-04481 Article The Transcriptome and Metabolome Reveal the Potential Mechanism of Lodging Resistance in Intergeneric Hybrids between Brassica napus and Capsella bursa-pastoris Zhang Libin 12† Miao Liyun 13† He Jianjie 1 Li Huaixin 1 Li Maoteng 1* Fernando Dilantha Academic Editor Voll Lars Matthias Academic Editor 1 College of Life Science and Technology, Huazhong University of Science and Technology, Wuhan 430074, China; libinzhang@hust.edu.cn (L.Z.); miaoliyun428@sxtcm.edu.cn (L.M.); jianjie_he@hust.edu.cn (J.H.); huaixin_lll@hust.edu.cn (H.L.) 2 Hubei Bioinformatics & Molecular Imaging Key Laboratory, Department of Bioinformatics and Systems Biology, College of Life Science and Technology, Huazhong University of Science and Technology, Wuhan 430074, China 3 College of Basic Medical Sciences, Shanxi University of Traditional Chinese Medicine, Jinzhong 030619, China * Correspondence: limaoteng426@hust.edu.cn † These authors have contributed equally to this work. 19 4 2022 5 2022 23 9 448118 3 2022 17 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Lodging is one of the main reasons for the reduction in seed yield and is the limitation of mechanized harvesting in B. napus. The dissection of the regulatory mechanism of lodging resistance is an important goal in B. napus. In this study, the lodging resistant B. napus line, YG689, derived from the hybridization between B. napus cv. Zhongyou 821 (ZY821) and Capsella bursa-pastoris, was used to dissect the regulation mechanism of hard stem formation by integrating anatomical structure, transcriptome and metabolome analyses. It was shown that the lignocellulose content of YG689 is higher than that of ZY821, and some differentially expressed genes (DEGs) involved in the lignocellulose synthesis pathway were revealed by transcriptome analyses. Meanwhile, GC–TOF–MS and UPLC–QTOF–MS identified 40, 54, and 31 differential metabolites in the bolting stage, first flower stage, and the final flower stage. The differential accumulation of these metabolites might be associated with the lignocellulose biosynthesis in B. napus. Finally, some important genes that regulate the metabolic pathway of lignocellulose biosynthesis, such as BnaA02g18920D, BnaA10g15590D, BnaC05g48040D, and NewGene_216 were identified in B. napus through the combination of transcriptomics and metabolomics data. The present results explored the potential regulatory mechanism of lignocellulose biosynthesis, which provided a new clue for the breeding of B. napus with lodging resistance in the future. Brassica napus lignocellulose transcriptome differentially expressed genes metabolome ==== Body pmc1. Introduction B. napus is one of the most important oil crops in the world, and the stem tissue is an advantageous energy material for the development of biodiesel. Crop lodging was mainly induced because of stem fall caused by strong winds and other factors, as well as root fall caused by poor root anchorage strength [1]. Stem strength, i.e., the bending and/or breaking strength of the culm, is important for stem lodging resistance [2]. The stem strength of crops is primarily determined by plant architecture (morphological traits and anatomical structure) [3]. In particular, the anatomical structure is a consequence of plant growth and development at the cellular level, such as cell division, cell growth, and cell spatial arrangement, and it is closely related to environmental factors. For example, the stem with a large diameter and a thick cell wall could increase the bending strength and lodging resistance of wheat, barley, and rice stems [2,4,5]. Plant height is considered to be an important morphological feature affecting crop lodging, but it is not the main factor determining crop lodging [6,7]. Lignocellulose is a major component of the mechanical strength of crop stems [8], and its content is related to the lodging resistance of crops [3,5,9]. Lignin and cellulose determine the mechanical strength of the stem [10], while the accumulation of starch could also increase the flexural strength and hardness of the stem [11]. Peng et al. [12] and Berry et al. [1] found that the increase in lignin and hemicellulose content could enhance the stalk strength and lodging resistance of wheat. Metabolomics provides a powerful way to study the metabolic phenotype of Brassica species [13,14,15,16]. Tan et al. [13] identified the dynamic metabolic changes from both seeds and silique walls that occur during oil accumulation by using gas chromatography coupled with mass spectrometry (GC–MS) and demonstrated that the oil content was independent of leaf photosynthesis and phloem transport during oil accumulation, but it required the metabolic influx from the silique wall. Kortesniemi et al. [16] investigated NMR metabolomics of ripened and developing oilseed rape (B. napus) and turnip rape (B. rapa) and found differences in the major lipids and the minor metabolites between the two species. High-throughput sequencing is one of the important tools used to study important agronomic traits and gene expression regulation in B. napus [17,18,19,20,21]. For example, Li et al. [18] identified 71 candidate genes with stem lodging-related traits through the genome-wide association study (GWAS) method, and constructed a gene co-expression network based on transcriptome sequencing, which revealed the genetic basis of stem lodging traits in B. napus. Kawakatsu et al. [21] reported the phenotypic and transcriptomic landscape of 61 rice (Oryza sativa) accessions with highly diverse below-ground traits grown in an upland field, which provided a useful resource for understanding the genomic and transcriptomic bases of phenotypic variation under upland field conditions. The metabolomics analysis is the quantitative detection of all metabolites and their biochemical states in specific organisms or tissues. Moreover, high-throughput sequencing has been combined with metabolomics data to explore the complex regulatory metabolism networks [22,23,24,25,26,27,28,29]. For example, Guo et al. [29] characterized the acclimation of maize (Zea mays B73) to drought and cold stresses using physiological measurements and comparative transcriptomics combined with metabolomics during the stress treatments and recovery stages, which presented a model showing that the plant response to the combined stress is multi-faceted and revealed an ABA-dependent maize acclimation mechanism to the stress combination. However, there is a lack of a systematic investigation of the lignocellulose-related gene expression networks through the combination of different omics technologies. Interspecific and intergeneric hybridization are widely used to create the new germplasm with valuable characteristics. Zhongyou 821 (ZY821) is a winter rapeseed variety with a high yield that is multi resistant with a wide adaptability, which was cultivated in China for many years [30]. Park [31] found that the content of erucic acid in the rapeseed oil of C. bursa-pastoris was significantly lower than that of other cruciferous plants. Moreover, C. bursa-pastoris displayed high resistance to black spot disease and Sclerotinia sclerotiorum [32,33] and could endure cold, salt, and drought stress [34], which was probably related to its lignified or wooden stems. Furthermore, the hybridizations between B. napus (2n = 38; female) and C. bursa-pastoris (male, 2n = 32) were performed in the fields by hand emasculation and pollination [33]. The stably inherited hybrids with the expected chromosome number and normal meiosis behaviors were screened. Coincidently, a hard stem material, YG689, was identified in its offspring population. YG689 was reported to contain high lignin content, a certain resistance to Sclerotinia sclerotiorum, and the obvious characteristics of its C. bursa-pastoris origin, such as wooden stems [33]. In addition, Shen et al. [35] obtained one doubled haploid (DH) population obtained from a cross between Y689 and B. napus var. Westar, and the QTL for the plant height (PH), branch initiation height (BIH), stem diameter (SD), and flowering time (FT) were obtained. Although some studies had been carried out for the YG689, the molecular regulatory mechanism and metabolic pathway of lodging resistance still remains elusive in YG689. In the present study, the regulatory mechanism for the stem lodging resistance of YG689 was studied by combining the anatomical structure, comparative transcriptomic and metabolomic analyses. The present results shed light on the molecular regulatory pathways in lignocellulose biosynthesis, which provides a new clue for the breeding of B. napus with lodging resistance in the future. 2. Results 2.1. The Anatomical Structure and Lignocellulose Analysis of ZY821 and YG689 The stems of ZY821 and YG689 in the early flowering and maturation stages were used for the morphological and anatomical analyses. As shown in Figure 1C, the stem strength and dry weight of YG689 were significantly higher than that of ZY821. These results indicate that the YG689 had stronger stem lodging resistance than that of ZY821. To explore the difference in stem anatomical structure, the stems of YG689 and ZY821 were evenly divided into five sections from bottom to top (Figure 1B). It was revealed that the lateral distribution of lignin in YG689 was obviously wider than that of ZY821 (Figure 2). The stem anatomical structures between YG689 and ZY821 at the maturation stage were also compared, and similar results were obtained (Figure S1). To further analyze the difference in the lignocellulose components in stems between YG689 and ZY821, the content of lignocellulose in the YG689 and ZY821 stems at the seedling, bolting and budding, early flowering, terminal flowering, and maturation stages were measured. It was revealed that the lignin content of YG689 was significantly higher than that of ZY821 at all stages (p < 0.05), and the cellulose content of YG689 was significantly higher than that of ZY821 at the bolting and budding stage and the terminal flowering stage (p < 0.01) (Figure 1D). The content of hemicellulose was also significantly higher in YG689 than that of ZY821 (p < 0.05) except at the seedling stage. The lignin monomers in YG689 and ZY821 stems at different developmental stages were further analyzed by using the GC–MS. It was found that the G monolignol was the main monomer, and the content in YG689 was higher than that of ZY821 at the seedling stage (Table S1). Further analyses showed that the S monolignol also existed, but the content of the S monolignol was lower than that of the G monolignol in both YG689 and ZY821 (Table S1). 2.2. Transcriptome Analysis of YG689 and ZY821 The transcriptome of YG689 and ZY821 for the stem at the initial flowering stage were analyzed by RNA-Seq technology, where 99,472,340 (SRA number: SRX1522099) and 134,015,130 (SRA number: SRX1142564) clean reads, after removing adaptor sequences and low-quality reads, were generated, respectively. It was revealed that 88.35% (87,884,725) of YG689 and 90.60% (121,415,423) of ZY821 clean reads were mapped to the B. napus genome, among which 79.08% (78,663,759) and 61.44% (82,335,196) of clean reads were uniquely mapped to the B. napus genome. The trinity program was then applied for the de novo assembly of the clean reads, and 2697 novel genes were identified by comparing the assembled transcripts with the genome annotation information of B. napus. The genes were then annotated to the COG, NR, Swiss-Prot, GO, and KEGG databases (Table S2). Among these novel genes, 339 were annotated by COG (12.57%), 1509 were annotated by GO (55.95%), 418 were annotated by KEGG (15.50%), 1348 (49.98%) were annotated with Swiss-Prot, and 2284 (84.69%) were annotated with NR. To reveal the global differential gene expression profiles between YG689 and ZY821, the DEGs, by setting the gene expression level of ZY821 as a control, were analyzed, and the up- or down-regulated genes in YG689 were observed. In total, 8644 DEGs (3818 up-regulated and 4826 down-regulated genes) were identified (Figure 3A). Thirteen DEGs were randomly selected for RT-qPCR validation. The majority of the expression trends of selected DEGs were consistent with the RNA-seq results (Figure 3B), which indicated the reliability of transcriptome sequencing results. The DEGs between YG689 and ZY821 were further annotated with Gene Ontology terms, which could be classified into three categories and 51 subcategories (Figure S2). For example, the “cellular process”, “metabolic process”, and “single-organism process” in the biological process were enriched in most of the DEGs. The DEGs were also analyzed by the Cytoscape Enrichment Map based on the GO annotation, and a total of 4873 DEGs were enriched in the biological process category (Figure S3). The overlapping terms between up-regulated and down-regulated genes were “metabolic process”, “response to stimulus”, “biological regulation”, “developmental process”, “localization” and “cellular process”. Importantly, we found that the up-regulated DEGs in YG689 were mainly assigned to the “rhythmic process”, “cellular component biogenesis”, and “signaling” terms, which suggested the activities of the cell metabolism and signal transductions were much more active in YG689 than in ZY821. To better understand the biological functions of DEGs between YG689 and ZY821, the DEGs were further assigned for the KEGG analysis (Supplementary Dataset S1 with the listed top-20 pathways) and some important pathways associated with lignocellulose synthesis were found, including “arginine and proline metabolism”, “cysteine and methionine metabolism”, “carotenoid biosynthesis”, and “glutathione metabolism”. Moreover, the DEGs involved in these KEGG pathways were localized on the B. napus genome (Figure 4). Among them, 30 DEGs were found to be significantly associated with lignocellulose synthesis (Table S3). For example, GH9B17, TPS1, UGE4, and GAUT13 were up-regulated in the “starch and sucrose metabolism” pathway in YG689. (Figure 5). Twelve DEGs related to the lignocellulose synthesis were randomly selected for further expression analyses in the stem of YG689, ZY821, and TN070 (another B. napus line with a soft stem) in bolting, early flowering, and final flowering stages. It was shown that nine DEGs were significantly up-regulated in the early-flowering stage of YG689, and the up-regulation of BnaA09g42650 persisted to the final flowering stage and reached the highest value (Figure 3C). For example, BnaA09g42650 is homologous to PRX17 in Arabidopsis, and the transcription factor AGL15 participates in the transition of vegetative growth to reproductive growth, as well as the formation of lignification tissues, by directly regulating PRX17. Therefore, we speculated that the hybridization between C. bursa-pastoris and ZY821 leads to the differential expression of lignin synthesis-related genes, which induced the stem traits of YG689. Further analyses revealed that 624 DEGs could not be compared to the B. napus reference genome, 108 of which could be compared to the C. bursa-pastoris genome. Twenty-three DEGs have more than 40% of the homology rate and six DEGs have more than 80% of the homology rate. The 23 DEGs were subjected to KEGG analysis and five new genes (NewGene_6090, NewGene_3494, NewGene_1078, NewGene_2975, and NewGene_3483) were annotated. NewGene_6090 is homologous to Carubv10019973m (Vacuolar ATP synthase subunit A, VHA-A) of C. bursa-pastoris and participated in the oxidative phosphorylation pathway; NewGene_3494 is homologous to Carubv10003976m (RNA polymerase II large subunit, NRPB1) of C. bursa-pastoris and participates in the purine or pyrimidine metabolism pathway; NewGene_1078 is homologous to Carubv10011493m (2-oxoglutarate, 2OG) of C. bursa-pastoris and is involved in cysteine and methionine metabolism; NewGene_2975 is homologous to Carubv10001560m (syntaxin of plants 132, SYP132) of C. bursa-pastoris and is involved in SNARE interactions in the vesicular transport pathway; and NewGene_3483 is homologous to Carubv10008883m (beta glucosidase 40, BGLU40) of C. bursa-pastoris and is involved in starch and sucrose metabolism or phenylpropane biosynthesis. Notably, another 516 DEGs could not be compared to the B. napus genome and the C. bursa-pastoris genome, which indicates that these DEGs might be novel genes that derived from the chromosome exchange between B. napus and C. bursa-pastoris. 2.3. Metabolome Analysis of Stem of YG689 and ZY821 To reveal the differential metabolites in the stem of YG689 and ZY821, the metabolomics of the stems of YG689 and ZY821 in the bolting, early flowering, and terminal flowering stages were also analyzed by using the GC–TOF–MS and UPLC–QTOF–MS techniques. A total of 449 peaks were obtained by the GC–TOF–MS analysis, and the PCA score map of the metabolic spectrum showed that all the YG689 and ZY821 samples were in the 95% confidence interval (Figure S4). Twenty differential metabolites between YG689 and ZY821 at the bolting stage were identified (Table S4), and the content of 19 metabolites in YG689 were lower than that of ZY821, which included saccharides (sucrose, L-threose, ribulose-5-phosphate, D- galactose, raffinose, and tagatose), amino acids and derivatives (L-valine, β-alanine and threo-beta-hyrdoxyaspartate) and other products of plant metabolism in B. napus. Twenty-six differential metabolites were identified in the early flowering stage (Table S4), and the content of eight metabolites in YG689 is higher than that of ZY821, which included maleamate, elaidic acid, succinic acid, malonic acid, L-valine, 5, 6-dihydrouracil, erythrose and mannitol. Seven differential metabolites between YG689 and ZY821 at the final flowering stage were obtained (Table S4), and four metabolites in YG689 were higher than of ZY821, including glutaconic acid, inosine, nornicotine, and leucrose. The UPLC–QTOF–MS was also used to detect the differential metabolites between YG689 and ZY821, and 25, 36, and 39 differential metabolites were observed in the bolting, early flowering, and final flowering stages, respectively (Table S5). In the bolting stage, the content of proline was increased in YG689, while 5-L-glutamyl-L-alanine, deoxyadenosine, and kaempferol were decreased in YG689 (Table S5). In the early flowering stage, the content of L-threonate, aconitate, and (S)-2-hydroxyglutarate increased in YG689, while cytidine, guanosine, and cytidine 2′, 3′-cyclic phosphate decreased in YG689 (Table S5). In the final flowering stage, the content of beta-D-fructose 6-phosphate, D-gluconate and L-isoleucine increased in YG689, while the content of N,N′-Diacetylchitobiose, alpha-linolenic acid, L-leucine and succinate decreased in YG689 (Table S5). Furthermore, the metabolites identified by GC–TOF–MS and UPLC–QTOF–MS were integrated together, and 40, 54, and 31 differential metabolites between YG689 and ZY821 in the bolting, initial flowering, and final flowering stages were identified, respectively. These differential metabolites were involved in nine important metabolic pathways (Table S6), including glycolysis; the TCA cycle; the pentose phosphate pathway; fructose and mannose metabolism; galactose metabolism; phenylalanine, tyrosine, and tryptophan synthesis; starch and sucrose metabolism; glycine, serine, and threonine metabolism; and cysteine and methionine metabolism. For example, cis-aconitate and succinate were common differential metabolites in three developmental stages and were involved in TCA cycle pathways. Finally, the related metabolic pathway of lignocellulose synthesis was constructed by integrating DEGs and differential metabolites in the stem of YG689 and ZY821 (Figure 6), which consisted of 14 distinct metabolites and four important DEGs that might regulate the lignocellulose synthesis. 2.4. The Validation of DEGs Regulating the Metabolic Pathway of Lignocellulose Synthesis The level of secondary metabolites is associated with gene expression, protein modification, and the response to environmental changes in the growth and development process of plant. A total of four DEGs (BnaA02g18920D, BnaA10g15590D, NewGene_216, and BnaC05g48040D) were identified to be the key enzyme genes and were involved in the regulation of lignocellulose synthesis (Figure 6). The RT-qPCR validation results were shown in Figure S5. For example, the expression of BnaA02g18920D (TPS1) increased, which may facilitate the production of Trehalose 6-phosphate (T6P) and the development of a hard stem in YG689. In addition, BnaA10g15590D is homologous with the ASP2 gene of Arabidopsis and participates in phenylalanine biosynthesis. BnaA10g15590D was up-regulated in YG689, which may stimulate the level of phenylalanine biosynthesis in the metabolic pathway of lignocellulose. Nevertheless, BnaC05g48040D (MS2) was down-regulated in YG689, which indicates the decrease in the phenylalanine and L-methionine levels in the metabolic pathway of lignocellulose. Meanwhile, some new genes that were expressed in YG689 derived from the C. bursa-pastoris genome were also identified. For example, NewGene_216 was predicted to be trehalose-phosphate synthase 7, which catalyzes UDP glucose to form trehalose-6-phosphate and participates in the synthesis of trehalose. Interestingly, although the expression of NewGene_216 increased, the content of trehalose decreased in YG689. We guessed that the content of trehalose was transformed by other pathways. Therefore, we speculated that these DEGs might affect the lignocellulose synthesis in the stem tissues of YG689, thus affecting the stem hardness in B. napus. 3. Discussion Distant hybridization is an important way to create new germplasm resources [36]. For example, Liu et al. [37] transferred the clubroot-resistance genes from clubroot-resistant Chinese cabbages to B. napus by distant hybridization combined with embryo rescue. Gong et al. [38] obtained the hybrids with powdery mildew (PM) resistance through hybridization between the B. napus cultivar ‘Zhongshuang11′ and the PM-resistant B. carinata. Lodging is one of the main reasons that influences mechanized harvesting, which results in yield reduction [39]. Long et al. [40] reported that Oryza longistaminata has a strong stem and a high biomass productivity, and 12, 11, and 3 QTLs for the stem diameter (SD), stem length (SL), and breaking strength (BS), respectively, were obtained in the mapping population that was obtained between line 93–11 and O. longistaminata. In the previous studies, a germplasm of YG689 with high fertility and lodging resistance was obtained in the offspring that were derived from B. napus and C. bursa-pastoris, and further analyses revealed that YG689 had a high content of lignin [33]. Intensive research on the lodging resistance mechanism of YG689 is helpful to cultivate the new B. napus lines with lodging resistance characteristics. Transcriptomic analyses have been widely used to find the candidate genes or study the regulatory mechanisms of important agronomic traits. For example, four candidate genes that regulate the lignin content were identified by the integration of GWAS and transcriptome sequencing, which provides insight into the genetic control of lodging and lignin in B. napus [17]. The unique QTLs for stem lodging-related traits (plant height, branch initiation height, and stem diameter) were found by Shen et al. [35] in B. napus, and some genes (including ESK1 and CESA6) involved in lodging resistance have been identified [18]. Li et al. [41] reported that the CESA9 conserved-site mutation could affect its association with the CESA complexes and cause the low-DP (degree of polymerization) cellulose synthesis, which significantly enhanced plant lodging resistance and biomass enzymatic saccharification in rice. In the present study, some DEGs that were closely related to lignocellulose synthesis were identified between ZY821 and YG689, and most of these DEGs were located in the main QTLs regions reported by Shen et al. [35]. Meanwhile, 624 DEGs could not be mapped to the B. napus genome, and 108 DEGs could be mapped to the C. bursa-pastoris genome. Some genes that were associated with the lignocellulose biosynthesis were included. For example, BnaA09g42650 is homologous with PRX17 from Arabidopsis. PRX17 encodes a cell wall-localized class III peroxidase and is involved in lignified tissue formation. The transcription factor AGL15 (agamous-like15) directly regulates PRX 17 and participates in the transition from vegetative growth to reproductive growth and the formation of lignified tissue in A. thaliana [42]. Interestingly, 516 of 624 DEGs could be neither mapped to the B. napus nor mapped to the C. bursa-pastoris genome, which indicates that these new genes might derive from the genome recombination between B. napus and C. bursa-pastoris. Coincidently, Zhang et al. [43] identified 37 HE (homologous exchange) events in the progeny of a nascent allotetraploid (AADD) from two diploid progenitors of hexaploid bread wheat. The obtained HEs are highly enriched within gene bodies, giving rise to novel recombinant genes. Furthermore, the generation of chimeric genes was detected in the HEs of the allopolyploid Brassica, rice, Arabidopsis suecica, banana, and peanut [43], which provides a mechanism for the generation of new genes and new proteins in nascent allopolyploids. Recently, the integration analysis of the transcriptome and metabolome has been applied to reveal the regulatory pathways of specific agronomic traits in B. napus [44,45,46,47,48,49]. For example, Jia et al. [44] performed the metabolomic and transcriptomic analyses of the yellow-flowered rapeseed cultivar ZS11 and the white-flowered inbred line WP, and it was shown that the white petal color in WP flowers is primarily due to decreased lutein and zeaxanthin contents, and BnNCED4b might play a key role in white petal formation. Tan et al. [45] investigated the gene expression profiles and metabolite content by the integration analysis of the transcriptome and metabolome in the seeds of B. napus. It was revealed that the expression of major carbohydrate metabolism-regulating genes was significantly correlated with carbohydrate content during seed maturation. In the present study, we integrated the DEGs and differential metabolites to construct a specific pathway of the lignocellulose metabolism in YG689. Notably, a total of 14 distinct metabolites and four DEGs were found to be involved in the regulation of lignocellulose synthesis. For example, the content of cis-aconitate and succinate was higher in the stem of YG689 than in ZY821 in three developmental stages. Importantly, cis-aconitate and succinate are involved in the TCA cycle pathway. Therefore, we speculate that the more active tricarboxylic acid cycle pathway promotes lignin synthesis in YG689 in comparison with ZY821. TPS1 is a gene coding for an enzyme that catalyzes the production of Trehalose 6-phosphate (T6P). It was reported that TPS1 is mainly expressed in axillary buds and the subtending vasculature, as well as in the leaf and stem vasculature [50]. A recent study showed that TPS1 is associated with the traits of plant height, peduncle length, and biomass in wheat [51]. Moreover, it was shown that the loss of TPS1 in A. thaliana impaired high-temperature-mediated hypocotyl growth [52]. Our study found that TPS1 was highly expressed in the stem tissue and up-regulated in YG689 compared with ZY821. Further integration analyses of the transcriptome and metabolome revealed that TPS1 was involved in the regulation of lignocellulose synthesis, which broadens our understanding of key genes regulating important agronomic traits of crops. In addition, as the most abundant pectic glycan, Homogalacturonan (HG) functions as a cell wall structural and signaling molecule essential for plant growth, development, and responses to pathogens [53]. GAUT13 was reported to de novo synthesize HG in the absence of exogenous HG acceptors [53]. In this study, GAUT13 was found to be up-regulated in YG689 (Table S3), which suggests the functional significance of GAUT13 in the lignocellulose synthesis of B. napus. However, the exact mechanism of GAUT13 regulating the lignocellulose synthesis in B. napus needs to be further explored in the future. Lignocellulose is a major component of the mechanical strength of the crop stem tissue and is related to the lodging resistance of crops. The present results are not only useful for understanding the potential regulatory mechanism of lignocellulose biosynthesis, but they also suggest a novel strategy for breeding new varieties with lodging resistance traits, ultimately increasing rapeseed yield in the future. 4. Materials and Methods 4.1. Plant Materials The seeds of YG689, ZY821, C. bursa-pastoris, and TN070 (the B. napus line with a soft stem) are provided by Professor Zaiyun Li of Huazhong Agricultural University, China. YG689 was selected from the successive cytological and fertility selection in the offspring derived from the hybridization between B. napus var. ZY821 and C. bursa-pastoris (Figure 1A). The materials were planted in the experimental field of Huazhong Agricultural University from 2013 to 2016. The seeds were generally sown in late September of the year and the plants and seeds were collected in early May of following year. The row spacing was 40 cm and the plant spacing was 20 cm. 4.2. Anatomical Structure and Lignocellulose Content Analysis of YG689 and ZY821 Stem samples of YG689 and ZY821 at the early flowering and mature stages with three replicates were evenly divided into five segments from the base to the top (Figure 1B). The middle part of the segment was taken for anatomic observation. Some of the transected materials were stained with 1% resorcinol (dissolved in 95% alcohol) for 2 min, then stained with concentrated hydrochloric acid for 1 min, and finally were observed and photographed by the stereomicroscope (Olympus MVX10, Japan). Meanwhile, the stem samples of YG689 and ZY821 at the seedling, bolting and budding, early flowering, final flowering, and maturation stages with three replicates were collected for the lignocellulose total content analysis. The seedling, bolting and budding, early flowering, final flowering, and maturation stages were 105, 142, 160, 178, and 215 days after sowing, respectively. The different developmental stages of stems were firstly blanched at 105℃ and dried at 60 ℃, and 5 mg of YG689 and ZY821 stalk powder were put into 10 mL test tubes for lignocellulose total content measurements, respectively. The acetyl bromide method was used to extract and detect lignin [54], and the content was calculated according to the Bouguer–Lambert–Beer law method [55]. The lignin monomer was prepared as previously described [56], and its contents were determined by GC–MS analysis. The cellulose and hemicellulose were extracted and detected according to the previous literature [57]. Fifteen whole plant materials of YG689 and ZY821 were randomly selected, and the stem length was measured after removing the roots. Furthermore, the 20-cm stems were cut from the stem base and were put on the stem strength tester (YYD-1) to measure the stem strength index. Stem fiber components were measured by near-infrared reflectance spectroscopy (NIRS) using NIR (FOSS, NIRS 5000) with WinISI software, according to previous reports [17]. Five random plants for YG689 and ZY821 were selected and their stems at 20 cm above the cotyledon scar were intercepted, dried, and ground into powder for measuring the fiber components. The phenotype values for acid detergent lignin (ADL), acid detergent fiber (ADF), and neutral detergent fiber (NDF) were speculated from NIRS spectra using NIRS calibrations for these traits, as described by Wei et al. [17]. 4.3. RNA Extraction and Transcriptome Sequencing Total RNAs were extracted from the stems of YG689 and ZY821 by the TriZol method (Invitrogen, Carlsbad, CA, USA), and the mRNAs were isolated from total RNA using Dynabeads oligo (dT) (Invitrogen). First- and second-strand cDNA were synthesized using Superscript II reverse transcriptase and random hexamer primers. Double-stranded cDNA was fragmented by nebulization and used to generate RNA-seq libraries, as previously described [58]. Three biological replicates of the cDNA libraries were sequenced using the Illumina Hiseq 2000 platform. The mRNA expression levels in YG689 and ZY821 were verified by using RT-qPCR. One microgram of total RNA was reverse-transcribed using SuperScript III reverse transcriptase and oligo (dT)18, according to the manufacturer’s instructions. The RT-qPCR reaction system was performed by using the TOYOBO SYBR® R Green Realtime PCR Master Mix (code No. QPK-201) kit. RNA-seq data showed that the mRNA level of actin in ZY821 and YG689, or in each growth stage (the bolting stage, early flowering stage, and terminal flowering stage) was stable. Therefore, actin was set as the reference gene in RT-qPCR experiments. The primers for mRNA RT-qPCR are listed in Table S7. The relative expression levels of these genes were measured by the 2-ΔΔCt method using RT-qPCR. 4.4. Differentially Genes Expression and Function Enrichment Analysis The differentially expressed genes (DEGs) between YG689 and ZY821 were identified using the expression levels of ZY821 transcripts as the control and were tested with the software package DESeq (version 1.12.3) [59] with a false discovery rate (FDR) of < 0.01 and a normalized fold change of ≥ 2. The GO enrichment analysis applied a hypergeometric test to find significantly enriched GO terms in DEGs comparing the genome background [60], where the calculating formula was the same as previously described [61], and the GO terms, with an adjusted p-value of 0.05, were defined as significantly enriched GO terms in DEGs. The enriched GO categories were visualized using the Cytoscape plug-in Enrichment Map (http://www.cytoscape.org/ (19/08/2021)). DEGs which could not be mapped to the B. napus reference genome [62] were aligned to the C. bursa-pastoris reference genome [63] to predict the origin and function of them. 4.5. Metabolome Analysis of YG689 and ZY821 Six biological replicates of each stem sample (0.05 g per sample) of YG689 and ZY821 in the bolting, early flowering, and terminal flowering stages were collected. A total of 36 stem samples were extracted for the GC–TOF–MS analysis, as previously described [64]. An Agilent 7890 GC system equipped with a Pegasus 4D TOFMS (LECO, St. Joseph, MI, USA) was used for the GC–TOF–MS analysis. Metabolite quantification was performed using a multiple reaction monitoring (MRM) method, as described [65]. Statistical significance was defined at p < 0.05, with highly significant values at p < 0.01. The VIP (variable importance in the projection) value (threshold > 1) of the first principal component of OPLS-DA model and the p value of the t-test (threshold 0.05) are used to identify the differentially expressed metabolites. Acknowledgments Not applicable. Supplementary Materials The following are available online at https://www.mdpi.com/article/10.3390/ijms23094481/s1. Click here for additional data file. Author Contributions Conceptualization, M.L.; experiments and data analysis, L.M. and L.Z.; manuscript writing, L.Z.; review and editing, M.L., J.H. and H.L. design of figures, J.H. and H.L. All authors have read and agreed to the published version of the manuscript. Funding This research was funded by the National Natural Science Foundation of China (32072098). Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 (A) The sketch map of YG689 generation from the hybridization of ZY821 and C. bursa-pastoris. (B) The morphology of ZY821 (Left) and YG689 (Right) at initial flowering stage. Bar = 50 cm. (C) The morphology analysis of ZY821 and YG689 at the mature stage (fifteen biological replicates). (D) Comparison of stem lignocellulose (lignin, cellulose, and hemicellulose) content of ZY821 and YG689 at five development stages (three biological replicates). * Significant at p < 0.05; ** significant at p < 0.01. Figure 2 Transverse stem sections of ZY821 and YG689 at initial flowering stage. Stained red characteristics represent the distribution of lignin. The plant is divided into five parts from the base to the top, in turn, and labeled 1, 2, 3, 4, and 5. Bar = 2 mm. Figure 3 (A) Heatmap analysis of differentially expressed genes between YG689 and ZY821. RPKM (reads per kb per million reads) was used to calculate gene expression level. The color key (0, 2, 4, 6, 8, 10) represents FPKM normalized log(10) transformed counts. Red represents high expression and blue represents low expression. Each row represents a gene. (B) Validation of differentially expressed genes by RT-qPCR. Thirteen DEGs were randomly chosen for RT-qPCR validation using ZY821 transcript expression levels as the control. The relative expression levels of each gene are expressed as the fold change between ZY821 and YG689. The B. napus ACT 7 actin gene is used as an internal control. Histogram indicated the relative expression level between ZY821 and YG689 from RT-qPCR results and red square indicated the relative expression level between ZY821 and YG689 from RNA-seq. (C) Heatmap analysis of lignocellulose-related DEGs among ZY821, YG689, and TN070. The color key represents relative expression levels normalized log 2-transformed counts. Red represents high expression and green represents low expression. Each row represents a gene (n = 3). Figure 4 The distribution of the DEGs involved in KEGG pathways on B. napus genome. A01-A10 indicates 10 chromosomes of A genome in B. napus; C01-C09 indicates 9 chromosomes of C genome in B. napus. Red dots indicate up-regulated genes and blue dots indicate down-regulated genes. Large dots indicate 30 DEGs closely related to lignocellulose synthesis. Figure 5 Analysis of metabolic pathways of the lignocellulose-related DEGs. The red arrows represent the up-regulated gene, and blue arrows represent the down-regulated gene. Figure 6 Metabolic pathway of lignocellulose by integrating DEGs, differentially accumulated metabolites between ZY821 and YG689. Red marks indicate up-regulated genes or up-accumulated metabolites in the pathway of lignocellulose synthesis. Blue marks indicate down-regulated genes or down-accumulated metabolites in the pathway of lignocellulose synthesis. NewGene_216 and NewGene_6060 are predicted new genes for B. napus by transcriptome analysis. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Berry P.M. Spink J.H. Gay A.P. Craigon J. A comparison of root and stem lodging risks among winter wheat cultivars J. Agric. Sci. 2003 141 191 202 10.1017/S002185960300354X 2. Zuber U. Winzeler H. Messmer M.M. Keller M. Stampet P. Morphological traits associated with lodging resistance of spring wheat (Triticumae stivum L.) J. Agron. Crop Sci 1999 182 17 24 10.1046/j.1439-037x.1999.00251.x 3. Zhang Y. Liu P. Zhang X. Zheng Q. Chen M. Ge F. Li Z. Sun W. Guan Z. Liang T. Multi-Locus Genome-Wide Association Study Reveals the Genetic Architecture of Stalk Lodging Resistance-Related Traits in Maize Front. Plant Sci. 2018 9 611 10.3389/fpls.2018.00611 29868068 4. Kashiwagi T. Ishimaru K. Identification and functional analysis of a locus for improvement of lodging resistance in rice Plant. Physiol. 2004 134 676 683 10.1104/pp.103.029355 14739343 5. Pinthus M.J. Lodging in wheat, barley and oats: The phenomenon, its causes and preventative measures Adv. Agron. 1973 25 210 263 6. Ma J.F. Yamaji N. Silicon uptake and accumulation in higher plants Trends Plant Sci. 2006 11 392 397 10.1016/j.tplants.2006.06.007 16839801 7. Islam M.S. Peng S. Visperas R.M. Ereful N. Bhuiya M. Julfiquar A.W. Lodging-related morphological traits of hybrid rice in a tropical irrigated ecosystem Field Crop Res. 2007 101 240 248 10.1016/j.fcr.2006.12.002 8. Ma Q.H. Xu Y. Lin Z.B. He P. Cloning of cDNA encoding COMT from wheat which is differentially expressed in lodging-sensitive and-resistant cultivars J. Exp. Bot. 2002 53 2281 2282 10.1093/jxb/erf102 12379798 9. Ookawa T. Inoue K. Matsuoka M. Ebitani T. Takarada T. Yamamoto T. Ueda T. Yokoyama T. Sugiyama C. Nakaba S. Increased lodging resistance in long- culm, low-lignin gh2 rice for improved feed and bioenergy production Sci. Rep. 2014 4 6567 10.1038/srep06567 25298209 10. Ookawa T. Ishihara K. Varietal differences of physical characteristics of the culm related to lodging in paddy rice Jpn. J. Crop Sci. 1992 61 419 425 10.1626/jcs.61.419 11. Kashiwagi T. Madoka Y. Hirotsu N. Ishimaru K. Locus prl5 improves lodging resistance of rice by delaying senescence and increasing carbohydrate reaccumulation Plant Physiol. Biochem. 1992 44 152 157 10.1016/j.plaphy.2006.02.004 16647265 12. Peng D. Chen X. Yin Y. Lu K. Yang W. Tang Y. Wang Z. Lodging resistance of winter wheat (Triticumae stivum L.) lignin accumulation and its related enzymes activities due to the application of paclobutrazol or gibberellin acid Field Crop Res. 2014 157 1 7 10.1016/j.fcr.2013.11.015 13. Tan H. Xie Q. Xiang X. Li J. Zheng S. Xu X. Guo H. Ye W. Dynamic metabolic profiles and tissue specific source effects on the metabolome of developing seeds of Brassica napus PLoS ONE 2015 10 e0124794 10.1371/journal.pone.0124794 25919591 14. Tan H. Xiang X. Tang J. Wang X. Nutritional functions of the funiculus in Brassica napus seed maturation revealed by transcriptome and dynamic metabolite profile analyses Plant Mol. Biol. 2016 92 539 553 10.1007/s11103-016-0530-3 27539000 15. Borisjuk L. Neuberger T. Schwender J. Heinzel N. Sunderhaus S. Fuchs J. Hay J.O. Tschiersch H. Braun H.P. Denolf P. Seed architecture shapes embryo metabolism in oilseed rape Plant Cell. 2013 25 1625 1640 10.1105/tpc.113.111740 23709628 16. Kortesniemi M. Vuorinen A.L. Sinkkonen J. Yang B. Rajala A. Kallio H. NMR metabolomics of ripened and developing oilseed rape (Brassica napus) and turnip rape (Brassica rapa) Food Chem. 2015 172 63 70 10.1016/j.foodchem.2014.09.040 25442524 17. Wei L. Jian H. Lu K. Yin N. Wang J. Duan X. Li W. Liu L. Xu X. Wang R. Genetic and transcriptomic analyses of lignin- and lodging-related traits in Brassica napus Theor. Appl. Genet. 2017 130 1961 1973 10.1007/s00122-017-2937-x 28634809 18. Li H. Cheng X. Zhang L. Hu J. Zhang F. Chen B. Xu K. Gao G. Li H. Li L. An Integration of Genome-Wide Association Study and Gene Co-expression Network Analysis Identifies Candidate Genes of Stem Lodging-Related Traits in Brassica napus Front. Plant Sci. 2018 9 796 10.3389/fpls.2018.00796 29946333 19. Wei C. Zhu L. Wen J. Yi B. Ma C. Tu J. Shen J. Fu T. Morphological, transcriptomics and biochemical characterization of new dwarf mutant of Brassica napus Plant Sci. 2018 270 97 113 10.1016/j.plantsci.2018.01.021 29576090 20. Miller C.N. Harper A.L. Trick M. Wellner N. Werner P. Waldron K.W. Bancroft I. Dissecting the complex regulation of lodging resistance in Brassica napus Mol. Breed. 2018 38 30 10.1007/s11032-018-0781-6 29568228 21. Kawakatsu T. Teramoto S. Takayasu S. Maruyama N. Nishijima R. Kitomi Y. Uga Y. The transcriptomic landscapes of rice cultivars with diverse root system architectures grown in upland field conditions Plant J. 2021 106 1177 1190 10.1111/tpj.15226 33751672 22. Dobrowolska I. Businge E. Abreu I.N. Moritz T. Egertsdotter U. Metabolome and transcriptome profiling reveal new insights into somatic embryo germination in Norway spruce (Picea abies) Tree Physiol. 2017 37 1752 1766 10.1093/treephys/tpx078 28985382 23. Huang H. Yao Q. Xia E. Gao L. Metabolomics and Transcriptomics Analyses Reveal Nitrogen Influences on the Accumulation of Flavonoids and Amino Acids in Young Shoots of Tea Plant (Camellia sinensis L.) Associated with Tea Flavor J. Agric. Food Chem. 2018 66 9828 9838 10.1021/acs.jafc.8b01995 30198713 24. Li Y. Wang W. Feng Y. Tu M. Wittich P.E. Bate N.J. Messing J. Transcriptome and metabolome reveal distinct carbon allocation patterns during internode sugar accumulation in different sorghum genotypes Plant Biotechnol. J. 2019 17 472 487 10.1111/pbi.12991 30051585 25. Reem N.T. Chen H.Y. Hur M. Zhao X. Wurtele E.S. Li X. Li L. Zabotina O. Comprehensive transcriptome analyses correlated with untargeted metabolome reveal differentially expressed pathways in response to cell wall alterations Plant Mol. Biol. 2018 96 509 529 10.1007/s11103-018-0714-0 29502299 26. Verhoeven K.J.F. Verbon E.H. van Gurp T.P. Oplaat C. Ferreira de Carvalho J. Morse A.M. Stahl M. Macel M. McIntyre L.M. Intergenerational environmental effects: Functional signals in offspring transcriptomes and metabolomes after parental jasmonic acid treatment in apomictic dandelion New Phytol. 2018 217 871 882 10.1111/nph.14835 29034954 27. Kumar R. Gyawali A. Morrison G.D. Saski C.A. Robertson D.J. Cook D.D. Tharayil N. Schaefer R.J. Beissinger T.M. Sekhon R.S. Genetic Architecture of Maize Rind Strength Revealed by the Analysis of Divergently Selected Populations Plant Cell Physiol. 2021 62 1199 1214 10.1093/pcp/pcab059 34015110 28. Yu M. Wang M. Gyalpo T. Basang Y. Stem lodging resistance in hulless barley: Transcriptome and metabolome analysis of lignin biosynthesis pathways in contrasting genotypes Genomics 2021 113 935 943 10.1016/j.ygeno.2020.10.027 33127582 29. Guo Q. Li X. Niu L. Jameson P.E. Zhou W. Transcription-associated metabolomic adjustments in maize occur during combined drought and cold stress Plant Physiol. 2021 186 677 695 10.1093/plphys/kiab050 33582802 30. Zhao J. Buchwaldt L. Rimmer S.R. Sharpe A. McGregor L. Bekkaoui D. Hegedus D. Patterns of differential gene expression in Brassica napus cultivars infected with Sclerotinia sclerotiorum Mol Plant Pathol. 2009 10 635 649 10.1111/j.1364-3703.2009.00558.x 19694954 31. Park R.J. The occurrence of mustard oil glucosides in Lepidium hyssopifolium Desv., L. bonariense (L.) and Capsella bursa pastoris (L.) Medic Aust. J. Chem. 1967 20 799 2801 10.1071/CH9672799 32. Sigareva M.A. Earle E.D. Regeneration of plants from protoplasts of Capsella bursa-pastoris and somatic hybridization with rapid cycling Brassica oleracea Plant Cell Rep. 1999 18 412 417 10.1007/s002990050595 33. Chen H.F. Wang H. Li Z.Y. Production and genetic analysis of partial hybrids in intertribal crosses between Brassica species (B. rapa, B. napus) and Capsella bursa-pastoris Plant Cell Rep. 2007 26 1791 1800 10.1007/s00299-007-0392-x 17569050 34. Liu S. Wang X. Fan Z. Pang Y. Sun X. Wang X. Tanga K. Molecular cloning and characterization of a novel cold-regulated gene from Capsella bursa-pastoris DNA Seq. 2004 15 262 268 10.1080/10425170400002421 15620213 35. Shen Y. Xiang Y. Xu E. Ge X. Li Z. Major Co-localized QTL for Plant Height, Branch Initiation Height, Stem Diameter, and Flowering Time in an Alien Introgression Derived Brassica napus DH Population Front. Plant Sci. 2018 9 390 10.3389/fpls.2018.00390 29643859 36. Chen Z.J. Genetic and epigenetic mechanisms for gene expression and phenotypic variation in plant polyploids Ann. Rev. Plant Biol. 2007 58 377 406 10.1146/annurev.arplant.58.032806.103835 17280525 37. Liu Y. Xu A. Liang F. Yao X. Wang Y. Liu X. Zhang Y. Dalelhan J. Zhang B. Qin M. Screening of clubroot-resistant varieties and transfer of clubroot resistance genes to Brassica napus using distant hybridization Breed Sci. 2018 68 258 267 10.1270/jsbbs.17125 29875610 38. Gong Q. Dai C.Y. Zhang X.H. Wang X.L. Huang Z. Xu A.X. Dong J.G. Yu C.Y. Towards breeding of rapeseed (Brassica napus) with alien cytoplasm and powdery mildew resistance from Ethiopian mustard (Brassica carinata) Breed Sci. 2020 70 387 395 10.1270/jsbbs.20017 32714062 39. Kendall S.L. Holmes H. White C.A. Clarke S.M. Berry P.M. Quantifying lodging-induced yield losses in oilseed rape Field Crop Res. 2017 211 106 113 10.1016/j.fcr.2017.06.013 40. Long W. Dan D. Yuan Z. Chen Y. Jin J. Yang W. Zhang Z. Li N. Li S. Deciphering the Genetic Basis of Lodging Resistance in Wild Rice Oryza longistaminata Front. Plant Sci. 2020 11 628 10.3389/fpls.2020.00628 32547576 41. Li F. Xie G. Huang J. Zhang R. Li Y. Zhang M. Wang Y. Li A. Li X. Xia T. OsCESA9 conserved-site mutation leads to largely enhanced plant lodging resistance and biomass enzymatic saccharification by reducing cellulose DP and crystallinity in rice Plant Biotechnol. J. 2017 15 1093 1104 10.1111/pbi.12700 28117552 42. Cosio C. Ranocha P. Francoz E. Burlat V. Zheng Y. Perry S.E. Ripoll J.J. Yanofsky M. Dunand C. The class III peroxidase PRX17 is a direct target of the MADS-box transcription factor AGAMOUS-LIKE15 (AGL15) and participates in lignified tissue formation New Phytol. 2017 213 250 263 10.1111/nph.14127 27513887 43. Zhang Z. Gou X. Xun H. Bian Y. Ma X. Li J. Li N. Gong L. Feldman M. Liu B. Homoeologous exchanges occur through intragenic recombination generating novel transcripts and proteins in wheat and other polyploids Proc. Natl. Acad. Sci. USA 2020 117 14561 14571 10.1073/pnas.2003505117 32518116 44. Jia L. Wang J. Wang R. Duan M. Qiao C. Chen X. Ma G. Zhou X. Zhu M. Jing F. Comparative transcriptomic and metabolomic analyses of carotenoid biosynthesis reveal the basis of white petal color in Brassica napus Planta 2021 253 8 10.1007/s00425-020-03536-6 33387047 45. Tan H. Zhang J. Qi X. Shi X. Zhou J. Wang X. Xiang X. Correlation analysis of the transcriptome and metabolome reveals the regulatory network for lipid synthesis in developing Brassica napus embryos Plant Mol. Biol. 2019 99 31 44 10.1007/s11103-018-0800-3 30519824 46. Wang W. Pang J. Zhang F. Sun L. Yang L. Zhao Y. Yang Y. Wang Y. Siddique K.H.M. Integrated transcriptomics and metabolomics analysis to characterize alkali stress responses in canola (Brassica napus L.) Plant Physiol. Biochem. 2021 166 605 620 10.1016/j.plaphy.2021.06.021 34186284 47. Courbet G. D’Oria A. Maillard A. Jing L. Pluchon S. Arkoun M. Pateyron S. Paysant Le Roux C. Diquélou S. Ourry A. Comparative Omics Analysis of Brassica napus Roots Subjected to Six Individual Macronutrient Deprivations Reveals Deficiency-Specific Genes and Metabolomic Profiles Int. J. Mol. Sci. 2021 22 1679 10.3390/ijms222111679 33562367 48. D’Oria A. Jing L. Arkoun M. Pluchon S. Pateyron S. Trouverie J. Etienne P. Diquélou S. Ourry A. Transcriptomic, Metabolomic and Ionomic Analyses Reveal Early Modulation of Leaf Mineral Content in Brassica napus under Mild or Severe Drought Int. J. Mol. Sci. 2022 23 781 10.3390/ijms23020781 35054964 49. Raza A. Su W. Hussain M.A. Mehmood S.S. Zhang X. Cheng Y. Zou X. Lv Y. Integrated Analysis of Metabolome and Transcriptome Reveals Insights for Cold Tolerance in Rapeseed (Brassica napus L.) Front. Plant Sci. 2021 12 721681 10.3389/fpls.2021.721681 34691103 50. Fichtner F. Barbier F.F. Annunziata M.G. Feil R. Olas J.J. Mueller-Roeber B. Stitt M. Beveridge C.A. Lunn J.E. Regulation of shoot branching in arabidopsis by trehalose 6-phosphate New Phytol. 2021 229 2135 2151 10.1111/nph.17006 33068448 51. Lyra D.H. Griffiths C.A. Watson A. Joynson R. Molero G. Igna A.A. Hassani-Pak K. Reynolds M.P. Hall A. Paul M.J. Gene-based mapping of trehalose biosynthetic pathway genes reveals association with source- and sink-related yield traits in a spring wheat panel Food Energy Secur. 2021 10 e292 10.1002/fes3.292 34594548 52. Hwang G. Kim S. Cho J.Y. Paik I. Kim J.I. Oh E. Trehalose-6-phosphate signaling regulates thermoresponsive hypocotyl growth in Arabidopsis thaliana EMBO Rep. 2019 20 e47828 10.15252/embr.201947828 31393060 53. Engle K.A. Amos R.A. Yang J.Y. Glushka J. Atmodjo M.A. Tan L. Huang C. Moremen K.W. Mohnen D. Multiple Arabidopsis galacturonosyltransferases synthesize polymeric homogalacturonan by oligosaccharide acceptor-dependent or de novo synthesis Plant J. 2021 109 1441 1456 10.1111/tpj.15640 34908202 54. Van A.R. Vanholme R. Storme V. Mortimer J.C. Dupree P. Boerjan W. Lignin biosynthesis perturbations affect secondary cell wall composition and saccharification yield in Arabidopsis thaliana Biotechnol. Biofuels. 2013 6 46 23622268 55. Chang X.F. Chandra R. Berleth T. Beatson R.P. Rapid, microscale, acetylbromide-based method for high-throughput determination of lignin content in Arabidopsis thaliana J. Agric. Food Chem. 2008 56 6825 6834 10.1021/jf800775f 18666780 56. Chen F. Tobimatsu Y. Jackson L. Nakashima J. Ralph J. Dixon R.A. Novel seed coat lignins in the Cactaceae: Structure, distribution and implications for the evolution of lignin diversity Plant J. 2013 73 201 211 10.1111/tpj.12012 22957702 57. Besseau S. Hoffmann L. Geoffroy P. Lapierre C. Pollet B. Legrand M. Flavonoid accumulation in Arabidopsis repressed in lignin synthesis affects auxin transport and plant growth Plant Cell. 2007 19 148 162 10.1105/tpc.106.044495 17237352 58. Wang X. Elling A.A. Li X. Li N. Peng Z. He G. Sun H. Qi Y. Liu X.S. Deng X.W. Genome-wide and organ-specific landscapes of epigenetic modifications and their relationships to mRNA and small RNA transcriptomes in maize Plant Cell. 2009 21 1053 1069 10.1105/tpc.109.065714 19376930 59. Anders S. Huber W. Differential expression analysis for sequence count data Genome Biol. 2010 11 R106 10.1186/gb-2010-11-10-r106 20979621 60. Fu W.Q. Zhao Z.G. Ge X.H. Ding L. Li Z.Y. Anatomy and transcript profiling of gynoecium development in female sterile Brassica napus mediated by one alien chromosome from Orychophragmus violaceus BMC Genom. 2014 15 61 10.1186/1471-2164-15-61 61. Yan X. Dong C. Yu J. Liu W. Jiang C. Liu J. Hu Q. Fang X. Wei W. Transcriptome profile analysis of young floral buds of fertile and sterile plants from the self-pollinated offspring of the hybrid between novel restorer line NR1 and Nsa CMS line in Brassica napus BMC Genom. 2013 14 26 10.1186/1471-2164-14-26 23324545 62. Chalhoub B. Denoeud F. Liu S. Parkin I.A. Tang H. Wang X. Chiquet J. Belcram H. Tong C. Samans B. Plant genetics. Early allopolyploid evolution in the post-Neolithic Brassica napus oilseed genome Science 2014 345 950 953 10.1126/science.1253435 25146293 63. Slotte T. Hazzouri K.M. Ågren J.A. Koenig D. Maumus F. Guo Y.L. Steige K. Platts A.E. Escobar J.S. Newman L.K. The Capsella rubella genome and the genomic consequences of rapid mating system evolution Nat Genet. 2013 45 831 835 10.1038/ng.2669 23749190 64. Sun H.Z. Wang D.M. Wang B. Wang J.K. Liu H.Y. Guan L. Liu J.X. Metabolomics of four biofluids from dairy cows: Potential biomarkers for milk production and quality J. Proteome Res. 2015 14 1287 1298 10.1021/pr501305g 25599412 65. Chen W. Gao Y. Xie W. Gong L. Lu K. Wang W. Li Y. Liu X. Zhang H. Dong H. Genome-wide association analyses provide genetic and biochemical insights into natural variation in rice metabolism Nat. Genet. 2014 46 714 721 10.1038/ng.3007 24908251

PMC009xxxxxx/PMC9099623.txt

==== Front Int J Environ Res Public Health Int J Environ Res Public Health ijerph International Journal of Environmental Research and Public Health 1661-7827 1660-4601 MDPI 10.3390/ijerph19094997 ijerph-19-04997 Article Effects of Changes in Seasonal Weather Patterns on the Subjective Well-Being in Patients with CAD Enrolled in Cardiac Rehabilitation https://orcid.org/0000-0001-6945-4834 Martinaitiene Dalia * Raskauskiene Nijole Okamura Tomonori Academic Editor Laboratory of Behavioral Medicine, Neuroscience Institute, Lithuanian University of Health Sciences, Vyduno al 4, LT 00135 Palanga, Lithuania; nijole.raskauskiene@lsmuni.lt * Correspondence: dalia.martinaitiene@lsmuni.lt; Tel.: +370-460-30012 20 4 2022 5 2022 19 9 499717 2 2022 16 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Objective: We examined whether seasonal and monthly variations exist in the subjective well-being of weather-sensitive patients with coronary artery disease (CAD) during cardiac rehabilitation. Methods: In this cross-sectional study, 865 patients (30% female, age 60 ± 9) were recruited within 2–3 weeks of treatment for acute coronary syndrome and during cardiac rehabilitation. The patients completed the Palanga self-assessment diary for weather sensitivity (PSAD-WS) daily, for an average of 15.5 days. PSAD-WS is an 11-item (general) three-factor (psychological, cardiac, and physical symptoms) questionnaire used to assess weather sensitivity in CAD patients. Weather data were recorded using the weather station “Vantage Pro2 Plus”. Continuous data were recorded eight times each day for the weather parameters and the averages of the data were linked to the respondents’ same-day diary results. Results: Weather-sensitive (WS) patients were found to be more sensitive to seasonal changes than patients who were not WS, and they were more likely to experience psychological symptoms. August (summer), December (winter), and March (spring) had the highest numbers of cardiac symptoms (all p < 0.001). In summary, peaks of symptoms appeared more frequently during the transition from one season to the next. Conclusion: This study extends the knowledge about the impact of atmospheric variables on the general well-being of weather-sensitive CAD patients during cardiac rehabilitation. coronary artery diseases weather rehabilitation subjective well-being weather sensitivity ==== Body pmc1. Introduction Despite significant advancements in prevention and control, cardiovascular diseases (CVDs) continue to be a leading cause of health problems and deaths worldwide [1], including in Lithuania [2]. CVD treatment requires a comprehensive approach due to the significant effects it has on healthcare services. While traditional risk factors such as smoking, alcohol use, hypertension, high cholesterol levels, and obesity are fundamental in explaining a substantial number of CVDs, numerous environmental factors have been discovered to have significant impacts on the risk, development, and severity of CVD [3], for example, the weather. Weather is described as the condition of the atmosphere at any particular time and place. Weather is constantly changing, which partly explains why epidemiologists have conducted little research on the relationships between atmospheric conditions and human health. The prospects for reducing health risks are limited, in contrast to the preventive possibilities that readily flow from the previously mentioned traditional risk factors [4]. Seasonal differences in cardiovascular events, such as acute myocardial infarction, hypertensive urgency, and progressive heart failure have also been well documented and confirmed in large-scale epidemiological studies in various geographic and climatic regions worldwide [5,6,7]. A recent review revealed that seasonality influenced the pattern of nearly every subtype of CVD and that CVD seasonality was most pronounced in individuals living in milder climates, who were the least prepared for extreme weather variations [5]. The effect of season on CVD is complex and, according to the literature, it is modulated by a number of variables. The proposed risk factors include environmental factors such as temperature and exposure to sunlight [8]. Parameters that may increase the risk of acute myocardial infarction include atmospheric variables such as low and high temperature, low atmospheric pressure, high humidity, and reduced exposure to sunlight; the significance of each parameter varies among different seasons [7]. The impact of weather and climate on human health is becoming an increasingly important factor in the context of the current trend towards global and specific regional climatic conditions [9]. Climate change is a gradual change in average weather conditions that changes (mostly increasing) the frequency and intensity of extreme weather events. It is expected that the adverse effects of climate change and the advent of more extreme weather events will become more evident; therefore, the need to find practical ways to reduce the seasonality of CVD [5] and to evaluate the predictors of the development of high weather sensitivity to avoid the risk of the complications provoked by the unfavorable weather conditions [10] has been emphasized. Most studies on the effects of seasonality on CVD have been performed retrospectively and, usually, hospital or emergency aid services databases have been used to identify acute events associated with CVD. Some authors have suggested that well-conducted studies of limited geographical areas with clearly defined parameters of interest and specific risk groups were better than retrospective analyses of large datasets that, so far, have provided confusing data [3,11]. Rehabilitation is one of the main aspects included in CVD medical care. Cardiac rehabilitation programs aim to reduce the risk of another heart event, to monitor and control the current heart condition, and to improve the health and quality of life of patients with CVDs [12]. Cross-sectional studies conducted during rehabilitation programs have also provided additional evidence of a link between weather conditions and health and/or well-being, as well as provided more information to health professionals and the general public. Awareness about the potential influence of weather conditions on the well-being of patients with heart disease may contribute to the planning and implementation of actions leading to improved medical care services and preventative measures that help to avoid worsening of well-being in the future. Previously, we conducted a study aimed at evaluating the association between the subjective well-being of patients with coronary artery disease (CAD) and daily weather parameters [13]. We conducted a weather sensitivity survey in Lithuania, a country in northeastern Europe with four different seasons and different winter–summer conditions. Using the same data, in this study, we aim to determine if seasonal and monthly variations exist in the subjective well-being of weather-sensitive CAD patients during cardiac rehabilitation and if this variation is related to meteorological parameters. 2. Materials and Methods 2.1. Sample and Procedure A detailed description of the study methodologies is presented in our recent study [13]. We used the STROBE cross sectional checklist when writing the report for this cross-sectional study [14]. Briefly, we enrolled 865 patients with CAD attending a rehabilitation program at the Palanga Clinic of the Lithuanian University of Health Sciences Neuroscience Institute (LUHS NI) from June 2008 to October 2012 (Figure 1). Inclusion criteria included an established diagnosis of CAD and age of 18 years or older. Patients were excluded from the study if they had undergone coronary artery bypass graft surgery, cognitive disorientation, communicative disabilities, or other severe diseases, or did not speak Lithuanian fluently. Patients were referred to the rehabilitation clinic within two weeks of treatment for acute coronary syndromes. The duration of cardiac rehabilitation varied based on the diagnosis, from 14 to 20 days. All patients got standard secondary prevention of CAD treatment according to the existing guidelines, including cardiologists’ prescribed medication, physical therapy, risk factor management, and nutritional and psychological counseling. Moreover, physical therapy is also comprised of daily 1.5-h outside walking sessions. Within three days of admission to the rehabilitation clinic, patients were assessed for demographics (age and gender) and clinical characteristics (including NYHA functional class). 2.2. Well-Being To evaluate well-being, all patients completed the Palanga self-assessment diary for weather sensitivity (PSAD-WS). PSAD-WS is a valid and reliable 11-item (general) three-factor questionnaire for collecting information regarding weather sensitivity in patients with CAD [15]. PSAD-WS was created as a self-assessment diary consisting of a list of symptoms. Questionnaire validation was performed on the same sample of patients with CAD. The three factors or symptom subscales reflected: (1) psychological symptoms (apathy, indolence, lack of energy, weakness, daytime sleepiness, and bad mood); (2) cardiac symptoms (palpitation, irregular heartbeat, heartache, and stabbing pain in the chest area); and (3) physical symptoms (joint pain and numbness in the extremities). There were four different PSAD-WS scores defined: a score concerning psychological symptoms, a score concerning cardiac symptoms, a score concerning physical symptoms, and a total score. The total score was the sum of all 11 items [15]. Each day, for a period from 8 to 21 days, depending on the length of the rehabilitation program (average 15.5 ± 3.1 days), all patients (n = 865) completed the PSAD-WS questionnaire about their well-being on the last day, including the nighttime. They were only required to indicate whether one or more of the symptoms listed in the self-assessment diary existed on the day of completing the questionnaire. We did not ask patients to relate their well-being to weather conditions. The severity of symptoms was rated according to 0 (not expressed at all), 1 (expressed), or 2 (strongly expressed). During the study period (between June 2008 and October 2012), a total of 13,327 well-being measurements were taken. 2.3. Weather Sensitivity We asked patients the following question to assess their self-perceived weather sensitivity: “Do you feel the weather changes?” The possible answers were “NO” or “YES”. When patients replied “YES”, they were classified as weather sensitive (WS). 2.4. Study Location and Weather Data During the study period, the weather variable observations were carried out using a “Vantage Pro2 Plus” weather station, located on the roof of the Palanga Clinic of the LUHS NI (in the same location where the respondents attended the rehabilitation program). Palanga city is located in northwest Lithuanian on the eastern shore of the Baltic Sea and belongs to the Lithuanian coastal climate region (55°58′ N, 21°03′ E). The coastal climate region belongs to the Baltic coastal climate region, with climatic indexes that are the most different from the other three Lithuanian climate regions and are closer to Europe’s marine northwest climate. Eight times each day with three-hour intervals (at 0, 3, 6, 9, 12, 15, 18, and 21 h), continuous data of the weather parameters (atmospheric pressure in hectopascal, the temperature in Celsius, relative humidity in percentage, wind speed in meters per second, wind direction in degrees (in a clockwise direction from true north (0 to 360°)), and solar radiation in watt per square meter) were recorded and automatically transmitted to the database using sensors by radio, and then stored in a matrix form. The averages of the weather parameter data recorded eight times daily were calculated and linked to the patients’ same-day self-assessment diary responses. The results referred to more than four years and included all seasons. The classification of seasons was based on the dates when the patients completed the self-assessment diaries. According to the meteorological season calendar, spring begins on 1 March, summer on 1 June, autumn on 1 September, and winter on 1 December. 2.5. Statistical Analysis Demographic variables and clinical conditions were summarized by descriptive analyses. For descriptive purpose, the participants were grouped based on the presence or absence of weather sensitivity (self-reported), and then compared with clinical and demographic variables, using χ2 tests and odds ratio (OR). The correlations between the PSAD-WS subscales and weather parameters were assessed using the Spearman correlation coefficient. Line graphs were used to present the distributions visually. ANOVA was conducted to examine if different patient characteristics (reported weather sensitivity, age group, gender, and NYHA functional class) had overall effects on PSAD-WS concepts (discriminate validity). The effects are reported as an F statistic and its associated degrees of freedom and p-value. The following form of statistical analysis was used: days and the expression of patients’ well-being in those days were taken as the units of analysis. Daily scores of PSAD-WS were aggregated from records on individual patients at current day. The results of our previous study [13] showed that the patients who described themselves as being weather-sensitive had almost two times greater possibility to fill above the median on PSAD-WS than patients who described themselves as non-weather-sensitive. The main outcome variable used in the analysis was the cardiac symptoms PSAD-WS subscale. Four multiple regression models (forward stepwise method) were developed for the seasons to predict patient well-being according to the cardiac symptom subscale. The independent variables for all models were: gender; age; interaction (gender × temperature); psychological symptoms subscale; physical symptoms subscale; and the daily mean of atmospheric pressure, temperature, relative humidity, and solar radiation. The results from the regression analyses were reported as beta coefficients and R2. Statistical analyses were performed using the SPSS Statistical Software (version 17.0, SPSS Inc., Chicago, IL, USA); a p-value of less than 0.05 was considered to be statistically significant. 3. Results A total of 865 patients with CAD were enrolled in the study (Figure 1): 609 (70%) males and 256 (30%) females with a mean age of 60 ± 9 years, and a range from 32 to 88 years. Almost half of the patients’ responses to the question “Do you feel the weather changes?”) indicated that they were WS (n = 410, 47.3%). Table 1 shows the demographic and clinical characteristics of all study participants, stratified by self-reported weather sensitivity. Female patients (OR = 2 and 95% CI 1.7–3.0) and patients over 60, adjusted for gender (OR = 1.6 and 95% CI 1.5–1.7) were more likely to identify as WS. The possibility of being WS increased with a higher NYHA functional class. The wording of a single question can introduce substantial differences and may emphasize only one dimension of weather sensitivity. Within our study, we chose a multidimensional assessment of weather sensitivity. There was a significant effect (Table 2) of patient’s weather sensitivity type (no or yes) on all PSAD-WS symptom subscale scores. There was a significant effect of age group (p = 0.020) and gender (p < 0.001) only on the physical symptoms score. The NYHA functional class affected physical symptoms (p = 0.001) and psychological symptoms (p < 0.001) scores, but not cardiac symptom (p = 0.250) scores. The effects of weather parameters on the PSAD-WS total score and symptom subscale scores were found to be weak. Meanwhile, the analysis of the female group revealed only one significant association between temperature and cardiac symptoms. Furthermore, there was interaction (gender x temperature). The associations between temperature and cardiac symptoms differed between male and female patients: at higher temperatures, there were more cardiac symptoms in female patients, whereas at lower temperatures, there were more symptoms in male patients (Table 3). The mean value distributions of temperature, solar radiation, relative humidity, and barometric pressure, according to month, during the study period, are shown in Figure 2. The percentages of measurements per month, of patients with CAD who reported one or more symptoms on the PSAD-WS total scale and symptom subscales were visualized according to reported WS (Figure 3). There were more reported symptoms in the group describing themselves as WS as compared with patients describing themselves as not WS on the PSAD-WS total scale and in three of the symptom subscales. In addition, the percentage of measurements with symptoms fluctuated throughout the year, and this fluctuation was more pronounced in the WS group. The WS group clearly showed expressed peaks, which were more pronounced in the separate symptom groups. The highest incidences of symptoms were reported in August (summer), December (winter), and March (spring); the lowest incidences of symptoms were reported in January, July, and November (PSAD-WS total scale). Analyses of individual symptom groups showed that WS patients were dominated by a wide range of psychological symptoms (40–80% of measurements per month), with pronounced peaks in August (summer), April (spring), and October (autumn). Cardiac symptoms (30–60% of measurements per month) were more pronounced in March (spring), August (summer), and December (winter); physical symptoms (30–60% of measurements per month) were more pronounced in March and May (spring) and September (autumn). The data show that peaks in symptoms, especially cardiac symptoms, are more evident during the transition from one season to the next. Looking at the weather data during the study period, we can observe more pronounced changes in weather parameters during the peaks of symptoms, i.e., March to April, increase in solar radiation, increase in atmospheric pressure, and significant decrease in humidity (Figure 2) and August, rapid decrease in solar radiation and increase in humidity. Seasonal Associations between Subjective Well-Being and Weather Parameters: Multivariate Approach (among WS Patients) Analyses were performed only among patients who described themselves as WS (n = 410, 6361 days of well-being measurements, average 15.3 day for each patient). As different symptoms groups showed different seasonal distribution; in this case, we paid main attention only to cardiac symptoms (outcome) (Table 4). In spring, the month with the highest number of cardiac symptoms reported by WS patients was March, whereas the month with the lowest number was May (p < 0.001) (Table 4). The regression analysis revealed a significant interaction between gender and temperature. The reported cardiac symptoms were independently associated with lower NYHA functional class, higher temperature (for female patients (interaction)), and lower solar radiation. In summer, the highest number of cardiac symptoms were reported in August and the lowest number were reported in June (p < 0.001). More cardiac symptoms were reported by female patients, by younger than 60-year-old patients with a higher NYHA functional class. Lower atmospheric pressure was an independent factor for the presence of cardiac symptoms. Lowering atmospheric pressure was found to be an independent predictor of the presence of cardiac symptoms. In autumn, the lowest number of cardiac symptoms were reported in October, whereas the highest number of cardiac symptoms were reported in September (p < 0.001). More cardiac symptoms were reported by female patients and by patients with lower NYHA functional class, and a higher number of symptoms were associated with lower atmospheric pressure. In winter, the highest number of cardiac symptoms were reported in December, and the lowest number of cardiac symptoms were reported in January (p < 0.001). For male and female patients, the chance of having cardiac symptoms in the cold period was the same. It was associated with patients older than 60 years of age and higher NYHA functional class with the presence of psychological symptoms. 4. Discussion The purpose of this study was to determine if seasonal and monthly variations existed in the subjective well-being of WS patients with CAD during cardiac rehabilitation and if this variation was related to meteorological parameters. We found that WS patients reported being more sensitive to seasonal changes than patients who were not WS. The highest number of symptoms were reported in August (summer), December (winter), and March (spring) (PSAD-WS total scale and cardiac symptoms subscale), suggesting that the peak of symptoms was more pronounced during the transition from one season to the next, when the fluctuations in weather parameters were more pronounced. Meanwhile, the peaks in the different symptom groups were unequal, and psychological symptoms were most common. Peaks in psychological symptoms appeared in August, April, and October, and most of the physical symptoms were reported by patients who were WS in March, May, and September. It should be noted that Lithuania is a country in northeastern Europe with four different seasons and different winter–summer conditions; our study was performed in the Lithuanian coastal climate region, which is characterized by a cool spring and cool summer; a warm, often without permanent snow cover in winter; a warm and rainy autumn; and small fluctuations in daily and annual temperature [16]. We were unable to find studies that examined the subjective well-being of patients with CAD during the rehabilitation period; therefore, direct comparisons with previous studies are limited. The findings of this study on subjective well-being are consistent with previous results from large-scale epidemiological studies conducted in various geographic and climatic regions worldwide, i.e., CVDs are affected by seasonal variations [5,6,8,17,18]. However, the seasonal variations do not appear to be universal. Most studies have indicated that the winter season had the highest rate of CVD-related hospitalizations and mortality [17,18,19,20] and that event rates were typically 10–20 percent higher than in the summer [5]. It has been suggested that seasonality in CVD was particularly prominent in people who live in milder climates and, therefore, are less prepared for extreme weather changes [5]. Contrary to the general misconception, the majority of temperature-related deaths occur at milder, non-optimal temperatures [19]. In a study in the Czech Republic, excess deaths due to ischemic heart disease (IHD) during hot spells were found to be mainly among people with chronic illnesses whose health had already been compromised. Meanwhile, cardiovascular changes induced by cold stress may result in deaths from acute coronary events rather than chronic IHD [21]. The seasonality of CVD is likely to be caused by the complex interplay between human vulnerability and environmental conditions [5]. Some potential risk factors suggested by researchers are temperature, sunlight exposure, air pollution, some characteristics of the population (age, sex, location, and socioeconomic status), lifestyle (physical activity, smoking, and eating habits), infections, hormones, and vitamin D [9,17]. Weather sensitivities are not always straightforward and are often identified as an accelerator or catalyst rather than a causal factor. In a large nationwide study in Sweden, low air temperature, low atmospheric air pressure, high wind velocity, and shorter sunshine duration were associated with the risk of heart attacks, with the most evident association observed for air temperature [22]. Another nationwide study in the Republic of Slovenia, a South-Central European country [23], found that daily average temperature, atmospheric pressure, and relative humidity had relevant and significant influences on the incidence of acute coronary syndrome for the entire population. In a study performed in Lithuania [24], some patterns of relative humidity, cloud cover, and daily changes in atmospheric pressure and relative humidity were associated with the risk of some types of strokes. Our study revealed that subjective cardiac symptoms were independently associated with lower atmospheric pressure in summer and autumn and lower solar radiation in spring. In contrast to previous studies, which have shown that increased morbidity and mortality for CVD were most associated with temperature, both high and low [1,25], we found only one significant association with temperature and only in spring. This might be explained by the fact that study participants were already recovering from an acute event, were enrolled in a rehabilitation program (which included care, medicines, a different diet, and a different daily routine than at home), and spent more time inside. Meteorological changes are also thought to impact other risk factors of CAD control. For instance, cold days may limit physical activity and promote excessive or high-calorie food intake, alcohol consumption, smoking, etc., which may also contribute to the seasonality of symptoms. Furthermore, these observations support those of previous studies reporting that the significance of each weather parameter varied among different seasons. In a study performed in Italy, a higher risk of primary percutaneous coronary intervention was found with lower minimum atmospheric pressure in the preceding days, lower rainfall in winter, greater changes in atmospheric pressure in spring, and higher temperatures in summer [26]. In most cases, the physiological responses that occur at a specific temperature, humidity, sunlight, and other values or changes in those values are well understood [6,7,27]. Non-optimal temperatures have a variety of effects on humans’ physiological systems, as well as interact with pre-existing diseases and chronic disorders. Even when body temperature remains normal, thermoregulation strains the cardiovascular system [28]. The higher or lower the temperature and the longer the exposure, the more work is required of the cardiovascular system to maintain an optimal temperature. Present theories suggest that lower atmospheric pressure, which affects the sympathetic nervous system and the immune system, increases blood pressure, and that changes in atmospheric pressure can affect atherosclerotic plaques, which cause the plaques to rupture [29,30]. As in previous studies [5,9,23], our study showed that seasonal differences between age and gender groups were unequal. Almost in all seasons, females experienced more cardiac symptoms than males, meanwhile, in winter, the risk of having cardiac symptoms for males and females was the same. Furthermore, the only one association with temperature in spring had a significant gender interaction, suggesting that, in spring, females were more likely to experience cardiac symptoms at higher temperatures and males were more likely to experience symptoms at lower temperatures. Gender differences have also been found in previous weather-health related studies [27,31,32], which were assumed to be due to the different thermoregulatory systems, different physiological structures of males and females (differences in body fatness and distribution, body surface area, and mass), and the varied impacts of bioclimatic conditions on neurohormonal systems [31,32,33]. We found that, in summer, more cardiac symptoms were associated with patients who were younger than 60 years old, meanwhile, in winter, cardiac symptoms were associated with patients who were older than 60 years old. This is consistent with other studies’ results showing that, during a cold period, the risk of acute coronary events [34], acute myocardial infarction [23], and CHD mortality [35] increased with age. Our study showed that WS patients were more likely to experience psychological symptoms during cardiac rehabilitation. Psychological factors are known to affect the biological processes associated with CAD progression [36]. Data suggests that the incidence of depression in people with heart failure is 20% higher than in healthy people [37]. Psychologically distressed people or people with sensitive nervous systems are another well-known WS group [38]. Although psychological symptoms were not thoroughly examined in this study, it can be assumed that environmental factors would contribute more significantly to the seasonal well-being of patients during cardiac rehabilitation if individuals experience significant mental distress symptoms. Our results extend the knowledge about the impact of atmospheric variables on humans’ health and well-being. Even though we analyzed the subjective well-being of patients during rehabilitation, the general findings of our study were in line with the findings of the majority of studies, which have also noted seasonality in CVD and associations with weather parameters such as atmospheric pressure, solar radiation, and temperature. The seasonality of the perceived symptoms may indicate that environmental factors are involved in cardiovascular diseases. Seasonal variations in various symptoms may aid in interpreting the association between the symptoms and other factors in cardiac rehabilitation across different seasons. Knowing about patients’ weather sensitivity can help physicians to understand and interpret patients’ complaints. If patients describe themselves as weather sensitive, physicians should be aware that these patients will be more responsive to atmospheric conditions, especially with changing seasons. If individuals experience significant mental distress symptoms, environmental factors may contribute more to the seasonal well-being of patients during cardiac rehabilitation. Strength and Limitations The present study involved a relatively large sample size of CAD patients who completed their rehabilitation in the same clinic under the same conditions. The results spanned four years and included all seasons. Separate monthly and seasonal analyses were performed. Despite these strengths, there were some limitations. We used only available outdoor atmospheric parameter data to evaluate the impact of climatic variables on the well-being of patients and did not assess the time that patients spent outside. The effect of climatic factors on various well-being symptoms could have been moderated by indoor conditions and time spent outdoors. Some risk factors, such as socioeconomic status and lifestyle, were not taken into consideration. 5. Conclusions Our results confirm the importance of seasonal atmospheric state variability in the general well-being of weather-sensitive CAD patients during cardiac rehabilitation. Seasonal variations in some cardiac, psychological, or physical symptoms could affect patients undergoing cardiac rehabilitation, especially if patients describe themselves as weather sensitive. This pattern changes slightly according to age and gender. If patients experience significant mental distress symptoms, their sensitivity to environmental factors will be higher. This study extends the knowledge about the impact of atmospheric variables on the general well-being of weather-sensitive CAD patients during cardiac rehabilitation. Acknowledgments We would cordially thank all the staff members at the Laboratory of Behavioral Medicine of the Neuroscience Institute of Lithuanian University of Health Science for their strong support of this study. We would also like to thank the patients for their time and willingness to contribute to this study. Author Contributions D.M. and N.R., conceptualization, methodology, formal analysis, writing—review and editing; D.M., writing the original draft, investigation; N.R., data curation, visualization. All authors have read and agreed to the published version of the manuscript. Funding This research received no external funding. Institutional Review Board Statement The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Lithuanian Bioethics Committee (certificates no. BE-2-21, BE-2-26 and P-82) on 12 April 2007. Informed Consent Statement Written informed consent was obtained from all patients involved in the study. Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Flowchart of the study. * PSAD-WS, Palanga self-assessment diary for weather sensitivity; ** WS, weather sensitive. Figure 2 Mean values of weather variables according to month, during the study period (from June 2008 to October 2012). Error bars denote a 95% interval. Figure 3 Percentages of measurements per month, of patients with CAD who reported one or more subjective well-being symptoms on the PSAD-WS total scale and symptom subscales: (A) Weather sensitive patients; (B) not weather sensitive patients. ijerph-19-04997-t001_Table 1 Table 1 Demographic and clinical characteristics of all patients at inclusion and stratified by self-report weather sensitivity. Variable n (%) Not Weather Sensitive n = 455 Weather Sensitive n = 410 Odds Ratio (95% CI) Age, years: pfor trend < 0.01 <50 133 (15.4) 85 (18.7) 48 (11.7) 1 51–60 280 (32.4) 161 (35.5) 119 (29) 1.3 (0.8–1.9) 61–70 324 (37.4) 147 (32.2) 177 (43.2) 2.1 (1.4–3.2) >70 128 (14.8) 62 (13.6) 66 (16.1) 1.9 (1.1–3.1) Gender: pfor trend < 0.001 Male 609(70) 356 (78.3) 243 (61.7) 1 Female 256 (30) 99 (21.7) 157 (38.3) 2.2 (1.7–3.0) NYHA class: pfor trend < 0.001 I 54 (6.2) 44 (9.7) 10 (2.4) 1 II 552 (63.9) 301 (66.1) 251 (61.2) 3.6 (1.8–7.4) III 259 (29.9) 110 (24.2) 149 (36.4) 5.9 (2.8–12.2) NYHA, New York Heart Association. ijerph-19-04997-t002_Table 2 Table 2 Analysis of variance by PSAD-WS subscales (F statistic). PSAD-WS Subscales Psychological Symptoms Cardiac Symptoms Physical Symptoms Weather sensitive (no vs. yes) F(1,864) = 27.1 p < 0.001 F(1,864) = 6.3 p = 0.012 F(1,864) = 30 p < 0.001 Age groups <50, 51–60, 61–70, 70+ F(3,862) = 2.2 p = 0.093 F(3,862) = 1.3 p = 0.27 F(3,862) = 3.3 p = 0.020 Gender, female vs. male F(1,864) = 2.2 p = 0.139 F(1,864) = 3.3 p = 0.070 F(1,864) = 21.8 p < 0.001 NYHA class F(3,864) = 12 p < 0.001 F(3,864) = 1.37 p = 0.25 F(3,864) = 5.6 p = 0.001 ijerph-19-04997-t003_Table 3 Table 3 Spearman correlation coefficients (r) between weather parameters and daily PSAD-WS, according to gender. Atmospheric Pressure hPa Temperature °C Relative Humidity % Solar Radiation W/m2 Males PSAD-WS total 0.005 −0.043 ** 0.043 ** −0.042 ** Psychological symptoms −0.009 −0.036 ** 0.042 ** −0.029 * Cardiac symptoms 0.014 −0.024 * 0.017 −0.026 * Physical symptoms 0.025 * −0.051 ** 0.041 ** −0.056 ** Females PSAD-WS total 0.005 0.026 −0.024 −0.027 Psychological symptoms 0.014 0.019 −0.027 −0.027 Cardiac symptoms −0.019 0.031 * −0.004 −0.023 Physical symptoms 0.008 0.007 0.008 −0.007 * p < 0.05 and ** p < 0.01. ijerph-19-04997-t004_Table 4 Table 4 Multiple linear regression analysis predicting the reporting cardiac symptoms scores separately by four seasons (significant standardized regression coefficients β). Dependent Variable Sum of Cardiac Symptoms Independent Variables Spring Summer Autumn Winter Gender (1 = M; 2 = W) 0.104 0.171 Interaction: Gender (1 = M; 2 = W) × (temperature) 0.107 Age ≥60 vs. <60 years −0.057 0.072 NYHA class −0.085 0.169 −0.249 0.091 Solar radiation −0.064 Atmospheric pressure −0.082 −0.089 May vs. March, April −0.124 August vs. June, July 0.058 October vs. September, November −0.096 December vs. January, February 0.167 Sum psychological symptoms subscale 0.379 0.285 0.369 0.544 Sum physical symptoms subscale 0.199 0.291 0.201 Model R2 = 0.266 F = 93.73 p < 0.001 R2 = 0.287 F = 71.01 p < 0.001 R2 = 0.242 F = 74.51 p < 0.001 R2 = 0.369 F = 171.34 p < 0.001 Number of subjects included in the analysis n = 410 WS patients, 6361 measurements. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Bhatnagar A. Environmental determinants of cardiovascular disease Circ. Res. 2017 121 162 180 10.1161/CIRCRESAHA.117.306458 28684622 2. Health Information Centre of Institute of Hygiene Causes of Death 2020 Available online: https://hi.lt/uploads/pdf/leidiniai/Statistikos/Mirties_priezastys/Mirties_priezastys_2020.pdf (accessed on 22 December 2021) 3. McArthur K. Dawson J. Walters M. What is it with the weather and stroke? Expert. Rev. Neurother. 2010 10 243 249 10.1586/ern.09.154 20136380 4. Schneider S. Root T. Mastrandrea M. Health Encyclopedia of Climate and Weather 2nd ed. Oxford University Press Oxford, UK 2011 10.1093/acref/9780199765324.001.0001 5. Stewart S. Keates A.K. Redfern A. McMurray J.J.V. Seasonal variations in cardiovascular disease Nat. Rev. Cardiol. 2017 14 654 664 10.1038/nrcardio.2017.76 28518176 6. Mikić M. Šestak A. Volarić M. Rudan S. Trtica Majnarić L. Seasonality of the cardiac biomarker troponin in the Eastern Croatian population J. Clin. Med. 2018 7 520 10.3390/jcm7120520 7. Furukawa Y. Meteorological factors and seasonal variations in the risk of acute myocardial infarction Int. J. Cardiol. 2019 294 13 14 10.1016/j.ijcard.2019.07.089 31399297 8. Pell J.P. Cobbe S.M. Seasonal variations in coronary heart disease QJM 1999 92 689 696 10.1093/qjmed/92.12.689 10581331 9. Balbus J. Crimmins A. Gamble J.L. Easterling D.R. Kunkel K.E. Saha S. Sarofim M.C. Ch. 1: Introduction: Climate Change and Human Health The Impacts of Climate Change on Human Health in the United States: A Scientific Assessment U.S. Global Change Research Program Washington, DC, USA 2016 25 42 10. Uyanaeva A.I. Tupitsyna Y.Y. Rassulova M.A. Turova E.A. Lvova N.V. Ajrapetova N.S. The influence of the climatic and weather conditions on the mechanisms underlying the formation of enhanced meteosensitivity (a literature review) Vopr. Kurortol. Fizioter. Lech. Fiz. Kult. 2016 93 52 57 10.17116/kurort2016552-57 28635705 11. Bishop-Williams K.E. Berrang-Ford L. Sargeant J.M. Pearl D.L. Lwasa S. Namanya D.B. Edge V.L. Cunsolo A. IHACC Research Team Bwindi Community Hospital Understanding Weather and Hospital Admissions Patterns to Inform Climate Change Adaptation Strategies in the Healthcare Sector in Uganda Int. J. Environ. Res. Public Health 2018 15 2402 10.3390/ijerph15112402 12. Bellmann B. Lin T. Greissinger K. Rottner L. Rillig A. Zimmerling S. The beneficial effects of cardiac rehabilitation Cardiol. Ther. 2020 9 35 44 10.1007/s40119-020-00164-9 31997145 13. Martinaitiene D. Raskauskiene N. Weather-related subjective well-being in patients with coronary artery disease Int. J. Biometeorol. 2021 65 1299 1312 10.1007/s00484-020-01942-9 32494961 14. von Elm E. Altman D.G. Egger M. Pocock S.J. Gøtzsche P.C. Vandenbroucke J.P. STROBE Initiative. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: Guidelines for reporting observational studies J. Clin. Epidemiol. 2008 61 344 349 10.1016/j.jclinepi.2007.11.008 18313558 15. Stroputė D. Raškauskienė N. Bunevičius R. Psychometric properties of the Palanga self-assessment diary for weather sensitivity in patients with coronary artery disease Biol. Psychiatry Psychopharmacol. 2014 16 3 9 16. Galvonaitė A. Kilpys J. Kitrienė Z. Valiukas D. Lietuvos kurortų klimatas. Vilnius 2015 13 Available online: http://www.meteo.lt/documents/20181/103901/Kurortu_klimatas_internetui.pdf/085a6384-0340-4a56-a25d-a7b0fdaf7ec9 (accessed on 20 December 2021) 17. Fares A. Winter cardiovascular diseases phenomenon N. Am. J. Med. Sci. 2013 5 266 279 10.4103/1947-2714.110430 23724401 18. Afoakwah C. Nghiem S. Scuffham P. Stewart S. Byrnes J. Weather Trumps Festivity? More Cardiovascular Disease Events Occur in Winter than in December Holidays in Queensland, Australia Int. J. Environ. Res. Public Health 2021 18 10158 10.3390/ijerph181910158 34639460 19. Gasparrini A. Guo Y. Hashizume M. Lavigne E. Zanobetti A. Schwartz J. Tobías A. Tong S. Rocklöv J. Forsberg B. Mortality risk attributable to high and low ambient temperature: A multicountry observational study Lancet 2015 386 369 375 10.1016/S0140-6736(14)62114-0 26003380 20. Vaičiulis V. Jaakkola J.J.K. Radišauskas R. Tamošiūnas A. Lukšienė D. Niilo R.I.R. Association between winter cold spells and acute myocardial infarction in Lithuania 2000–2015 Sci. Rep. 2021 11 17062 10.1038/s41598-021-96366-9 34426618 21. Davídkovová H. Plavcová E. Kynčl J. Kyselý J. Impacts of hot and cold spells differ for acute and chronic ischaemic heart diseases BMC Public Health 2014 14 480 10.1186/1471-2458-14-480 24886566 22. Mohammad M.A. Koul S. Rylance R. Fröbert O. Alfredsson J. Sahlén A. Witt N. Jernberg T. Muller J. Erlinge D. Association of Weather with Day-to-Day Incidence of Myocardial Infarction: A SWEDEHEART Nationwide Observational Study JAMA Cardiol. 2018 3 1081 1089 10.1001/jamacardio.2018.3466 30422202 23. Ravljen M. Bilban M. Kajfež-Bogataj L. Hovelja T. Vavpotič D. Influence of daily individual meteorological parameters on the incidence of acute coronary syndrome Int. J. Environ. Res. Public Health 2014 11 11616 11626 10.3390/ijerph111111616 25396770 24. Vencloviene J. Radisauskas R. Kranciukaite-Butylkiniene D. Tamosiunas A. Vaiciulis V. Rastenyte D. Association between stroke occurrence and changes in atmospheric circulation BMC Public Health 2021 21 42 10.1186/s12889-020-10052-5 33407282 25. Parsons K. Human Thermal Environments: The Effects of Hot, Moderate, and Cold Environments on Human Health, Comfort, and Performance 3rd ed. CRC Press Boca Raton, FL, USA 2002 10.1201/b1675014 26. Versaci F. Biondi-Zoccai G. Giudici A.D. Mariano E. Trivisonno A. Sciarretta S. Valenti V. Peruzzi M. Cavarretta E. Frati G. Climate changes and ST-elevation myocardial infarction treated with primary percutaneous coronary angioplasty Int. J. Cardiol. 2019 294 1 5 10.1016/j.ijcard.2019.07.006 31301864 27. Schneider A. Schuh A. Maetzel F.K. Ruckerl R. Breitner S. Peters A. Weather-induced ischemia and arrhythmia in patients undergoing cardiac rehabilitation: Another difference between men and women Int. J. Biometeorol. 2008 52 535 547 10.1007/s00484-008-0144-9 18228048 28. Havenith G. Temperature regulation, heat balance, and climatic stress Extreme Weather Events and Public Health Responses Kirch W. Bertollini R. Menne B. Springer Berlin, Germany 2005 69 80 29. Charach G. Friedman I. Nochomovich H. Rogowski O. Charach L. Steinvil R. Angel Y. Argov O. Whether atmospheric pressure changes predict variations in blood pressure J. Cardiol. Curr. Res. 2017 9 00330 10.15406/jccr.2017.09.00330 30. Gunes H. Kandis H. Saritas A. Dikici S. Buyukkaya R. The relationship between ischemic stroke and weather conditions in Duzce, Turkey World J. Emerg. Med. 2015 6 207 211 10.5847/wjem.j.1920-8642.2015.03.008 26401182 31. Graham T.E. Thermal, metabolic, and cardiovascular changes in men and women during cold stress Med. Sci. Sports Exerc. 1988 20 S185 S192 10.1249/00005768-198810001-00017 3057320 32. Mazza M. Di Nicola M. Catalano V. Callea A. Martinotti G. Harnic D. Bruschi A. Battaglia C. Janiri L. Description and validation of a questionnaire for the detection of meteoropathy and meteorosensitivity: The METEO-Q Compr. Psychiatry 2012 53 103 106 10.1016/j.comppsych.2011.02.002 21489419 33. Lan L. Lian Z. Liu W. Liu Y. Investigation of gender difference in thermal comfort for Chinese people Eur. J. Appl. Physiol. 2008 102 471 480 10.1007/s00421-007-0609-2 17994246 34. Hodzic E. Perla S. Iglica A. Vucijak M. Seasonal Incidence of Acute Coronary Syndrome and Its Features Mater Sociomed 2018 30 10 14 10.5455/msm.2018.30.10-14 29670472 35. Douglas A.S. Allan T.M. Rawles J.M. Composition of seasonality of disease Scott. Med. J. 1991 36 76 82 10.1177/003693309103600304 1925507 36. Kop W.J. Chronic and acute psychological risk factors for clinical manifestations of coronary artery disease Psychosom. Med. 1999 61 476 487 10.1097/00006842-199907000-00012 10443756 37. Khayyam-Nekouei Z. Neshatdoost H. Yousefy A. Sadeghi M. Manshaee G. Psychological factors and coronary heart disease ARYA Atheroscler. 2013 9 102 111 23690809 38. Yackerson N.S. Zilberman A. Todder D. Kaplan Z. The influence of several changes in atmospheric states over semi-arid areas on the incidence of mental health disorders Int. J. Biometeorol. 2011 55 403 410 10.1007/s00484-010-0350-0 20668888

PMC009xxxxxx/PMC9099624.txt

==== Front Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells11091435 cells-11-01435 Article The Phenotype of the Adipocytes Derived from Subcutaneous and Visceral ADMSCs Is Altered When They Originate from Morbidly Obese Women: Is There a Memory Effect? https://orcid.org/0000-0003-4936-3120 Mikłosz Agnieszka 1* https://orcid.org/0000-0002-4954-3431 Łukaszuk Bartłomiej 1 Supruniuk Elżbieta 1 https://orcid.org/0000-0002-6828-6397 Grubczak Kamil 2 https://orcid.org/0000-0003-1084-5747 Starosz Aleksandra 2 Kusaczuk Magdalena 3 https://orcid.org/0000-0001-5229-1805 Naumowicz Monika 4 https://orcid.org/0000-0002-7407-8156 Chabowski Adrian 1 Bunnell Bruce A. Academic Editor 1 Department of Physiology, Medical University of Bialystok, Mickiewicza 2C Street, 15-222 Bialystok, Poland; bartlomiej.lukaszuk@umb.edu.pl (B.Ł.); elzbieta.supruniuk@umb.edu.pl (E.S.); adrian.chabowski@umb.edu.pl (A.C.) 2 Department of Regenerative Medicine and Immune Regulation, Medical University of Bialystok, Waszyngtona 13 Street, 15-269 Bialystok, Poland; kamil.grubczak@umb.edu.pl (K.G.); aleksandra.starosz@umb.edu.pl (A.S.) 3 Department of Pharmaceutical Biochemistry, Medical University of Bialystok, Mickiewicza 2A Street, 15-222 Bialystok, Poland; magdalena.kusaczuk@umb.edu.pl 4 Department of Physical Chemistry, Faculty of Chemistry, University of Bialystok, K. Ciolkowskiego 1K Street, 15-245 Bialystok, Poland; monikan@uwb.edu.pl * Correspondence: agnieszka.miklosz@umb.edu.pl; Tel.: +48-85-746-55-85 23 4 2022 5 2022 11 9 143525 3 2022 21 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Adipose tissue is an abundant source of mesenchymal stem cells (ADMSCs). Evidence has suggested that depot-specific ADMSCs (obtained from subcutaneous or visceral adipose tissue–subADMSCs or visADMSCs, respectively) account for differential responses of each depot to metabolic challenges. However, little is known about the phenotype and changes in metabolism of the adipocytes derived from ADMSCs of obese individuals. Therefore, we investigated the phenotypic and metabolic characteristics, particularly the lipid profile, of fully differentiated adipocytes derived from ADMSCs of lean and obese (with/without metabolic syndrome) postmenopausal women. We observed a depot-specific pattern, with more pronounced changes present in the adipocytes obtained from subADMSCs. Namely, chronic oversupply of fatty acids (present in morbid obesity) triggered an increase in CD36/SR-B2 and FATP4 protein content (total and cell surface), which translated to an increased LCFA influx (3H-palmitate uptake). This was associated with the accumulation of TAG and DAG in these cells. Furthermore, we observed that the adipocytes of visADMSCs origin were larger and showed smaller granularity than their counterparts of subADMSCs descent. Although ADMSCs were cultured in vitro, in a fatty acids-deprived environment, obesity significantly influenced the functionality of the progenitor adipocytes, suggesting the existence of a memory effect. ADMSCs adipocytes metabolic syndrome obesity subcutaneous and visceral adipose tissue lipid metabolism ==== Body pmc1. Introduction Over the past few decades, the prevalence of obesity has increased dramatically, and its consequences represent a major public health concern. A chronic imbalance between energy intake and energy expenditure leads to adipose tissue (AT) dysfunction, including adipocyte hypertrophy, disequilibrium between lipogenesis and lipolysis, abnormal secretion of numerous adipokines and cytokines, and infiltration of fat tissue with inflammatory M1 macrophages [1]. The dysfunctional AT is unable to efficiently process the circulating metabolites, which increases the risk of metabolic disorders. White adipose tissue is a highly heterogeneous organ distributed throughout the body, but its principal depots are the subcutaneous adipose tissue (SAT) and visceral adipose tissue (VAT) [2]. Although morphologically similar, SAT and VAT show distinct characteristics, they differ in terms of gene expression pattern, endocrine and metabolic activity, insulin sensitivity, vascularization, innervation and infiltration by immune cells [3,4]. This functional heterogeneity and diverse anatomical distribution translate into their divergent involvement in the formation of metabolic disorders. In fact, VAT expansion is a major predictive factor for the development of metabolic abnormalities, such as insulin resistance (IR), lipodystrophy and type 2 diabetes mellitus (T2DM), whereas the enlargement of subcutaneous fat improves insulin sensitivity, reduces metabolic complications, and therefore, is considered to be a protective [5]. However, in obesity, superfluous energy supply exceeds the storage capacity of SAT, resulting in lipid deposition in VAT and non-adipose tissues (ectopic lipid accumulation). Therefore, the comparison of adipose depots with respect to their distinct anatomic location is crucial to clarify their role in obesity-related complications. Postmenopausal women are one of the groups most prone to develop metabolic disorders as a result of the loss of endogenous sex hormone production. Menopause changes the phenotype of adipose tissue causing adipocyte hypertrophy, which translates to tissue hypoxia, triggering its inflammation, and fibrosis. Moreover, inadequate estrogen-mediated ability to store excess lipids in SAT may contribute to their deposition in less favourable anatomic locations [6,7]. Still the protective role of SAT against the negative effects of obesity is highly debated [8]. Adipose-derived mesenchymal stem cells (ADMSCs) are a new source of stem cells that can be easily isolated from adipose tissue [1]. Importantly, ADMSCs of subcutaneous or visceral ancestry exhibit morphological and molecular differences. Subcutaneous ADMSCs are characterized by a higher proliferation rate and greater potential towards adipogenic differentiation, moreover, they display an increased capacity to store lipids and pronounced ability for adiponectin secretion compared with visceral ADMSCs [9,10]. In addition to the inherent properties of ADMSCs [11,12], also the microenvironment surrounding the cells [12,13] may underlie the biological functions of adipose depots. In response to excessive caloric intake, pluripotent ADMSCs undergo transformation to preadipocytes; thus, ADMSCs constitute an almost unlimited source of these cells [14]. However, obesity disrupts ADMSCs functioning by reducing their lipid storage ability, compromising their regenerative potential, modifying their apoptosis susceptibility and endocrine/paracrine function (e.g., by stimulating cytokine secretion) [15,16,17,18]. For example, visceral ADMSCs in obese humans secrete more pro-inflammatory cytokines compared to subcutaneous ADMSCs, which is consistent with the stronger pro-inflammatory pattern adopted by VAT during the development of the metabolic dysfunction [16,19]. Still, the distinct effects of nutrient overflow on the phenotype of ADMSCs have only recently been started to be unravelled. Because, the differences between the particular white adipose tissue depots may well come from the ADMSCs properties, we decided to thoroughly investigate the intrinsic characteristics of the adipocytes differentiated from subcutaneous and visceral ADMSCs. Knowing that adipocytes’ functionality may be lost during obesity, we obtained the cells (ADMSCs) from lean as well as morbidly obese (with and without metabolic syndrome) postmenopausal female subjects and differentiated them into mature adipocytes. Particularly, we determined the effects of obesity and metabolic syndrome on the expression and cellular localization of several protein carriers responsible for the uptake of long chain fatty acids (LCFAs). Furthermore, we examined the expression of proteins involved in lipid turnover, in conjunction with the cellular lipid profile in the adipocytes differentiated from the ADMSCs. 2. Materials and Methods 2.1. The Origin of Human Adipose-Derived Mesenchymal Stem Cells (ADMSCs) Subcutaneous (abdominal region) and visceral (omental region) white adipose tissue biopsies (approximately 2–4 g) were obtained from postmenopausal female subjects treated at the First Department of General and Endocrine Surgery at the University Hospital in Białystok. The study protocol was approved by the Ethics Committee of the Medical University of Bialystok (permission R-I-002/187/2017), in agreement with the Declaration of Helsinki. All patients were informed about the properties of the study and gave their written consent before participating in the study. Enrolled patients underwent clinical examination, anthropometric measurements and appropriate laboratory tests (Table S1). Major exclusion criteria were acute inflammatory process and a history of malignancy. The lean group consisted of four women of normal BMI value (19–24.9 kg/m2) who underwent elective laparoscopic cholecystectomy. Patients with obesity were divided into morbidly obese female individuals of BMI > 40 kg/m2 without metabolic syndrome (n = 4) and with metabolic syndrome (n = 4) who underwent sleeve gastrectomy. Immediately following dissection, the samples were placed in phosphate-buffered saline (PBS, PAN-Biotech, Aidenbach, Germany) and transported to the laboratory for ADMSCs isolation. Based on the available literature we have developed an isolation protocol for ADMSCs as described previously [20]. Briefly, adipose tissue samples were washed in PBS, cleaned from visible blood vessels and then minced and digested in collagenase (250 U/mL collagenase NB 4G Proved Grade, Serva, Heidelberg, Germany) at 37 °C for approximately 1 h, until complete digestion. Mature adipocytes from the stromal vascular fraction were removed using a 500 µm strainer. After centrifuging, the cell pellet was resuspended in erythrocyte lysis buffer (Thermo Fisher Scientific, Waltham, MA, USA), filtered through a 200 µm and then 20 µm strainers and centrifuged again for 5 min at 600× g. The ADMSCs-rich pellet was suspended in mesenchymal stem cells medium containing growth supplements (MSCM, ScienCell Research Laboratories, Carlsbad, CA, USA), 5% fetal bovine serum (FBS, Thermo Fisher Scientific, Waltham, MA, USA) and antibiotics (PAN-Biotech, Aidenbach, Germany). The viability and cell purity of isolated ADMSCs was determined via by flow cytometry as described below. The culture medium was changed every 2–3 days until cells reached the confluency of 80–90%; then they were detached with trypLE (Thermo Fisher Scientific, Waltham, MA, USA). ADMSCs were then collected at a density of 5 × 105 − 5 × 106 cells in stem cell cryopreservation medium (Stem-Cell banker DMSO Free, Takara Bio, Mountain View, CA, USA) and stored at −80 °C. 2.2. ADMSCs Differentiation towards Adipocytes 2.2.1. Adipogenesis of ADMSCs For further experiments, cells from 2 to 4 passages were thawed and seeded in MSCM on the appropriate plates. Once the cell confluency reached approximately 90%, adipogenesis was induced by a differentiation medium containing supplements (MADM, Mesenchymal Stem Cell Adipogenic Differentiation Medium, ScienCell Research Laboratories, Carlsbad, CA, USA), 5% FBS and 1% penicillin/streptomycin solution. The medium was replaced every 3–4 days, and the progress of adipogenesis was monitored by microscopic observation of lipid vacuoles in the cells. After 14–21 days, the adipocytes were harvested and subjected to specific analyses. 2.2.2. Assessment of the Accumulated Lipids The amount of accumulated lipids was measured in mature adipocytes after fixing with 10% formalin (Sigma Aldrich, St. Louis, MO, USA) for 30 min at RT. Then, the cells were incubated with 0.5% Oil Red-O solution (Sigma Aldrich) for 1 h, and the overstaining was washed several times with PBS. The images were created using an inverted microscope (Olympus, magnification ×400). Next, the accumulated Oil Red-O was extracted with 100% isopropanol and the absorbance was measured at 510 nm using a microplate reader (Synergy H1 Hybrid Reader, BioTek, Santa Clara, CA, USA). Undifferentiated ADMSCs stained with Oil Red-O served as control. Oil Red-O concentration was determined based on known standards of Oil Red-O (0.02 mg/mL to 2.5 mg/mL). Additionally, the average size and the number of lipid droplets per cell were evaluated based on the microphotographs using OLYMPUS cellSens Standard 1.18 software. 2.3. Flow Cytometry Characterization Adipocytes collected after cell culture were subjected to immunostaining with monoclonal antibodies: anti-FATP1 (ACSVL5, mouse anti-human) (R&D Systems, Minneapolis, MN, USA); anti-FATP4 (ACSVL4, rabbit anti-human); anti-FABP4 (A-FABP, rabbit anti-human); and anti-CD36 (SR-B2; rabbit anti-human) (Abcam, Cambridge, UK). It should be noted that prior to staining procedures, the adipocytes were divided into two separate groups, one for extracellular staining only, and another permeablized with FACS Permeabilizing Solution 2 (BD Bioscience, Franklin Lakes, NJ, USA) for additional detection of intracellular markers. Approximately 100,547 of the collected cells were used, the number of cells recovered was 45,689 for extracellular events and 15,105 for intracellular events. Following incubation at room temperature, in dark conditions, cells were washed in phosphate-buffered saline (PBS with no calcium and magnesium; Corning, Corning, NY, USA). Subsequently, secondary detection antibodies were used to bind to antibodies bound to adipocytes proteins: goat anti-rabbit (Alexa Fluor 488); and goat anti-mouse (Alexa Fluor 647) (Invitrogen, MA, USA). Incubation was followed by double centrifugation in PBS to wash out unbound antibodies. Finally, cells were fixed using CellFIX (BD Biosciences) and stored in 4 °C until acquisition using a FACS Calibur flow cytometer (BD Biosciences; San Jose, CA, USA). Processing of the flow cytometric data was performed with the use of FlowJo software (TreeStar Inc., Ashland, OR, USA). The studied proteins were analyzed within adipocytes distinguished initially on the basis of morphological properties–forward scatter (FSC; relative size) and side scatter (SSC; relative granularity/complexity) (Figure S1). Lack of 7AAD-related fluorescence was used to gate viable cells. The obtained results were presented as frequencies of selected markers within adipocytes or mean fluorescence intensity (MFI) of a specific marker within adipocytes. For the assessment of relative changes in size and granularity/complexity, adipocytes were divided into four separate subgroups with increasing values of each morphological parameter. 2.4. RNA Isolation and Quantitative Real Time RT-PCR Total cellular RNA was extracted using TRIzol Reagent (Sigma Aldrich, Saint Louis, MO, USA) in accordance with the manufacturer’s instructions. RNA concentration and purity were assessed by spectrophotometry (at an absorbance OD ratio of 260/280 and 260/230). Reverse transcription was performed using the EvoScript universal cDNA master kit (Roche Molecular Systems, Boston, MA, USA) with 1 µg of total RNA. Quantitative real time polymerase chain reaction (qRT-PCR) was carried out in duplicate using the LightCycler 96 System Real-Time thermal cycler with FastStart essential DNA green master (Roche Molecular Systems) as the detection dye. The following reaction parameters were applied in a thermal cycler: 15 s denaturation at 94 °C, 15 s annealing at 57 °C for RPLO13A and TBC1D1, 58 °C for TBC1D4 and CD36/SR-B2, 59 °C for FATP1 and FATP4, and 62 °C for FABPpm, and then a 15 s extension at 72 °C for 45 cycles. The primer sequences are listed in Table S1. The primers’ efficiency was analysed using the standard curve method. Gene expression was calculated according to the Pfaffl method [21], normalizing to the housekeeper gene (RPLO13A). 2.5. Immunoblotting Routine Western blotting procedures were used to detect protein content in total lysate [20]. In brief, ADMSCs lysates were prepared using ice-cold radioimmunoprecipitation assay (RIPA) buffer containing a mix of protease and phosphatase inhibitors (Roche Diagnostics GmbH, Mannheim, Germany). The total protein concentration was assayed using the BCA method with bovine serum albumin (BSA) as a standard. Then, lysates were reconstituted in Laemmli buffer (Bio-Rad, Hercules, CA, USA), and equal amounts of the proteins (30 µg per sample) were loaded on Criterion TGX Stain-Free Precast Gels (Bio-Rad, Hercules, CA, USA) for sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Size-separated proteins were transferred onto polyvinylidene difluoride (PVDF) membranes. After blocking in 5% non-fat dry milk for 1 h, membranes were incubated overnight at 4 °C with the corresponding primary antibodies, i.e., TBC1D1 (1:500, cat. no. H00023216-B02P, Novus Biologicals, CO, USA), AS160/TBC1D4 (1:1000, cat. no. # 07-741, Merck Millipore, CA, USA), CD36/SR-B2 (1:500, cat. no. sc-7309, Santa Cruz Biotechnology, Inc., Dallas, TX, USA), FATP4 (1:1000, cat. no. ab200353, Abcam, Cambridge, UK), FATP1 (1:500, cat. no. sc-25541, Santa Cruz Biotechnology, Inc., Dallas, TX, USA), FABPpm (1:4000, ab180162, Abcam, Cambridge, UK), β-HAD (1:1000, cat. no. sc-271495, Santa Cruz Biotechnology, Inc., Dallas, TX, USA), DGAT1 (1:500, cat. no. sc-271934, Santa Cruz Biotechnology, Inc., Dallas, TX, USA), ATGL (1:500, cat. no. sc-365278, Santa Cruz Biotechnology, Inc., Dallas, TX, USA), FASN (1:1000, cat. no. ab128870, Abcam, Cambridge, UK), GAPDH (1:1000, sc-47724, Santa Cruz Biotechnology, Inc., Dallas, TX, USA). Thereafter, bound antibodies were detected with suitable anti-rabbit or anti-goat IgG horseradish peroxidase-conjugate secondary antibodies (1:3000, Santa Cruz Biotechnology, Dallas, TX, USA). The protein bands were imaged by chemiluminescence using Clarity Western ECL Substrate (Bio-Rad, Hercules, CA, USA), and signal intensities were quantified densitometrically using a ChemiDoc visualisation system (Bio-Rad, Hercules, CA, USA). The protein expression (optical density arbitrary units) was normalised to GAPDH expressions and was related to the control subADMSCs group. 2.6. 9,10-[3H]-Palmitic Acid Uptake Before the analysis, differentiated adipocytes were starved in a serum-free medium for 3 h. Then the cells were incubated with Krebs–Ringer-HEPES buffer supplemented with palmitic acid (Sigma Aldrich, St. Louis, MO, USA) bound to fatty acids-free bovine serum albumin (BSA, Sigma Aldrich) with the radiolabelled 9,10-[3H] palmitic acid (Perkin Elmer) at the specific activity of 1 μCi mL for 5 min at 37 °C/5% CO2. Afterwards, the reaction was terminated by addition of ice-cold PBS and finally the cells were solubilised in 0.05 N NaOH. Radioactivity was measured using a Packard TRI-CARB 1900 TR scintillation counter, and was normalised to the protein concentrations. 2.7. Lipid Content Quantification (Gas Liquid Chromatography) Lipids from the ADMSCs were extracted using chloroform-methanol solution according to the Folch method and separated into different fractions using thin-layer chromatography (TLC). Individual fatty acids (FAs) from each fraction were methylated and then esters were quantitatively determined in relation to corresponding retention times of standards by means of GLC method (Hawlett-Packard 5890 Series II gas chromatograph, HP-INNOWax capillary column). The concentrations of FFA, DAG and TAG were assessed as the sum of the individual FAs in each fraction. 2.8. Statistical Analysis The numbers of patients (n–number of patients per group) included in the analysis are mentioned in the legends of the appropriate Figures/Tables. The measurements were made in triplicate or duplicate (see Figure legends) and the arithmetic means were used for subsequent investigation. Statistical analyses were performed with R (ver. 3.6.3) or GraphPad Prism (9.0.0) programs. The Shapiro–Wilk test (test for normality) and Fligner–Killeen tests (test for homogeneity of variances) were used to determine the subsequent application of parametric or non-parametric methods. The data were analysed using three-way ANOVA (Table S6). In addition to the above, the data were analysed with either the Student’s t-test, or Wilcoxon rank sum test. 3. Results 3.1. Characterization of Human ADMSCs First, using specific markers, we characterized the ADMSCs immunophenotype by flow cytometry. As previously reported by us [20], the ADMSCs showed high expressions of CD105, CD73, and CD90 (the proportion of positive cells exceeded 99%) and lacked the expression of CD45 and lineage markers; negative markers were expressed in less than 1% of the cells. Additionally, high expression of CD10 and low expression of CD105 distinguished subADMSCs from visADMSCs. Thereafter, the differentiation into mesoderm derivatives, i.e., osteocytes, chondrocytes, and adipocytes, confirmed the multilineage potential of the isolated cells. The effectiveness of the above was confirmed upon microscopic inspection and by immunocytochemistry [20]. Moreover, neither obesity nor the presence of the metabolic syndrome affects the viability of ADMSCs derived from both subcutaneous and visceral adipose tissues (Figure 1A). To confirm adipogenesis in stem cells obtained from patient donors with different metabolic status, the cells were stained with Oil Red O (ORO) for the identification of lipid droplets. Once differentiated, all the cells presented a large number of lipid-filled vacuoles that appeared red under an inverted light microscope. Quantitative analysis of ORO demonstrated a pronounced build-up of intracellular lipids in the adipocytes when compared with undifferentiated precursor cells (Figure 1B). The average cell size was relatively stable in the adipocytes of subADMSC ancestry (Figure 1C). In the cells originating from visceral tissue, we observed greater cell areas in the cells from visObese(-) and visObese(+) groups when compared with visLean. We observed no statistically significant changes between the analysed groups with respect to the number of lipid droplets per cell (Figure 1C). The mean average cell size was unchanged in the adipocytes Further, based on the results obtained from flow cytometry analysis, FABP4 mean fluorescence intensity (MFI) was diminished in adipocytes derived from subADMSCs isolated from obese women with metabolic syndrome, when compared to the lean control group. However, this does not apply to adipocytes originating from visADMSCs. Moreover, the adipocytes obtained from the visADMSCs of the lean patients appeared to have a lower expression of FABP4 compared to their subcutaneous counterparts (Figure 1D). 3.2. SLC27A4/FATP4 and CD36/SR-B2, but Not FATP1 or FABPpm, Contribute to Increased Fatty Acid Uptake in Obese Adipocytes Derived from subADMSCs The rate of cellular fatty acid (FA) uptake is short- and long-term regulated. Short-term regulation (i.e., minutes) occurs via reversible recycling of CD36/SR-B2 or FATP4 from intracellular compartments to the plasma membrane, which is dependent on Rab-GTPase activating proteins (RabGAPs), namely TBC1D1 and/or TBC1D4 (AS160) [22]. As we detected virtually no changes in the amount of TBC1D1 at both mRNA and protein levels, it seems that only AS160 had a regulatory role in the adipocytes (Figure 2A–C). An examination of TBC1D4 levels revealed its greater expression in the adipocytes derived from the obese patients (with and without metabolic syndrome) when compared to their lean analogues. Herein, the mRNA and protein level of AS160 was increased by roughly 40% and 60% in the case of subcutaneous tissue. In visADMSCs, the protein expression of TBC1D4 was also markedly elevated (+76% and 79% for visObese(-) and visObese(+) vs. visLean, p < 0.05). In addition to the above-mentioned alterations, we observed a lower TBC1D4 protein content in the cells of visceral origin when compared to their subcutaneous counterparts (Figure 2A–C). Long-term regulation of the rate of cellular FA uptake, which occurs in obesity (high fatty acid supply), involves changes in the gene transcription and/or protein abundance. In our study, CD36/SR-B2 mRNA was not altered in obese adipocytes compared with lean cells derived from both sub- and visADMSCs. Interestingly, the presence of metabolic syndrome significantly augmented the transcript level of this transporter (+74% and +87% for subObese(+) vs. subObese(-) and visObese(+) vs. visObese(-), p < 0.05, Figure 3A). However, the protein abundance was solely increased in subADMSCs mature adipocytes that originated from obese but metabolically healthy patients compared to the lean control group (+59%), whereas both obese groups exhibited higher CD36/SR-B2 levels in the adipocytes derived from visADMSC (+117% and +166 % for visObese(-) and visObese(+) vs visLean, respectively, p < 0.05, Figure 3B). Furthermore, we observed no changes in the expression of FABPpm (neither at mRNA nor protein level) with respect to any of the investigated groups (Figure 3A,B). Similarly, neither mRNA nor the total protein expression of FATP1 were significantly altered in any of the examined groups in the adipocytes derived from subADMSCs (Figure 3A,B). On the contrary, the adipocytes of visceral provenance displayed a characteristic pattern, i.e., the fat cells derived from ADMSCs of obese patients had a significantly lower total FATP1 protein content when compared with their analogues stemming from the lean patients (p < 0.05). The examination of FATP4 mRNA and total protein content revealed changes, but only in the adipocytes differentiated from subADMSCs. We observed a greater FATP4 transcript content in the case of the cells derived from the obese patients (with and without metabolic syndrome) when compared with their counterparts descending from the lean individuals (+73% and +50% for subObese(-) and subObese(+) vs. subLean, p < 0.05, Figure 3A). The above-mentioned pattern was also apparent in the total protein content; the cells derived from obese women had a significantly greater expression of FATP4 compared to the subLean group (+115% and +103% for subObese(-) and subObese(+) vs. subLean, p < 0.05, Figure 3B). Moreover, we observed some differences in the protein content with respect to the cells’ tissue of origin. In general, the adipocytes descending from visADMSCs showed lower CD36/SR-B2 and FATP4 protein content than their counterparts derived from subADMSCs (Figure 3B). To examine the changes in subcellular localization of FA transporters, we performed flow cytometry analysis. The surface expression of the two most abundant FA transporters in adipocytes, i.e., CD36/SR-B2 and FATP4 was significantly higher in the cells differentiated from subADMSCs of obese, metabolically unhealthy women (Figure 4). It may suggest that both the transporters under high fatty acid supply (as in obesity) are permanently relocated to the adipocytes surface, increasing FA uptake and further progressive lipid accumulation. In contrast, obesity itself did not change the subcellular location of FATP1 in the adipocytes derived from both adipose tissues. The rate of long-chain fatty acid influx was determined by 3H-palmitate uptake. In line with the results described above, the adipocytes originated from subADMSCs of obese patients with metabolic syndrome showed a 145% increase in 3H-palmitate uptake, when compared with the adipocytes from lean ADMSCs (Figure 2D). Likewise, palmitate uptake was increased in the adipocytes stemming from obese patients with metabolic syndrome of visceral ancestry. However, the above-mentioned changes did not reach the level of statistical significance (p > 0.05). In addition to the above, we observed a significantly lower palmitate uptake while comparing visceral and subcutaneous cells derived from obese patients with metabolic syndrome (−44%, p < 0.05, Figure 2D). 3.3. Effect of ADMSCs Tissue Origin and the Metabolic Status of the Patient Donor on Adipocyte Size and Granularity Using the flow cytometry analysis (forward scatter (FSC) = positive correlation with the size of cells) we found that the adipocytes differentiated from visADMSCs were larger than those from subADMSCs (Figure 5A). The size of mature adipocytes and the size distribution were significantly different between morbidly obese and lean postmenopausal women with respect to visADMSCs-derived cells. For example, the adipocytes from the visObese(+) group had greater cell volume compared to their visLean and visObese(-) counterparts. In addition, a higher incidence of intermediate adipocyte sizes was noted for ranges II (275–550) and III (550–825) (Figure 5A). On the other hand, the metabolic status of the donor patient did not affect the cell size or the size distribution in the adipocytes differentiated from subADMSCs. Furthermore, analysis of the side scatter (SSC) data allowed for the estimation of adipocyte granularity. The cells of visceral origin showed less granularity, indicating their tendency to accumulate lipids in several spacious vacuoles. In all the studied groups, the granularity of the adipocytes appeared to decrease with the size of the cells (Figure 5B). 3.4. Enhanced FA Uptake Promotes Lipid Accumulation Together with Changes in the Composition of Different FA Species Mainly in the Adipocytes Derived from subADMSCs of Morbidly Obese Women with Metabolic Syndrome The total fatty acid content of FFA was increased only in the adipocytes differentiated from subADMSCs of obese patients with metabolic syndrome (+45% subObese(+) vs. subLean, p < 0.05, Figure 6A). This was mostly caused by an increase in the concentration of saturated fatty acid species, i.e., palmitic acid (C16:0), while no changes in unsaturated FA levels were observed. When comparing the adipocytes of different tissue descent we observed a lower total FFA content in the cells derived from visADMSCs, but only in the case of obese patients with metabolic syndrome (−66% for visObese(+) vs. subObese(+), p < 0.05, Figure 6A). This was due to their lower saturated (C16:0), and also unsaturated fatty acids content (UNSFA), i.e., palmitooleic (C16:1), arachidonic (20:4n6) (Table S3). The adipocytes derived from the ADMSCs of obese patients with metabolic syndrome also had a greater total DAG concentration when compared with the cells of other patient type descent (Figure 6B). However, this was evident only in the case of the adipocytes of subcutaneous origin (+66% for subObese(+) vs. subLean, p < 0.05; and +37% for subObese(+) vs. subObese(-), p < 0.05). Tissue comparison revealed that the adipocytes stemming from the visADMSCs of obese patients with metabolic syndrome had a lower total DAG level compared with their counterparts derived from the subADMSCs (−47% for visObese(+) vs. subObese(+), p < 0.05). This was most likely caused by a decrease in the amount of unsaturated, i.e., palmitooleic (C16:1) and oleic (18:1n9c), fatty acids content (Table S4). Similarly to the total FFA and DAG levels, TAG content in the cells differentiated from subADMSCs was significantly increased in the case of patients with metabolic syndrome in comparison with the lean adipocytes (+30% for subObese(+) vs. subLean, p < 0.05, Figure 6C). This was probably due to a greater concentration of saturated fatty acid species with unchanged UNSFA in this group. Moreover, we did not observe any statistically significant changes in the total amount of TAG in the cells of visceral origin with respect to the patient type. However, a comparison of the adipocytes derived from the sub- and visADMSCs revealed a lower total concentration of TAG in the cells of visceral provenance (−68% and −74% for visLean vs. subLean and for visObese(+) vs. subObese(+), respectively, p < 0.05). A more detailed analysis of the individual content of distinct FA species within the adipocytes revealed lower amounts of saturated fatty acids such as palmitic acid (C16:0) as well as unsaturated fatty acids content, i.e., oleic acid (18:1n9c), linoleic acid (18:2n6c) and linolenic acid (C18:9n3) (Table S5). 3.5. LCFAs Synthesis and Utilization Are Enhanced in the Adipocytes Derived from ADMSCs of Morbidly Obese Women The expression of fatty acid synthase (FASN), an enzyme involved in de novo lipogenesis as well as diacylglycerol acyltransferase (DGAT1) and which catalyzes the final step in the biosynthesis of TAGs, was markedly elevated in the adipocytes derived from both sub- and visADMSCs of obese patients (with or without metabolic syndrome). In the case of the cells of subcutaneous origin, the amount of FASN was greater by 81% and 97%, whereas DGAT1 expression was increased by 84% and 112% (for subObese(-) and subObese(+) vs. subLean, p < 0.05, Figure 7). The above-mentioned pattern was also visible in the cells that descended from visADMSCs (FASN: +132% and +105%; DGAT1: +180% and +302%, for visObese(-) and visObese(+) vs. visLean, p < 0.05). In addition to the above-mentioned changes, we observed that FASN and DGAT1 expression was lower in the cells that descend from visceral fat depots in comparison to their counterparts of subcutaneous origin. Western Blot analysis of protein expression of beta-hydroxyacyl CoA dehydrogenase (β-HAD) revealed that the adipocytes derived from subADMSCs had an increased amount of β-HAD in the Obese(+) group, an even greater change was observed in the cells from the Obese(-) group, compared with the lean controls (+36% and +108% for subObese(+) and subObese(-) vs. subLean, p < 0.05, Figure 7). The cells of visceral provenance in the obese group without metabolic syndrome also showed the greatest expression of β-HAD protein when compared with the lean controls (+47% for visObese(-) vs. visLean, p < 0.05). The level of adipose triglyceride lipase (ATGL) in the adipocytes derived from subADMSCs was significantly increased solely in the Obese(-) group in comparison with the control group (Figure 7). Surprisingly, in the cells of visADMSCs provenance we observed a decreased amount of ATGL protein in the obese patients compared with the lean subjects. We also observed a difference in the amount of ATGL between the cells of subcutaneous and visceral origin. The latter had a lower cellular level of ATGL, which was especially visible in the case of obese patients (−83% for visObese(-) vs. subObese(-), p < 0.05; and −90% for visObese(+) vs. subObese(+), p < 0.05). 4. Discussion A growing body of evidence points to the existence of metabolic and functional differences between individual fat depots. In consequence, ADMSCs obtained from different locations vary with respect to their morphology, function and biochemical/metabolic properties, as well as gene expression patterns. These characteristics are stable and are maintained after the ADMSCs have been isolated and cultured in vitro [23]. Although the role of the cells constituting white adipocyte tissue (WAT) has been well recognized in obesity, the metabolism of adipocytes differentiated from human mesenchymal stem cells is still being elucidated. Importantly, WAT is a highly heterogeneous tissue in which mature adipocytes account for only 15–30% of the total adipose cell fraction, the rest is stromal vascular fraction (SVF) [24,25]. Thus, a study of the adipocytes derived from the tissue stem cells should allow the observation of ‘pure’ (i.e., not obfuscated by other cellular fractions, nor the tissue milieu) adipocyte phenotypes. Here, we decided to compare the makeup and lipid profile of the adipocytes originated from ADMSCs of lean and morbidly obese women (with or without metabolic syndrome) to better understand intrinsic factors behind each condition. In adipose tissue, the cellular uptake of LCFAs is facilitated by several membrane-associated proteins, including CD36/SR-B2, FABPpm and FA transport proteins (FATP1 and FATP4) [26]. Our data show distinct expression profiles of fatty acid handling proteins in the adipocytes of sub- and visADMSCs origin. The levels of two of the most abundant FA transporters, CD36/SR-B2 and FATP4, were distinctly greater in the adipocytes differentiated from subADMSCs than in those from visADMSCs. These results may imply different fatty acid handling in various adipose tissue depots. Previously, it was suggested that both CD36/SR-B2 and FATP4 are important fatty acids transporters, especially in the case of low extracellular FAs concentration [27,28]. This could be the case in our in vitro experiment (low fatty acid concentration in medium). In line with that notion, we observed a higher total TAG content in the subLean group in comparison with their counterparts from the visLean group. However, it is not an obvious effect, since palmitate uptake was similar in both groups. Obesity was associated with greater protein expression of CD36/SR-B2 in the adipocytes originated from ADMSCs of both depots and FATP4 (only in ADMSCs derived from subcutaneous depots). This is consistent with the higher CD36/SR-B2 level observed in vivo by Bonen et al. for VAT and SAT of obese individuals [29], and indicates the fundamental role of subcutaneous adipose tissue in the FA turnover. Our finding is also supported by the results from the examination of monozygotic twins provided by Gertow and co-workers [30]. The authors reported moderate to strong positive correlation between the amount of mRNA for the transporters and the volume of subcutaneous fat (+0.62 for FATP4, and +0.71 for CD36/SR-B2). Interestingly, only a weak correlation (or no correlation at all) was noted for the transporters and intra-abdominal (visceral) fat size (+0.3 for FATP4, and +0.14 for CD36/SR-B2) [30]. Moreover, the adipocytes from the subObese(+) group are characterized by the redistribution of CD36/SR-B2 and FATP4 towards the plasma membrane, as evidenced by the flow cytometry. This translates into the pronounced increased influx of the LCFA (radio-isotope labeled palmitic acid uptake) observed in the subObese(+) group (Figure 8). Such a lipid oversupply is believed to trigger the metabolic complications observed in obese patients with metabolic syndrome. Recently, we have demonstrated that AS160/TBC1D4 and its structural homolog TBC1D1 could be involved in the cellular redistribution of fatty acid transporters. Silencing of TBC1D4, but not TBC1D1, in the adipocytes led to a greater translocation of fatty acid transporters (mostly CD36/SR-B2) into the plasma membrane [20]. Interestingly, in the current study we observed an increased expression of AS160/TBC1D4 in the adipocytes derived from ADMSCs of obese patients (especially from individuals with metabolic syndrome). This should translate into lower plasmalemmal expression of fatty acid transporters and smaller LCFA uptake by the cells, since AS160/TBC1D4 serves as a negative regulator for the translocation of transporters to the cell surface [31]. However, this prediction is in contrast to the actual observations mentioned above (specifically the data from flow cytometry for CD36/SR-B2, FATP4, and palmitic acid uptake). The discrepancy is perplexing and not straightforward to explain. Still, recent review papers by Glatz et al. may shed some light on the topic [22,32]. The authors point to the existence of two mechanisms regulating CD36/SR-B2 mediated FA uptake in skeletal muscle and cardiomyocytes. Long-term regulation is triggered by high fatty acids oversupply (such as occurs in obese patients in vivo) and results in increased CD36/SR-B2 gene transcription. Short-term regulation, on the other hand, is initiated by low fatty acids supply (as observed in vitro) and relies on CD36/SR-B2 subcellular recycling [22,32]. We postulate, therefore, that adipocytes from the obese group have an activated long-term overexpression of fatty acid transporters (CD36/SR-B2, FATP4 at mRNA and protein level) that cannot be exactly counterbalanced by their decreased plasmalemmal translocation (by TBC1D4 overexpression). The net result of this interplay of factors is an increased amount of FA transporters on the cell surface and propensity towards greater fatty acids intake (larger palmitate uptake). On the other hand, neither FABPpm nor FATP1 total protein content was changed in subADMSCs, whereas FATP1 expression was even diminished in the adipocytes from visADMSCs. These results, along with the fact that mRNA expression of FATP1 is relatively low in human adipose tissue, suggest that this transporter does not play a crucial role in LCFAs uptake in obese patients. In line with this notion, Binnert et al. found that the transcript level of FATP1 was not correlated with BMI in male subjects’ subcutaneous adipose tissue [33]. Nevertheless, there are very few reports concerning the fatty acid transporters in human adipocytes differentiated from ADMSCs, and further investigations are required to directly confirm the above-mentioned suppositions. Adipocytes primarily store and release free fatty acids (FFA) to support local and systemic metabolic demands. The majority, i.e., approximately 80%, of body fat is deposited in the subcutaneous area where the excess of FFA and glycerol is stockpiled as TAG. On the other hand, visceral fat accounts for about 10–20% of total body fat in lean men and less than 10% in healthy women [34]. Studies have shown that subcutaneous adipose tissue increases in volume by hyperplasia rather than by hypertrophy. The opposite is true for visceral fat [35]. This tendency was also maintained in our study. Overall, the adipocytes differentiated from visADMSCs were larger than those of subADMSCs provenance. Considering the metabolic status of the donor patient, obesity significantly increased cell volumes, but only in the adipocytes derived from visADMSCs. Further cytometric analyses demonstrated smaller granularity of the cells of visceral origin, which points to their propensity to accumulate lipids in a few spacious vacuoles. This agrees also with the evaluation using micro-photographs of the average cell size and number of lipid droplets per cell (greater cell size of the adipocytes in visObese(-) and visObese(+) with the unchanged number of lipid droplets suggests an increased size of the latter). Nevertheless, the larger adipocytes in the visceral depot have a decreased capacity to take up an excess of FFAs compared with the new smaller adipocytes differentiated from subADMSCs. This notion is supported by an increased TAG accumulation in the fully differentiated adipocytes of subcutaneous ancestry. Kim et at. reported that subcutaneous ADMSCs had a higher capacity to proliferate and differentiate into adipogenic lineages than those from visceral ADMSCs [36]. Accordingly, in the present study, we found a higher expression of FABP4, an adipogenic marker, in adipocytes derived from lean subADMSCs, which was confirmed in the study by Baglioni and co-workers [11]. The adipocytes of subcutaneous origin (subADMSCs) seem to better reflect the metabolic status of donor patients compared with visceral-derived ADMSCs, since we found that subObese(+) had more lipids (TAG, DAG, and FFA) in comparison with subLean (Figure 8). This is quite interesting because, in general, visceral fat is considered to be more (patho)physiologically active with respect to the induction of cardiovascular and metabolic complications of obesity [5]. However, the progression of the disorder runs rather in the opposite direction (first subcutaneous then visceral fat accumulation and malfunction). In overweight individuals, superfluous lipids are deposited mostly in subcutaneous tissue [24,37]. With progress towards obesity and metabolic syndrome development the lipids are shifted towards visceral storage sites. At some point, hypertrophic adipocytes in the visceral locations approach the limit of their storage capacity. Thus, further overflow of VAT with FFAs results in their ectopic deposition in adjacent organs, e.g., in the liver, advancing the pathological condition even farther [24]. The expression of enzymes involved in lipid synthesis is also in line with that notion. The examined proteins (FASN and DGAT1) presented clearly greater levels in the adipocytes obtained from subADMSCs than those from visADMSCs, which may underlie the higher lipid-storing capacity of the cells of subcutaneous deposits (Figure 8). This finding is supported by previous results of other authors, in which the genes engaged in lipogenesis, such as FASN, acetyl-CoA carboxylase α (ACACA), and stearoyl-CoA desaturase (SCD1), were upregulated in differentiated adipocytes derived from SAT of lean individuals [38]. While the concomitant increases in lipolytic enzymes, i.e., ATGL and β-HAD, in the Obese(-) group appear to prevent augmented deposition of lipid fractions compared with the subLean group, the development of metabolic syndrome reduced these compensatory cellular mechanisms and led to lipid accumulation (FFA, DAG and TAG) within the adipocytes. Furthermore, the cells differentiated from visADMSCs of both the obese groups had lower levels of the lipolytic enzymes than their counterparts obtained from subADMSCs, which indicates greater lipid turnover rate in the cells of subcutaneous provenance. Arner et al. reports that greater lipid turnover is observed in obesity when fat tissue is composed of numerous small adipocytes (hyperplasia) [24], and this was the case in our study. Nevertheless, most studies report that the progress of obesity is associated with a decreased lipid turnover in the cells [24]. In the present study, only the level of ATGL dropped with the advancement of obesity as evidenced in the cells originating from visADMSCs. The exact role of ATGL in metabolic abnormalities, however, is inconclusive since current literature reported either unchanged [39], increased [40] or decreased [41] levels of ATGL protein in obese humans or mice, with different patterns in male, female and mixed studied groups. Based on the data obtained, we may conclude that the adipocytes of the ADMSCs ancestry preserve several characteristics of mature fat cells in vivo. Particularly, the cells from the Obese(+) group (and to a lesser extent the Obese(-) group) reflect the phenotype observed in vivo in the early stages of obesity. 5. Conclusions In summary, our study is the first to characterize in detail the phenotype of the adipocytes differentiated from the stem cells (ADMSCs) obtained from lean and obese patients (with and without metabolic syndrome). Particular emphasis was put on their broadly investigated lipid profile. We report several interesting findings regarding the differences in the examined parameters with respect to the tissue of origin and donor patients’ metabolic status. Several of the observed patterns reflect those that were previously reported in mature adipocytes in vivo. First of all, we observed that the cells of visADMSCs provenance were larger and displayed smaller granularity than their counterparts of subADMSCs ancestry. This points to their propensity to accumulate lipids in a few spacious vacuoles. Moreover, obesity led to further increases in the cell size and a drop in their granularity as evidenced by comparing the visObese(+) with visLean group. Secondly, we identified CD36/SR-B2 and FATP4 as the two most important FA transporters in the investigated adipocytes. Moreover, obesity appears to be associated with greater expression of the transporters (CD36/SR-B2 and FATP4) at the transcript (mRNA) and protein (total and cell surface) level. Overall, we found increased lipid concentration (TAG) in the cells of subcutaneous origin when compared with their counterparts from the other depot. Interestingly, the cells were also characterized by a boosted lipid synthesis (greater protein expression of FASN and DGAT1). Besides, obesity and metabolic syndrome were associated with further accumulation of TAG, DAG, and FFA in the cells that stemmed from subADMSCs accompanied by an increase in FASN and DGAT1 protein expression. Additionally, the cells of subcutaneous ancestry displayed greater lipid turnover than their visceral counterparts. Altogether, these data highlight that obesity significantly alters the phenotype of adipocytes differentiated from ADMSCs, and the progression of obesity towards metabolic syndrome further exacerbates these differences. These, differential responses of the adipocytes suggest that there is some memory effect of obesity influencing the functionality of progenitor adipocytes isolated from different fat depots. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/cells11091435/s1, Figure S1: Gating strategy for the determination of the size and granularity of ADMSC-derived adipocytes by flow cytometry; Table S1: The primers used for real-time PCR, Table S2: Clinical characteristics of patients (tissue donors), Table S3: FFA—Fatty acid composition (nmol/mg of protein), Table S4: DAG—Fatty acid composition (nmol/mg of protein), Table S5: TAG—Fatty acid composition (nmol/mg of protein), Table S6: Summary table of ANOVA analysis. Click here for additional data file. Author Contributions Conceptualization, A.M.; Data curation, A.M., B.Ł. and E.S.; Formal analysis, A.M. and B.Ł.; Funding acquisition, A.M.; Investigation, A.M., B.Ł., E.S., K.G. and A.S.; Methodology, A.M., B.Ł., E.S. and M.N.; Project administration, A.M.; Resources, A.M.; Software, B.Ł. and K.G.; Supervision, A.M.; Validation, A.M., B.Ł. and E.S.; Visualization, E.S., B.Ł. and M.K.; Writing—original draft, A.M., B.Ł. and E.S.; Writing—review & editing, A.M., B.Ł., E.S., M.K., M.N. and A.C. All authors have read and agreed to the published version of the manuscript. Funding This work was supported by the National Science Centre (grant no. 2016/23/D/NZ3/01660) and Medical University of Bialystok (grant numbers SUB/1/DN/21/005/1118 and SUB/1/DN/20/009/1118). Institutional Review Board Statement The study was approved by the Ethics Committee of the Medical University of Bialystok (permission number: R-I-002/187/2017, date of approval: 25 May 2017). All procedures were designed, conducted, and reported in compliance with the Declaration of Helsinki 1975, according to the guidelines for Good Clinical Practice. All subjects gave their informed consent to participate in the study. Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The data presented in this study are available on request. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Viability of cultured ADMSCs and their differentiation into adipocytes. (A) Viability of ADMSCs derived from both subcutaneous and visceral adipose tissues. (B) Quantification of neutral lipids content in the undifferentiated and differentiated ADMSCs derived from lean, obese(-) and obese(+) individuals. Representative images of undifferentiated and adipogenically differentiated ADMSCs are shown. Lipid droplets inside the cytoplasm were stained with Oil Red O solution [200× magnification]. (C) Quantification of the average cell size [µm2] and the number of lipid droplets per cell. Bars and whiskers represent the mean and SD, respectively. Number of patients equals three (measurements taken in triplicate). a–difference vs. Lean group in the studied tissue; b–difference vs. Obese(-) group in the studied tissue; * significantly different from undifferentiated cells, p < 0.05. Designation of the groups: Obese(-)–obese without metabolic syndrome patients; Obese(+)–obese with metabolic syndrome patients. (D) Representative histograms of FABP4 expression in adipocytes differentiated form sub- and visADMSCs obtained from lean and obese subjects. Figure 2 Gene (A) and protein expression (B) of AS160/TBC1D4 and its structural homolog TBC1D1 in adipocytes differentiated from ADMSCs derived from SAT and VAT of lean and morbidly obese women. Values are expressed in arbitrary units; the mean in adipocytes differentiated from the subADMSCs’ lean control was set at 1 or 100, respectively. (C) Representative Western Blot images are shown. (D) Measurement of 3H-palmitate uptake in fully differentiated adipocytes from subADMSCs and visADMSCs of lean and morbidly obese women. Values are expressed in DPM per mg of protein. a–difference vs. Lean group in the studied tissue; b–difference vs. Obese(-) group in the studied tissue; c–difference between adipocytes differentiated from visADMSCs vs. subADMSCs within the patient metabolic status. Data are presented as mean ± SD (n = 4 for each study group, measurements taken in duplicate for LCFA uptake and n = 3 for WB analysis). p < 0.05. Designation of the groups: Obese(-)–obese without metabolic syndrome patients; Obese(+)–obese with metabolic syndrome patients. Figure 3 Quantification of (A) mRNA and (B) protein abundance of CD36/SR-B2, FABPpm, FATP1 and FATP4 in fully differentiated adipocytes derived from subADMSCs and visADMSCs of lean and morbidly obese women. Values are expressed in arbitrary units; the mean in adipocytes differentiated from the subADMSCs’ lean control was set at 1 or 100, respectively. Representative Western Blot images are shown. a–difference vs. Lean group in the studied tissue; b–difference vs. Obese(-) group in the studied tissue; c–difference between adipocytes differentiated from visADMSCs vs. subADMSCs within the patient metabolic status. Data are presented as mean ± SD (n = 4 for each study group, measurements taken in duplicate for rtPCR method and n = 3 for WB analysis). p < 0.05. Designation of the groups: Obese(-)–obese without metabolic syndrome patients; Obese(+)–obese with metabolic syndrome patients. Figure 4 Flow cytometric analysis of cell surface expression of FA transporters in adipocytes derived from subADMSCs and visADMSCs of lean and morbidly obese women. The quantification of the cell populations staining positively for CD36/SR-B2, FATP1 and FATP4. Representative plots are shown. a–difference vs. Lean group in the studied tissue; b–difference vs. Obese(-) group in the studied tissue; c–difference between adipocytes differentiated from visADMSCs vs. subADMSCs within the patient metabolic status. Data are presented as mean ± SD (n = 3 for each study group, measurements taken in duplicate). p < 0.05. Designation of the groups: Obese(-)–obese without metabolic syndrome patients; Obese(+)–obese with metabolic syndrome patients. Figure 5 The size and the granularity of adipocytes differentiated from subADMSCs and visADMSCs of lean and morbidly obese women. The cells were characterized based on the forward scatter (FSC, relative to adipocytes size) and side scatter (SSC, relative to adipocytes internal structure) during flow cytometry analysis. (A) Adipocytes relative size and size distribution. (B) Adipocytes relative granularity and granularity distribution. Numbers indicate ranges used to classify the cells to different groups based on the cell size (I: 0–275, II: 275–550, III: 550–825, IV: 825–1100) or granularity (I: 0–250, II: 250–500, III: 500–750, IV: 750–1100). *—p < 0.05; **—p < 0.01; ***—p < 0.001; ****—p < 0.0001. The inner horizontal line represents the median. Whiskers: 25–75 percentile. n = 3 for each study group, measurements taken in triplicate. Designation of the groups: Obese(-)–obese without metabolic syndrome patients; Obese(+)–obese with metabolic syndrome patients. Figure 6 Lipid content and FA-profile in differentiated adipocytes derived from subADMSCs and visADMSCs of lean and morbidly obese women. The amount of total saturated fatty acid species (SFA), unsaturated fatty acid species (UNSFA) and FA distribution in (A) FFA, (B) DAG, (C) TAG. Values are expressed in nmol per mg of protein. a–difference vs. Lean group in the studied tissue; b–difference vs. Obese(-) group in the studied tissue; c–difference between adipocytes differentiated from visADMSCs vs. subADMSCs within the patient metabolic status. Bars and whiskers represent the mean and SD (n = 4 for each study group, measurements taken in triplicate). For the percentage fatty acid composition, numbers in brackets inside bars represent SD. p < 0.05. Designation of the groups: Obese(-)–obese without metabolic syndrome patients; Obese(+)–obese with metabolic syndrome patients. Figure 7 Protein content of enzymes involved in fatty acid synthesis and utilization (FASN, DGAT1, β-HAD, ATGL) in adipocytes differentiated from subADMSCs and visADMSCs of lean and morbidly obese women. Values are expressed in arbitrary units; the mean in adipocytes differentiated from the subADMSCs’ lean control was set at 100. Representative Western Blot images are shown. a–difference vs. Lean group in the studied tissue; b–difference vs. Obese(-) group in the studied tissue; c–difference between adipocytes differentiated from visADMSCs vs. subADMSCs within the patient metabolic status. Data are presented as mean ± SD (n = 3 for each study group, measurements taken in duplicate). p < 0.05. Designation of the groups: Obese(-)–obese without metabolic syndrome patients; Obese(+)–obese with metabolic syndrome patients. Figure 8 Graphical summary of the lipid profile of mature adipocytes derived from ADMSCs of lean and obese (with/without metabolic syndrome) postmenopausal women. A depot-specific pattern, with more pronounced changes present in the adipocytes obtained from subADMSCs was observed. Namely, morbid obesity triggered a relocation of fatty acid transporters i.e., CD36/SR-B2 and FATP4, from intracellular vesicles to the cell surface, which caused an increased LCFA influx (3H-palmitate uptake) and promoted the accumulation of lipids in these cells. This was accompanied by a greater lipid turnover rate in the cells of subcutaneous provenance (greater protein expression of lipolytic enzymes, i.e., ATGL and β-HAD). On the other hand, the adipocytes of visADMSCs origin had lower lipid content and expression of lipid metabolizing enzymes, i.e., ATGL, β-HAD and FASN, when compared to their counterparts of subADMSCs ancestry. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Mikłosz A. Nikitiuk B.E. Chabowski A. Using adipose-derived mesenchymal stem cells to fight the metabolic complications of obesity: Where do we stand? Obes. Rev. 2022 23 e13413 10.1111/obr.13413 34985174 2. Korac A. Srdic-Galic B. Kalezic A. Stancic A. Otasevic V. Korac B. Jankovic A. Adipokine signatures of subcutaneous and visceral abdominal fat in normal-weight and obese women with different metabolic profiles Arch. Med. Sci. 2021 17 323 336 10.5114/aoms/92118 33747267 3. Zhou Q. Fu Z. Gong Y. Seshachalam V.P. Li J. Ma Y. Liang H. Guan W. Lin S. Ghosh S. Metabolic Health Status Contributes to Transcriptome Alternation in Human Visceral Adipose Tissue During Obesity Obesity 2020 28 2153 2162 10.1002/oby.22950 32985130 4. Ibrahim M.M. Subcutaneous and visceral adipose tissue: Structural and functional differences Obes. Rev. 2010 11 11 18 10.1111/j.1467-789X.2009.00623.x 19656312 5. Ferrara D. Montecucco F. Dallegri F. Carbone F. Impact of different ectopic fat depots on cardiovascular and metabolic diseases J. Cell. Physiol. 2019 234 21630 21641 10.1002/jcp.28821 31106419 6. Barrea L. Vetrani C. Caprio M. El Ghoch M. Frias-Toral E. Mehta R.J. Mendez V. Moriconi E. Paschou S.A. Pazderska A. Nutritional management of type 2 diabetes in subjects with obesity: An international guideline for clinical practice Crit. Rev. Food Sci. Nutr. 2021 1 13 10.1080/10408398.2021.1980766 7. Abildgaard J. Ploug T. Al-Saoudi E. Wagner T. Thomsen C. Ewertsen C. Bzorek M. Pedersen B.K. Pedersen A.T. Lindegaard B. Changes in abdominal subcutaneous adipose tissue phenotype following menopause is associated with increased visceral fat mass Sci. Rep. 2021 11 14750 10.1038/s41598-021-94189-2 34285301 8. Yokokawa H. Fukuda H. Saita M. Goto K. Kaku T. Miyagami T. Takahashi Y. Hamada C. Hisaoka T. Naito T. An association between visceral or subcutaneous fat accumulation and diabetes mellitus among Japanese subjects Diabetol. Metab. Syndr. 2021 13 44 10.1186/s13098-021-00646-3 33853648 9. Ritter A. Friemel A. Roth S. Kreis N.N. Hoock S.C. Safdar B.K. Fischer K. Möllmann C. Solbach C. Louwen F. Subcutaneous and Visceral Adipose-Derived Mesenchymal Stem Cells: Commonality and Diversity Cells 2019 8 1288 10.3390/cells8101288 10. Fang J. Chen F. Liu D. Gu F. Wang Y. Adipose tissue-derived stem cells in breast reconstruction: A brief review on biology and translation Stem Cell Res. Ther. 2021 12 8 10.1186/s13287-020-01955-6 33407902 11. Baglioni S. Cantini G. Poli G. Francalanci M. Squecco R. Franco A. Borgogni E. Frontera S. Nesi G. Liotta F. Functional differences in visceral and subcutaneous fat pads originate from differences in the adipose stem cell PLoS ONE 2012 7 e36569 10.1371/journal.pone.0036569 22574183 12. Abkenar I.K. Rahmani-Nia F. Lombardi G. The effects of acute and chronic aerobic activity on the signaling pathway of the inflammasome NLRP3 complex in young men Medicina 2019 55 105 10.3390/medicina55040105 30991661 13. Alviz L. Tebar-García D. Lopez-Rosa R. Galan-Moya E.M. Moratalla-López N. Alonso G.L. Nava E. Llorens S. Pathogenic microenvironment from diabetic–obese visceral and subcutaneous adipocytes activating differentiation of human healthy preadipocytes increases intracellular fat, effect of the apocarotenoid crocetin Nutrients 2021 13 1032 10.3390/nu13031032 33806806 14. Tang Q.Q. Lane M.D. Adipogenesis: From stem cell to adipocyte Annu. Rev. Biochem. 2012 81 715 736 10.1146/annurev-biochem-052110-115718 22463691 15. Silva K.R. Liechocki S. Carneiro J.R. Claudio-Da-Silva C. Maya-Monteiro C.M. Borojevic R. Baptista L.S. Stromal-vascular fraction content and adipose stem cell behavior are altered in morbid obese and post bariatric surgery ex-obese women Stem Cell Res. Ther. 2015 6 72 10.1186/s13287-015-0029-x 25884374 16. Silva K.R. Côrtes I. Liechocki S. Carneiro J.R.I. Souza A.A.P. Borojevic R. Maya-Monteiro C.M. Baptista L.S. Characterization of stromal vascular fraction and adipose stem cells from subcutaneous, preperitoneal and visceral morbidly obese human adipose tissue depots PLoS ONE 2017 12 e0174115 10.1371/journal.pone.0174115 28323901 17. Posada-González M. Villagrasa A. García-Arranz M. Vorwald P. Olivera R. Olmedillas-López S. Vega-Clemente L. Salcedo G. García-Olmo D. Comparative Analysis Between Mesenchymal Stem Cells From Subcutaneous Adipose Tissue and Omentum in Three Types of Patients: Cancer, Morbid Obese and Healthy Control Surg. Innov. 2021 29 9 21 10.1177/15533506211013142 33929270 18. Suárez-Cuenca J.A. De La Peña-Sosa G. De La Vega-Moreno K. Banderas-Lares D.Z. Salamanca-García M. Martínez-Hernández J.E. Vera-Gómez E. Hernández-Patricio A. Zamora-Alemán C.R. Domínguez-Pérez G.A. Enlarged adipocytes from subcutaneous vs. visceral adipose tissue differentially contribute to metabolic dysfunction and atherogenic risk of patients with obesity Sci. Rep. 2021 11 1831 10.1038/s41598-021-81289-2 33469087 19. Lefevre C. Chartoire D. Ferraz J.C. Verdier T. Pinteur C. Chanon S. Pesenti S. Vieille-Marchiset A. Genestier L. Vidal H. Obesity activates immunomodulating properties of mesenchymal stem cells in adipose tissue with differences between localizations FASEB J. 2021 35 e21650 10.1096/fj.202002046RR 33993539 20. Mikłosz A. Łukaszuk B. Supruniuk E. Grubczak K. Moniuszko M. Choromańska B. Myśliwiec P. Chabowski A. Does TBC1D4 (AS160) or TBC1D1 Deficiency Affect the Expression of Fatty Acid Handling Proteins in the Adipocytes Differentiated from Human Adipose-Derived Mesenchymal Stem Cells (ADMSCs) Obtained from Subcutaneous and Visceral Fat Depots? Cells 2021 10 1515 10.3390/cells10061515 34208471 21. Pfaffl M.W. A new mathematical model for relative quantification in real-time RT-PCR Nucleic Acids Res. 2001 29 e45 10.1093/nar/29.9.e45 11328886 22. Glatz J.F.C. Nabben M. Luiken J.J.F.P. CD36 (SR-B2) as master regulator of cellular fatty acid homeostasis Curr. Opin. Lipidol. 2022 33 103 111 10.1097/MOL.0000000000000819 35125400 23. Badimon L. Cubedo J. Adipose tissue depots and inflammation: Effects on plasticity and resident mesenchymal stem cell function Cardiovasc. Res. 2017 113 1064 1073 10.1093/cvr/cvx096 28498891 24. Arner P. Rydén M. Human white adipose tissue: A highly dynamic metabolic organ J. Intern. Med. 2021 291 611 621 10.1111/joim.13435 25. Wang T. Sharma A.K. Wolfrum C. Novel insights into adipose tissue heterogeneity Rev. Endocr. Metab. Disord. 2021 23 5 12 10.1007/s11154-021-09703-8 34935088 26. Thompson B.R. Lobo S. Bernlohr D.A. Fatty acid flux in adipocytes: The in’s and out’s of fat cell lipid trafficking Mol. Cell. Endocrinol. 2010 318 24 33 10.1016/j.mce.2009.08.015 19720110 27. Nickerson J.G. Alkhateeb H. Benton C.R. Lally J. Nickerson J. Han X.X. Wilson M.H. Jain S.S. Snook L.A. Glatz J.F.C. Greater transport efficiencies of the membrane fatty acid transporters FAT/CD36 and FATP4 compared with FABPpm and FATP1 and differential effects on fatty acid esterification and oxidation in rat skeletal muscle J. Biol. Chem. 2009 284 16522 16530 10.1074/jbc.M109.004788 19380575 28. Stahl A. A current review of fatty acid transport proteins (SLC27) Pflug. Arch. Eur. J. Physiol. 2004 447 722 727 10.1007/s00424-003-1106-z 29. Bonen A. Tandon N.N. Glatz J.F.C. Luiken J.J.F.P. Heigenhauser G.J.F. The fatty acid transporter FAT/CD36 is upregulated in subcutaneous and visceral adipose tissues in human obesity and type 2 diabetes Int. J. Obes. 2006 30 877 883 10.1038/sj.ijo.0803212 30. Gertow K. Pietiläinen K.H. Yki-Järvinen H. Kaprio J. Rissanen A. Eriksson P. Hamsten A. Fisher R.M. Expression of fatty-acid-handling proteins in human adipose tissue in relation to obesity and insulin resistance Diabetologia 2004 47 1118 1125 10.1007/s00125-004-1417-4 15168018 31. Mikłosz A. Łukaszuk B. Żendzian-Piotrowska M. Brańska-Januszewska J. Ostrowska H. Chabowski A. Challenging of AS160/TBC1D4 Alters Intracellular Lipid milieu in L6 Myotubes Incubated With Palmitate J. Cell. Physiol. 2017 232 2373 2386 10.1002/jcp.25632 27714805 32. Glatz J.F.C. Luiken J. From fat to FAT (CD36/SR-B2): Understanding the regulation of cellular fatty acid uptake Biochimie 2017 136 21 26 10.1016/j.biochi.2016.12.007 28013071 33. Binnert C. Koistinen H.A. Martin G. Andreelli F. Ebeling P. Koivisto V.A. Laville M. Auwerx J. Vidal H. Fatty acid transport protein-1 mRNA expression in skeletal muscle and in adipose tissue in humans Am. J. Physiol.-Endocrinol. Metab. 2000 279 1072 1079 10.1152/ajpendo.2000.279.5.E1072 11052962 34. Wajchenberg B.L. Lé B. Wajchenberg O. Subcutaneous and Visceral Adipose Tissue: Their Relation to the Metabolic Syndrome Endocr. Rev. 2000 21 697 738 11133069 35. Joe A.W.B. Lin Y. Even Y. Vogl A.W. Rossi F.M.V. Depot-specific differences in adipogenic progenitor abundance and proliferative response to high-fat diet Stem Cells 2009 27 2563 2570 10.1002/stem.190 19658193 36. Kim B. Lee B. Kim M.K. Gong S.P. Park N.H. Chung H.H. Kim H.S. No J.H. Park W.Y. Park A.K. Gene expression profiles of human subcutaneous and visceral adipose-derived stem cells Cell Biochem. Funct. 2016 34 563 571 10.1002/cbf.3228 27859461 37. Zacharia A. Saidemberg D. Mannully C.T. Kogan N.M. Shehadeh A. Sinai R. Zucker A. Bruck-Haimson R. Goldstein N. Haim Y. Distinct infrastructure of lipid networks in visceral and subcutaneous adipose tissues in overweight humans Am. J. Clin. Nutr. 2020 112 979 990 10.1093/ajcn/nqaa195 32766878 38. Małodobra-Mazur M. Cierzniak A. Pawełka D. Kaliszewski K. Rudnicki J. Dobosz T. Metabolic Differences between Subcutaneous and Visceral Adipocytes Differentiated with an Excess of Saturated and Monounsaturated Fatty Acids Genes 2020 11 1092 10.3390/genes11091092 39. Berndt J. Kralisch S. Klöting N. Ruschke K. Kern M. Fasshauer M. Schön M.R. Stumvoll M. Blüher M. Adipose triglyceride lipase gene expression in human visceral obesity Exp. Clin. Endocrinol. Diabetes 2008 116 203 210 10.1055/s-2007-993148 18072017 40. Gaidhu M.P. Anthony N.M. Patel P. Hawke T.J. Ceddia R.B. Dysregulation of lipolysis and lipid metabolism in visceral and subcutaneous adipocytes by high-fat diet: Role of ATGL, HSL, and AMPK Am. J. Physiol.-Cell Physiol. 2010 298 961 971 10.1152/ajpcell.00547.2009 41. Yao-Borengasser A. Varma V. Coker R.H. Ranganathan G. Phanavanh B. Rasouli N. Kern P.A. Adipose triglyceride lipase expression in human adipose tissue and muscle. Role in insulin resistance and response to training and pioglitazone Metabolism 2011 60 1012 1020 10.1016/j.metabol.2010.10.005 21129760

PMC009xxxxxx/PMC9099625.txt

==== Front Cancers (Basel) Cancers (Basel) cancers Cancers 2072-6694 MDPI 10.3390/cancers14092135 cancers-14-02135 Article Functional Outcomes in Head and Neck Cancer Patients Riechelmann Herbert 1 https://orcid.org/0000-0001-9176-7415 Dejaco Daniel 1* Steinbichler Teresa Bernadette 1 Lettenbichler-Haug Anna 1 Anegg Maria 1 Ganswindt Ute 2 https://orcid.org/0000-0001-7025-9729 Gamerith Gabriele 3 https://orcid.org/0000-0003-0518-3144 Riedl David 4 Bloemena Elisabeth Academic Editor 1 Department of Otorhinolaryngology—Head and Neck Surgery, Medical University of Innsbruck, 6020 Innsbruck, Austria; herbert.riechelmann@i-med.ac.at (H.R.); teresa.steinbichler@i-med.ac.at (T.B.S.); praxis@hno-am-dez.at (A.L.-H.); maria.anegg@student.i-med.ac.at (M.A.) 2 Department of Radiation-Oncology, Medical University of Innsbruck, 6020 Innsbruck, Austria; ute.ganswindt@i-med.ac.at 3 Internal Medicine V, Department of Hematology & Oncology, Medical University Innsbruck, 6020 Innsbruck, Austria; gabriele.gamerith@i-med.ac.at 4 Department of Psychiatry, Psychotherapy, Psychosomatics and Medical Psychology, Medical University of Innsbruck, 6020 Innsbruck, Austria; david.riedl@tirol-kliniken.at * Correspondence: daniel.dejaco@i-med.ac.at; Tel.: +43-512-504-23142 25 4 2022 5 2022 14 9 213516 3 2022 19 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Simple Summary With increasing long-term survivorship of head and neck cancer (HNC), the functional outcomes are gaining importance. Recently, a tool for the rapid clinical assessment of the functional status in HNC-patients based on observable clinical criteria, termed “HNC-Functional InTegrity (FIT) Scales, was validated. Here, the functional outcomes of 681 newly diagnosed HNC-patients were reported using the HNC-FIT Scales. A normal/near-normal outcome in all six functional domains was observed in 61% of patients, with individual scores of 79% for food intake, 89% for breathing, 84% for speech, 89% for pain, 92% for mood, and 88% for neck and shoulder mobility. Clinically relevant impairment in at least one functional domain was observed in 30% of patients, and 9% had loss of function in at least one functional domain. Thus, clinically relevant persistent functional deficits in at least one functional domain must be expected in 40% of HNC-patients. The treatment of these functional deficits is an essential task of oncologic follow-up. Abstract With the increase in long-term survivorship of head and neck cancer (HNC), the functional outcomes are gaining importance. We reported the functional outcomes of HNC patients using the HNC-Functional InTegrity (FIT) Scales, which is a validated tool for the rapid clinical assessment of functional status based on observable clinical criteria. Patients with newly diagnosed HNC treated at the Medical University of Innsbruck between 2008 and 2020 were consecutively included, and their status in the six functional domains of food-intake, breathing, speech, pain, mood, and neck and shoulder mobility was scored by the treating physician at oncological follow-up visits on a scale from 0 (loss of function) to 4 (full function). HNC-FIT scales were available for 681 HNC patients at a median of 35 months after diagnosis. The response status was complete remission in 79.5%, 18.1% had recurrent or persistent disease, and 2.4% had a second primary HNC. Normal or near-normal scores (3 and 4) were seen in 78.6% for food intake, 88.7% for breathing, 83.7% for speech, 89% for pain, 91.8% for mood, and 87.5% for neck and shoulder mobility. A normal or near-normal outcome in all six functional domains was observed in 61% of patients. Clinically relevant impairment (score 1–2) in at least one functional domain was observed in 30%, and 9% had loss of function (score 0) in at least one functional domain. The main factors associated with poor functional outcome in a multivariable analysis were recurrence or persistent disease, poor general health (ASA III and IV), and higher T stage. Particularly, laryngeal and hypopharyngeal tumors impaired breathing and speech function, and primary radiation therapy or concomitant systemic therapy and radiotherapy worsened food intake. Clinically relevant persistent functional deficits in at least one functional domain must be expected in 40% of the patients with HNC. The treatment of these functional deficits is an essential task of oncologic follow-up. head and neck neoplasms health status functional outcomes surveys and questionnaires questionnaire design ==== Body pmc1. Introduction Recent advances have significantly improved the survival of patients with head and neck cancer (HNC). This has led to more long-term survivors [1,2,3]. Although survival is the most important outcome for HNC patients [4,5,6], other dimensions of treatment outcome such as physical status and functional abilities, psychological status and wellbeing, social interactions, and economic status are becoming increasingly important as a result of this trend [7,8]. These dimensions of outcome are most often measured with quality of life (QoL) instruments, where QoL means the patient’s subjective perception of their state and abilities in these domains [9,10]. Currently, several instruments are available to assess health-related QoL in HNC patients [11,12,13,14,15,16,17,18,19,20]. However, those QoL measurements are subjected to various psychological factors [21,22], response shifts [23,24], and social desirability biases [25,26,27,28,29], which do not necessarily reflect the severity of functional impairment and symptoms [30,31]. In contrast, functional endpoints measure the degree to which patients can perform an activity. Specifically in the head and neck region, functional endpoints include seeing, hearing, smelling, tasting, speech, breathing, eating, and neck and shoulder mobility. The measurements of these functions can be patient-reported, observer-rated, or measured by objective tests such as a barium swallow [10,32]. Unlike QoL measures, functional endpoints strive for objectivity and inter-individual comparability. Patients with the same functional endpoint should also have the same functional scores. Therefore, it makes sense to link functional scores to external criteria that are observable. There are few instruments for the standardized assessment of HNC-related functional endpoints [16,33,34,35]. We recently developed and validated the Head and Neck Functional Integrity Scales (HNC-FIT Scales) for German speaking HNC-patients. The HNC-FIT Scales represent an instrument for a rapid clinician-rated assessment of functional status during routine oncologic follow-up. Functional domains commonly affected in HNC patients are scored. The scores are calibrated using observable external criteria. They are clinically oriented and indicate, for example, whether the patient is dependent on a tracheostomy, feeding tube, opiates, or antidepressants at each follow-up visit [36]. Here, we report the functional outcome of patients with incident HNC treated at the Department of Otorhinolaryngology, Head and Neck Surgery, Medical University of Innsbruck between 2008 and 2020 using the HNC-FIT Scales. The effects of several clinical covariates including age, gender, tumor site, stage, and treatment modality on functional outcome were investigated. Our objective was to describe the outcomes of HNC survivors in frequently affected functional domains and to identify factors associated with favorable functional outcome. 2. Materials and Methods 2.1. Study Population Patients with newly diagnosed carcinoma of the head and neck that were treated at the Department of Otorhinolaryngology, Head and Neck Surgery, Medical University of Innsbruck between 2008 and 2020 were consecutively included. The exclusion criteria were carcinoma of the thyroid, esophagus, eye, or brain or spinal cord, as well as melanoma, sarcoma, lymphoma, other non-carcinomas, carcinomas of head and neck skin, benign neoplasms, intraepithelial neoplasia, dysplasia, inflammatory pseudotumors, and head and neck metastases from distant primary tumor sites. The patient data were prospectively recorded in the clinical tumor registry of the department. Demographic data were recorded at initial diagnosis and included gender, age in years, and the American Society of Anesthesiology (ASA) physical status score as a simple measure of general health status [37]. ASA scores were dichotomized into ASA I/II and ASA III/IV. Further host factors included smoking history (≤10 vs. >10 pack years) [38] and alcohol consumption (daily vs. less than daily). The recorded medical data included tumor site, which was grouped into oral cavity, oropharynx, hypopharynx, larynx, and other sites. The UICC-TNM staging that was valid at the time of the initial diagnosis was used. For clinical T-, N-, and UICC stage, only the first numerical digit with no further subclassification was used. The clinical response status at the last follow-up visit was grouped into complete remission, recurrent, or persistent disease, as well as second primary HNC. The study was conducted according to the principles of the Declaration of Helsinki (59th version, 21 October 2008). The research protocol was approved by the Ethics Committee of the Medical University of Innsbruck (ethics committee number 1182/2019). All patients had given written consent for their data to be used anonymously. 2.2. Treatment The treatment modality was recommended by an interdisciplinary head and neck tumor board that is in line with National Comprehensive Cancer Network (NCCN) Guidelines. The treatments included upfront surgical resection, radiotherapy (RT), and systemic therapy (ST). The treatment modalities were categorized into surgery only, upfront surgery and postoperative radiotherapy (PORT), upfront surgery and postoperative concomitant ST and RT, primary concomitant ST and RT, and primary RT. The surgical procedures included transoral laser microsurgery, transfacial or transcervical tumor resections, pedicled or free-flap reconstructions, and uni- or bilateral selective or comprehensive neck dissections. RT was applied as a first-line treatment or as PORT in advanced disease if no high-risk factors were present. High risk factors included involved margins or extracapsular lymph node extension [39]. Patients with high risk factors received postoperative ST and RT, as did patients with advanced disease treated without upfront surgery. RT was usually carried out in conventional fractions with 1.8–2.0 Gy daily, five times a week, as three-dimensional conformal radiotherapy or intensity-modulated radiation therapy. The total dose in the area of an untreated primary tumor or in the region of an untreated primary lymph node metastases was 70–72 Gy. In regions associated with a high risk of existing or persisting tumor cells, the dose was 60 to a maximum of 66 Gy, and in areas of physiological anatomical lymphatic drainage, the dose was about 50 Gy. For postoperative patients with high-risk features and for patients with advanced disease who received primary non-surgical treatment, concomitant systemic therapy (ST/RT) was applied [40,41]. Concomitant systemic therapy consisted of either cisplatin, at 100 mg/m2 on days 1, 22, and 43, or 25 mg/m2 on days 1–4 and 29–32 [42]. Alternatively, mitomycin C, at 10 mg/m2 (max. 15 mg total), was used on days 1 and 29, and 5-fluorouracil, at 600 mg/m2 (24 h infusion), was used on days 1–5 and 29–33 [43] and was prescribed for patients not suitable for cisplatin treatment. Alternatively, RT was combined with cetuximab with a loading dose of 400 mg/m2 once a week before the start of radiotherapy followed by 250 mg/m2 weekly for the duration of the RT in frail patients [44]. Depending on their needs, all patients were offered rehabilitation treatment focusing on swallowing and nutrition, speech, physical therapy, psycho-oncological counseling, assistive devices, and comprehensive interdisciplinary programs. 2.3. Head and Neck Functional InTegrity Scales The HNC-FIT scale is a matrix of six verbal rating scales reflecting the functional domains of food intake, breathing, speech, pain, mood, and neck and shoulder mobility. Only these higher-level functional domains are recorded, e.g., food intake, but not related specific functions and conditions such as chewing, swallowing, salivating, tasting, aspiration, xerostomia, trismus, or dental problems [45]. Each functional domain is divided into five functional levels (Supplementary Figure S1). These levels are not solely based on the patient’s or examiner’s subjective assessment, such as no, mild, moderate, or severe. Rather, they are anchored to observable external criteria. For example, breathing function is anchored to the external criterion of the need for a tracheotomy, and food-intake function is anchored to dependence on a feeding tube. Physicians received detailed instructions on how to complete the HNC-FIT scales (Supplementary text S1 Instructions for clinicians). When taking the interim oncologic history during oncologic follow-up, the treating physician interviewed the patient and marked the respective functional levels in this matrix. This structured face-to-face interview lasted a median of 1.2 min and could be entirely integrated into the taking of the interim medical history [36]. Verbal ratings were numerically coded from 0 (loss of function) to 4 (normal function) Low scores reflected disease related functional impairments. These numeric codes were transferred to the clinical tumor registry by a medical documentation assistant after each follow-up visit. For data evaluation, the relative count (percent) of patients at each level of each functional domain was calculated. In the validation study, most of the healthy control subjects examined reported level 4, corresponding to normal function for all functions recorded, with few reporting level 3, which corresponds to a slightly impaired near-normal function [36]. A functional level <3 did not occur for any of the recorded functional domains in healthy controls. Therefore, for the evaluation we divided the data into levels 0–2, which corresponds to a clinically important dysfunction, and level 3 and 4, which correspond to a near-normal or normal function. The functional data from the last oncologic follow-up visit were evaluated in this study. 2.4. Data Analysis Time intervals between the first diagnosis and the last HNC-FIT scale assessment were grouped into <24 months, >=24 months, <60 months, and >=60 months. The absolute and relative patient counts were tabulated for the five levels of the six functional domains (Table 1). The effects of the categorical covariates including gender; age group; ASA-Score; smoking history; alcohol consumption; tumor site; T-, N-, and UICC-stage; treatment modality; and p16 positivity on the dichotomized functional outcome (count of patients with integrity scores 0 to 2 vs. integrity scores 3 and 4) of the six functional domains were compared with Chi-square tests (Table 2). Factors with p-values < 0.2 in the Chi-square tests were included as independent variables in a binary logistic regression model. The dichotomized functional outcome (count of patients with integrity scores 0 to 2 vs. integrity scores 3 and 4) were modeled as the dependent variable. A main effects model with indicator coding for categorical variables with backstep elimination was used. The epsilon criterion to detect collinearity was set to 10−5; the model fit was tested with the Hosmer–Lemeshow test. The Hosmer–Lemeshow tests indicated good model fits, and there was no collinearity problem. The response status was analyzed separately because, unlike the factors, it was not known at the time of initial diagnosis and could also be considered in the choice of therapy. The statistical analyses were performed using SPSS 27 (IBM, Armonk, NY, USA). 3. Results 3.1. Study Population During the observation period, 1236 patients with incident HNC were treated at our department. Data from at least one HNC-FIT scale assessment were available in 681 patients, which comprises the study population. Of these, 137 (20%) were female. UICC stage I and II disease were observed in 251 patients (36.8%). Of these, 106 patients had stage I or II laryngeal carcinoma. Further demographic and medical data are outlined in Table 2. In the study population, 79.5% were in complete remission, 18.1% had recurrent or persistent disease, and 2.4% suffered from a second head and neck primary at their last follow-up visit. The median time interval between the initial diagnosis and the last follow-up visit in 681 patients was 46 (95%CI 42 to 50) months (46). The last follow-up was less than 24 months after the date of initial diagnosis in 36%, between 24 and 60 months in 38%, and >=60 months in 26%. 3.2. Functional Status at Last Follow-Up At the last follow-up visit, severe limitations with functional integrity scores of 0 to 2 most frequently affected the functional domain of food intake (21.5%), followed by the functional domain of speech (16.3%; Figure 1). Of 681 patients, 14.3% depended on a gastrostomy or feeding tube, 10.2% on a tracheotomy, 10% were not able to communicate by telephone, 6.7% depended on opioids, 8.2% were on antidepressants, and 4.1% were not able to drive a car or comb their hair due to head and neck stiffness (Table 1). The frequencies of patients with impaired function (integrity scores 0–2) vs. normal or near-normal function (integrity scores 3 and 4) were analyzed with Chi-square tests. The normal or near-normal functional outcome was recorded, e.g., for food intake in 78% of male patients and 80% of female patients (Chi-square p > 0.05). The ASA score, tumor site, T-stage, treatment modality, and response status most frequently affected the functional outcome in a univariate analysis (Table 2). The relative frequencies of the normal or near-normal outcomes (integrity scores 3 and 4) of the six functional domains by tumor site, T-stage, and treatment modality are depicted in Figure 2. 3.3. Binary Logistic Regression Analyses Patients with a complete response at the last follow-up visit had a better functional outcome for food intake (OR 2.8; 95%CI 1.5 to 5.3; p = 0.001), pain (OR 6.9; 95%CI 3.6 to 13.4; p < 0.001) and neck and shoulder mobility (OR 2.6; 95%CI 1.4 to 4.8; p = 0.003) when compared with patients with recurrent or persistent disease (see Supplementary Tables S1–S6). Considering the outcome results for the individual functional domains independent of response status, the following results emerged. In the univariate analysis of food intake, p-values of <0.2 were observed for age, ASA score, smoking, drinking, tumor site, T-stage, N-stage, UICC-stage, and treatment modality (Table 2). The odds ratios needed to achieve normal or near-normal food intake were calculated. In this multivariable model, patients with ASA scores of I/II had 2.4 times higher odds to achieve a normal or near-normal outcome for food intake than patients with ASA scores of III/IV (p = 0.001). Further significant factors in this multivariable model included T-stage, tumor site, and treatment modality (Table 3). In the univariate analysis for breathing function, p-values of <0.2 were observed for gender, ASA score, smoking, tumor site, T-stage, UICC-stage, p16 positivity, and treatment modality (Table 2). In this multivariable model, patients with ASA scores of I/II had a 2.9 times better chance to achieve a normal or near-normal outcome for breathing than patients with ASA scores of III/IV (p = 0.001). Further significant factors in this multivariable model included T-stage and tumor site; however, the p-value for p16 status missed the 0.05 limit (Table 4). In the univariate analysis, p-values of <0.2 for speech function were observed for gender, age, ASA score, smoking, tumor site, T-stage, UICC-stage, p16 positivity, and treatment modality (Table 1). In the logistic regression model, patients with ASA scores of I/II had a 2.7 times better chance to achieve a normal or near-normal outcome for speech than patients with ASA scores of III/IV (p = 0.002). Notably, primary concomitant radiotherapy/systemic therapy, which served as the reference treatment modality, resulted in the highest percentage of a normal or near-normal speech function (Table 5). In the univariate analysis, p-values of <0.2 for pain were observed for treatment modality, tumor site, T-stage, N-stage, UICC-stage, and ASA score (general health condition). In the backstep logistic regression model, all variables except ASA score (p = 0.01) were excluded. The OR to have no or near-normal disease-related pain was 2.9 (95%CI 1.6 to 5.2; p < 0.001). In the univariate analysis, p-values of <0.2 for mood were observed for gender and ASA score (general health condition). In the logistic regression model, only gender remained in the model (p = 0.02). The OR for men to report a normal or near-normal mood was 2.2 (95%CI 1.1 to 4.4). In the univariate analysis, p-values of <0.2 for neck and shoulder mobility were observed for treatment modality, tumor site, T stage, N stage truncated, UICC stage, p16, and ASA score. In the logistic regression model, patients treated only with surgery (reference primary ST/RT; OR 4.2; 95%CI 1.3 to 14.0; p = 0.02) had a better chance of reporting normal or near-normal neck and shoulder mobility. 3.4. Poorest Functional Otucome If the functional domain with the poorest outcome would affect the patient the most, the functional score of the domain with the poorest outcome was determined for each patient. A functional score of zero (worst possible outcome) in at least one domain was observed in 62 patients, whereas 185 patients had a normal functional outcome in all six domains (Table 6). Normal or near-normal function (functional integrity) in all six functional domains was observed in 61% of 681 HNC patients. In other words, almost 40% of HNC patients had at least one serious functional impairment. 4. Discussion Survival, health-related quality of life, treatment-related toxicity, and functional outcome are distinct albeit related outcome parameters in patients with head and neck cancer [32]. We examined the outcomes in six essential commonly affected functional domains in HNC patients [36] by using the disease specific HNC-FIT scales, an intentionally minimalistic, validated assessment tool that enables a time-saving assessment of these functions by the clinician using observable external criteria [36]. An additional major advantage of the tool is the standardized and combined assessment, thereby minimizing the risk to under- or overestimate impairments or to overlook essential ones (Table 6). By reporting the percentages of affected patients and dichotomizing the results into normal (scores 1–2) versus impaired (scores 3–4), these results can be more easily compared with other studies and the exact meaning of the functional results becomes directly apparent (Table 1). The evaluation was performed in 681 survivors out of 1236 HNC patients that were hospitalized and diagnosed with head and neck carcinomas in our department between 2008 and 2020. Even though this cohort can be considered representative for a tertiary oncology referral center in Central Europe, because of their survival, the patients in this study inherently represent a cohort with positive prognostic factors, which was not representative of HNC patients at the time of initial diagnosis (data not shown). The median follow-up time of the survivors was 46 (95%CI 42 to 50) months [46]—so, the majority of patients may be regarded as long-term survivors. All patients were offered a comprehensive rehabilitation program [47] and most patients took advantage of this opportunity. Not surprisingly, patients with a complete response at last follow-up had a significantly better outcome in food intake (p = 0.001), pain (p < 0.001) and neck and shoulder mobility (p = 0.003) than patients with recurrent or persistent disease. Notably, pain was associated with recurrence/persistence in the multivariable logistic regression (OR 6.9; 95%CI 3.6 to 13.4). Given that response status was the only factor recorded at the last follow-up, its effect on functional outcome was considered separately from the other demographic and medical factors recorded at diagnosis. This did not lead to a relevant bias in the effect estimates of the latter factors (see Supplementary Tables S1–S6). The influence of demographic and medical factors such as T stage, tumor site, and treatment modality on important domains of functional outcomes were synoptically depicted in star diagrams (Figure 2). Interestingly, general health had an impact on all functional domains studied (Supplementary Figure S2), possibly reflecting a higher resilience with better general health status. While all functional domains deteriorated with increasing T stage (Figure 2a), the tumor site primarily affected food intake, breathing, and speech, with particularly poor outcomes in hypopharyngeal cancer (Figure 2b). The functional domain most affected by treatment modality was food intake, with better results in patients treated with upfront surgery, whereas better speech outcomes were observed in patients treated with concomitant systemic treatment and radiotherapy (Figure 2c). In line with previous publications, food intake was most frequently affected in HNC survivors [48]. This functional domain covers various functions or symptoms of nutrition (eating and drinking) including swallowing, dysphagia, or trismus [45]. For the food-intake domain, the observable external criteria included dependence on a feeding tube and normality of diet. In line with the results of a previous study [49], 21.4% of patients were either dependent on a PEG or feeding tube or could only eat liquid or soft food. In a logistic regression with backward elimination, the frequency of normal or near-normal food intake depended on the general health condition (ASA score), smoking history, T-stage, tumor site, and treatment modality (Table 3). These factors have already been identified in previous studies [21,50,51,52,53]. Although food intake after primary RT and primary ST/RT was better in this study than in previous reports [54,55], it was still worse than after upfront surgery (Figure 2c), which is also consistent with previous publications [53,56,57,58]. However, this evaluation included many patients who had received three-dimensional conformal radiotherapy and not intensity-modulated radiation therapy, which yields better swallowing outcomes [59,60]. The observable external criteria for breathing function included dependence on a tracheotomy and dyspnea in patients without a tracheotomy. Concerning breathing function, the frequency of normal or near-normal breathing significantly depended on the general health condition, T-stage, and tumor site (Table 4), with hypopharyngeal and laryngeal cancers being associated with the worst outcomes for breathing (Figure 2b). This is in line with a previous report [55]. Although not significant, p16 positive patients had a trend for better odds to achieve normal or near-normal breathing [61]. In line with previous publications, treatment modality had a comparatively low impact on breathing [62,63,64]. Phonation and articulation are the key requirements for the understandability of speech. Speech intelligibility in telephone conversations served as an external criterion for this functional domain. Speech was the functional domain most affected by tumor stage; however, consistent with previous data, tumor stage had a major impact on several functional domains [10,49,53,65]. Interestingly, T-stage had a substantially higher impact on functional outcome than N-stage or UICC-stage in all functional domains. Poor understandability of speech in patients with laryngeal and hypopharyngeal cancers has been reported previously [21]. When compared with primary ST/RT, patients treated with upfront surgery had markedly poorer odds to achieve normal or near-normal speech intelligibility. This is in line with previous reports [21,55,66,67]. Normal or near-normal speech was particularly rare in patients receiving upfront surgery followed by concomitant ST/RT (Figure 2c). Pain was classified as a body function in the HNC-FIT scale following the International Classification of Functioning, Disability, and Health [68]. The use of pain medication served as the external criterion for the grading of pain. A standardized pain therapy was carried out in cooperation with the Pain Clinic of the Medical University of Innsbruck in accordance with WHO guidelines [69,70]. Interestingly, 89% of HNC patients reported no or almost no pain and 76.4% reported taking no pain medication at all. This low prevalence of pain in HNC-survivors is not consistent with previously reported data. In most pain studies in HNC survivors, chronic pain, often requiring opiates, is a common health problem [71,72,73,74]. The comparatively low consumption of analgesics may be due to regional differences of analgesic consumption [75] and social desirability bias. In logistic regression, only the ASA score (p < 0.001) and T stage (p = 0.01) had a significant effect on the proportion of patients with no or almost no pain. This is consistent with previous publications [76,77]. The use of antidepressants served as an objective external criterion for depressed mood. In the study population, 87.5% reported normal or near-normal mood. This is in line with recent publications in cancer patients [78,79,80], which reported moderate to severe depression in 10–15% of cancer survivors. This compares to an estimated 4.4% of the general population worldwide [81]. However, HNC patients experienced the highest rates of depressive disorder of all oncology patients [82]. Suicide rates among patients with HNC in the USA are more than three times higher than in the general population [83]. In a logistic regression analysis, men had a higher chance to achieve normal or near-normal mood than women (OR 2.2; 95%CI 1.1 to 4.4; p = 0.02), which is in line with a previous publication [84]. A higher prevalence of depression was observed when depression scales were used [85,86]. Bamonti and coworkers reported a correlation between pain and depression [87], and this was also observed in this study. Another function often impaired by HNC treatments is neck and shoulder mobility. In the represented cohort, 87.5% of the patients had normal or near-normal mobility with a prevalence of patients treated only with surgery (Figure 2), especially when compared to patients treated with primary ST/RT (OR 4.2; 1.3 to 14.0; p = 0.02). Do and co-authors reported, based on the Neck Pain and Disability Scale (NDI), values of more than 20 in patients with spinal accessary nerve injury after head and neck cancer surgery, which corresponds with mildly impaired neck mobility. Range of motion assessments suggest a reduction in head and neck range of motion below 20% in most HNC patients [88,89,90]. Overall, we observed a better functional outcome in HNC survivors than expected based on previous publications on this topic. In many previous studies on functional outcome, patients with advanced tumor stages and/or functionally problematic tumor sites such as the oral cavity, oropharynx, or hypopharynx were selected [91]. In this study group, 251 patients (36.8%) had UICC stage I and II disease and 106 had stage I or II laryngeal carcinoma with a functionally favorable prognosis. Suarez-Cunqueiro and co-authors, for instance, reported normal speech (i.e., absence of problems) in 36.2% and normal swallowing in 24.6% of 851 patients treated with radical surgery for oral and oropharyngeal cancer [92]. Oozer and co-authors used the Performance Status Scale for Head and Neck Cancer Patients in 79 laryngectomies with a mean age of 64 years. Speech was at least understandable most of the time in 62% and normalcy of diet, including full diet with no restrictions, was reported in 76% [93]. In 240 patients with advanced unresectable oropharyngeal and hypopharyngeal carcinoma treated with an aggressive concomitant ST/RT regimen, 25% of patients were dependent on a feeding tube two years after treatment [94]. Of 166 patients, who were disease free five years following the start of treatment in the RTOG9003 trials, 13 (7.8%) were feeding tube dependent [95]. In 181 surgically treated HNC patients, List and coworkers observed a normalcy of diet with a score of >50 in 52% and an understandability of speech of >50 in 77% [14]. Two additional factors, which potentially influence the better functional outcome observed in this study, should be discussed. Firstly, the inclusion criteria for the present study were comparatively exclusive. Rather than including all upper aerodigestive tract cancers, only HNC patients typically treated by head and neck surgeons at our institution were included. Consequently, patients suffering from esophageal cancer who typically showed significant functional impairment, for example, were excluded. Secondly, approximately one third of the included patients’ tumor sites were categorized as “other site” (see Table 2). Patients subsumed to this group mostly suffered from salivary gland or sinonasal cancer. The better functional outcome observed in this study (Table 2, Table 3, Table 4 and Table 5) might be caused by the design of the HNC-FIT Scales itself. While cancer of the oral cavity, pharynx, and larynx typically impairs functions included by the HNC-FIT Scales, cancer of the salivary glands or paranasal sinuses also leads to functional impairment. However, these impairments might not be identified by the HNC-FIT Scales (e.g., impaired facial movement after radical parotidectomy or impaired nasal breathing after paranasal sinus radiation). The main advantage of the HNC-FIT scales as compact, rapid instruments is also their main limitation. By restricting the scales to the functions and symptoms most frequently mentioned in publications, many important functional domains are ignored [36]. One limitation of this study is that not all HNC-related functions and symptoms were explicitly captured by the HNC-FIT scales due to its restriction to higher-level, commonly affected functional domains. In particular, xerostomia, a frequent and often distressing symptom, was mapped into the higher-level function food intake. Other important, but at best only indirectly covered functional outcomes, include aesthetic appearance [96,97], socioeconomic conditions and occupational concerns, fatigue, social eating, sexuality, hearing loss, lymph edema, sleep disturbance, or psychological distress and anxiety [98]. Although the primary aim during the development of the HNC-FIT Scales was to capture the functional outcomes in HNC patients as objectively as possible by anchoring equidistance verbal ratings to objectifiable external criteria [36], certain biases are inherent to the scale (e.g., HNC patients after total laryngectomy will be classified with low functional integrity despite good physical condition). This is the price of being able to capture some sort of functional landscape in many patients with reasonable effort. However, there are abundant data on individual functions, symptoms, or disease-related QoL that take these aspects into account [20,68,99,100]. Another limitation of the present study is that no QoL data were raised. Although the HNC-FIT scales were intentionally designed to objectively assess functional outcomes, the inclusion of subjective single function-assessment tools to assess QoL (e.g., EQ-5D [101]) would have improved the studies quality. However, for previous comparisons of the HNC-FIT scales with the patient-reported EORTC QoL H&N35 questionnaire [18], close correlations were observed [36]. Finally, the HNC-FIT Scales were originally developed and validated for German speaking HNC-patients only [36]. Thus, both the generalizability of the observations presented here as well as the applicability of the HNC-FIT Scales for languages other than German are hampered. Although an English translation was provided as part of the original validation study [36], to date no formal translation and cross-cultural validation for English has been performed. 5. Conclusions The HNC-FIT scales provide an overview of six relevant functional domains in HNSCC patients. The results are consistent with numerous previous studies that have considered individual symptoms or functions separately. According to multivariable analysis, general health status, tumor site, tumor stage, and treatment modality had the strongest impact on functional outcome, with treatment modality being the only selectable factor. Applying the HNC-FIT scale as a plain tool for the rapid assessment of functional outcomes in HNC patients might aid both patients and physicians. The scale allows for a quick capture and clear presentation of key functional results, filling a gap in HNC outcome assessment. If relevant functional impairments are identified by this simple screening tool, various, more detailed single function assessment tools may supplement the HNC FIT-scales. Thus, an adequate and timely functional rehabilitation may be initiated. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/cancers14092135/s1, Supplemental Figure S1. The Head and Neck Carcinoma Functional InTegrity (HNC-FIT) Scales; Supplemental Figure S2. Star plot representing percentage (star axes) of HNC-patients with normal or near-normal functional outcome (functional integrity) in 6 functional domains depending on general health status; Supplemental text S1. Instructions for clinicians on how to complete the Head and Neck Cancer Functional InTegrity Scales; Supplemental Table S1: Results of binary logistic regression including response status at last follow up for the functional domain food intake. Table S2: Results of binary logistic regression including response status at last follow up for the functional domain breathing. Table S3: Results of binary logistic regression including response status at last follow up for the functional domain speech. Table S4: Results of binary logistic regression including response status at last follow up for the functional domain pain. Table S5: Results of binary logistic regression including response status at last follow up for the functional domain mood. Table S6: Results of binary logistic regression including response status at last follow up for the functional domain neck and shoulder mobility. Click here for additional data file. Author Contributions Conceptualization, H.R., D.D., A.L.-H. and D.R.; methodology, H.R. and D.R.; validation, D.D., T.B.S. and M.A.; formal analysis, H.R., D.D. and D.R.; investigation, D.D., T.B.S. and A.L.-H.; data curation, H.R., A.L.-H. and M.A.; writing—original draft preparation, H.R.; writing—review and editing, H.R., D.D., D.R., U.G., G.G. and D.R.; visualization, H.R.; supervision H.R., U.G., G.G.; project administration, H.R.; All authors have read and agreed to the published version of the manuscript. Funding This research received no external funding. Institutional Review Board Statement The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Ethics Committee of the Medical University of Innsbruck, Austria (protocol code 1182/2019 and 28 May 2020). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The data presented in this study are available on request from the corresponding author. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Outcome in six functional domains (x-axis) and respective number of HNC-patients (y-axis) at a median of three years after initial diagnosis as reported by the treating physicians using the HNC-FIT Scale. For related verbal ratings and relative patient count in percent, see Table 1. Figure 2 Star plots representing percentage (star axes) of HNC-patients with normal or near-normal functional outcomes (functional integrity) in six functional domains depending on T-stage (a), tumor site (b), and treatment modality (c). The further peripherally, the better (Mobility: neck and shoulder mobility). cancers-14-02135-t001_Table 1 Table 1 Absolute & relative frequencies (percent) of functional outcomes of 681 HNC-survivors at a median of 4 years after initial diagnosis. The number of patients per functional domain and level as well as the relative proportion in percent are indicated. Functional Domain Verbal Rating Integrity Score Count Percent Food intake No oral feeding; only via gastrostomy tube 0 38 5.6% Gastrostomy tube needed; some oral feeding possible 1 59 8.7% No gastrostomy tube, oral diet, but only liquid/soft food 2 49 7.2% No gastrostomy tube, diet/swallowing near normal 3 122 17.9% Normal 4 413 60.6% Breathing Tracheostoma, needs cuffed cannula 0 5 0.7% Tracheostoma, speech cannula/no cannula 1 65 9.5% No tracheostoma, breathing difficulties at rest 2 7 1.0% No tracheostoma, breathing difficulties only on exertion 3 84 12.3% Normal 4 520 76.4% Speech Not possible, without phonation 0 27 4.0% Difficult to understand, no phone calls 1 41 6.0% Telephoning possible 2 43 6.3% Easy to understand, but pronunciation/voice changed 3 218 32.0% Normal 4 352 51.7% Pain Pain despite opiate therapy 0 7 1.0% Controlled with opiates 1 39 5.7% Regularly needs non-opioid analgesics 2 29 4.3% Needs analgesics from time to time 3 86 12.6% Normal 4 520 76.4% Mood Suicidal thoughts 0 1 0.1% Very depressed despite antidepressants 1 10 1.5% With antidepressants overall normal mood, very depressed without antidepressants 2 45 6.6% Occasionally depressed, no antidepressants needed 3 112 16.4% Normal 4 513 75.3% Mobility 1 Stiff neck and/or shoulder, hardly any movement possible 0 9 1.3% Can hardly comb hair, looking backwards in car not possible 1 19 2.8% Combing with problems, looking backwards in car difficult 2 57 8.4% Combing and looking backwards in car slightly restricted 3 135 19.8% Normal 4 461 67.7% 1 Neck and shoulder mobility. cancers-14-02135-t002_Table 2 Table 2 Relative counts (percent) of 681 patients with incident head and neck cancer with normal or near-normal outcome in HNC functional integrity scales (integrity score 3 and 4) stratified by patients- and disease-related factors. Factor Factor-Level n= Food Breathing Speech Pain Mood Mobility 1 Gender male 544 78% 87% * 2 81% ** 89% 93% * 87% female 137 80% 94% 93% 88% 88% 89% Age 3 <=50 91 85% 92% 95%* 95% 93% 91% 51–60 213 81% 88% 85% 88% 93% 87% 61–70 210 74% 86% 80% 90% 90% 86% 71–80 129 75% 91% 80% 87% 91% 85% >80 38 84% 90% 84% 84% 95% 95% ASA 4 ASA I/II 321 82% *** 93% *** 89% *** 93% *** 93% 88% ASA III/IV 162 66% 79% 70% 79% 88% 82% Smoking <10 pack years 219 81% * 91% 85% 89% 91% 87% >=10 pack years 266 73% 86% 79% 88% 91% 86% Drinking <daily 322 78% 89% 82% 89% 90% 86% daily 120 69% 87% 78% 88% 93% 83% Tumor site Lips and oral cavity 94 76% *** 98% *** 90% *** 94% 90% 87% Oropharynx 225 70% 97% 90% 87% 92% 85% Hypopharynx 41 59% 49% 54% 81% 90% 81% Larynx 177 87% 76% 70% 89% 93% 89% Others 5 144 90% 97% 96% 92% 91% 92% T stage T0 6 32 81% *** 97% *** 97% *** 91% *** 94% 94% *** T1 220 91% 95% 91% 95% 93% 94% T2 205 82% 93% 89% 92% 92% 87% T3 106 70% 75% 67% 75% 88% 82% T4 118 58% 81% 72% 86% 92% 80% N stage N0 329 88% *** 89% 83% 91% * 92% 91% * N1 110 80% 91% 89% 94% 95% 90% N2 221 65% 86% 82% 84% 90% 82% N3 21 67% 91% 86% 86% 95% 86% UICC Stage I 163 95% *** 94% * 89% 96% ** 94% 95% ** Stage II 88 81% 91% 86% 90% 90% 86% Stage III 127 84% 88% 84% 90% 95% 91% Stage IV 303 67% 86% 80% 85% 90% 83% p16 negative 301 74% 85% *** 76% *** 86% 91% 84% positive 50 79% 97% 94% 91% 92% 91% Observation 7 <=24 months 247 72% *** 84% 80% * 82% *** 90% 88% 24–60 months 258 87% 93% 89% 93% 92% 89% 60+ months 176 76% 88% 82% 94% 93% 85% Response 8 Complete remission 540 83% *** 92% *** 86% *** 93% *** 92% 90% *** Residual disease 123 61% 78% 76% 72% 90% 76% Second primary 16 56% 69% 63% 94% 81% 81% Treatment 9 Surgery only 251 92% *** 92% 84% 96% *** 94% 94% ** Upfront surgery & PORT 144 81% 87% 83% 89% 92% 87% Upfront surgery & ST/RT 51 71% 80% 73% 90% 88% 84% Prim. ST/RT 192 63% 90% 88% 83% 90% 82% Prim. RT only 34 71% 82% 82% 79% 88% 85% 1 Neck and shoulder mobility; 2 The content of this cell means that 87% of 544 male patients had normal or near-normal breathing at last follow-up; 3 Age at diagnosis in years; 4 American Society of Anesthesiologists score (general health condition); 5 This group includes salivary gland cancer and sinonasal cancer; 6 Cancer of unknow primary; 7 Interval (months) between diagnosis and last functional assessment; 8 Response status at last follow-up; 9 First line treatment. ST: systematic therapy, RT: radiotherapy, ST/RT: concomitant systemic therapy and radiotherapy; *: p < 0.05, **: p < 0.005; ***, p < 0.0005 (Chi-square test at the respective functional domain). cancers-14-02135-t003_Table 3 Table 3 Binary logistic model of factors influencing food intake in 427 patients with newly diagnosed HNC. Dependent variable was normal or near-normal food intake function (integrity score 3 and 4) vs. impaired food intake (integrity score 0–2). Factors with p values < 0.2 in Chi-square tests were included. Odds ratios (OR) indicate the chance to achieve normal or near normal food intake (the higher, the better). Factor Factor-Level B (±SE) Sig. OR (95%CI) Compared to ASA 1 ASA I/II 0.82 (±0.27) 0.002 2.3 (1.3 to 3.9) ASA III/IV T stage T1 1.43 (±0.42) <0.001 4.2 (1.8 to 9.5) T4 T2 1.06 (±0.34) 0.002 2.9 (1.5 to 5.6) T4 T3 0.68 (±0.38) 0.068 2.0 (0.95 to 4.1) T4 Tumor site Lips and oral cavity −0.52 (±0.53) 0.328 0.6 (0.2 to 1.7) Hypopharynx Oropharynx 0.06 (±0.45) 0.902 1.1 (0.4 to 2.5) Hypopharynx Larynx 1.05 (±0.52) 0.042 2.9 (1.0 to 7.9) Hypopharynx Others 2 1.16 (±0.66) 0.081 3.2 (0.9 to 12.9) Hypopharynx Treatment 3 Surgery only 1.31 (±0.4) 0.001 3.7 (1.7 to 8.1) Primary ST/RT Upfront surgery & PORT 0.32 (±0.33) 0.323 1.4 (0.7 to 2.7) Primary ST/RT Upfront surgery & ST/RT −0.40 (±0.43) 0.350 0.7 (0.3 to 1.6) Primary ST/RT Prim. RT only 0.34 (±0.62) 0.584 1.4 (0.4 to 4.7) Primary ST/RT 1 Society of Anesthesiologists score (general health condition); 2 This group includes salivary gland cancer and sinonasal cancer; 3 First line treatment. PORT: postoperative radiotherapy, ST/RT: concomitant systemic therapy and radiotherapy, RT: radiotherapy. cancers-14-02135-t004_Table 4 Table 4 Binary logistic model of factors influencing breathing in 416 patients with newly diagnosed HNC. Dependent variable was normal or near-normal breathing function (integrity score 3 and 4) vs. impaired breathing function (integrity score 0–2). Factors with p values < 0.2 in Chi-square tests were included. Odds ratios (OR) indicate the chance to achieve normal or near normal breathing function (the higher, the better). Factor Factor-Level B (±SE) Sig. OR (95%CI) Compared to ASA 1 ASA I/II 1.08 (±0.39) 0.006 2.9 (1.3 to 6.3) ASA III/IV T stage T1 2.0 (±0.61) 0.001 7.3 (2.2 to 24.6) T4 T2 1.37 (±0.52) 0.01 3.9 (1.3 to 11.1) T4 T3 0.02 (±0.49) 0.958 1 (0.3 to 2.7) T4 Tumor site Lips and oral cavity 3.25 (±0.84) <0.001 25.8 (4.9 to 135.4) Hypopharynx Oropharynx 3.21 (±0.59) <0.001 24.9 (7.7 to 79.9) Hypopharynx Larynx 1.11 (±0.49) 0.024 3.1 (1.2 to 8.1) Hypopharynx Others 2 3.07 (±1.12) 0.006 21.7 (2.4 to 196) Hypopharynx p16 Status p16-positive 1.01 (±0.6) 0.092 2.7 (0.8 to 9) p16-negative 1 Society of Anesthesiologists score (general health condition); 2 This group includes salivary gland cancer and sinonasal cancer. cancers-14-02135-t005_Table 5 Table 5 Binary logistic model of factors influencing speech function in 416 patients with newly diagnosed HNC. Dependent variable was normal or near-normal speech function (integrity score 3 and 4) vs. impaired speech (integrity score 0–2). Factors with p values < 0.2 in Chi-square tests were included. Odds ratios (OR) indicate the chance to achieve normal or near normal speech function (the higher, the better). Factor Factor-Level B (±SE) Sig. OR (95%CI) Compared to ASA 1 ASA I/II 1.02 (±0.32) 0.002 2.7 (1.4 to 5.2) ASA III/IV T stage T1 2.77 (±0.56) <0.001 15.9 (5.2 to 48.4) T4 T2 2.36 (±0.48) <0.001 10.6 (4.1 to 27.7) T4 T3 0.78 (±0.44) 0.08 2.1 (0.9 to 5.2) T4 Tumor site Lips and oral cavity 1.94 (±0.63) 0.002 6.9 (2 to 24.2) Hypopharynx Oropharynx 1.85 (±0.49) <0.001 6.3 (2.3 to 16.8) Hypopharynx Larynx 0.5 (±0.5) 0.314 1.6 (0.6 to 4.4) Hypopharynx Others 2 3.2 (±1.15) 0.005 24.5 (2.5 to 233.5) Hypopharynx Treatment 3 Surgery only −1.92 (±0.5) <0.001 0.1 (0.05 to 0.3) Primary ST/RT Upfront surgery & PORT −0.67 (±0.47) 0.152 0.5 (0.2 to 1.2) Primary ST/RT Upfront surgery & ST/RT −1.93 (±0.59) 0.001 0.1 (0 to 0.4) Primary ST/RT Prim. RT only −0.58 (±0.7) 0.405 0.5 (0.1 to 2.2) Primary ST/RT 1 Society of Anesthesiologists score (general health condition); 2 This group includes salivary gland cancer and sinonasal cancer; 3 First line treatment. PORT: postoperative radiotherapy, ST/RT: concomitant systemic therapy and radiotherapy, RT: radiotherapy. cancers-14-02135-t006_Table 6 Table 6 Poorest outcome in at least one of six functional domains (food intake, breathing, speech, pain, mood, and neck and shoulder mobility) in 681 patients with incident HNC. A normal or near-normal outcome in all six functional domains was observed in 61% of HNC-patients; however, 30% had a relevant impairment in at least one functional domain and 9.1% had maximal functional impairment (worst possible outcome) in at least one functional domain. Worst Outcome Frequency Percent Worst 62 9.1 Poor 119 17.5 Impaired 85 12.5 Near normal 230 33.8 Normal 185 27.2 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Chow L.Q.M. Head and neck cancer N. Engl. J. Med. 2020 382 60 72 10.1056/NEJMra1715715 31893516 2. León X. Orús C. Casasayas M. Neumann E. Holgado A. Quer M. Trends in disease-specific survival of head and neck squamous cell carcinoma patients treated in a single institution over a 30-year period Oral Oncol. 2021 115 105184 10.1016/j.oraloncology.2021.105184 33581504 3. Johnson D.E. Burtness B. Leemans C.R. Lui V.W.Y. Bauman J.E. Grandis J.R. Head and neck squamous cell carcinoma Nat. Rev. Dis. Primers 2020 6 92 10.1038/s41572-020-00224-3 33243986 4. List M.A. Stracks J. Colangelo L. Butler P. Ganzenko N. Lundy D. Sullivan P. Haraf D. Kies M. Goodwin W. How Do head and neck cancer patients prioritize treatment outcomes before initiating treatment? J. Clin. Oncol. 2000 18 877 884 10.1200/JCO.2000.18.4.877 10673531 5. Brotherston D.C. Poon I. Le T. Leung M. Kiss A. Ringash J. Balogh J. Lee J. Wright J.R. Patient preferences for oropharyngeal cancer treatment de-escalation Head Neck 2012 35 151 159 10.1002/hed.22930 22431201 6. List M.A. Rutherford J.L. Stracks J. Pauloski B.R. Logemann J.A. Lundy D. Sullivan P. Goodwin W. Kies M. Vokes E.E. Prioritizing treatment outcomes: Head and neck cancer patients versus nonpatients Head Neck 2004 26 163 170 10.1002/hed.10367 14762885 7. Spilker B. Introduction Quality of Life Assessments in Clinical Trials Spilker B. Raven Press New York, NY, USA 1990 3 9 8. Ringash J. Bernstein L.J. Devins G. Dunphy C. Giuliani M. Martino R. McEwen S. Head and Neck Cancer Survivorship: Learning the Needs, Meeting the Needs Semin. Radiat. Oncol. 2018 28 64 74 10.1016/j.semradonc.2017.08.008 29173757 9. Schipper H.C.J. Powell V. Definitions and Conceptual Issues Quality of Life Assessments in Clinical Trials Spilker B. Raven Press New York, NY, USA 1990 11 24 10. List M.A. Bilir S.P. Functional outcomes in head and neck cancer Semin. Radiat. Oncol. 2004 14 178 189 10.1053/j.semradonc.2003.12.008 15095263 11. Pusic A. Liu J.C. Chen C.M. Cano S. Davidge K. Klassen A. Branski R. Patel S. Kraus D. Cordeiro P.G. A systematic review of patient-reported outcome measures in head and neck cancer surgery Otolaryngol. Head Neck Surg. 2007 136 525 535 10.1016/j.otohns.2006.12.006 17418246 12. Ojo B. Genden E.M. Teng M.S. Milbury K. Misiukiewicz K.J. Badr H. A systematic review of head and neck cancer quality of life assessment instruments Oral Oncol. 2012 48 923 937 10.1016/j.oraloncology.2012.03.025 22525604 13. Hassan S.J. Weymuller E.A. Jr. Assessment of quality of life in head and neck cancer patients Head Neck 1993 15 485 496 10.1002/hed.2880150603 8253555 14. List M.A. Ritter-Sterr C. Lansky S.B. A performance status scale for head and neck cancer patients Cancer 1990 66 564 569 10.1002/1097-0142(19900801)66:3<564::AID-CNCR2820660326>3.0.CO;2-D 2364368 15. Cella D.F. Tulsky D.S. Gray G. Sarafian B. Linn E. Bonomi A. Silberman M. Yellen S.B. Winicour P. Brannon J. The Functional Assessment of Cancer Therapy scale: Development and validation of the general measure J. Clin. Oncol. 1993 11 570 579 10.1200/JCO.1993.11.3.570 8445433 16. List M.A. D’Antonio L.L. Cella D.F. Siston A. Mumby P. Haraf D. Vokes E. The Performance Status Scale for Head and Neck Cancer Patients and the Functional Assessment of Cancer Therapy-Head and Neck Scale. A study of utility and validity Cancer 1996 77 2294 2301 10.1002/(SICI)1097-0142(19960601)77:11<2294::AID-CNCR17>3.0.CO;2-S 8635098 17. Aaronson N.K. Ahmedzai S. Bergman B. Bullinger M. Cull A. Duez N.J. Filiberti A. Flechtner H. Fleishman S.B. de Haes J.C. The European Organization for Research and Treatment of Cancer QLQ-C30: A quality-of-life instrument for use in international clinical trials in oncology J. Natl. Cancer Inst. 1993 85 365 376 10.1093/jnci/85.5.365 8433390 18. Bjordal K. De Graeff A. Fayers P.M. Hammerlid E. van Pottelsberghe C. Curran D. Ahlner-Elmqvist M. Maher E.J. Meyza J.W. Bredart A. A 12 country field study of the EORTC QLQ-C30 (version 3.0) and the head and neck cancer specific module (EORTC QLQ-H&N35) in head and neck patients. EORTC Quality of Life Group Eur. J. Cancer 2000 36 1796 1807 10974628 19. Bjordal K. Hammerlid E. Ahlner-Elmqvist M. De Graeff A. Boysen M. Evensen J.F. Biorklund A. de Leeuw J.R. Fayers P.M. Jannert M. Quality of life in head and neck cancer patients: Validation of the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire-H&N35 J. Clin. Oncol. 1999 17 1008 1019 10071296 20. Høxbroe Michaelsen S. Grønhøj C. Høxbroe Michaelsen J. Friborg J. von Buchwald C. Quality of life in survivors of oropharyngeal cancer: A systematic review and meta-analysis of 1366 patients Eur. J. Cancer 2017 78 91 102 10.1016/j.ejca.2017.03.006 28431302 21. Campbell B.H. Marbella A. Layde P.M. Quality of life and recurrence concern in survivors of head and neck cancer Laryngoscope 2000 110 895 906 10.1097/00005537-200006000-00003 10852502 22. Dunne S. Mooney O. Coffey L. Sharp L. Desmond D. Timon C. O’Sullivan E. Gallagher P. Psychological variables associated with quality of life following primary treatment for head and neck cancer: A systematic review of the literature from 2004 to 2015 Psycho-Oncology 2017 26 149 160 10.1002/pon.4109 26918648 23. Sprangers M.A. Schwartz C.E. The challenge of response shift for quality-of-life-based clinical oncology research Ann. Oncol. 1999 10 747 749 10.1023/A:1008305523548 10470418 24. Ilie G. Bradfield J. Moodie L. Lawen T. Ilie A. Lawen Z. Blackman C. Gainer R. Rutledge R.D.H. The Role of Response-Shift in Studies Assessing Quality of Life Outcomes Among Cancer Patients: A Systematic Review Front. Oncol. 2019 9 783 10.3389/fonc.2019.00783 31482068 25. Vernon S.W. Abotchie P.N. McQueen A. White A. Eberth J.M. Coan S.P. Is the accuracy of self-reported colorectal cancer screening associated with social desirability? Cancer Epidemiol. Biomark. Prev. 2012 21 61 65 10.1158/1055-9965.EPI-11-0552 22144501 26. Mondal H. Mondal S. Social desirability bias: A confounding factor to consider in survey by self-administered questionnaire Indian J. Pharmacol. 2018 50 143 144 10.4103/ijp.IJP_15_17 30166752 27. Protzko J. Zedelius C.M. Schooler J.W. Rushing to Appear Virtuous: Time Pressure Increases Socially Desirable Responding Psychol. Sci. 2019 30 1584 1591 10.1177/0956797619867939 31603024 28. Hitchcott P.K. Penna M.P. Fastame M.C. Age Trends in Well-Being and Depressive Symptoms: The Role of Social Desirability Psychiatr. Q 2020 91 463 473 10.1007/s11126-020-09711-y 31983016 29. Meisters J. Hoffmann A. Musch J. Controlling social desirability bias: An experimental investigation of the extended crosswise model PLoS ONE 2020 15 e0243384 10.1371/journal.pone.0243384 33284820 30. Huguenin P.U. Taussky D. Moe K. Meister A. Baumert B. Lütolf U.M. Glanzmann C. Quality of life in patients cured from a carcinoma of the head and neck by radiotherapy: The importance of the target volume Int. J. Radiat. Oncol. Biol. Phys. 1999 45 47 52 10.1016/S0360-3016(99)00128-5 10477005 31. Kirsh E. Naunheim M. Holman A. Kammer R. Varvares M. Goldsmith T. Patient-reported versus physiologic swallowing outcomes in patients with head and neck cancer after chemoradiation Laryngoscope 2019 129 2059 2064 10.1002/lary.27610 30582151 32. Ringash J. Bernstein L.J. Cella D. Logemann J. Movsas B. Murphy B. Trotti A. Wells N. Yueh B. Ridge J. Outcomes toolbox for head and neck cancer research Head Neck 2015 37 425 439 10.1002/hed.23561 24318166 33. Parke S.C. Oza S. Shahpar S. Ngo-Huang A. Herbert A. Barksdale T. Gerber L. Identifying Gaps in Research on Rehabilitation for Patients With Head and Neck Cancer: A Scoping Review Arch. Phys. Med. Rehabil. 2019 100 2381 2388 10.1016/j.apmr.2019.03.022 31082380 34. Osborn H.A. Goldsmith T.A. Varvares M.A. Assessing functional outcomes in head and neck surgical oncology Head Neck 2019 41 2051 2057 10.1002/hed.25656 30698897 35. Chen A.Y. Frankowski R. Bishop-Leone J. Hebert T. Leyk S. Lewin J. Goepfert H. The development and validation of a dysphagia-specific quality-of-life questionnaire for patients with head and neck cancer: The M. D. Anderson dysphagia inventory Arch. Otolaryngol. Head Neck Surg. 2001 127 870 876 11448365 36. Dejaco D. Riedl D. Gasser S. Schartinger V.H. Innerhofer V. Gottfried T. Steinbichler T.B. Riechelmann F. Moschen R. Galvan O. A Tool for Rapid Assessment of Functional Outcomes in Patients with Head and Neck Cancer Cancers 2021 13 5529 10.3390/cancers13215529 34771691 37. Reid B.C. Alberg A.J. Klassen A.C. Koch W.M. Samet J.M. The American Society of Anesthesiologists’ class as a comorbidity index in a cohort of head and neck cancer surgical patients Head Neck 2001 23 985 994 10.1002/hed.1143 11754504 38. Fakhry C. Zhang Q. Gillison M.L. Nguyen-Tân P.F. Rosenthal D.I. Weber R.S. Lambert L. Trotti A.M. 3rd Barrett W.L. Thorstad W.L. Validation of NRG oncology/RTOG-0129 risk groups for HPV-positive and HPV-negative oropharyngeal squamous cell cancer: Implications for risk-based therapeutic intensity trials Cancer 2019 125 2027 2038 10.1002/cncr.32025 30913305 39. Bernier J. Domenge C. Ozsahin M. Matuszewska K. Lefebvre J.L. Greiner R.H. Giralt J. Maingon P. Rolland F. Bolla M. Postoperative irradiation with or without concomitant chemotherapy for locally advanced head and neck cancer N. Engl. J. Med. 2004 350 1945 1952 10.1056/NEJMoa032641 15128894 40. Maihoefer C. Schuttrumpf L. Macht C. Pflugradt U. Hess J. Schneider L. Woischke C. Walch A. Baumeister P. Kirchner T. Postoperative (chemo) radiation in patients with squamous cell cancers of the head and neck-clinical results from the cohort of the clinical cooperation group “Personalized Radiotherapy in Head and Neck Cancer” Radiat. Oncol. 2018 13 123 10.1186/s13014-018-1067-1 29970111 41. Schuttrumpf L. Marschner S. Scheu K. Hess J. Rietzler S. Walch A. Baumeister P. Kirchner T. Ganswindt U. Zitzelsberger H. Definitive chemoradiotherapy in patients with squamous cell cancers of the head and neck-results from an unselected cohort of the clinical cooperation group “Personalized Radiotherapy in Head and Neck Cancer” Radiat. Oncol. 2020 15 7 10.1186/s13014-019-1452-4 31906998 42. Rades D. Seidl D. Janssen S. Strojan P. Karner K. Bajrovic A. Hakim S.G. Wollenberg B. Schild S.E. Comparing two lower-dose cisplatin programs for radio-chemotherapy of locally advanced head-and-neck cancers Eur. Arch. Otorhinolaryngol. 2017 274 1021 1027 10.1007/s00405-016-4326-5 27687678 43. Budach V. Stromberger C. Poettgen C. Baumann M. Budach W. Grabenbauer G. Marnitz S. Olze H. Wernecke K.D. Ghadjar P. Hyperfractionated accelerated radiation therapy (HART) of 70.6 Gy with concurrent 5-FU/Mitomycin C is superior to HART of 77.6 Gy alone in locally advanced head and neck cancer: Long-term results of the ARO 95-06 randomized phase III trial Int. J. Radiat. Oncol. Biol. Phys. 2015 91 916 924 10.1016/j.ijrobp.2014.12.034 25670541 44. Bonner J.A. Harari P.M. Giralt J. Azarnia N. Shin D.M. Cohen R.B. Jones C.U. Sur R. Raben D. Jassem J. Radiotherapy plus cetuximab for squamous-cell carcinoma of the head and neck N. Engl. J. Med. 2006 354 567 578 10.1056/NEJMoa053422 16467544 45. Crowder S.L. Douglas K.G. Yanina Pepino M. Sarma K.P. Arthur A.E. Nutrition impact symptoms and associated outcomes in post-chemoradiotherapy head and neck cancer survivors: A systematic review J. Cancer Surviv. 2018 12 479 494 10.1007/s11764-018-0687-7 29556926 46. Schemper M. Smith T.L. A note on quantifying follow-up in studies of failure time Control. Clin. Trials 1996 17 343 346 10.1016/0197-2456(96)00075-X 8889347 47. Rodriguez A.M. Komar A. Ringash J. Chan C. Davis A.M. Jones J. Martino R. McEwen S. A scoping review of rehabilitation interventions for survivors of head and neck cancer Disabil. Rehabil. 2019 41 2093 2107 10.1080/09638288.2018.1459880 29976091 48. Wan Leung S. Lee T.F. Chien C.Y. Chao P.J. Tsai W.L. Fang F.M. Health-related quality of life in 640 head and neck cancer survivors after radiotherapy using EORTC QLQ-C30 and QLQ-H&N35 questionnaires BMC Cancer 2011 11 128 10.1186/1471-2407-11-128 21486431 49. Kondo T. Sugauchi A. Yabuno Y. Kobashi H. Amano K. Aikawa T. Kogo M. Okura M. Performance status scale for head and neck scores for oral cancer survivors: Predictors and factors for improving quality of life Clin. Oral Investig. 2019 23 1575 1582 10.1007/s00784-018-2587-7 50. Rogers S.N. Lowe D. Yueh B. Weymuller E.A. Jr. The physical function and social-emotional function subscales of the University of Washington Quality of Life Questionnaire Arch. Otolaryngol. Head Neck Surg. 2010 136 352 357 10.1001/archoto.2010.32 20403851 51. Veldhuis D. Probst G. Marek A. Noack V. Ural A. Adamietz I. Dazert S. Minovi A. Tumor site and disease stage as predictors of quality of life in head and neck cancer: A prospective study on patients treated with surgery or combined therapy with surgery and radiotherapy or radiochemotherapy Eur. Arch. Otorhinolaryngol. 2016 273 215 224 10.1007/s00405-015-3496-x 25575842 52. Ryzek D.F. Mantsopoulos K. Kunzel J. Grundtner P. Zenk J. Iro H. Psychogios G. Early stage oropharyngeal carcinomas: Comparing quality of life for different treatment modalities BioMed Res. Int. 2014 2014 421964 10.1155/2014/421964 24719863 53. Chaukar D.A. Walvekar R.R. Das A.K. Deshpande M.S. Pai P.S. Chaturvedi P. Kakade A. D’Cruz A.K. Quality of life in head and neck cancer survivors: A cross-sectional survey Am. J. Otolaryngol. 2009 30 176 180 10.1016/j.amjoto.2008.05.001 19410123 54. Nguyen N.P. Moltz C.C. Frank C. Vos P. Smith H.J. Karlsson U. Dutta S. Midyett F.A. Barloon J. Sallah S. Dysphagia following chemoradiation for locally advanced head and neck cancer Ann. Oncol. 2004 15 383 388 10.1093/annonc/mdh101 14998839 55. Rinkel R.N. Verdonck-de Leeuw I.M. Doornaert P. Buter J. de Bree R. Langendijk J.A. Aaronson N.K. Leemans C.R. Prevalence of swallowing and speech problems in daily life after chemoradiation for head and neck cancer based on cut-off scores of the patient-reported outcome measures SWAL-QOL and SHI Eur. Arch. Otorhinolaryngol. 2016 273 1849 1855 10.1007/s00405-015-3680-z 26071622 56. Crombie A.K. Farah C.S. Batstone M.D. Health-related quality of life of patients treated with primary chemoradiotherapy for oral cavity squamous cell carcinoma: A comparison with surgery Br. J. Oral Maxillofac. Surg. 2014 52 111 117 10.1016/j.bjoms.2013.09.014 24148700 57. Chen A.M. Daly M.E. Luu Q. Donald P.J. Farwell D.G. Comparison of functional outcomes and quality of life between transoral surgery and definitive chemoradiotherapy for oropharyngeal cancer Head Neck 2015 37 381 385 10.1002/hed.23610 24431059 58. Howard J. Masterson L. Dwivedi R.C. Riffat F. Benson R. Jefferies S. Jani P. Tysome J.R. Nutting C. Minimally invasive surgery versus radiotherapy/chemoradiotherapy for small-volume primary oropharyngeal carcinoma Cochrane Database Syst. Rev. 2016 12 Cd010963 10.1002/14651858.CD010963.pub2 27943254 59. Rathod S. Gupta T. Ghosh-Laskar S. Murthy V. Budrukkar A. Agarwal J. Quality-of-life (QOL) outcomes in patients with head and neck squamous cell carcinoma (HNSCC) treated with intensity-modulated radiation therapy (IMRT) compared to three-dimensional conformal radiotherapy (3D-CRT): Evidence from a prospective randomized study Oral Oncol. 2013 49 634 642 10.1016/j.oraloncology.2013.02.013 23562564 60. Kraaijenga S.A. Oskam I.M. van der Molen L. Hamming-Vrieze O. Hilgers F.J. van den Brekel M.W. Evaluation of long term (10-years+) dysphagia and trismus in patients treated with concurrent chemo-radiotherapy for advanced head and neck cancer Oral Oncol. 2015 51 787 794 10.1016/j.oraloncology.2015.05.003 26027851 61. Ringash J. Fisher R. Peters L. Trotti A. O’Sullivan B. Corry J. Kenny L. Van Den Bogaert W. Wratten C. Rischin D. Effect of p16 Status on the Quality-of-Life Experience During Chemoradiation for Locally Advanced Oropharyngeal Cancer: A Substudy of Randomized Trial Trans-Tasman Radiation Oncology Group (TROG) 02.02 (HeadSTART) Int. J. Radiat. Oncol. Biol. Phys. 2016 97 678 686 10.1016/j.ijrobp.2016.03.017 27209505 62. Trivedi N.P. Swaminathan D.K. Thankappan K. Chatni S. Kuriakose M.A. Iyer S. Comparison of quality of life in advanced laryngeal cancer patients after concurrent chemoradiotherapy vs total laryngectomy Otolaryngol. Head Neck Surg. 2008 139 702 707 10.1016/j.otohns.2008.06.002 18984267 63. El-Deiry M. Funk G.F. Nalwa S. Karnell L.H. Smith R.B. Buatti J.M. Hoffman H.T. Clamon G.H. Graham S.M. Trask D.K. Long-term quality of life for surgical and nonsurgical treatment of head and neck cancer Arch. Otolaryngol. Head Neck Surg. 2005 131 879 885 10.1001/archotol.131.10.879 16230590 64. Tschiesner U. Schuster L. Strieth S. Harréus U. Functional outcome in patients with advanced head and neck cancer: Surgery and reconstruction with free flaps versus primary radiochemotherapy Eur. Arch. Otorhinolaryngol. 2012 269 629 638 10.1007/s00405-011-1642-7 21643935 65. Dwivedi R.C. St Rose S. Chisholm E.J. Youssefi P. Hassan M.S. Khan A.S. Elmiyeh B. Kerawala C.J. Clarke P.M. Nutting C.M. Evaluation of factors affecting post-treatment quality of life in oral and oropharyngeal cancer patients primarily treated with curative surgery: An exploratory study Eur. Arch. Otorhinolaryngol. 2012 269 591 599 10.1007/s00405-011-1621-z 21553147 66. Meyer T.K. Kuhn J.C. Campbell B.H. Marbella A.M. Myers K.B. Layde P.M. Speech intelligibility and quality of life in head and neck cancer survivors Laryngoscope 2004 114 1977 1981 10.1097/01.mlg.0000147932.36885.9e 15510026 67. Vartanian J.G. Kowalski L.P. Acceptance of major surgical procedures and quality of life among long-term survivors of advanced head and neck cancer Arch. Otolaryngol. Head Neck Surg. 2009 135 376 379 10.1001/archoto.2009.5 19380360 68. World Health Organization International Classification of Functioning, Disability and Health: ICF World Health Organization Geneva, Switzerland 2001 Volume 2019 69. World Health Organization WHO Guidelines for the Pharmacological and Radiotherapeutic Management of Cancer Pain in Adults and Adolescents World Health Organization Geneva, Switzerland 2018 70. Blasco M.A. Cordero J. Dundar Y. Chronic Pain Management in Head and Neck Oncology Otolaryngol. Clin. N. Am. 2020 53 865 875 10.1016/j.otc.2020.05.015 32684285 71. Dugué J. Humbert M. Bendiane M.K. Bouhnik A.D. Babin E. Licaj I. Head and neck cancer survivors’ pain in France: The VICAN study J. Cancer Surviv. 2021 16 119 131 10.1007/s11764-021-01010-0 33635450 72. Cho S.F. Rau K.M. Shao Y.Y. Yen C.J. Wu M.F. Chen J.S. Chang C.S. Yeh S.P. Chiou T.J. Hsieh R.K. Patients with head and neck cancer may need more intensive pain management to maintain daily functioning: A multi-center study Support. Care Cancer Off. J. Multinatl. Assoc. Support. Care Cancer 2019 27 1663 1672 10.1007/s00520-018-4404-x 73. Cramer J.D. Johnson J.T. Nilsen M.L. Pain in Head and Neck Cancer Survivors: Prevalence, Predictors, and Quality-of-Life Impact Otolaryngol. Head Neck Surg. 2018 159 853 858 10.1177/0194599818783964 29943677 74. Havard M. Esslinger H. Mierzwa M. Kharofa J. Takiar V. Increased Pain Reporting by Head and Neck Cancer Patients at Radiation Oncology Consultation: A Quality-of-Life Analysis Laryngoscope 2021 131 326 332 10.1002/lary.28784 32609888 75. Humphreys K. Avoiding globalisation of the prescription opioid epidemic Lancet 2017 390 437 439 10.1016/S0140-6736(17)31918-9 28792397 76. Gourin C.G. McAfee W.J. Neyman K.M. Howington J.W. Podolsky R.H. Terris D.J. Effect of comorbidity on quality of life and treatment selection in patients with squamous cell carcinoma of the head and neck Laryngoscope 2005 115 1371 1375 10.1097/01.mlg.0000167983.32017.64 16094107 77. Wells M. Swartzman S. Lang H. Cunningham M. Taylor L. Thomson J. Philp J. McCowan C. Predictors of quality of life in head and neck cancer survivors up to 5 years after end of treatment: A cross-sectional survey Support. Care Cancer Off. J. Multinatl. Assoc. Support. Care Cancer 2016 24 2463 2472 10.1007/s00520-015-3045-6 26660345 78. Götze H. Friedrich M. Taubenheim S. Dietz A. Lordick F. Mehnert A. Depression and anxiety in long-term survivors 5 and 10 years after cancer diagnosis Support. Care Cancer Off. J. Multinatl. Assoc. Support. Care Cancer 2020 28 211 220 10.1007/s00520-019-04805-1 79. Naser A.Y. Hameed A.N. Mustafa N. Alwafi H. Dahmash E.Z. Alyami H.S. Khalil H. Depression and Anxiety in Patients With Cancer: A Cross-Sectional Study Front. Psychol. 2021 12 585534 10.3389/fpsyg.2021.585534 33935849 80. Chen A.M. Daly M.E. Vazquez E. Courquin J. Luu Q. Donald P.J. Farwell D.G. Depression among long-term survivors of head and neck cancer treated with radiation therapy JAMA Otolaryngol.-Head Neck Surg. 2013 139 885 889 10.1001/jamaoto.2013.4072 23949013 81. World Health Organization Depression and Other Common Mental Disorders: Global Health Estimates World Health Organization Geneva, Switzerland 2017 82. Lydiatt W.M. Moran J. Burke W.J. A review of depression in the head and neck cancer patient Clin. Adv. Hematol. Oncol. 2009 7 397 403 19606075 83. Kam D. Salib A. Gorgy G. Patel T.D. Carniol E.T. Eloy J.A. Baredes S. Park R.C. Incidence of Suicide in Patients With Head and Neck Cancer JAMA Otolaryngol.-Head Neck Surg. 2015 141 1075 1081 10.1001/jamaoto.2015.2480 26562764 84. Maciejewski O. Smeets R. Gerhards F. Kolk A. Kloss F. Stein J.M. Kasaj A. Koch F. Grosjean M. Riediger D. Gender specific quality of life in patients with oral squamous cell carcinomas Head Face Med. 2010 6 21 10.1186/1746-160X-6-21 20727183 85. Chhabria K.S. Carnaby G.D. Psychometric validation of the Center for Epidemiological Studies Depression Scale in Head and Neck Cancer patients Oral Oncol. 2017 75 158 162 10.1016/j.oraloncology.2017.11.010 29224813 86. Karnell L.H. Funk G.F. Christensen A.J. Rosenthal E.L. Magnuson J.S. Persistent posttreatment depressive symptoms in patients with head and neck cancer Head Neck 2006 28 453 461 10.1002/hed.20370 16320360 87. Bamonti P.M. Moye J. Naik A.D. Pain is associated with continuing depression in cancer survivors Psychol. Health Med. 2018 23 1182 1195 10.1080/13548506.2018.1476723 29901408 88. van Hinte G. Wetzels J.G.H. Merkx M.A.W. de Haan A.F.J. Koole R. Speksnijder C.M. Factors influencing neck and shoulder function after oral oncology treatment: A five-year prospective cohort study in 113 patients Support. Care Cancer Off. J. Multinatl. Assoc. Support. Care Cancer 2019 27 2553 2560 10.1007/s00520-018-4534-1 30430300 89. Mücke T. Koschinski J. Wagenpfeil S. Wolff K.D. Kanatas A. Mitchell D.A. Deppe H. Kesting M.R. Functional outcome after different oncological interventions in head and neck cancer patients J. Cancer Res. Clin. Oncol. 2012 138 371 376 10.1007/s00432-011-1106-x 22143103 90. Eickmeyer S.M. Walczak C.K. Myers K.B. Lindstrom D.R. Layde P. Campbell B.H. Quality of life, shoulder range of motion, and spinal accessory nerve status in 5-year survivors of head and neck cancer PM R J. Inj. Funct. Rehabil. 2014 6 1073 1080 10.1016/j.pmrj.2014.05.015 24880060 91. Nilsen M.L. Mady L.J. Hodges J. Wasserman-Wincko T. Johnson J.T. Burden of treatment: Reported outcomes in a head and neck cancer survivorship clinic Laryngoscope 2019 129 E437 E444 10.1002/lary.27801 30648277 92. Suarez-Cunqueiro M.M. Schramm A. Schoen R. Seoane-Leston J. Otero-Cepeda X.L. Bormann K.H. Kokemueller H. Metzger M. Diz-Dios P. Gellrich N.C. Speech and swallowing impairment after treatment for oral and oropharyngeal cancer Arch. Otolaryngol. Head Neck Surg. 2008 134 1299 1304 10.1001/archotol.134.12.1299 19075126 93. Oozeer N.B. Owen S. Perez B.Z. Jones G. Welch A.R. Paleri V. Functional status after total laryngectomy: Cross-sectional survey of 79 laryngectomees using the Performance Status Scale for Head and Neck Cancer J. Laryngol. Otol. 2010 124 412 416 10.1017/S0022215109992192 20003587 94. Staar S. Rudat V. Stuetzer H. Dietz A. Volling P. Schroeder M. Flentje M. Eckel H.E. Mueller R.P. Intensified hyperfractionated accelerated radiotherapy limits the additional benefit of simultaneous chemotherapy—Results of a multicentric randomized German trial in advanced head-and-neck cancer Int. J. Radiat. Oncol. Biol. Phys. 2001 50 1161 1171 10.1016/S0360-3016(01)01544-9 11483325 95. Beitler J.J. Zhang Q. Fu K.K. Trotti A. Spencer S.A. Jones C.U. Garden A.S. Shenouda G. Harris J. Ang K.K. Final results of local-regional control and late toxicity of RTOG 9003: A randomized trial of altered fractionation radiation for locally advanced head and neck cancer Int. J. Radiat. Oncol. Biol. Phys. 2014 89 13 20 10.1016/j.ijrobp.2013.12.027 24613816 96. Graboyes E.M. Hand B.N. Ellis M.A. Huang A.T. Nilsen M.L. Pipkorn P. Marsh C.H. Maurer S. Day T.A. Sterba K.R. Validation of a Novel, Multidomain Head and Neck Cancer Appearance- and Function-Distress Patient-Reported Outcome Measure Otolaryngol. Head Neck Surg. 2020 163 979 985 10.1177/0194599820927364 32482151 97. Fingeret M.C. Yuan Y. Urbauer D. Weston J. Nipomnick S. Weber R. The nature and extent of body image concerns among surgically treated patients with head and neck cancer Psycho-Oncology 2012 21 836 844 10.1002/pon.1990 21706673 98. Ichikura K. Yamashita A. Sugimoto T. Kishimoto S. Matsushima E. Persistence of psychological distress and correlated factors among patients with head and neck cancer Palliat. Support. Care 2016 14 42 51 10.1017/S1478951515000711 26089108 99. Kisser U. Adderson-Kisser C. Coenen M. Stier-Jarmer M. Becker S. Sabariego C. Harreus U. The development of an ICF-based clinical guideline and screening tool for the standardized assessment and evaluation of functioning after head and neck cancer treatment Eur. Arch. Otorhinolaryngol. 2017 274 1035 1043 10.1007/s00405-016-4317-6 27695934 100. Tschiesner U. Rogers S. Dietz A. Yueh B. Cieza A. Development of ICF core sets for head and neck cancer Head Neck 2010 32 210 220 10.1002/hed.21172 19572286 101. Beck A.C.C. Kieffer J.M. Retèl V.P. van Overveld L.F. Takes R.P. van den Brekel M.W. van Harten W.H. Stuiver M.M. Mapping the EORTC QLQ-C30 and QLQ-H&N35 to the EQ-5D for head and neck cancer: Can dis-ease-specific utilities be obtained? PLoS ONE 2019 14 e0226077 31834892

PMC009xxxxxx/PMC9099626.txt

==== Front J Clin Med J Clin Med jcm Journal of Clinical Medicine 2077-0383 MDPI 10.3390/jcm11092541 jcm-11-02541 Article Bereaved Family Members’ Perspectives of Good Death and Quality of End-of-Life Care for Malignant Pleural Mesothelioma Patients: A Cross-Sectional Study Nagamatsu Yasuko 1* Sakyo Yumi 1 Barroga Edward 1 Koni Riwa 2 Natori Yuji 3 Miyashita Mitsunori 4 Bertolaccini Luca Academic Editor 1 Graduate School of Nursing Science, St. Luke’s International University, 10-1 Akashi-cho, Chuo-ku, Tokyo 104-0044, Japan; yumi-sakyo@slcn.ac.jp (Y.S.); edward-barroga@slcn.ac.jp (E.B.) 2 St. Luke’s International Hospital, 9-1 Akashi-cho, Chuo-ku, Tokyo 104-8560, Japan; riwakoni@luke.ac.jp 3 Hirano Kameido Himawari Clinic, 7-10-1 Kameido, Koto-ku, Tokyo 136-0071, Japan; natori@himawari-clinic.jp 4 Department of Palliative Nursing, Health Sciences, Tohoku University Graduate School of Medicine, 2-1 Seiryomachi, Aoba-ku, Sendai 980-8575, Japan; miya@med.tohoku.ac.jp * Correspondence: sarah-nagamatsu@slcn.ac.jp; Tel./Fax: +81-3-5550-2262 01 5 2022 5 2022 11 9 254116 3 2022 28 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Objective: This study investigated whether malignant pleural mesothelioma (MPM) patients achieved good deaths and good quality of end-of-life care compared with other cancer patients from the perspective of bereaved family members in Japan. Methods: This cross-sectional study was part of a larger study on the achievement of good deaths of MPM patients and the bereavement of their family members. Bereaved family members of MPM patients in Japan (n = 72) were surveyed. The Good Death Inventory (GDI) was used to assess the achievement of good death. The short version of the Care Evaluation Scale (CES) version 2 was used to assess the quality of end-of-life care. The GDI and CES scores of MPM patients were compared with those of a Japanese cancer population from a previous study. Results: MPM patients failed to achieve good deaths. Only 12.5% of the MPM patients were free from physical pain. The GDI scores of most of the MPM patients were significantly lower than those of the Japanese cancer population. The CES scores indicated a significantly poorer quality of end-of-life care for the MPM patients than the Japanese cancer population. The total GDI and CES scores were correlated (r = 0.55). Conclusions: The quality of end-of-life care for MPM patients remains poor. Moreover, MPM patients do not achieve good deaths from the perspective of their bereaved family members. mesothelioma asbestos rare lung disease palliative care good death quality of care Japan Society for the Promotion of Science Grants-in-Aid for Scientific Research (JSPS KAKENHI)16H05579 Ministry of Health, Labor and Welfare, Japan210901-01 This work was supported by the Japan Society for the Promotion of Science Grants-in-Aid for Scientific Research (JSPS KAKENHI), grant number 16H05579 and grants-in-aid from the Ministry of Health, Labor and Welfare, Japan, grant number 210901-01. ==== Body pmc1. Introduction Malignant pleural mesothelioma (MPM) is a rare fatal malignancy caused mainly by asbestos [1]. The number of people with MPM who die each year in Japan is about 1550, and that number is growing [2]. It is estimated that Japan will have 66,000–100,000 deaths from mesothelioma between the years 2003 and 2050 [3,4]. The median survival from the time of diagnosis in Japan is 7.9 months [5]. MPM causes a series of debilitating physical symptoms, such as chest pain, dyspnea, fatigue, anorexia, insomnia, constipation, and sweating [6,7,8,9,10,11]. Psychological issues, such as uncertainty, lack of control [12], memory problems, difficulties in concentrating, feeling that problems cannot be solved [13], depression, anxiety, fear, and isolation [8], all negatively affect the quality of life of MPM patients. Finally, there is additional psychological distress for victims of the asbestos industry [14]. Suffering from asbestos-related disease causes fear of premature death [15]. MPM patients in Japan reportedly suffer from physical and psychological distress [16], and their quality of life is impaired [9]. Lamentably, the quality of life of MPM patients in the terminal stage, particularly their achievements of good deaths and good quality of end-of-life care, has been scarcely researched and thus remains poorly understood. Unfortunately, there are barriers to conducting research on MPM patients in their terminal stage. These include their small population, and the short lengths of time between disease diagnosis, debilitation, and death. Moreover, conducting research on terminally ill patients imposes unnecessary burdens on them. Therefore, many studies are conducted with bereaved family members [13,17,18,19,20] to evaluate the patients’ achievements of good deaths and the quality of their end-of-life care. This study aimed to investigate whether MPM patients achieved good deaths and good quality of end-of-life care compared with other cancer patients in Japan from the perspective of bereaved family members. The data for the other cancer patients in Japan were taken from a previous study [21]. 2. Methods 2.1. Study Design, Participants, and Setting This study used a cross-sectional survey design to examine the achievements of good death and good quality of end-of-life care for MPM patients from the perspective of bereaved family members. The inclusion criteria for bereaved family members were as follows: (1) had lost a loved one to MPM, (2) had a loved one who had been diagnosed with MPM after 2008 when the first evidence-based chemotherapy succeeded in prolonging the survival of MPM patients, and (3) could respond to a self-administered questionnaire written in Japanese. The exclusion criterion was a bereaved family member who had experienced a loss within six months. This research is part of a larger study which also investigated the complicated grief of the bereaved family members of MPM patients. According to the previous study, the diagnosis of complicated grief should be made at least six months after the death of a family member [22]. A request for cooperation was sent to the advocacy group of the Japan Association of Mesothelioma and Asbestos-Related Disease Victims and their Families. The association has 15 branches across Japan and works with approximately 700 victims of asbestos-related diseases and their families. The association sent the informed consent information and questionnaires to 109 bereaved family members in November 2016. Those agreeing to participate returned the completed questionnaires via postal mail by March 2017. 2.2. Outcomes The primary outcomes were the achievements of a good death and good quality of end-of-life care for MPM patients. The secondary outcome was the presence of the common symptoms of MPM. 2.3. Instruments 2.3.1. Information of Patients and Bereaved Family Members The following information was provided by the bereaved family members about the deceased patients: sex, age at diagnosis, survival and received treatments, receipt of two types of insurance compensation benefits, and place of death. The information about the bereaved family members included the following: age, relationship to the patient, time of bereavement, experience of end-of-life discussion with the patient, timing of patient’s death, financial impact of patient’s MPM on family, and level of anger toward asbestos. The bereaved family members were also asked about their satisfaction with care on diagnosis, when the patient became critical, and when the patient died. 2.3.2. Good Death Inventory Achievement of good death was measured using the Good Death Inventory (GDI), which had internal consistency (α = 0.74–0.95) and acceptable test–retest reliability (intra-class correlation coefficient = 0.38–0.72) [17]. The GDI was validated to evaluate the achievement of good death from the perspective of bereaved family members in Japan [17]. The GDI has 18 items consisting of 10 core items and 8 optional items, and is answered using a seven-point Likert scale (1 = absolutely disagree, 7 = absolutely agree). The possible scores range from 18 to 126, and a high score indicates the achievement of a good death. 2.3.3. Care Evaluation Scale The short version of the Care Evaluation Scale (CES) version 2 (Cronbach’s α = 0.96) was used to evaluate the quality of end-of-life care in Japan [23]. The CES consists of 10 items. The bereaved family members answered using a six-point Likert scale (1 = highly disagree, 6 = highly agree). A high total CES score indicates a good quality of end-of-life care. 2.3.4. Symptoms The presence of the common symptoms of MPM, namely, pain, dyspnea, anorexia, fatigue, anxiety, dysphagia, constipation, nausea, insomnia, edema, and palpitation, was asked with respect to two time points. These time points were (1) at the end of chemotherapy (only for the bereaved family members of patients who received chemotherapy—i.e., when chemotherapy was stopped, being no longer effective), and (2) at the final critical stage (i.e., when the patient entered the critical stage). The bereaved family members checked the items of symptoms the MPM patients experienced. These two time points enabled the comparison of the present results with previous results that reported on the care needs of patients because of their severe symptoms [16]. 2.4. Missing Data Mean imputation was conducted for the missing data of GDI and CES scores according to the instructions for the tools. 2.5. Comparison of Study Data A nationwide project to evaluate hospice and palliative care in Japan was previously conducted by Miyashita et al. and reported as the Japan Hospice and Palliative care Evaluation (J-HOPE) study [21]. This project evaluated the end-of-life care of cancer patients from the perspective of bereaved family members in nationwide designated cancer centers, inpatient palliative care units (PCUs), and home hospices. The study focused on care satisfaction, the structure and process of care, and the achievement of a good death. This previous study compared the data according to the last place of care. Data from this previous study were provided to us by Dr. Miyashita, who is a co-author of the present study. There were 8398 questionnaire responses from family members that were analyzed by Miyashita et al. [24]. 2.6. Statistical Analysis The scores of each scale were calculated using a previously reported scoring procedure [17,23]. The scores of the measurement tool items in GDI and CES were totaled and compared with those of cancer patients in the J-HOPE study [21]. The GDI and CES mean scores in the J-HOPE study [21] were calculated according to the place of death and compared with the GDI and CES mean scores in the present study. The achievements of good death (measured using GDI) and good quality of end-of-life care (measured using CES) scores in the present MPM study and the previous J-HOPE study were compared using the binominal test. The GDI and CES total scores in the present MPM study and the previous J-HOPE study were compared using a one-sample t-test. The correlations between the GDI and the CES were examined. Thereafter, the GDI scores and the patients’ and bereaved participants’ information were examined. Sex, receiving treatments, approval for compensation, experience of end-of-life discussion with patients, and satisfaction of care were treated as dichotomous variables. Finally, the coefficients and their 95% confidence intervals estimated by multiple regression analysis were used to assess the correlations between the GDI and CES scores and the clinical social factors. A p-value of < 0.05 was considered to indicate a statistically significant difference. Statistical analysis was performed using SPSS version 27. 2.7. Ethical Consideration This study was approved by the Research Ethics Committee of St. Luke’s International University (16-A035). It was conducted based on the ethical principles of avoiding harm, voluntary participation, anonymity, and the protection of privacy and personal information. 3. Results Of the 109 questionnaires distributed to the bereaved family members through the related victims and family advocacy group, 74 (67.9%) were completed and returned via postal mail by the end of March 2017. Two bereaved family member respondents who had experienced a loss within the last six months were excluded. Thus, a total of 72 questionnaires were analyzed. 3.1. Characteristics of Malignant Pleural Mesothelioma Patients and Bereaved Family Members As shown in Table 1, 81.9% of the deceased MPM patients were men, and their mean age at diagnosis was 66.9 years. The treatment modalities they received were chemotherapy (70.8%), palliative care (56.9%), and surgery (19.4%). A large minority (48.6%) died in the respiratory ward, followed by the PCU or hospice (33.3%). Only 13.9% died at home. The mean survival time was 14.5 months from the time of diagnosis. The majority of the bereaved family members (72.2%) was spouses of the MPM patients, and the mean bereavement time was 45.2 months. 3.2. Achievement of Good Death The obtained data revealed that MPM patients failed to achieve good deaths. The mean total GDI score of the MPM patients was 61.9 ± 15.7, which was significantly lower than the 81.1 of the J-HOPE cancer patients. Figure 1 shows the comparison of the percentage scores of MPM patients and J-HOPE cancer patients for the GDI items for the achievement of good death. The lowest percentages of achievement by the MPM patients in the 10 core items of the GDI were for the items “being free from physical distress” (12.5%) followed by “feeling that life is completed” (18.1%) and “having some pleasure in daily life” (27.8%). The binominal test showed that the percentages regarding the achievement of a good death in the MPM patients were significantly lower than those in the J-HOPE cancer patients in all items, except for the following four items: “being independent in daily activities”, “knowing what to expect about the future condition”, “living in calm circumstances”, and “supported by religion”. The greatest gaps in the achievement of good death between the MPM patients and the J-HOPE cancer patients were for “being free from physical distress”, which was true for 12.5% of the MPM patients compared with 64.7% of the J-HOPE cancer patients, followed by “not exposing one’s physical and mental weakness to family”, “dying a natural death”, and “feeling life is completed”. 3.3. Quality of End-of-Life Care The total scores of CES in the MPM patients and the J-HOPE cancer patients were significantly different, as shown in Figure 2. The mean total score of CES in the MPM patients was 70.3 ± 16.0, which was significantly lower than the 75.8 in the J-HOPE cancer patients. The binominal test showed that all the scores of the CES items indicated a significantly poorer quality of end-of-life care in the MPM patients than in the J-HOPE cancer patients except in the items “cost”, “coordination and consistency”, and “explanation to family by physician”. 3.4. Symptoms The percentages of MPM patients who experienced symptoms at the end of chemotherapy are shown in Figure 3, and the same percentages at the final critical stage are shown in Figure 4. More than half of the MPM patients experienced pain, dyspnea, anorexia, and anxiety at the end of chemotherapy. When the MPM patients reached the final critical stage, symptoms such as fatigue and dysphasia followed. 3.5. Factors Associated with a Good Death The GDI and CES total scores were significantly associated (correlation coefficient ρ = 0.554, p = 0.0001), indicating that the patients who received better end-of-life care were more likely to achieve good deaths. The multiple regression analysis results are shown in Table 2. The final regression model for predicting good death showed that higher GDI scores were significantly related to the surveyed family member being female, the patient dying later than expected, and satisfaction with care when the patient became critical. 3.6. Factors Associated with Quality of End-of-Life Care The final regression model for predicting good death (Table 3) showed that higher CES scores were significantly related to the following: satisfied with the care received when the patient died, and Received chemotherapy. 4. Discussion In this study, we described the extent to which Japanese MPM patients achieved good deaths and their good quality of end-of-life care. The findings were compared with those of a large cohort of Japanese cancer patients from the J-HOPE study [21]. The present results demonstrate a lack of good deaths among MPM patients. The three main findings of this study are as follows: (1) there was a remarkable lack of good deaths among the MPM patients; (2) there was an enormous burden of symptoms in the MPM patients; and (3) the quality of end-of-life care in the MPM patients was poorer than that in the J-HOPE cancer patients. The CES score was correlated with the GDI score, consistent with the findings of Miyashita et al. [17]. The final regression model showed that a higher GDI score was significantly related to the surveyed family member being female, the patient dying later than expected, and satisfaction with the care received when the MPM patient became critical. 4.1. Poor Achievement of Good Death This study showed an extreme lack of good deaths among the MPM patients. The lowest score from among the 10 GDI core items was for the item “being free from physical distress” (12.5%), which was significantly lower than the 62.9% score for the Japanese cancer population [21]. Symptom management is difficult in MPM patients, possibly because (1) MPM progresses rapidly and causes a variety of severe symptoms [6,9,25,26]; and (2) MPM results in anger and negative feelings of injustice [7,14,16], which tend to complicate the patient’s physiological distress more than other malignancies. Additionally, MPM has the potential to cause spiritual pain. Some studies have advocated care to ease the spiritual pain of MPM patients [27,28]. Only 18.1% of the MPM patients in the present study had the “feeling that life is completed”, which was significantly lower than the figure of 49.9% among the cancer population [21]. The possible reasons are as follows: (1) In this current study, the mean age of diagnosis was 66.9 years, and the mean survival time was only 14.5 months. The patients died relatively young, and they had very little time to complete their lives and face their deaths. (2) As the cause of MPM was asbestos and not one’s own doing, the patient may have felt that death from MPM was unfair. For patients with MPM, “Dying without awareness that one is dying” (4.2%) was, for the most part, not possible. Patients were told at the time of their diagnosis that their disease was incurable [7]. Only 11.1% of the MPM patients felt “supported by religion”; however, this percentage was not significantly different from the 19.6% of the cancer population [21]. As Ando et al. [29] reported, religious care is not very common in Japan. The multiple regression analysis showed that the family member surveyed being female, the patient dying later than expected, and satisfaction with care when the patient became critically ill were related to the GDI score. It is not clear why the family member surveyed being female was related to a higher GDI score. One possibility is that a higher number of Japanese females do not work and focus on caregiving; however, we did not ask about the jobs of the bereaved family members. It is necessary to investigate the relationship between the gender of the family member and the achievement of a good death. Carr [30] reported that the interval between the onset of terminal illness and death provided opportunities for people to plan their end-of-life care. However, an MPM diagnosis leaves a much shorter time for patients than in most cases, especially for those who died sooner than expected, reducing their capacity to prepare for good deaths. The satisfaction with care when patients become critical is related to the achievement of a good death, which is consistent with the findings in the “Good Death” study by Miyashita et al. [17]. For patients with MPM to achieve a good death, preparation for the acute exacerbation of the disease and the implementation of physical, psychological, and spiritual care in a timely manner are crucial. 4.2. Heavy Symptom Burden The present results show that the MPM patients experienced various kinds of symptoms. As shown in other published studies [6,9,25,26,31], pain, dyspnea, anorexia, and fatigue were the major symptoms exhibited by the MPM patients. The major symptoms of MPM patients are similar to the major symptoms of lung cancer patients, with a high prevalence of pain, fatigue, dyspnea, anorexia, and anxiety [6,32]. An important outcome of the present study was that it revealed the high prevalence of the various symptoms of MPM patients at the end of chemotherapy. For symptom management in MPM, several studies have recommended the introduction of palliative care in the early stages of MPM [26,33]. Unfortunately, similarly to cancer patients [34], MPM patients often refuse palliative care because of their denial of the fatal nature of the disease. They are thus unwilling to end their anticancer treatment and enter palliative care. Advanced care planning is encouraged; however, this is challenging for MPM patients, who have short prognoses. Horne et al. reported that discussions about end-of-life care planning following the disclosure of a terminal prognosis caused a feeling of abandonment [35]. 4.3. Poor Quality of End-of-Life Care The present results show a poor quality of end-of-life care for MPM patients in Japan and significantly worse care than for other cancer patients. The possible reasons for this poor quality of end-of-life care could be (1) the limited availability of treatment for MPM, which has recently improved in Japan [36]; and (2) the health providers’ lack of knowledge and skills regarding the treatment and care of MPM patients [8]. As the multiple regression analysis showed that “Satisfaction with the care received when the patient died” and “Received chemotherapy” were related to the CES score, improvements in end-of-life care are recommended through (1) the assurance of quality care on the death bed, and (2) the provision of continuous end-of-life care to patients who do not receive chemotherapy. 4.4. Implications for Care and Further Research The MPM patients experienced various symptoms at the end of chemotherapy and when they entered the final critical stage. Medical professionals need to understand that MPM patients develop various symptoms in the early stages of the disease, even when treated with chemotherapy. Thus, medical professionals need to inform MPM patients regarding the possible symptoms that they will encounter and advise them on how to prepare, which may be challenging for patients. To support MPM patients at this difficult time, transition care is crucial. The care for MPM patients must include (1) symptom management from the earliest stage; (2) care for psychological, social, and spiritual pain; and (3) care for their families as provided by a multidisciplinary team, consisting of a patient and family advocacy group, and a lawyer [10,27,28]. 4.5. Limitations This study has some limitations. First, not all of the bereaved family members of the deceased MPM patients were contacted, as Japan has no registration system for MPM patients. Therefore, this study had a small sample. Second, as the participants were members of the advocacy group, it is uncertain whether the results are representative of the general population of bereaved family members of deceased MPM patients. The patients and family advocacy group, with their network of medical staff and hospitals, may have represented bereaved family members who are less distressed by the care their loved ones receive, thus representing a biased group. Third, the mean number of months of bereavement was 45.2; therefore, the participants may have had recall bias or forgotten key factors. Finally, this study was a cross-sectional study, and therefore, no causal relationships were established. To overcome the limitations regarding representativeness, it is necessary to conduct census surveys based on an MPM registration system, as this will allow representative random samplings. 5. Conclusions This cross-sectional study revealed the remarkably rare achievement of a good death among MPM patients in Japan. The MPM patients experienced an enormous burden from their symptoms and were seldom free of physical distress. Another challenge faced by MPM patients in the achievement of a good death was the sense of life completion, which was difficult for patients with MPM caused by asbestos. The quality of end-of-life care of MPM patients was poorer than that of other cancer patients. The GDI score of the MPM patients was closely correlated with their CES score. Further research and interventions are urgently required, aimed at achieving a good death for MPM patients by providing quality continuous care, including (1) symptom management from the earliest stage; (2) care for psychological, social, and spiritual pain; and (3) care for their families as provided by a multidisciplinary team. Acknowledgments We appreciate the participants and Sarah E. Porter for editorial assistance. Author Contributions Conceptualization and design of the investigational plan, Y.N. (Yasuko Nagamatsu), Y.S., R.K., Y.N. (Yuji Natori) and M.M.; Data curation, Y.N. (Yasuko Nagamatsu); Formal analysis, Y.N. (Yasuko Nagamatsu), Y.S., Y.N. (Yuji Natori) and M.M.; Funding acquisition, Y.N. (Yasuko Nagamatsu); Investigation, Y.N. (Yasuko Nagamatsu) and Y.S.; Methodology, Y.N. (Yasuko Nagamatsu), Y.S., E.B. and M.M.; Project administration, Y.N. (Yasuko Nagamatsu); Resources, R.K. and Y.N. (Yuji Natori); Supervision, Y.N. (Yasuko Nagamatsu); Validation, Y.N. (Yasuko Nagamatsu); Visualization, Y.N. (Yasuko Nagamatsu) and E.B.; Writing—original draft, Y.N. (Yasuko Nagamatsu) and E.B.; Writing—review and editing, Y.N. (Yasuko Nagamatsu), Y.S., E.B., R.K., Y.N. (Yuji Natori) and M.M. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki and approved by the Research Ethics Committee of St. Luke’s International University (16-A035). Informed Consent Statement This research was conducted based on the ethical principles of avoiding harm, voluntary participation, anonymity, and the protection of privacy and personal information. The purpose, procedures, and confidentiality of the study were explained in written format. The participants were informed that nonparticipation would not disadvantage them. Answering the questionnaire and sending it to the authors was regarded as written informed consent to participate in the study. Informed consent was obtained from all the subjects for the publication of their details. Data Availability Statement The datasets generated and analyzed from this study are not publicly available to protect the anonymity of the participants but are available from the corresponding author, Yasuko Nagamatsu, upon reasonable request. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Comparison of the percentages of MPM patients and J-HOPE cancer patients concerning GDI items for the achievement of a good death. Sum of “somewhat agree”, “agree”, and “absolutely agree” responses. Data of cancer patients were from the J-HOPE national survey of Japanese cancer patients (reference [21]). Weighted means of GDI scores in general cancer patients in Japan (reference [21]) were calculated according to the place of death. Core and optional items were established by factor analysis (reference [17]). * p < 0.05, ** p < 0.005, *** p < 0.001. Figure 2 Comparison of the percentages of MPM and J-HOPE cancer patients with regard to CES items for achieving good quality end-of-life care. Sum of “somewhat agree”, “agree”, and “absolutely agree”. The weighted means of CES scores in general cancer patients in Japan were calculated according to the place of death. Data are from the J-HOPE study (Reference [21]). ** p < 0.005, *** p < 0.001. Figure 3 Percentages of MPM patients experiencing symptoms at the end of chemotherapy (n = 51). Figure 4 Percentages of MPM patients experiencing symptoms at the final critical stage (n = 72). jcm-11-02541-t001_Table 1 Table 1 Comparison of the characteristics of malignant pleural mesothelioma patients and cancer patients, and their bereaved participants. Disease MPM Cancer * n = 72 Place of Death Designated Cancer Center (n = 2794) Palliative Care Unit (n = 5312) Home Hospice (n = 292) Patients n % n % n % n % Sex Men 59 81.9 1820 65.1 2906 54.7 181 62 Women 13 18.1 973 34.8 2364 44.5 111 38 Primary cancer site Pleura ** 72 100 - - - - - - Lung 0 0 688 24.6 1246 23.5 63 21.6 Stomach 0 0 395 14.1 635 12 36 12.3 Colorectum/rectum 0 0 260 9.3 651 12.3 54 18.5 Liver 0 0 279 10 281 5.3 18 6.2 Gall bladder/bile duct 0 0 165 5.9 201 3.8 14 4.8 Pancreas 0 0 243 8.7 398 7.5 18 6.2 Esophagus 0 0 112 4 184 3.5 8 2.7 Breast 0 0 83 3 266 5 8 2.7 Others - - 513 18.4 1389 26.2 69 23.7 Source of asbestos exposure Occupation 49 68.1 Neighboring factory 17 23.6 School 1 1.4 Family 1 1.4 Unknown 4 5.4 Treatment Surgery 14 19.4 (includes multiple treatments) Extrapleural pneumonectomy 12 16.7 Pleurectomy decoration 2 2.8 Chemotherapy 51 70.8 Radiotherapy 15 20.8 Palliative care 41 56.9 Compensated Workmen’s accident compensation insurance 47 65.3 (some had both types) Asbestos-related health damage relief system 56 77.8 Place of death Respiratory ward 35 48.6 Palliative care unit/hospice 24 33.3 Home 10 13.9 Other 3 4.2 Age at diagnosis (years) Range: 36–92 Mean ± SD 66.9 ± 9.6 69.8 ± 11.5 70.9 ± 12.1 71.8 ± 13.0 Survival (months) 0.5–69 14.5 ± 14.1 Bereaved family members n % n % n % n % Sex Men 15 20.8 825 29.5 1694 31.9 60 20.6 Women 57 79.2 1696 60.7 3556 67.1 228 78.1 Relationship with patient Spouse 52 72.2 1535 54.9 2506 47.2 165 56.5 Child 20 17.8 672 24.1 1809 34.1 78 26.7 Son/daughter-in-law 0 0 181 6.5 353 6.7 34 11.6 Parent 0 0 49 1.8 100 1.9 4 1.4 Sibling 0 0 56 2 310 5.8 6 2.1 Others 0 0 32 1.2 188 3.5 4 1.4 Experience of end-of-life discussion with patient Yes 27 37.5 No 44 61.1 Timing of patient’s death Much sooner than expected 31 43.1 Sooner than expected 25 34.7 Moderate 9 12.5 Later than expected 5 6.9 Much later than expected 2 2.8 Satisfaction with care on diagnosis Satisfied 29 40.3 Not satisfied 43 59.7 When patient became critical Satisfied 31 38.9 Not satisfied 41 61.1 When patient died Satisfied 47 65.3 Not satisfied 25 34.7 Financial impact of patient’s MPM on family Significant impact 12 16.7 Some impact 15 20.8 Moderate impact 20 27.8 Minor impact 15 20.8 No impact 10 13.9 Level of anger toward asbestos Very angry 56 77.8 Angry 11 15.3 Moderately angry 4 5.6 Slightly angry 1 1.4 Not angry at all 0 0 Age (in years) Range: 32–82 Mean ± SD 62.5 ± 12.2 60.4 ± 12.5 59.3 ± 12.8 60.6 ± 12.1 Time since bereavement (months) 9–110 45.2 ± 27.2 12.4 ± 3.5 11.8 ± 3.7 12.2 ± 6.6 * Cited from the J-HOPE study (reference [21]). ** Pleural mesothelioma was classified as “Others” in the J-HOPE study. MPM = malignant pleural mesothelioma. jcm-11-02541-t002_Table 2 Table 2 Multiple regression model predicting good death (n = 72). Dependent Variable: GDI Total Score (F = 9.098, p = 0.0001, Adjusted R2 = 0.260) Model B SE β t 95% CI p-Value Constant 41.724 4.769 8.794 32.202–51.246 0.001 Satisfied with care received when patient became critical 11.597 3.278 0.370 3.538 5.053–18.141 0.001 Female bereaved family member 11.061 4.028 0.284 2.746 3.018–19.103 0.008 Patient died later than expected 3.270 1.556 0.220 2.102 0.164–6.376 0.039 Abbreviations: F, overall F-test for regression; R2, correlation of determination; B, unstandardized coefficient; SE, standard error; β, standardized coefficient (beta); t, independent-sample t test; CI, confidence interval. Note: The variables included were as follows: patient’s age on diagnosis; sex of patient; survival; whether the patient received certified workmen’s accident compensation insurance; whether the patient was certified for asbestos-related health damage relief system; whether the patient received surgery; whether the patient received chemotherapy; whether the patient received palliative care; age of bereaved family member; sex of bereaved family member; timing of patient’s death; bereaved family members’ level of anger toward asbestos; the financial impact of the patient’s MPM on the family; whether bereaved family members were satisfied with the care received on diagnosis; whether bereaved family members were satisfied with the care received when the patient became critical; whether family members were satisfied with the care received at the point of death; the relationship of patient and bereaved family members; and whether family members had an end-of-life discussion with the patient. jcm-11-02541-t003_Table 3 Table 3 Multiple regression model predicting quality end-of-life care (n = 72). Dependent Variable: CES Total Score (F = 34.558, p = 0.0001, Adjusted R2 = 0.493) Model B SE β t 95% CI p-Value Constant 30.545 1.807 16.907 26.939–34.152 0.001 Satisfied with the care received when the patient died 13.272 1.727 0.664 7.683 9.824–16.720 0.001 Received chemotherapy 4.048 1.832 0.191 2.209 0.391–7.705 0.031 Abbreviations: same as Table 2. Note: same as Table 2. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Bianchi C. Bianchi T. Malignant Mesothelioma: Global Incidence and Relationship with Asbestos Ind. Health 2007 45 379 387 10.2486/indhealth.45.379 17634686 2. Japan Ministry of Health Labor & Welfare Yearly Changes (from 1995 to 2018) in Number of Death from Mesothelioma by Prefecture (Based on Vital Statistics) 2019 Available online: https://www.mhlw.go.jp/toukei/saikin/hw/jinkou/tokusyu/chuuhisyu17/index.html (accessed on 25 May 2021) 3. Myojin T. Azuma K. Okumura J. Uchiyama I. Future trends of mesothelioma mortality in Japan bas ed on a risk function Ind. Health 2012 50 197 204 10.2486/indhealth.MS1184 22453207 4. Murayama T. Takahashi K. Natori Y. Kurumatani N. Estimation of future mortality from pleural malignant mesothelioma in Japan based on an age-cohort model Am. J. Ind. Med. 2005 49 1 7 10.1002/ajim.20246 16362942 5. Gemba K. Fujimoto N. Kato K. Aoe K. Takeshima Y. Inai K. Kishimoto T. National survey of malignant mesothelioma and asbestos exposure in Japan Cancer Sci. 2012 103 483 490 10.1111/j.1349-7006.2011.02165.x 22126592 6. Hollen P.J. Gralla R.J. Liepa A.M. Symanowski J.T. Rusthoven J.J. Adapting the lung cancer symptom scale (LCSS) to mesothelioma: Using the LCSS-Meso conceptual model for validation Cancer 2004 101 587 595 10.1002/cncr.20315 15274072 7. Clayson H. Seymour J. Noble B. Mesothelioma from the Patient’s Perspective Hematol. Clin. N. Am. 2005 19 1175 1190 10.1016/j.hoc.2005.09.003 8. Moore S. Darlison L. Tod A.M. Living with mesothelioma. A literature review Eur. J. Cancer Care 2010 19 458 468 10.1111/j.1365-2354.2009.01162.x 9. Nagamatsu Y. Oze I. Aoe K. Hotta K. Kato K. Nakagawa J. Hara K. Kishimoto T. Fujimoto N. Quality of life of survivors of malignant pleural mesothelioma in Japan: A cross sectional study BMC Cancer 2018 18 350 10.1186/s12885-018-4293-x 29587685 10. Woolhouse I. Bishop L. Darlison L. De Fonseka D. Edey A. Edwards J. Faivre-Finn C. Fennell D.A. Holmes S. Kerr K.M. British Thoracic Society Guideline for the investigation and management of malignant pleural mesothelioma Thorax 2018 73 i1 i30 10.1136/thoraxjnl-2017-211321 29444986 11. Muers M.F. Stephens R.J. Fisher P. Darlison L. Higgs C.M. Lowry E. Nicholson A.G. O’Brien M. Peake M. Rudd R. Active symptom control with or without chemotherapy in the treatment of patients with malignant pleural mesothelioma (MS01): A multicentre randomised trial Lancet 2008 371 1685 1694 10.1016/S0140-6736(08)60727-8 18486741 12. Arber A. Spencer L. Parker A. ‘It’s all bad news’: The first 3 months following a diagnosis of malignant pleural mesothelioma BMJ Support. Palliat. Care 2011 1 A14 A15 10.1136/bmjspcare-2011-000020.41 13. Innamorati M. Tamburello S. Tamburello A. Casale S. Cont C. Guglielmucci F. Granieri A. Quality of life and personality traits in patients with malignant pleural mesothelioma and their first-degree caregivers Neuropsychiatr. Dis. Treat. 2013 9 1193 1202 10.2147/NDT.S48965 23983468 14. Hughes N. Arber A. The lived experience of patients with pleural mesothelioma Int. J. Palliat. Nurs. 2008 14 66 71 10.12968/ijpn.2008.14.2.28597 18414321 15. Barak Y. Achiron A. Rotstein Z. Elizur A. Noy S. Stress associated with asbestosis: The trauma of waiting for death Psycho-Oncol. 1988 7 126 128 10.1002/(SICI)1099-1611(199803/04)7:2<126::AID-PON291>3.0.CO;2-A 16. Nagamatsu Y. Horiuchi S. Natori Y. The stages and difficulties of patients with malignant pleural mesothelioma J. Hum. Care Stud. 2012 12 69 81 17. Miyashita M. Morita T. Sato K. Hirai K. Shima Y. Uchitomi Y. Good Death Inventory: A Measure for Evaluating Good Death from the Bereaved Family Member’s Perspective J. Pain Symptom Manag. 2008 35 486 498 10.1016/j.jpainsymman.2007.07.009 18. Addington-Hall J. O’Callaghan A. A comparison of the quality of care provided to cancer patients in the UK in the last three months of life in in-patient hospices compared with hospitals, from the perspective of bereaved relatives: Results from a survey using the VOICES questionnaire Palliat. Med. 2009 23 190 197 10.1177/0269216309102525 19251834 19. Rhodes R.L. Mitchell S.L. Miller S.C. Connor S.R. Teno J.M. Bereaved family members’ evaluation of hospice care: What factors influence overall satisfaction with services? J. Pain Symptom Manag. 2008 35 365 371 10.1016/j.jpainsymman.2007.12.004 20. Downey L. Curtis J.R. Lafferty W.E. Herting J.R. Engelberg R.A. The quality of dying and death questionnaire (QODD): Empirical domains and theoretical perspectives J. Pain Symptom Manag. 2010 39 9 22 10.1016/j.jpainsymman.2009.05.012 21. Miyashita M. Morita T. Sato K. Tsuneto S. Shima Y. A Nationwide Survey of Quality of End-of-Life Cancer Care in Designated Cancer Centers, Inpatient Palliative Care Units, and Home Hospices in Japan: The J-HOPE Study J. Pain Symptom Manag. 2015 50 38 47.e3 10.1016/j.jpainsymman.2015.01.007 22. Shear M.K. Simon N. Wall M. Zisook S. Neimeyer R. Duan N. Reynolds C. Lebowitz B. Sung S. Ghesquiere A. Complicated grief and related bereavement issues for DSM-5 Depress. Anxiety 2011 28 103 117 10.1002/da.20780 21284063 23. Miyashita M. Aoyama M. Nakahata M. Yamada Y. Abe M. Yanagihara K. Shirado A. Shutoh M. Okamoto Y. Hamano J. Development the Care Evaluation Scale Version 2.0: A modified version of a measure for bereaved family members to evaluate the structure and process of palliative care for cancer patient BMC Palliat. Care 2017 16 8 10.1186/s12904-017-0183-2 28114917 24. Miyashita M. Morita T. Tsuneto S. Sato K. Shima Y. The Japan HOspice and Palliative Care Evaluation Study (J-HOPE Study): Study Design and Characteristics of Participating Institutions Am. J. Hosp. Palliat. Med. 2008 25 223 232 10.1177/1049909108315517 25. Clayson H. The Experience of Mesothelioma in Northern England University of Sheffield Sheffield, UK 2007 26. Mercadante S. Degiovanni D. Casuccio A. Symptom burden in mesothelioma patients admitted to home palliative care Curr. Med. Res. Opin. 2016 32 1985 1988 10.1080/03007995.2016.1226165 27532369 27. Ahmedzai S. Clayson H. Supportive and Palliative Care in Mesothelioma University Press New York, NY, USA 2006 28. Abrahm J.L. Palliative Care for the Patient with Mesothelioma Semin. Thorac. Cardiovasc. Surg. 2009 21 164 171 10.1053/j.semtcvs.2009.06.006 19822289 29. Ando M. Kawamura R. Morita T. Hirai K. Miyashita M. Okamoto T. Shima Y. Value of religious care for relief of psycho-existential suffering in Japanese terminally ill cancer patients: The perspective of bereaved family members Psycho-Oncol. 2010 19 750 755 10.1002/pon.1625 30. Carr D. Death and dying in the contemporary United States: What are the psychological implications of anticipated death? Soc. Pers. Psychol. Compass 2012 6 184 195 10.1111/j.1751-9004.2011.00416.x 31. Bibby A.C. De Fonseka D. Morley A.J. Keenan E. Addeo A. Smith S. Edey A.J. Maskell N.A. Exploring the characteristics of patients with mesothelioma who chose active symptom control over chemotherapy as first-line treatment: A prospective, observational, single centre study BMC Palliat. Care 2017 16 71 10.1186/s12904-017-0255-3 29221485 32. Wysham N.G. Cox C.E. Wolf S.P. Kamal A.H. Symptom Burden of Chronic Lung Disease Compared with Lung Cancer at Time of Referral for Palliative Care Consultation Ann. Am. Thorac. Soc. 2015 12 1294 1301 10.1513/AnnalsATS.201503-180OC 26161449 33. Van Zandwijk N. Clarke C. Henderson D. Musk A.W. Fong K. Nowak A. Loneragan R. McCaughan B. Boyer M. Feigen M. Guidelines for the diagnosis and treatment of malignant pleural mesothelioma J. Thorac. Dis. 2013 5 E254 E307 24416529 34. Miyashita M. Hirai K. Morita T. Sanjo M. Uchitomi Y. Barriers to referral to inpatient palliative care units in Japan: A qualitative survey with content analysis Support. Care Cancer 2007 16 217 222 10.1007/s00520-007-0215-1 17318594 35. Horne G. Payne S. Seymour J. Do patients with lung cancer recall physician-initiated discussions about planning for end-of-life care following disclosure of a terminal prognosis? BMJ Support. Palliat. Care 2016 9 197 201 10.1136/bmjspcare-2015-001015 36. Nagamatsu Y. Oze I. Aoe K. Hotta K. Kato K. Nakagawa J. Hara K. Kishimoto T. Fujimoto N. Physician requests by patients with malignant pleural mesothelioma in Japan BMC Cancer 2019 19 383 10.1186/s12885-019-5591-7 31023248

PMC009xxxxxx/PMC9099627.txt

==== Front Int J Environ Res Public Health Int J Environ Res Public Health ijerph International Journal of Environmental Research and Public Health 1661-7827 1660-4601 MDPI 10.3390/ijerph19095059 ijerph-19-05059 Article Driver’s Visual Attention Characteristics and Their Emotional Influencing Mechanism under Different Cognitive Tasks Liu Yaqi 12 https://orcid.org/0000-0003-2418-7394 Wang Xiaoyuan 123* Chen Longfei 2 Liu Shijie 2 Han Junyan 2 Shi Huili 2 Zhong Fusheng 2 Tchounwou Paul B. Academic Editor 1 School of Transportation and Vehicle Engineering, Shandong University of Technology, Zibo 255049, China; 4019030007@mails.qust.edu.cn 2 College of Electromechanical Engineering, Qingdao University of Science & Technology, Qingdao 266000, China; chenlongfei@mails.qust.edu.cn (L.C.); 4019030012@mails.qust.edu.cn (S.L.); hanjunyan@mails.qust.edu.cn (J.H.); shihuili@qust.edu.cn (H.S.); bh136@qust.edu.cn (F.Z.) 3 Collaborative Innovation Center for Intelligent Green Manufacturing Technology and Equipment of Shandong, Qingdao 266000, China * Correspondence: wangxiaoyuan@qust.edu.cn; Tel.: +86-138-6445-5865 21 4 2022 5 2022 19 9 505902 3 2022 19 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). The visual attention system is the gateway to the human information processing system, and emotion is an important part of the human perceptual system. In this paper, the driver’s visual attention characteristics and the influences of typical driving emotions on those were explored through analyzing driver’s fixation time and identification accuracy to different visual cognitive tasks during driving. The results showed that: the increasing complexity of the cognitive object led to the improvement of visual identification speed. The memory and recall process increased drivers’ fixation time to cognitive objects, and the recall accuracy decreased with the increase in time interval. The increase in the number of cognitive objects resulted in the driver improving the visual identification speed for the cognitive object at the end of the sequence consciously. The results also showed that: the visual cognitive efficiency was improved in the emotional states of anger and contempt, and was decreased in the emotional states of surprise, fear, anxiety, helplessness and pleasure, and the emotional state of relief had no significant effect on the visual cognitive efficiency. The findings reveal the driver’s visual information processing mechanism to a certain extent, which are of great significance to understand the inner micro-psychology of driver’s cognition. drivers visual cognition attention emotion traffic safety ==== Body pmc1. Introduction Among the many factors that are capable of explaining root causes of traffic accidents, human factors are important causes of preventable road-traffic incidents [1]. Rational control of driving behavior is an effective means to reduce human-caused traffic accidents [2]. The generation of drivers’ conscious behavior contains four stages: perception, cognition, decision making and action [3]. Among them, perception plays a decisive role in the decision making and is a key link in the generation of driving behavior. The sensory organs related to driver perception are mainly the visual organs, and more than 80% of the traffic information perceived by the driver originates from the visual channel [4,5]. The attention system guides the selective concentration of visual organs, and improves the awareness level of drivers to local stimuli [6]. The physiological structure of the visual system determines the limited breadth and depth of the driver’s visual perception of the traffic environment. Due to the limitation of attention resources, drivers often cannot process multiple cognitive activities at the same time [7]. The real traffic environment is complex and changeable, and missing key environmental information may lead to traffic accidents directly [8]. During the driving process, one of the main tasks of the driver is to allocate the limited physical and psychological resources reasonably from the spatial and temporal dimensions. Over the past decades, the rapid development of intelligent technology and the widespread implementation of Advanced Driving Assistance Systems (ADAS)s have effectively enhanced and extended the driver’s ability to perceive information in driving activities [9]. For example, the blind-spot monitoring system (BSM) is able to alert the driver of obstacles or oncoming traffic within the safety range behind. In more recent years, with the development of connected vehicle technology, it is foreseeable that drivers can obtain more comprehensive traffic environment information during driving in the future. However, it should be noted that human perception is limited (as mentioned above), and the environmental information delivered through the in-vehicle system is not always perceived and reasonably applied by drivers. Sometimes, too much environmental information provided can even disrupt the driver’s normal behavioral planning. Therefore, it is necessary to investigate the efficiency of drivers’ visual information perception under different cognitive tasks, as it is important to understand the boundaries of drivers’ information processing and to build a more intelligent and rational human–vehicle interaction system. The physiological–psychological characteristics of drivers can show significant differences in different environments [10,11]. Except for external factors, the changes in the physiological–psychological characteristics of drivers due to individual factors also cannot be ignored [12]. Of the many individual factors, emotion is an individual’s attitude experience towards objective things and the corresponding physiological–psychological response [13]. In emotional activities, individuals not only accept the influence of stimuli on themselves, but also regulate their own responses to stimuli [14]. Drivers with different emotional states have significant differences in their feelings, preferences and needs for external stimuli, which in turn lead to changes in their perception and cognitive abilities [15]. In recent years, with the increasing of vicious traffic accidents caused by negative emotions such as road rage and anxiety, as well as the rapid development of affective computing and cognitive psychology, a growing interest has been seen in exploring the effects of driver’s emotions on driver’s physiological–psychological characteristics. Scholars in related fields have generally recognized that driving emotion is an important factor that cannot be ignored in traffic safety research [16]. In a nutshell, emotion is an important part of the human perceptual system, and the visual attention system is the basis of the driver’s information perception. It is of great significance to explore drivers’ visual attention characteristics and the influence of emotions on visual attention for improving the level of human–vehicle intelligent interaction in road traffic systems and improving road traffic safety. 2. Literature Review In view of the important impact of drivers’ visual attention on traffic safety, scholars in related fields have conducted many studies on drivers’ eye movement, visual distraction, visual attention, etc. Xu, Y. et al. explored the relationship between the driver’s eye movement and the construction conflict through collecting and analyzing the driver’s eye movement data in simulated construction conflicts at different speeds [17]. Rahman, H. et al. presented five machine learning models and three deep learning architectures to classify a driver’s cognitive load based on driver’s eye movement signals [18]. Onkhar, V. et al. studied the effect of drivers’ eye contact on pedestrians’ perceived safety to cross the road, and demonstrated how drivers’ eye contact affects pedestrians’ perceived safety as a function of time in a dynamic scenario [19]. Li, N. and Busso, C. defined regression models with elastic net regularization and binary classifiers to separately estimate the cognitive and visual distraction levels and proposed a novel joint visual–cognitive distraction space to characterize driver behaviors [20]. Karthaus, M. et al. investigated the effects of acoustic and visual distracting stimuli on responses to critical events. The results demonstrate the high impact of distraction on driving performance in critical traffic situations and indicate a driving-related inhibition deficit in young and old drivers [21]. For the take-over of automated driving systems, the percentage of face orientation to distraction area and time to boundary at take-over timing were proposed by Li, Q. et al. to accurately evaluate the degree of visual distraction based on merely face orientation under naturalistic non-driving related tasks and to evaluate take-over performance, respectively [22]. Grahn, H. and Kujala, T. argued that visual distraction by secondary in-car tasks is a major contributing factor in traffic incidents and studied the effects of touch screen size, user interface design and subtask boundaries on in-car tasks’ visual demand and visual distraction potential [23]. Reimer, B. et al. assessed the sensitivity of visual attention and driving performance for detecting changes in driver cognitive workload across different age groups. The results showed that the degree of gaze concentration with added cognitive demand is not related to age and the driving performance measures did not show a consistent relationship with the objective demand level [24]. Muñoz, M. et al. investigated distinguishing patterns in drivers’ visual attention allocation and the results suggested that differences in glance allocation strategies serve as an effective evaluator of the visual demand of a vehicle interface [25]. Louw, T. and Merat, N. assessed drivers’ visual attention distribution during automation and on approach to a critical event, and examined whether such attention changes followed repeated exposure to an impending collision [26]. Lemonnier, S. et al. focused on three top-down factors that influence the collection of visual information: the value of visual information for the ongoing task, their bandwidth and the familiarity with the environment. Effects were found for each of the three factors in agreement with Wickens’ theoretical framework and with previous studies [27]. Young, K. et al. examined the nature of observable visual and/or manual secondary task interruptions in real-world driving. It was found that drivers interrupt only a small percentage of the secondary tasks they were engaged in, and the number of interruptions made to secondary tasks was found to differ according to some task characteristics [28]. Liu, Q. et al. tested the driver’s visual parameters when the vehicles run under the pothole repair environment and the results showed that psychology and the gaze frequency, gaze duration and saccade speed of drivers on pothole sections were significantly increased while the saccade range was reduced [29]. The driving emotion was concerned and taken as an important research object since the phenomenon that the drivers in malignant emotional states prefer to choose aggressive driving behavior, which would more likely lead to traffic accidents [30]. Further studies showed that driving emotions are inextricably linked to many risk-related factors [31]. When it comes to driving emotions, the most work in this area was oriented toward the driving emotion generation mechanism, the driving emotion recognition, and the impacts of emotions on drivers’ physiological and psychological characteristics. Barnard, M and Chapman, P. studied the relations of fear, trait anxiety, physiological and attentional reactions to accident risk. Analysis of the data suggested that fear increased with increasing accident risk, the eye movements indicated different patterns of performance according to different dangerous situations and the trait anxiety was only associated with higher rates of disliking driving and use of maladaptive coping mechanisms on questionnaires [32]. Roseborough, J. and Wiesenthal, D. examined the effect of various punishments (i.e., police enforcement, collision with a roadside object, collision with another vehicle, collision with a roadside object and police enforcement, collision with other vehicle and police enforcement) on witnesses’ feelings of anger and happiness on roadways. Analyses indicated that perceived punishment by police reduced feelings of anger and increased feelings of happiness compared to the other four forms of punishment [33]. Paschero, M. et al. proposed an Emotion Recognition System based on classical neural networks and neuro-fuzzy classifiers. In comparison with Multi-Layer Perceptron trained by EBP algorithm, the proposed Neuro-fuzzy classifiers showed very short training times, allowing applications with easy and automated setup procedures [34]. Wang, X. et al. established an online identification model for the driving emotions of joy, anger, sadness and fear based on the factor analysis method, the fuzzy comprehensive evaluation and the PAD emotional model [35]. Fairclough S and Dobbins C found that an ensemble classification model provided an accuracy rate of 73.12% for the binary classification of episodes of high vs. low anger based upon a combination of features derived from driving (e.g., vehicle speed) and cardiovascular psychophysiology (heart rate, heart rate variability, and pulse transit time) [36]. Chan, M. and Singhal, A. explored the behavioral and event-related potential (ERP) effects elicited by auditory presented words of different emotional valence during driving (dual-task) and non-driving (single-task) conditions. The results demonstrate that emotion-related auditory distraction can differentially affect driving performance depending on the valence of the emotional content [37]. Wang, X. et al. used multiple-electrocardiogram (ECG) feature fusion to recognize the driver’s emotion. Based on the back-propagation network and the Dempster–Shafer evidence method, the proposed model can recognize drivers’ anxiety with an accuracy rate of 92.89% [38]. Kadoya, Y. et al. examined the association between the taxi drivers’ on-duty emotional states and driving speed in real driving situations. The results revealed that negative emotions of taxi drivers (angry and sad) have significant impacts on increasing driving speed, a neutral emotional state is related to decreased speed, while a happy and relaxed emotional state shows no significant impact [39]. In summary, many scholars have conducted extensive and in-depth research on drivers’ visual attention and driving emotions, and have achieved fruitful research results. However, previous studies have rarely explored the driver’s visual cognitive process from the perspective of limited attentional resources, and paid less attention to the influences of emotions on drivers’ visual attention characteristics. To understand driver’s visual attention characteristics and their emotional influencing mechanism under different cognitive tasks, more systematic reviews and empirical research are required. The purpose of this study is to explore the visual attention characteristics of drivers when dealing with different cognitive tasks, and to reveal the influence mechanism of different emotions on drivers’ visual attention characteristics. This paper contains two parts of research. In study 1, the visual attention characteristics of drivers in response to different cognitive tasks were studied through designing and implementing visual identification tasks, visual working memory tasks and multiple visual identification tasks in virtual driving. In study 2, the effects of eight typical driving emotions (anger, surprise, fear, anxiety, helplessness, contempt, relief and pleasure) on the visual attention characteristics of drivers were examined based on the experimental framework of visual attention characteristic data collection proposed in Study 1 and the experimental framework of driving emotion activation and measurement proposed in our previous study [16,40]. 3. Study 1—Driver’s Visual Attention Characteristics in Different Cognitive Tasks 3.1. Materials and Methods 3.1.1. Participants Sixty-seven drivers (35 males and 32 females) aged from 21 to 48 (M = 26.36, SD = 4.87) were recruited to participate in this study. All of the participants were licensed drivers and their driving experience ranged from 2 to 15 years (M = 3.93, SD = 2.84). 3.1.2. Collection of Visual Attention Characteristic Data The experiment referred to collecting the visual attention data of the participants when they respond to different visual cognitive tasks in virtual driving through the eye-tracking system. In the experiment, the participants were required to choose any lane to drive in the virtual road environment at a speed of 80 km/h to 120 km/h on a driving simulator. The virtual driving environment was set as a two-way four-lane highway, and the single-lane traffic flow was set at 300 pcu/h. The driving simulator included a driving control module and an environment display module. The participants manipulated the vehicle through the control module, and the environment display module displayed the corresponding environmental visualization information dynamically. Three independent screens for playing visual cognitive materials were set between the driving control module and the environment display module. The participants were required to complete the visual cognitive tasks in the shortest possible time while completing driving tasks (maintaining speed and not violating traffic rules), and their visual attention data were captured and recorded by the eye-tracking system. The design idea of visual attention characteristic data collection is shown in Figure 1. Visual cognitive tasks included the Visual Identification Task (VIT), Visual Working Memory Task (VWMT) and Multiple Visual Identification Task (MVIT). The VIT was to examine visual attention time and identification accuracy when faced with a single cognitive object of varying information capacity. The VWMT was to examine visual attention time and identification accuracy when faced with a single cognitive object of varying information capacity. The MVIT was to examine visual attention allocation characteristics and identification accuracy when faced with multiple cognitive objects. In the VIT, the visual cognitive materials were static pictures randomly combined with the diagrams of different types of vehicles (Figure 2a). Each vehicle diagram contained in the visual perception picture was defined as a basic information unit, and the pictures containing 3, 5 and 7 basic information units were defined as ternary pictures, quintuple pictures and seven-element pictures, respectively. The VWMT included Visual Identification and Memory Task (VIMT) and Visual Identification and Recall Task (VIRT). A VIMT and a VIRT constituted a Memory Task Unit (MTU). In the VWMT, the visual cognitive materials were all seven-element pictures composed of vehicle diagrams. In the MVIT, the visual cognitive materials were played simultaneously on the three independent screens, and the content played on each screen was a schematic diagram of a traffic sign, a signal light or a vehicle mode in Figure 2, respectively. Each traffic sign or signal light schematic included in the picture was also identified as a basic information unit, and the visual materials played in the MVIT were all unary information pictures. The materials played on the three screens at one time were a random combination of a traffic sign, a signal light and a vehicle model diagram, and the content attributes of the pictures played on each screen are different (that is, when the traffic signs were played on screen 1, the other two screens will no longer select the traffic signs, and so on). A further elaboration on the different visual cognitive tasks is shown in Table 1, and the display timeline of visual cognitive materials in VIT, VWMT and MVIT are shown in Figure 3, Figure 4 and Figure 5, respectively. Before the visual attention characteristic data collection, the participants were trained to be proficient in the operation of the driving simulator and to understand the meaning of the vehicle models, traffic signs and signal lights in Figure 2. For each participant, the three kinds of visual tasks were performed in sequence, and the interval between each kind of task was set to 15 s. During this interval, the participants did not interrupt the virtual driving, and the duration of each visual attention characteristic data collection was about 10 min. In the experiment, the eye-tracking system was used to record the distribution of the participants’ fixation points in real time. The fixation point distribution of a participant in the experiment is shown in Figure 6. 3.1.3. Data Preprocessing The visual characteristic data obtained from the experiment were imported into the data analysis software of the eye-tracking system. The data analysis software can display the fixation index of the participants at any time, and each fixation point represents 0.033 s of fixation duration. The driving simulator environment display screen and the independent screens 1 to 3 are demarcated as Area of Interest (AOI) 1 to 4, respectively. The duration of the participants’ fixation on the visual cognitive materials was obtained through counting the number of fixation points in each AOI in a specific time period. The identification accuracy of different visual materials was obtained by counting the picture identification results reported by the participants. The explanatory notes of the relevant parameters (symbols) obtained from the experiments are shown in Table 2. 3.2. Results and Discussions 3.2.1. Visual Attention Characteristics in VIT Figure 7 shows the fixation time and identification accuracy of the participants in the VIT for multi-information pictures. As shown in Figure 7a, as the number of the basic information units in the picture increases, the fixation time increases synchronously. The statistical results showed that the value range of VI3 was 0.875~1.343 s (M = 1.072, SD = 0.117), the value range of VI5 was 1.362~1.919 s (M = 1.624, SD = 0.143) and the value range of VI7 was 1.819~2.500 s (M = 2.153, SD = 0.161). As shown in Figure 7b, with the increase in the number of basic information units in the picture, participants’ fixation time to a single basic information unit decreased. The statistical results showed that the value range of SVI3 was 0.292~0.448 s (M = 0.357, SD = 0.039), the value range of SVI5 was 0.272~0.384 s (M = 0.325, SD = 0.029) and the value range was 0.260~0.357 s (M = 0.308, SD = 0.023). Figure 7c showed the average identification accuracy for different pictures (AVI3, AVI5 and AVI7) in the VIT. The statistical results showed that the value range of the identification accuracy rates was 0.333~1. Among them, AVI3, AVI5 and AVI7 were 0.930 (SD = 0.159), 0.905 (SD = 0.199) and 0.886 (SD = 0.221), respectively. One-way ANOVA was performed on SVI (Table 3). The results showed significant differences among SVI3, SVI5 and SVI7 (F = 45.172). Further, the results of multiple comparison analyses of SVI showed that there were significant differences between SVI3, SVI5 and SVI7 pairwise (Table 4). The above results demonstrated that the participants’ fixation time for a single basic information unit in a multi-information picture decreases with the increase in the number of basic information units. These results should be related to the fact that the participants performed virtual driving while completing the visual identification task. The virtual driving task simulated real driving activities and required participants not to focus their visual attention on a single cognitive object for a long time. As the complexity of the visual identification task increases, the participants would consciously improve the visual identification processing speed of a single cognitive object, thereby reducing the average attention time of the corresponding basic information unit. One-way ANOVA was performed on the identification accuracy of ternary, quintuple and seven-element pictures (AVI). The results (Table 5) showed no significant difference in the identification accuracy of different pictures (F = 0.886), but the identification accuracy showed a downward trend with the increase in picture information units. 3.2.2. Visual Attention Characteristics in VWMT The driver’s memory in the driving process is a typical working memory [41,42]. The VWMT aimed at stimulating the driver’s working memory process, and the VIMT corresponded to the information input stage, and VIRT corresponded to the information extraction and identification stage in the working memory process. Figure 8 shows the fixation time and identification accuracy for seven-element pictures in the VIMT. The visual materials in the VIMT were all seven-element pictures. Figure 8a shows the participants’ fixation time (SVIM) for a single basic information unit in the VIMT. The statistical results showed that the range of SVIM10 was 0.264~0.403 s (M = 0.325, SD = 0.028), the range of SVIM20 was 0.264~0.403 s (M = 0.325, SD = 0.027) and the range of SVIM30 was 0.267~0.409 s (M = 0.328, SD = 0.028). The results of one-way ANOVA on SVIM (Table 6) showed that there was no significant difference among SVIM10, SVIM20 and SVIM30. Figure 8b shows the average identification accuracy of the visual materials (AVIM) in the VIMT. The statistical results showed that the value of AVIM ranged from 0.333 to 1. Among them, when the display interval of the previous and subsequent picture was 10 s, the average identification accuracy (AVIM10) was 0.886 (SD = 0.206). When the display interval of the two pictures was 20 s, the average identification accuracy (AVIM20) was 0.866 (SD = 0.240). When the display interval of the two pictures was 30 s, the average identification accuracy (AVIM30) was 0.856 (SD = 0.248). The results of one-way ANOVA (Table 7) showed that there was no significant difference between AVIM10, AVIM20 and AVIM30 (F = 0.288). In the VIMT, participants were required to identify and memorize visual materials. This kind of memory was conscious memory, which required the participants to make a certain volitional effort [43]. In the VIT, the fixation time (SVI7) of the participants to the basic information unit in the seven-element picture was 0.308 s (SD = 0.023). In the VIMT, SVIM10, SVIM20 and SVIM30 were 0.325 s (SD = 0.028), 0.325 s (SD = 0.027) and 0.328 s (SD = 0.028), respectively. The one-way ANOVA results (Table 8) showed a significant difference between SVI7 and SVIM (F = 8.159). The results of multiple comparison analysis (Table 9) showed that SVI7 was different from SVIM10, SVIM20 and SVIM30, significantly. This indicated that the memory process increased the participants’ fixation time on visual materials significantly. In the VIT, AVI7 was 0.886 (SD = 0.221). In the VIMT, AVIM10, AVIM20 and AVIM30 were 0.885 (SD = 0.201), 0.866 (SD = 0.240) and 0.856 (SD = 0.248), respectively. The one-way ANOVA results showed (Table 10) no significant difference between AVI7 and AVIM (F = 0.284); that is, the memory process did not affect the identification accuracy. Figure 9a shows the average fixation time of the participants to the basic information unit in the VIRT (SVIR). The statistical results showed that the value range of SVIR10 was 0.267~0.406 s (M = 0.328, SD = 0.027), the value range of SVIR20 was 0.261~0.409 s (M = 0.326, SD = 0.029) and the value range of SVIR30 was 0.263~0.401 s (M = 0.327, SD = 0.029). One-way ANOVA results (Table 11) showed that there was no significant difference among SVIR10, SVIR20 and SVIR30 (F = 0.016). As shown in Figure 9b, with the increase in the display time interval between the previous and subsequent picture, AVIR kept decreasing. The value range of AVIR10, AVIR20 and AVIR30 was 0.333~1 (M = 0.886, SD = 0.214), 0~1 (SD = 0.284) and 0~1 (M = 0.766, SD = 0.291), severally. One-way ANOVA (Table 12) and multiple comparison analysis (Table 13) showed significant differences between pairs of AVIR10, AVIR20 and AVIR30. According to [44], the correct rate of recall drops to about 10% after 18 s. The recall accuracy for visual information in this experiment declined more slowly over time. After the visual information disappeared for 10 s, 20 s and 30 s, the recall accuracy was 88.6%, 76.6% and 63.2%, respectively. In the VIT, SVI7 was 0.308 s (SD = 0.023). In the VIRT, SVIR10, SVIR20 and SVIR30 were 0.328 s (SD = 0.027), 0.326 s (SD = 0.029) and 0.327 s (SD = 0.029), respectively. One-way ANOVA (Table 14) and multiple comparison analysis (Table 15) showed significant differences between SVI7 and SVIR. It was proved that the recall matching process increases participants’ fixation time to visual materials. In the VIT, the identification accuracy for seven-element pictures (AVI7) was 0.886 (SD = 0.221). In the VIRT, AVIR10, AVIR20 and AVIR30 were 0.886 (SD = 0.214), 0.766 (SD = 0.318) and 0.632 (SD = 0.291), respectively. The multiple comparison analysis results of AVI7, AVIR10, AVIR20 and AVIR30 (Table 16) showed that AVI7 was significantly different from AVIR20 and AVIR30, but not from AVIR10. It indicated that the recall accuracy in VIRT began to decline after the visual information disappeared 10 s. In VIMT and VIRT, the properties of the visual materials were the same, but the cognitive tasks that the participants need to complete were different. In order to examine the effects of different cognitive tasks on the participants’ fixation time, paired-sample T-tests were performed on SVIM10 and SVIR10, SVIM20 and SVIR20, and SVIM30 and SVIR30, respectively. The results showed (Table 17) no significant difference between the above-paired variables. It suggested that there was no significant difference in the effects of the two visual cognitive task attributes on the participants’ fixation time. 3.2.3. Visual Attention Characteristics in MVIT In the MVIT, participants were required to identify the visual materials played simultaneously on screens 1 to 3. Figure 10a shows the average fixation time to unary pictures on each screen (MulVI). The statistical results showed the range of fixation time on the unary picture in screen 1 (MulVI1) was 0.307~0.598 s (M = 0.468, SD = 0.070), the fixation time on the unary picture in screen 2 (MulVI2) ranged from 0.284 to 0.657 s (M = 0.449, SD = 0.081) and the value range for the fixation time on the unary picture in screen 3 (MulVI3) was 0.182 to 0.625 s (M = 0.376, SD = 0.083). The above results showed that MulVI3 was smaller than MulVI1 and MulVI2, distinctly. One-way ANOVA and multiple comparison analysis results (Table 18 and Table 19) showed that MulVI3 was significantly different from both MulVI1 and MulVI2. This result indicated that when there are a large number of visual tasks to be processed, drivers would consciously improve the identification speed of the cognitive object at the end of the sequence, and then reduce the fixation time of the corresponding visual task. Therefore, when dealing with multiple visual cognitive objects at the same time, the drivers’ attention time on the cognitive object at the end of the sequence would decrease with the compression of disposable time until an Attentional Blink (AB) occurs [45]. Figure 10b shows the identification accuracy of visual materials in the MVIT(AMulVI). The AMulVI1, AMulVI2 and AMulVI3 were 0.976 (SD = 0.096), 0.976 (SD = 0.096) and 0.982 (SD = 0.090), respectively. The one-way ANOVA (Table 20) showed there was no significant difference among AMulVI1, AMulVI2 and AMulVI3 (F = 0.090). 4. Study 2—Influences of Emotions on Driver’s Visual Attention Characteristics 4.1. Materials and Methods 4.1.1. Participants Forty-three drivers (35 males and 32 females) with normal vision were selected through social recruitment to participate in the study. The age distribution of the participants was 20 to 40 years old (M = 27.53, SD = 4.90). All the participants were licensed drivers and the driving experience was distributed from 1 to 12 years (M = 4.67, SD = 2.57). 4.1.2. Collection of Visual Attention Characteristics Data in Different Emotional States The purpose of this experiment was to collect data on the visual attention characteristics of participants in a neutral emotional state and eight typical emotional states. The main steps of the experiment included emotion activation, collection of visual attention characteristic data and evaluation of emotion activation efficacy. Emotion activation The neutral emotional state in this study referred to a state of mind in which the mood is calm and without any emotional swing. For the neutral state activation, the participants were asked to listen to a piece of soothing music before the start of the relevant test, and then participated in each test according to their personal habits and behavioral styles. For the other emotions, the emotion activation methods referred to the literature [20,29]. The activation of each emotion included primary activation and deep activation. The methods used in the primary activation included picture activation, reward activation, personal recall activation and competitive game. On the basis of the primary activation, the deep activation was carried out, and the method used was the video activation. 2. Collection of visual attention characteristic data The general idea of visual attention characteristics data collection was similar to Section 3.1.2 in study 1, but the experiments were simplified considering the timeliness of emotion activation efficacy. The experiment included visual cognitive tasks included the Visual Identification Task (VIT), Visual Working Memory Task (VWMT) and Multiple Visual Identification Task (MVIT). Each participant was required to complete a visual attention characteristic data collection under a neutral state and eight typical emotional states, respectively. In each visual attention characteristic data collection, the VIT, VWMT and MVIT were performed, sequentially, and the interval between the three types of tasks was 15 s. During this interval, the participants did not interrupt the virtual driving. In a single VIT, a total of 3 seven-element pictures were displayed on screen 2. In a single VWMT, a total of 3 sets of seven-element pictures (6 pictures in total) were played on screen 2. The participants were asked to identify and report whether the proportions of vehicle models contained in the previous and subsequent pictures were the same in the shortest possible time. In a single MVIT, screens 1, 2 and 3 displayed 3 groups of 9 pictures containing the vehicle types, traffic signs, and traffic light schematic diagrams (Figure 2) simultaneously. Figure 11 shows the visual materials display timeline of the single visual attention characteristic data collection in Study 2. 3. Evaluation of emotion activation efficacy Emotion activation efficacy evaluation was performed after the visual attention characteristics data collection. The overall idea was referred to [20,29], and the measurement tool used was the PAD scale (Figure S1 in Supplementary Materials). Before conducting relevant experiments, the connotation of the PAD scale was explained to the participants to ensure that they could use the scale to accurately describe their minds. After each visual attention characteristic data collection, the participants filled in the PAD scale once. The emotional state filled in the PAD scale corresponded to a point in the PAD space. The Euclidean distance between this point and the emotion coordinate was used to represent the activation strength of the corresponding emotion [24]. For example, the anxiety state filled in by a participant was (3, 2, 6), which corresponds to the point (−0.5, −0.75, 0.25) in the PAD space. The distance between the point (−0.5, −0.75, 0.25) and the coordinates of anxiety in the PAD space (−0.24, 0.08, −0.16) represented the activation efficacy of anxiety. The smaller the distance was, the higher the activation efficiency of anxiety was. The PAD scale can represent a total of 729 emotional states [20,29], corresponding to 729 points in the PAD space. The distances between the 729 points and the coordinates of 8 typical emotions were sorted, and the corresponding emotional activation efficacy of each point was assigned according to the distance distribution (Table S1 in Supplementary Materials). The assignment range was 0~5. The larger the value was, the higher the activation efficiency of the corresponding emotion was. 4.1.3. Data Preprocessing The visual data obtained in the above experiments were imported into the data analysis software of the eye-tracking system, and the fixation time of the participants on each visual material was obtained. The identification accuracy of different visual materials was obtained by counting the picture identification results reported by the participants. Based on the above method, the visual attention characteristics data of 43 participants under a neutral state and 8 typical emotions were obtained. The average fixation time of participants to 3 seven-element pictures in a single VIT was denoted as VI’. The average identification accuracy of 3 seven-element pictures in a single VIT was recorded as AVI’. The average fixation time of participants to 3 subsequent pictures in a single VWMT was denoted as VIR’. The average identification accuracy of 3 subsequent pictures in a single VWMT was recorded as AVIR’. It should be pointed out that the fixation time and identification accuracy of the previous pictures in VWMT were no longer recorded and analyzed. The average fixation time of participants to 9 unary pictures in a single MVIT was recorded as MulVI’. The average identification accuracy of 9 unary pictures in a single MVIT was recorded as AMulVI’. 4.2. Results and Discussion Figure 12 shows the proportion of activation efficacy levels for eight typical emotions, and Figure 13 shows the average activation efficacy of the emotions. According to statistics, the average activation efficacy of anger, surprise, fear, anxiety, helplessness, contempt, relief and pleasure were 1.56 (SD = 1.16), 1.26 (SD = 1.07), 1.28 (SD = 1.03), 1.37 (SD = 1.22), 1.21 (SD = 0.97), 1.14 (SD = 0.89), 1.77 (SD = 1.11) and 1.58 (SD = 1.14), respectively. 4.2.1. Influences of Different Emotions on Visual Attention Characteristics in VIT Figure 14 shows the visual attention characteristics of participants in different emotions in the VIT. Figure 14a,b were the fixation time (VI’) and identification accuracy (AVI’) to visual materials, severally. According to statistics, the average fixation time of the participants in the states of neuter, anger, surprise, fear, anxiety, helplessness, contempt, relief and pleasure to the visual materials was 2.124 s (SD = 0.199), 2.095 s (0.193), 2.114 s (SD = 0.202), 2.297 s (SD = 0.277), 2.332 s (SD = 0.255), 2.150 s (SD = 0.219), 2.095 s (SD = 0.204) and 2.129 s (SD = 0.217), respectively, the average identification accuracy of visual materials was 0.845 (SD = 0.245), 0.860 (SD = 0.244), 0.837 (SD = 0.245), 0.729 (SD = 0.311), 0.806 (SD = 0.254), 0.798 (SD = 0.301), 0.853 (SD = 0.233), 0.814 (SD = 0.233) and 0.837 (SD = 0.245), severally. In order to further test the influences of different emotions on visual attention characteristics in the VIT, the VI’ and AVI’ of the participants in the neutral state and in eight emotional states were tested by paired samples T-test (Table 21 and Table 22). According to the T-test results, there was a significant difference between the VI’ of participants in the angry state and neutral state (t = 4.416), and there was no significant difference between the AVI’ of participants in the angry state and neutral state (t = −1.431). It suggested that without affecting the identification accuracy of visual materials, the fixation time of angry participants on visual materials was significantly reduced; that is, the emotional state of anger improved the cognitive efficiency on visual materials in the VIT. The VI’ (t = −6.363) and AVI’ (t = 4.743) of participants in the fearful state were significantly different from those of participants in the neutral state. It showed that the fearful participants’ fixation time on visual materials increased significantly, but the identification accuracy of visual materials decreased significantly; that is, the cognitive efficiency of the fearful participants was significantly declined in the VIT. There was a significant difference between the VI’ of participants in the anxious state and neutral state (t = −7.449), and there was no significant difference between the AVI’ of participants in the anxious state and neutral state (t= 1.703). This indicated that the participants in the anxiety state spend more fixation time on visual material, while did not improve the identification accuracy; that is, the emotional state of anxiety led to a decrease in the cognitive efficiency of the participants in the VIT for visual materials. The VI’ (t = −3.235) and AVI’ (t = 2.610) of the helpless participants were significantly different from those in the neutral state. It showed that the emotional state of helplessness led to an increase in the participants’ fixation time on visual materials, but reduced the identification accuracy of visual materials; that is, the visual cognitive efficiency of helpless participants in VIT was significantly lower than that of neutral subjects. There was a significant difference between the VI’ of participants in the contempt and neutral state (t = 3.695), and there was no significant difference between the AVI’ of participants in the contempt and neutral state (t = −0.330). This showed that, without affecting the identification accuracy of visual materials, subjects in the emotional state of contempt spend less fixation time on visual materials; that is, the emotion of contempt improved the cognitive efficiency of participants in the VIT. The VI’ and AVI’ of the participants in the emotional states of surprise, relief and pleasure were not significantly different from that of the participants in the neutral state, indicating that the three emotions had no significant influence on participants’ fixation time and identification accuracy of visual materials in the VIT. 4.2.2. Influences of Different Emotions on Visual Attention Characteristics in VWMT Figure 15 shows the visual attention characteristics of participants with different emotions in the VWMT. Figure 15a,b were the fixation time (VIR’) and identification accuracy (AVIR’) to visual materials, severally. According to statistics, the average fixation time of the participants in the states of neuter, anger, surprise, fear, anxiety, helplessness, contempt, relief and pleasure to subsequent pictures was 2.288 s (SD = 0.252), 2.252 s (SD= 0.244), 2.307 s (SD = 0.259), 2.494 s (SD = 0.351), 2.479 s (SD = 0.329), 2.306 s (SD = 0.253), 2.279 s (SD = 0.237), 2.289 s (SD = 0.253) and 2.324 s (SD = 0.260), and the average identification accuracy of subsequent pictures was 0.690 (SD = 0.285), 0.760 (SD = 0.255), 0.535 (SD = 0.318), 0.535 (SD = 0.318), 0.612 (SD = 0.316), 0.620 (SD = 0.305), 0.682 (SD = 0.308), 0.721 (SD = 0.251) and 0.674 (SD = 0.321). To further test the influences of emotions on visual attention characteristics in the VWMT, the VIR’ and AVIR’ of the participants in the neutral state and in eight emotional states were tested by paired samples T-test (Table 23 and Table 24). According to the T-test results, the VIR’ (t = 5.072) and AVIR’ (t = −3.334) of the angry participants were significantly different from those in the neutral state. This showed that the emotion of anger not only reduced the participants’ fixation time on visual materials, but also improved the participants’ identification accuracy of visual materials; that is, the motion of anger significantly improved the participants’ cognitive efficiency of each subsequent picture in the VWMT. The VIR’ (t = −2.864) and AVIR’ (t = 6.043) of the participants in the surprised state were significantly different from those in the neutral state. It showed that the emotion of surprise increased the participants’ fixation time on the visual materials, while it reduced the identification accuracy to the visual materials; that is, the emotion of surprise significantly declined the cognitive efficiency of the participants in the VWMT. In the emotional state of fear, the participants’ VIR’ (t = −7.321) and AVIR’ (t = 5.547) were significantly different from those in the neutral state. It showed that the emotion of fear not only increases the participants’ fixation time on visual materials, but also reduced the identification accuracy of visual materials; that is, the emotion of fear significantly reduced the participants’ cognitive efficiency in the VWMT. The VIR’ (t = −6.361) and AVIR’ (t = 2.673) of anxious participants were significantly different from those of participants in the neutral state. This showed that the emotion of anxiety not only increases the participants’ fixation time on visual materials, but also reduces the identification accuracy of visual materials; that is, the emotion of anxiety significantly reduced the participants’ cognitive efficiency in the VWMT. In the emotional state of helplessness, the participants’ VIR’ (t = −5.126) and AVIR’ (t = 3.334) were significantly different from those in the neutral state. This indicated that the emotion of helplessness not only increases the participants’ fixation time on visual materials, but also reduces the identification accuracy of visual materials; that is, the emotion of helplessness led to participants’ cognitive efficiency to decline significantly in the VWMT. The VIR’ (t = −5.126) of the participants in the pleasant state was significantly different from that of the participants in the neutral state, and there was no significant difference between the AVIR’ (t = 0.496) of participants in the pleasant state and in the neutral state. It showed that the emotion of pleasure increased the participants’ fixation time on visual materials, but did not affect the participants’ identification accuracy of visual materials; that is, the emotion of pleasure declined the participants’ cognitive efficiency of each subsequent picture in the VWMT. Neither the VIR’ nor the AVIR’ of the participants in the emotional states of contempt and relief was significantly different from those of participants in the neutral state. It showed that the emotions of contempt and relief did not affect the participants’ cognitive efficiency of visual materials in the VWMT. 4.2.3. Influences of Different Emotions on Visual Attention Characteristics in MVIT Figure 16 shows the visual attention characteristics of participants with different emotions in the MVIT. Figure 16a,b were the fixation time (MulVI’) and identification accuracy (AMulVI’) to the visual materials, severally. According to statistics, the average fixation time of participants in the states of neuter, anger, surprise, fear, anxiety, helplessness, contempt, relief and pleasure to unary pictures was 0.420 s (SD = 0.058), 0.416 s (SD = 0.058), 0.433 s (SD = 0.064), 0.414 s (SD = 0.058), 0.458 s (SD = 0.069), 0.424 s (SD = 0.060), 0.416 s (SD = 0.056), 0.422 s (SD = 0.059) and 0.420 s (SD = 0.057), severally, and the average identification accuracy of unary pictures was 0.868 (SD = 0.098), 0.863 (SD = 0.137), 0.860 (SD = 0.137), 0.876 (SD = 0.122), 0.848 (SD = 0.116), 0.842 (SD = 0.114), 0.853 (SD = 0.120), 0.853 (SD = 0.083) and 0.840 (SD = 0.115), respectively. To further test the influences of emotions on visual attention characteristics in the MVIT, the MulVI’ and AMulVI’ of the participants in the neutral state and in eight emotional states were tested by paired samples T-test (Table 25 and Table 26). According to the T-test results, there was a significant difference between the MulVI’ of participants in the angry state and neutral state (t = 3.398), and there was no significant difference between the AMulVI’ of participants in the angry state and neutral state (t = 0.496). It suggested that without affecting the identification accuracy of visual materials, the fixation time of angry participants on visual materials was significantly reduced; that is, the emotion of anger improved the cognitive efficiency on visual materials in the MVIT. There was a significant difference between the MulVI’ of participants in the surprised state and neutral state (t = −5.267), and there was no significant difference between the AMulVI’ of participants in the surprised state and neutral state (t = 0.573). It suggested that the emotion of surprise increased the participants’ fixation time on the visual materials, while had no significant effect on the identification accuracy; that is, the emotion of surprise significantly declined the cognitive efficiency of the participants in the MVIT. Neither the MulVI’ (t = 0.916) nor the AMulVI’ (t = −0.724) of the participants in the emotional state of fear were significantly different from those of participants in the neutral state. It showed that the emotions of fear did not affect the participants’ cognitive efficiency of visual materials in the MVIT. There was a significant difference between the MulVI’ of participants in the anxious state and neutral state (t = −6.902), and there was no significant difference between the AMulVI’ of participants in the anxious state and neutral state (t = 1.838). This suggested that the emotion of anxiety increased the participants’ fixation time to visual materials, but did not significantly change the identification accuracy of visual materials; that is, the emotion of anxiety declined participants’ cognitive efficiency of visual materials in the MVIT. In the emotional state of helplessness, the participants’ MulVI’ (t = −2.327) and AMulVI’ (t = 3.177) were significantly different from those in the neutral state. This indicated that the emotion of helplessness not only increases the participants’ fixation time on visual materials, but also reduced the identification accuracy of visual materials; that is, the emotion of helplessness led to participants’ cognitive efficiency declined significantly in the MVIT. There was a significant difference between the MulVI’ of participants in the contempt and neutral state (t = 2.333), and there was no significant difference between the AMulVI’ of participants in the contempt and neutral state (t = 1.634). It suggested that the fixation time of participants in contempt on visual materials was significantly reduced, and the emotion of contempt had no significant effect on the identification accuracy of visual materials; that is, the emotion of contempt improved the cognitive efficiency on visual materials in the MVIT. Neither the MulVI’ (t = −0.722) nor the AMulVI’ (t = 1.431) of the participants in the emotional state of relief were significantly different from those of participants in the neutral state, which suggested that the emotions of relief did not affect the participants’ cognitive efficiency of visual materials in the MVIT. There was no significant difference between the MulVI’ of participants in the pleasant state and neutral state (t = 0.114), and there was a significant difference between the AMulVI’ of participants in the pleasant state and neutral state (t = 2.886). This suggested that the emotion of pleasure had no significant effect on participants’ fixation time on the visual materials, while reduced the identification accuracy significantly; that is, the emotion of pleasure significantly declined the cognitive efficiency of the participants in the MVIT. 4.2.4. Comprehensive Analysis and Discussion Based on the above data analysis results, it can be seen that anger reduced the driver’s fixation time on the visual materials in VIT, VWMT and MVIT, and improved the driver’s visual identification accuracy in the VWMT. This result demonstrated that the emotion of anger improves the driver’s visual perception ability, but it does not mean that angry drivers will have better cognitive and behavioral performance. Many previous studies have shown that anger can lead drivers to take risky driving behaviors, which, in turn, adversely affect driving safety [46,47]. However, some scholars have pointed out that anger leads to more frequent aggressive driving behaviors, but does not increase driving errors [48,49]. The findings in this paper may partly explain the phenomenon that anger increases aggressive driving behavior without increasing driving errors. That is, in the state of anger, the body would counteract the driving safety risk due to the expression of anger (risky driving behavior) by improving its own perception ability. At present, there are few research conclusions about the effect of surprise on the driver’s visual perception ability in related fields. The data analysis results in this paper showed that the emotion of surprise increased the driver’s visual material fixation time in the VWMT and MVIT, and reduced the driver’s identification accuracy of visual materials in the MVIT. This indicated that the emotion of surprise reduced the driver’s visual perception ability as a whole. According to the data analysis results, the emotion of fear increased the driver’s fixation time on visual material in the VIT and VWMT, and reduced the identification accuracy of visual materials. The above conclusions support the viewpoint that the emotion of fear increases driving errors proposed by Taylor, J. et al. [50]. The results in this paper that anxiety increases the fixation time on the visual materials in the VIT, VWMT and MVIT, and reduces the identification accuracy of visual materials in the VWMT, are consistent with previous research conclusions that anxiety increases driving errors and augments driving safety risks [51,52]. The effect of helplessness on the visual attention characteristics in the VIT and VWMT in this paper is similar to the emotion of fear, which largely supports the view that helplessness is a negative emotion similar to fear proposed by Fikretoglu, D. et al. [53]. However, it should be pointed out that the influence of helplessness on the visual attention characteristics in the MVIT is different from that of fear. Helplessness increased the driver’s fixation time on relevant visual materials and reduced the visual identification accuracy. According to the data analysis results, the emotion of contempt reduced the driver’s fixation time on visual material in the VIT and MVIT, and improved the cognitive efficiency of the visual materials. However, there are few research conclusions about the influence of contempt on the driver’s visual perception ability in related fields. There was no significant difference in the visual material identification results of the subjects in the emotional states of relief and neuter, indicating that the emotion of relief had no significant effect on the driver’s visual perception. Pleasure is generally considered to be a positive emotion. However, the emotion of pleasure led to the increase in driver’s fixation time on visual materials in the VWMT, and the decrease in driver’s visual identification accuracy in the MVIT. These results are consistent with Dolinski, D. et al.’s view that positive emotions are not necessarily positively related to safe driving [54]. 5. Conclusions The visual attention characteristic is the key factor to determine whether the driver can extract important information from the traffic environment and form an effective cognition of the environmental situation. Emotions are a special form of human reflection to objective reality. Drivers with different emotions have significant differences in their feelings, preferences and needs for external stimuli, which, in turn, lead to varying degrees of changes in their perceptual characteristics and cognitive abilities. In this paper, the drivers’ visual attention characteristics for different cognitive tasks and the effects of eight typical driving emotions on drivers’ visual attention characteristics were deeply studied through designing and implementing the Visual Identification Task (VIT), Visual Working Memory Task (VWMT) and Multiple Visual Identification Task (MVIT) in the virtual driving process. The findings of the present study have several important practical implications for improving the safety and intelligence of the road traffic system. On the one hand, information perception is the basis for the generation of conscious behavior in drivers. Ignoring the influence of the limited visual attention resources on driving behavior would lead to a driving behavior prediction model with the drawback of “taking the driver as absolute rationality”, i.e., holding that drivers have absolutely complete environmental knowledge and powerful computing ability, which is obviously a severe deviation of the reality. The results of this paper reveal the driver’s visual information processing mechanism to a certain extent, which contributes to an accurate understanding of the mechanism of generating driving behavior and can represent an interesting starting point in the improvement of driving behavior prediction. On the other hand, real-time communication of dynamic traffic environment information obtained by in-vehicle information sensing devices (radar, camera, etc.) to the driver is a common mode of assisted driving systems, which can expand the breadth and depth of information perception of the driver during driving. However, it should be noted that the in-vehicle system can seriously endanger driving safety by taking up too many perceptual and cognitive resources from the driver in an inappropriate manner. The results of this paper not only reveal the boundaries of drivers’ perception of different visual cognitive objects, but also examine the stability of drivers’ storage and extraction of visual information. These results can provide a scientific basis for the in-vehicle system to choose the proper time to provide the suitable amount of information to the driver and improve the interaction efficiency and intelligence of the in-vehicle human–computer interaction system. Finally, while many studies on driving emotions have provided evidence that emotions affect driving safety, the mechanisms by which emotions affect driving behavior are not yet clear. Some scholars have concluded that emotions can directly influence driving behavior. Some researchers argued that driving emotions can indirectly affect driving behaviors through driver’s perception, attitude and so on. The results of this study demonstrate the role of eight typical driving emotions on drivers’ visual attentional characteristics. Considering the important role of visual attention in influencing driving behavior, the results of this paper suggest that the influence of emotions on driving behavior is multifaceted and includes at least indirectly influencing driving behaviors through affecting perceptual characteristics. The research results can be applied for the vehicle security warning system and then the accuracy of driving behavior prediction would be improved. The current study has a few limitations. Firstly, the study is not representative of the entire population, since the participants were recruited using a convenience sample through social media. Nevertheless, recruiting participants in this way is considered a common practice worldwide. Secondly, considering feasibility and safety, data measurements are mostly conducted in specific environments or simulated driving environments, and data test results are susceptible to the influence of the environment. There are some differences between the simulated driving environment and the actual road conditions, and the proficiency of the driver using the driving simulator may also have some influence on the experimental results. Follow-up studies should pay attention to overcome the influence of environmental factors on the study results. Thirdly, static pictures are used to simulate the cognitive objects in driving activities in this paper, which limits the ability to reflect on the complex and changing real-world traffic environment. Designing more relevant experimental methods is the focus of subsequent research. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ijerph19095059/s1, Figure S1: PAD emotion scale; Table S1: Emotional activation efficacy level corresponding to each point in PAD space. Click here for additional data file. Author Contributions Conceptualization, X.W. and Y.L.; methodology, X.W. and Y.L.; software, Y.L. and L.C.; validation, X.W. and Y.L.; formal analysis, Y.L., S.L. and L.C.; investigation, Y.L., S.L. and J.H.; resources, X.W. and Y.L.; data curation, Y.L., L.C. and J.H.; writing—original draft preparation, Y.L.; writing—review and editing, X.W., Y.L. and H.S.; visualization, Y.L., H.S. and F.Z.; supervision, X.W.; project administration, X.W.; funding acquisition, X.W. All authors have read and agreed to the published version of the manuscript. Funding This research was funded by the Natural Science Foundation of Shandong Province, grant number ZR2020MF082; the Collaborative Innovation Center for Intelligent Green Manufacturing Technology and Equipment of Shandong Province, grant number IGSD-2020-012; the Qingdao Top Talent Program of Entrepreneurship and Innovation, grant number 19-3-2-11-zhc; and the National Key Research and Development Program, grant number 2018YFB1601500. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee at College of Electromechanical Engineering, Qingdao University of Science & Technology. Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Design idea of visual attention characteristic data collection. Figure 2 Visual cognitive materials: (a) Schematic diagram of vehicle types; (b) Schematic diagram of traffic signs and traffic lights. Figure 3 Display timeline of visual cognitive materials in VIT. Figure 4 Display timeline of visual cognitive materials in VWMT. Figure 5 Display timeline of visual cognitive materials in MVIT. Figure 6 Fixation point distribution of a participant. Note: In the “Distribution of visual fixation points”, each dot represents a visual fixation point. The numeric code represents the temporal order of the fixation points, and the two fixation points adjacent to each other in the temporal order are connected by a straight line. Figure 7 Drivers’ visual attention characteristics in VIT: (a) Fixation time for visual tasks; (b) Fixation time for basic information unit; (c) Identification accuracy for visual tasks. Figure 8 Drivers’ visual attention characteristics in VIMT: (a) Fixation time for basic information unit; (b) Identification accuracy for visual tasks. Figure 9 Drivers’ visual attention characteristics in VIRT: (a) Fixation time for basic information unit; (b) Identification accuracy for visual tasks. Figure 10 Drivers’ visual attention characteristics in MVIT: (a) Fixation time for basic information unit; (b) Identification accuracy for visual tasks. Figure 11 Visual materials display timeline of each visual attention characteristic data collection. Figure 12 Proportion of activation efficacy levels of eight typical emotions. Figure 13 Average activation efficacy of eight typical emotions. Figure 14 Visual attention characteristics of participants with different emotional states in VIT: (a) fixation time on visual materials; (b) identification accuracy of visual materials. Figure 15 Visual attention characteristics of participants with different emotional states in VWMT: (a) fixation time on visual materials; (b) identification accuracy of visual materials. Figure 16 Visual attention characteristics of participants with different emotional states in MVIT: (a) fixation time on visual materials; (b) identification accuracy of visual materials. ijerph-19-05059-t001_Table 1 Table 1 Elaboration on the different visual cognitive tasks. Tasks Description Visual Materials Display Screen VIT To identify and report the proportion of vehicle types in visual materials while completing the virtual driving. Ternary pictures/quintuple pictures/seven-element pictures Screen 2 VIMT To identify, report and memorize the proportions of vehicle types in the previous. Seven-element pictures Screen 2 VIRT To identify and report whether the proportion of vehicle types in the subsequent picture is the same as the previous picture. Seven-element pictures Screen 2 MVIT To identify and report the unary pictures on the 3 screens while completing driving in the order from left to right. Unary pictures Screen 1/Screen 2/Screen 3 ijerph-19-05059-t002_Table 2 Table 2 Explanatory notes of the relevant parameters (symbols) obtained from the experiments. Collective Name Symbols Description VI VI3, VI5, VI7 Average fixation time of each participant to each ternary picture, quintuple picture and seven-element picture in the VIT AVI AVI3, AVI5, AVI7 Average identification accuracy of each participant for ternary, quintuple and seven-element pictures in the VIT SVI SVI3, SVI5, SVI7 Fixation time to identify a basic information unit in a ternary picture, a quintuple picture and a seven-element picture in the VIT VIM VIM10, VIM20, VIM30 Average fixation time of each participants to each previous picture in the MTU with the display interval of 10, 20 and 30 s AVIM AVIM10, AVIM20, AVIM30 Average identification accuracy of each participant for the previous pictures in the MTU with the interval of 10, 20 and 30 s SVIM SVIM10, SVIM20, SVIM30 Fixation time to identify a basic information unit in the previous picture in the MTU with the interval of 10, 20 and 30 s VIR VIR10, VIR20, VIR30 Average fixation time of each participant to each subsequent picture in the MTU with the interval of 10, 20 and 30 s AVIR AVIR10, AVIR20, AVIR30 Average identification accuracy of each participant for the subsequent pictures in the MTU with the interval of 10, 20 and 30 s SVIR SVIR10, SVIR20, SVIR30 Fixation time to identify a basic information unit in the subsequent picture in the MTU with the interval of 10, 20 and 30 s MulVI MulVI1, MulVI2, MulVI3 Average fixation time (seconds) of each participant to pictures on screens 1, 2 and 3 in the MVIT AMulVI AMulVI1, AMulVI2, AMulVI3 Average identification accuracy of each participant for five unary pictures on screens 1, 2 and 3 in the MVIT ijerph-19-05059-t003_Table 3 Table 3 One-way ANOVA results for SVI. SVI SoS 1 df 2 MS 3 F Sig. 4 BG 5 0.086 2 0.043 45.172 0.000 WG 6 0.188 198 0.001 Tot. 7 0.274 200 1 SoS is the sum of squares. 2 df is the degree of freedom. 3 MS is the mean square. 4 Sig. is the significance. 5 BG means between groups. 6 WG means within groups. 7 Tot. means total. ijerph-19-05059-t004_Table 4 Table 4 Multiple comparison results for SVI. (I) (J) MD 1 (I–J) SE 2 Sig. 3 95% CI 4 LB 5 UB 6 SVI7 SVIR10 −0.020 * 7 0.004 0.000 −0.031 −0.008 SVIR20 −0.019 * 0.005 0.000 −0.031 −0.007 SVIR30 −0.019 * 0.005 0.000 −0.031 −0.007 SVIR10 SVI7 0.020 * 0.004 0.000 0.008 0.031 SVIR20 0.001 0.005 1.000 −0.012 0.014 SVIR30 0.000 0.005 1.000 −0.012 0.013 SVIR20 SVI7 0.019 * 0.005 0.000 0.007 0.031 SVIR10 −0.001 0.005 1.000 −0.014 0.012 SVIR30 0.000 0.005 1.000 −0.014 0.013 SVIR30 SVI7 0.019 * 0.005 0.000 0.007 0.031 SVIR10 0.000 0.005 1.000 −0.013 0.012 SVIR20 0.000 0.005 1.000 −0.013 0.014 1 MD is the mean difference. 2 SE is the standard error. 3 Sig. is the significance. 4 CI is the confidence interval. 5 LB is lower bound. 6 UB is upper bound. 7,* means the significance level 0.05. ijerph-19-05059-t005_Table 5 Table 5 One-way ANOVA results for AVI. AVI SoS 1 df 2 MS 3 F Sig. 4 BG 5 0.067 2 0.034 0.886 0.414 WG 6 7.532 198 0.038 Tot. 7 7.600 200 1 SoS is the sum of squares. 2 df is the degree of freedom. 3 MS is the mean square. 4 Sig. is the significance. 5 BG means between groups. 6 WG means within groups. 7 Tot. means total. ijerph-19-05059-t006_Table 6 Table 6 One-way ANOVA results for SVIM. SVIM SoS 1 df 2 MS 3 F Sig. 4 BG 5 0 2 0 0.166 0.847 WG 6 0.152 198 0.001 Tot. 7 0.152 200 1 SoS is the sum of squares. 2 df is the degree of freedom. 3 MS is the mean square. 4 Sig. is the significance. 5 BG means between groups. 6 WG means within groups. 7 Tot. means total. ijerph-19-05059-t007_Table 7 Table 7 One-way ANOVA results for AVIM. SVIM SoS 1 df 2 MS 3 F Sig. 4 BG 5 0.031 2 0.015 0.288 0.750 WG 6 10.630 198 0.054 Tot. 7 10.661 200 1 SoS is the sum of squares. 2 df is the degree of freedom. 3 MS is the mean square. 4 Sig. is the significance. 5 BG means between groups. 6 WG means within groups. 7 Tot. means total. ijerph-19-05059-t008_Table 8 Table 8 One-way ANOVA results for SVI7 and SVIM. SVIM SoS 1 df 2 MS 3 F Sig. 4 BG 5 0.017 3 0.006 8.159 0.000 WG 6 0.186 264 0.001 Tot. 7 0.204 267 1 SoS is the sum of squares. 2 df is the degree of freedom. 3 MS is the mean square. 4 Sig. is the significance. 5 BG means between groups. 6 WG means within groups. 7 Tot. means total. ijerph-19-05059-t009_Table 9 Table 9 Multiple comparison results for SVI7 and SVIM. (I) (J) MD 1 (I–J) SE 2 Sig. 3 95% CI 4 LB 5 UB 6 SVI7 SVIM10 −0.018 *7 0.004 0.001 −0.030 −0.006 SVIM20 −0.017 * 0.004 0.001 −0.029 −0.006 SVIM30 −0.020 * 0.004 0.000 −0.032 −0.008 SVIM10 SVI7 0.018 * 0.004 0.001 0.006 0.030 SVIM20 0.000 0.005 1.000 −0.012 0.013 SVIM30 −0.002 0.005 0.998 −0.015 0.011 SVIM20 SVI7 0.017 * 0.004 0.001 0.006 0.029 SVIM10 0.000 0.005 1.000 −0.013 0.012 SVIM30 −0.003 0.005 0.996 −0.015 0.010 SVIM30 SVI7 0.020 * 0.004 0.000 0.008 0.032 SVIM10 0.002 0.005 0.998 −0.011 0.015 SVIM20 0.003 0.005 0.996 −0.010 0.015 1 MD is the mean difference. 2 SE is the standard error. 3 Sig. is the significance. 4 CI is the confidence interval. 5 LB is lower bound. 6 UB is upper bound. 7,* means the significance level 0.05. ijerph-19-05059-t010_Table 10 Table 10 One-way ANOVA results for AVI7 and AVIM. SVIM SoS 1 df 2 MS 3 F Sig. 4 BG 5 0.045 3 0.015 0.284 0.837 WG 6 13.864 264 0.053 Tot. 7 13.909 267 1 SoS is the sum of squares. 2 df is the degree of freedom. 3 MS is the mean square. 4 Sig. is the significance. 5 BG means between groups. 6 WG means within groups. 7 Tot. means total. ijerph-19-05059-t011_Table 11 Table 11 One-way ANOVA results for SVIR. SVIM SoS 1 df 2 MS 3 F Sig. 4 BG 5 0.000 2 0.000 0.016 0.984 WG 6 0.161 198 0.001 Tot. 7 0.161 200 1 SoS is the sum of squares. 2 df is the degree of freedom. 3 MS is the mean square. 4 Sig. is the significance. 5 BG means between groups. 6 WG means within groups. 7 Tot. means total. ijerph-19-05059-t012_Table 12 Table 12 One-way ANOVA results for AVIR. SVIM SoS 1 df 2 MS 3 F Sig. 4 BG 5 1.991 2 0.996 14.107 0.000 WG 6 13.973 198 0.071 Tot. 7 15.965 200 1 SoS is the sum of squares. 2 df is the degree of freedom. 3 MS is the mean square. 4 Sig. is the significance. 5 BG means between groups. 6 WG means within groups. 7 Tot. means total. ijerph-19-05059-t013_Table 13 Table 13 Multiple comparison results for AVIR. (I) (J) MD 1 (I−J) SE 2 Sig. 3 95% CI 4 LB 5 UB 6 AVIR10 AVIR20 0.119 * 7 0.043 0.021 0.014 0.225 AVIR30 0.244 * 0.044 0.000 0.137 0.351 AVIR20 AVIR10 −0.119 * 0.043 0.021 −0.225 −0.014 AVIR30 0.124 * 0.050 0.041 0.004 0.245 AVIR30 AVIR10 −0.244 * 0.044 0.000 −0.351 −0.137 AVIR20 −0.124 * 0.050 0.041 −0.245 −0.004 1 MD is the mean difference. 2 SE is the standard error. 3 Sig. is the significance. 4 CI is the confidence interval. 5 LB is lower bound. 6 UB is upper bound. 7,* means the significance level 0.05. ijerph-19-05059-t014_Table 14 Table 14 One-way ANOVA results for SVI7 and SVIR. SVIM SoS 1 df 2 MS 3 F Sig. 4 BG 5 0.019 3 0.006 8.414 0.000 WG 6 0.196 264 0.001 Tot. 7 0.215 267 1 SoS is the sum of squares. 2 df is the degree of freedom. 3 MS is the mean square. 4 Sig. is the significance. 5 BG means between groups. 6 WG means within groups. 7 Tot. means total. ijerph-19-05059-t015_Table 15 Table 15 Multiple comparison results for SVI7 and SVIR. (I) (J) MD 1 (I−J) SE 2 Sig. 3 95% CI 4 LB 5 UB 6 SVI7 SVIR10 −0.020 * 7 0.004 0.000 −0.031 −0.008 SVIR20 −0.019 * 0.005 0.000 −0.031 −0.007 SVIR30 −0.019 * 0.005 0.000 −0.031 −0.007 SVIR10 SVI7 0.020 * 0.004 0.000 0.008 0.031 SVIR20 0.001 0.005 1.000 −0.012 0.014 SVIR30 0.000 0.005 1.000 −0.012 0.013 SVIR20 SVI7 0.019 * 0.005 0.000 0.007 0.031 SVIR10 −0.001 0.005 1.000 −0.014 0.012 SVIR30 0.000 0.005 1.000 −0.014 0.013 SVIR30 SVI7 0.019 * 0.005 0.000 0.007 0.031 SVIR10 0.000 0.005 1.000 −0.013 0.012 SVIR20 0.000 0.005 1.000 −0.013 0.014 1 MD is the mean difference. 2 SE is the standard error. 3 Sig. is the significance. 4 CI is the confidence interval. 5 LB is lower bound. 6 UB is upper bound. 7,* means the significance level 0.05. ijerph-19-05059-t016_Table 16 Table 16 Multiple comparison results for AVI7 and AVIR. (I) (J) MD 1 (I–J) SE 2 Sig. 3 95% CI 4 LB 5 UB 6 AVI7 AVIR10 0.000 0.038 1.000 −0.100 0.100 AVIR20 0.119 * 7 0.044 0.045 0.002 0.237 AVIR30 0.254 * 0.045 0.000 0.134 0.373 AVIR10 AVI7 0.000 0.038 1.000 −0.100 0.100 AVIR20 0.119 * 0.043 0.041 0.003 0.236 AVIR30 0.254 * 0.044 0.000 0.136 0.372 AVIR20 AVI7 −0.119 * 0.044 0.045 −0.237 −0.002 AVIR10 −0.119 * 0.043 0.041 −0.236 −0.003 AVIR30 0.134 * 0.050 0.046 0.002 0.267 AVIR30 AVI7 −0.254 * 0.045 0.000 −0.373 −0.134 AVIR10 −0.254 * 0.044 0.000 −0.372 −0.136 AVIR20 −0.134 * 0.050 0.046 −0.267 −0.002 1 MD is the mean difference. 2 SE is the standard error. 3 Sig. is the significance. 4 CI is the confidence interval. 5 LB is lower bound. 6 UB is upper bound. 7,* means the significance level 0.05. ijerph-19-05059-t017_Table 17 Table 17 Paired-sample T-test results for SVIM and SVIR. M 1 SD 2 SE 3 95% CI 4 t df 7 Sig. 8 (2-Tailed) LB 5 UB 6 SVIM10-SVIR10 −0.002 0.009 0.001 −0.004 0.000 −1.721 66 0.090 SVIM20-SVIR20 −0.001 0.010 0.001 −0.004 0.001 −1.167 66 0.247 SVIM30-SVIR30 0.001 0.010 0.001 −0.002 0.003 0.590 66 0.557 1 M is mean value. 2 SD is standard deviation. 3 SE is standard error. 4 CI is confidence interval. 5 LB is lower bound. 6 UB is upper bound. 7 df is degree of freedom. 8 Sig. is significance. ijerph-19-05059-t018_Table 18 Table 18 One-way ANOVA results for MulVI. SVIM SoS 1 df 2 MS 3 F Sig. 4 BG 5 0.314 2 0.157 26.168 0.000 WG 6 1.189 198 0.006 Tot. 7 1.503 200 1 SoS is the sum of squares. 2 df is the degree of freedom. 3 MS is the mean square. 4 Sig. is the significance. 5 BG means between groups. 6 WG means within groups. 7 Tot. means total. ijerph-19-05059-t019_Table 19 Table 19 Multiple comparison results for MulVI. (I) (J) MD 1 (I–J) SE 2 Sig. 3 95% CI 4 LB 5 UB 6 MulVI1 MulVI2 0.018 0.013 0.172 −0.008 0.045 MulVI3 0.092 * 7 0.013 0.000 0.065 0.118 MulVI2 MulVI1 −0.018 0.013 0.172 −0.045 0.008 MulVI3 0.073 * 0.013 0.000 0.047 0.100 MulVI3 MulVI1 −0.092 * 0.013 0.000 −0.118 −0.065 MulVI2 −0.073 * 0.013 0.000 −0.100 −0.047 1 MD is the mean difference. 2 SE is the standard error. 3 Sig. is the significance. 4 CI is the confidence interval. 5 LB is lower bound. 6 UB is upper bound. 7* means the significance level 0.05. ijerph-19-05059-t020_Table 20 Table 20 One-way ANOVA results for AMulVI. SVIM SoS 1 df 2 MS 3 F Sig. 4 BG 5 0.002 2 0.001 0.090 0.914 WG 6 1.742 198 0.009 Tot. 7 1.744 200 1 SoS is the sum of squares. 2 df is the degree of freedom. 3 MS is the mean square. 4 Sig. is the significance. 5 BG means between groups. 6 WG means within groups. 7 Tot. means total. ijerph-19-05059-t021_Table 21 Table 21 Paired-samples T-test results for VI’ of neuter and VI’ of eight emotions. VI’ M 1 SD 2 SE 3 95% CI 4 t Sig. 7 (2-Tailed) LB 5 UB 6 neuter-anger 0.029 0.043 0.007 0.016 0.042 4.416 0.000 neuter-surprise 0.011 0.068 0.010 −0.010 0.032 1.045 0.302 neuter-fear −0.173 0.178 0.027 −0.227 −0.118 −6.363 0.000 neuter-anxiety −0.207 0.183 0.028 −0.264 −0.151 −7.449 0.000 neuter-helplessness −0.025 0.051 0.008 −0.041 −0.010 −3.235 0.002 neuter-contempt 0.030 0.053 0.008 0.014 0.046 3.695 0.001 neuter-relief −0.015 0.065 0.010 −0.035 0.005 −1.481 0.146 neuter-pleasure −0.004 0.064 0.010 −0.024 0.016 −0.429 0.670 1 M is mean value. 2 SD is standard deviation. 3 SE is standard error. 4 CI is confidence interval. 5 LB is lower bound. 6 UB is upper bound. 7 Sig. is significance. ijerph-19-05059-t022_Table 22 Table 22 Paired-samples T-test results for AVI’ of neuter and AVI’ of eight emotions. AVI’ M 1 SD 2 SE 3 95% CI 4 t Sig. 7 (2-Tailed) LB 5 UB 6 neuter-anger −0.016 0.071 0.011 −0.037 0.006 −1.431 0.160 neuter-surprise 0.008 0.170 0.026 −0.045 0.060 0.298 0.767 neuter-fear 0.116 0.161 0.025 0.067 0.166 4.743 0.000 neuter-anxiety 0.039 0.149 0.023 −0.007 0.085 1.703 0.096 neuter-helplessness 0.047 0.117 0.018 0.011 0.082 2.610 0.013 neuter-contempt −0.008 0.154 0.024 −0.055 0.040 −0.330 0.743 neuter-relief 0.031 0.142 0.022 −0.013 0.075 1.431 0.160 neuter-pleasure 0.008 0.185 0.028 −0.049 0.065 0.274 0.785 1 M is mean value. 2 SD is standard deviation. 3 SE is standard error. 4 CI is confidence interval. 5 LB is lower bound. 6 UB is upper bound. 7 Sig. is significance. ijerph-19-05059-t023_Table 23 Table 23 Paired-samples T-test results for VIR’ of neuter and VIR’ of eight emotions. VIR’ M 1 SD 2 SE 3 95% CI 4 t Sig. 7 (2-Tailed) LB 5 UB 6 neuter-anger 0.035 0.045 0.007 0.021 0.049 5.072 0.000 neuter-surprise −0.019 0.045 0.007 −0.033 −0.006 −2.864 0.007 neuter-fear −0.207 0.185 0.028 −0.264 −0.150 −7.321 0.000 neuter-anxiety −0.191 0.197 0.030 −0.252 −0.131 −6.361 0.000 neuter-helplessness −0.019 0.056 0.009 −0.036 −0.001 −2.185 0.035 neuter-contempt 0.010 0.060 0.009 −0.009 0.028 1.070 0.291 neuter-relief −0.001 0.061 0.009 −0.020 0.017 −0.146 0.885 neuter-pleasure −0.037 0.047 0.007 −0.051 −0.022 −5.126 0.000 1 M is mean value. 2 SD is standard deviation. 3 SE is standard error. 4 CI is confidence interval. 5 LB is lower bound. 6 UB is upper bound. 7 Sig. is significance. ijerph-19-05059-t024_Table 24 Table 24 Paired-samples T-test results for AVIR’ of neuter and AVIR’ of eight emotions. AVIR’ M 1 SD 2 SE 3 95% CI 4 t Sig. 7 (2-Tailed) LB 5 UB 6 neuter-anger −0.070 0.137 0.021 −0.112 −0.028 −3.334 0.002 neuter-surprise 0.155 0.168 0.026 0.103 0.207 6.043 0.000 neuter-fear 0.155 0.183 0.028 0.099 0.211 5.547 0.000 neuter-anxiety 0.078 0.190 0.029 0.019 0.136 2.673 0.011 neuter-helplessness 0.070 0.137 0.021 0.028 0.112 3.334 0.002 neuter-contempt 0.008 0.212 0.032 −0.057 0.073 0.240 0.812 neuter-relief −0.031 0.216 0.033 −0.097 0.035 −0.942 0.352 neuter-pleasure 0.016 0.205 0.031 −0.048 0.079 0.496 0.623 1 M is mean value. 2 SD is standard deviation. 3 SE is standard error. 4 CI is confidence interval. 5 LB is lower bound. 6 UB is upper bound. 7 Sig. is significance. ijerph-19-05059-t025_Table 25 Table 25 Paired-samples T-test results for MulVI’ of neuter and MulVI’ of eight emotions. MulVI’ M 1 SD 2 SE 3 95% CI 4 t Sig. 7 (2-Tailed) LB 5 UB 6 neuter-anger 0.004 0.008 0.001 0.002 0.007 3.398 0.001 neuter-surprise −0.012 0.016 0.002 −0.017 −0.008 −5.267 0.000 neuter-fear 0.001 0.009 0.001 −0.002 0.004 0.916 0.365 neuter-anxiety −0.038 0.036 0.006 −0.049 −0.027 −6.902 0.000 neuter-helplessness −0.004 0.011 0.002 −0.007 −0.001 −2.327 0.025 neuter-contempt 0.004 0.010 0.002 0.000 0.007 2.333 0.025 neuter-relief −0.001 0.013 0.002 −0.005 0.003 −0.722 0.474 neuter-pleasure 0.000 0.013 0.002 −0.004 0.004 0.114 0.910 1 M is mean value. 2 SD is standard deviation. 3 SE is standard error. 4 CI is confidence interval. 5 LB is lower bound. 6 UB is upper bound. 7 Sig. is significance. ijerph-19-05059-t026_Table 26 Table 26 Paired-samples T-test results for AMulVI’ of neuter and AMulVI’ of eight emotions. AMulVI’ M 1 SD 2 SE 3 95% CI 4 t Sig. 7 (2-Tailed) LB 5 UB 6 neuter-anger 0.005 0.068 0.010 −0.016 0.026 0.496 0.623 neuter-surprise 0.008 0.089 0.014 −0.020 0.035 0.573 0.570 neuter-fear −0.008 0.070 0.011 −0.029 0.014 −0.724 0.473 neuter-anxiety 0.021 0.074 0.011 −0.002 0.043 1.838 0.073 neuter-helplessness 0.026 0.053 0.008 0.009 0.042 3.177 0.003 neuter-contempt 0.016 0.062 0.009 −0.004 0.035 1.634 0.110 neuter-relief 0.016 0.071 0.011 −0.006 0.037 1.431 0.160 neuter-pleasure 0.028 0.065 0.010 0.009 0.048 2.886 0.006 1 M is mean value. 2 SD is standard deviation. 3 SE is standard error. 4 CI is confidence interval. 5 LB is lower bound. 6 UB is upper bound. 7 Sig. is significance. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Wang X. Liu Y. Wang J. Wang J. Zhang J. Study on influencing factors selection of driver’s propensity Transp. Res. Part D Transp. Environ. 2019 66 35 48 10.1016/j.trd.2018.06.025 2. Yang S. Wang W. Jiang Y. Wu J. Zhang S. Deng W. What contributes to driving behavior prediction at unsignalized intersections? Transp. Res. C Emer. 2019 108 100 114 10.1016/j.trc.2019.09.010 3. Moran C. Bennett J. Prabhakharan P. The relationship between cognitive function and hazard perception in younger drivers Transp. Res. Part F Traffic Psychol. Behav. 2020 74 104 119 10.1016/j.trf.2020.08.008 4. Parkes A. Ward N. Bossi L. The potential of vision enhancement systems to improve driver safety/ (Le potentiel des systèmes d’amélioration de la vision visant à accroître la sécurité du conducteur) Le Trav. Hum. 1995 58 151 5. Underwood G. Chapman P. Brocklehurst N. Underwood J. Crundall D. Visual attention while driving: Sequences of eye fixations made by experienced and novice drivers Ergonomics 2003 46 629 646 10.1080/0014013031000090116 12745692 6. Fox E. de Fockert J. Inhibitory effects of repeating color and shape: Inhibition of return or repetition blindness? J. Exp. Psychol. Hum. Percept. Perform. 2001 27 798 10.1037/0096-1523.27.4.798 11518144 7. Forsman P. Vila B. Short R. Mott C. Van D.H. Efficient driver drowsiness detection at moderate levels of drowsiness Accid. Anal. Prev. 2013 50 341 350 10.1016/j.aap.2012.05.005 22647383 8. Kountouriotis G. Wilkie R. Gardner P. Merat N. Looking and thinking when driving: The impact of gaze and cognitive load on steering Transp. Res. Part F Traffic Psychol. Behav. 2015 34 108 121 10.1016/j.trf.2015.07.012 9. De Winter J. Happee R. Martens M. Stanton N. Effects of adaptive cruise control and highly automated driving on workload and situation awareness: A review of the empirical evidence Transp. Res. Part F Traffic Psychol. Behav. 2014 27 196 217 10.1016/j.trf.2014.06.016 10. Distefano N. Leonardi S. Pulvirenti G. Romano R. Boer E. Wooldridge E. Mining of the association rules between driver electrodermal activity and speed variation in different road intersections IATSS Res. 2021 in press 10.1016/j.iatssr.2021.12.002 11. Distefano N. Leonardi S. Pulvirenti G. Romano R. Merat N. Boer E. Woolridge E. Physiological and driving behaviour changes associated to different road intersections Eur. Transp. 2020 77 1 12 10.48295/ET.2020.77.4 12. Mehler B. Reimer B. Coughlin J. Dusek J. Impact of incremental increases in cognitive workload on physiological arousal and performance in young adult drivers Transp. Res. Rec. 2009 2138 6 12 10.3141/2138-02 13. Liu Y.Q. Wang X.Y. Guo Y.Q. The moderating effects of emotions on the relationship between self-reported individual traits and actual risky driving hehaviours Psychol. Res. Behav. Manag. 2021 14 423 10.2147/PRBM.S301156 33859507 14. Dureux A. Blini E. Grandi L.C. Close facial emotions enhance physiological responses and facilitate perceptual discrimination Cortex 2021 138 40 58 10.1016/j.cortex.2021.01.014 33677327 15. Wang X.Y. Guo Y.Q. Bai C.L. Yuan Q. Liu S.L. Han J.Y. Driver’s intention identification with involvement of emotional factors in two-lane roads IEEE Trans. Intell. Transp. Syst. 2020 22 6866 6874 10.1109/TITS.2020.2995837 16. Liu Y.Q. Wang X.Y. Differences in driving intention transitions caused by driver’s emotion evolutions Int. J. Environ. Res. Public Health 2020 17 6962 10.3390/ijerph17196962 17. Xu Y. Li S. Gao S. Tan D. Guo D. Wang Y. Recognition method of construction conflict based on driver’s eye movement Accid. Anal. Prev. 2018 113 193 201 10.1016/j.aap.2018.01.041 29427918 18. Rahman H. Ahmed M.U. Barua S. Funk P. Begum S. Vision-based driver’s cognitive load classification considering eye movement using machine learning and deep learning Sensors 2021 21 8019 10.3390/s21238019 34884021 19. Onkhar V. Bazilinskyy P. Dodou D. de Winter J. The effect of driver’s eye contact on pedestrians’ perceived safety Transp. Res. Part F Traffic Psychol. Behav. 2022 84 194 210 10.1016/j.trf.2021.10.017 20. Li N. Busso C. Predicting perceived visual and cognitive distractions of drivers with multimodal features IEEE Trans. Intell. Transp. Syst. 2014 16 51 65 10.1109/TITS.2014.2324414 21. Karthaus M. Wascher E. Falkenstein M. Getzmann S. The ability of young, middle-aged and older drivers to inhibit visual and auditory distraction in a driving simulator task Transp. Res. Part F Traffic Psychol. Behav. 2020 68 272 284 10.1016/j.trf.2019.11.007 22. Li Q. Hou L. Wang Z. Wang W. Zeng C. Yuan Q. Cheng B. Drivers’ visual-distracted take-over performance model and its application on adaptive adjustment of time budget Accid. Anal. Prev. 2021 154 106099 10.1016/j.aap.2021.106099 33770718 23. Grahn H. Kujala T. Impacts of touch screen size, user interface design, and subtask boundaries on in-car task’s visual demand and driver distraction Int. J. Hum. Comput. Stud. 2020 142 102467 10.1016/j.ijhcs.2020.102467 24. Reimer B. Mehler B. Wang Y. Coughlin J. A field study on the impact of variations in short-term memory demands on drivers’ visual attention and driving performance across three age groups Hum. Factors J. Hum. Factors Ergon. Soc. 2012 54 454 468 10.1177/0018720812437274 25. Muñoz M. Reimer B. Lee J. Mehler B. Fridman L. Distinguishing patterns in drivers’ visual attention allocation using Hidden Markov Models Transp. Res. Part F Traffic Psychol. Behav. 2016 43 90 103 10.1016/j.trf.2016.09.015 26. Louw T. Merat N. Are you in the loop? Using gaze dispersion to understand driver visual attention during vehicle automation Transp. Res. C Emer. 2017 76 35 50 10.1016/j.trc.2017.01.001 27. Lemonnier S. Désiré L. Brémond R. Baccino T. Drivers’ visual attention: A field study at intersections Transp. Res. Part F Traffic Psychol. Behav. 2020 69 206 221 10.1016/j.trf.2020.01.012 28. Young K. Osborne R. Grzebieta R. Williamson A. Haworth N. Senserrick T. Stephan K. Charlton J. Using naturalistic driving data to examine how drivers share attention when engaging in secondary tasks Saf. Sci 2020 129 104841 10.1016/j.ssci.2020.104841 29. Liu Q. Gao M. Zhang J. Mu X. Analysis on drivers’ vision-psychology under the influence of color difference after pavement pothole repair Teh. Vjesn. 2021 28 1894 1903 30. Lee K. Yoon H. Song J. Park K. Convolutional neural network-based classification of driver’s emotion during aggressive and smooth driving using multi-modal camera sensors Sensors 2018 18 957 10.3390/s18040957 31. Herrero-Fernández D. Fonseca-Baeza S. Angry thoughts in Spanish drivers and their relationship with crash-related events. The mediation effect of aggressive and risky driving Accid. Anal. Prev. 2017 106 99 108 10.1016/j.aap.2017.05.015 28600988 32. Barnard M. Chapman P. Are anxiety and fear separable emotions in driving? A laboratory study of behavioural and physiological responses to different driving environments Accid. Anal. Prev. 2016 86 99 107 10.1016/j.aap.2015.10.021 26536073 33. Roseborough J. Wiesenthal D. The influence of roadway police justice on driver emotion Transp. Res. Part F Traffic Psychol. Behav. 2018 56 236 244 10.1016/j.trf.2018.03.030 34. Paschero M. Del V.G. Benucci L. Rizzi A. Santello M. Fabbri G. Mascioli F.F. A Real Time Classifier for Emotion and Stress Recognition in a Vehicle Driver Proceedings of the 2012 IEEE International Symposium on Industrial Electronics Hangzhou, China 12 July 2012 35. Wang X.Y. Liu Y.Q. Wang F. Wang J.Q. Liu L.P. Wang J.H. Feature extraction and dynamic identification of drivers’ emotions Transp. Res. Part F Traffic Psychol. Behav. 2019 62 175 191 10.1016/j.trf.2019.01.002 36. Fairclough S. Dobbins C. Personal informatics and negative emotions during commuter driving: Effects of data visualization on cardiovascular reactivity & mood Int. J. Hum. Comput. Stud. 2020 144 102449 37. Chan M. Singhal A. Emotion matters: Implications for distracted driving Saf. Sci. 2015 72 302 309 10.1016/j.ssci.2014.10.002 38. Wang X. Guo Y. Ban J. Xu Q. Bai C. Liu S. Driver em.motion recognition of multiple-ECG feature fusion based on BP network and D-S evidence IET Intell. Transp. Syst. 2020 14 815 824 10.1049/iet-its.2019.0499 39. Kadoya Y. Watanapongvanich S. Khan M. How is emotion associated with driving speed? A study on taxi drivers in Japan Transp. Res. Part F Traffic Psychol. Behav. 2021 79 205 216 10.1016/j.trf.2021.04.020 40. Liu Y.Q. Wang X.Y. The analysis of driver’s behavioral tendency under different emotional stated based on a Bayesian Network IEEE Trans. Affect. Comput. 2020 in press 10.1109/TAFFC.2020.3027720 41. Ge Y. Sheng B.Y. Qu W.N. Xiong Y.X. Sun X.H. Zhang K. Differences in visual-spatial working memory and driving behavior between morning-type and evening-type drivers Accid. Anal. Prev. 2020 136 105402 10.1016/j.aap.2019.105402 31862644 42. Keogh R. Wicken M. Pearson J. Visual working memory in aphantasia: Retained accuracy and capacity with a different strategy Cortex 2021 143 237 253 10.1016/j.cortex.2021.07.012 34482017 43. Ramey M. Yonelinas A. Henderson J. Conscious and unconscious memory differentially impact attention: Eye movements, visual search, and recognition processes Cognition 2019 185 71 82 10.1016/j.cognition.2019.01.007 30665071 44. Peterson L. Peterson M. Miller A. Short-term retention and meaningfulness Can. J. Psychol. Can. Psychol. 1961 15 143 10.1037/h0083216 13734839 45. Martens S. Wyble B. The attentional blink: Past, present, and future of a blind spot in perceptual awarenes Neurosci. Biobehav. Rev. 2010 34 947 957 10.1016/j.neubiorev.2009.12.005 20025902 46. Roidl E. Frehse B. Oehl M. Höger R. The emotional spectrum in traffic situations: Results of two online-studies Transp. Res. Part F Traffic Psychol. Behav. 2013 18 168 188 10.1016/j.trf.2012.12.009 47. Ābele L. Haustein S. Møller M. Zettler I. Links between observed and self-reported driving anger, observed and self-reported aggressive driving, and personality traits Accid. Anal. Prev. 2020 140 105516 10.1016/j.aap.2020.105516 32244089 48. Zhang T.R. Chan A.H.S. Zhang W. Dimensions of driving anger and their relationships with aberrant driving Accid. Anal. Prev. 2015 81 124 133 10.1016/j.aap.2015.05.005 25984643 49. Precht L. Keinath A. Krems J. Effects of driving anger on driver behavior—Results from naturalistic driving data Transp. Res. Part F Traffic Psychol. Behav. 2017 45 75 92 10.1016/j.trf.2016.10.019 50. Taylor J. Deane F. Podd J. Driving fear and driving skills: Comparison between fearful and control samples using standardised on-road assessment Behav. Res. Ther. 2007 45 805 818 10.1016/j.brat.2006.07.007 16962560 51. Shahar A. Self-reported driving behaviors as a function of trait anxiety Accid. Anal. Prev. 2009 41 241 245 10.1016/j.aap.2008.11.004 19245881 52. Dula C. Adams C.L. Miesner M. Leonard R. Examining relationships between anxiety and dangerous driving Accid. Anal. Prev. 2010 42 2050 2056 10.1016/j.aap.2010.06.016 20728661 53. Fikretoglu D. Brunet A. Best S. Metzler T. Delucchi K. Weiss D. Marmar C. Peritraumatic fear, helplessness and horror and peritraumatic dissociation: Do physical and cognitive symptoms of panic mediate the relationship between the two? Behav. Res. Ther. 2007 45 39 47 10.1016/j.brat.2006.01.008 16516845 54. Dolinski D. Odachowska E. Beware when danger on the road has passed. The state of relief impairs a driver’s ability to avoid accidents Accid. Anal. Prev. 2018 115 73 78 10.1016/j.aap.2018.03.007 29549773

PMC009xxxxxx/PMC9099628.txt

==== Front Molecules Molecules molecules Molecules 1420-3049 MDPI 10.3390/molecules27092981 molecules-27-02981 Review Nano-Drug Delivery Systems Based on Different Targeting Mechanisms in the Targeted Therapy of Colorectal Cancer https://orcid.org/0000-0002-6334-0200 Wang Ke Shen Ruoyu Meng Tingting Hu Fuqiang Yuan Hong * Nazemi Ali Academic Editor College of Pharmaceutical Science, Zhejiang University, 866 Yuhangtang Road, Hangzhou 310058, China; wk1194229090@163.com (K.W.); 3150101386@zju.edu.cn (R.S.); mengtt@zju.edu.cn (T.M.); hufq@zju.edu.cn (F.H.) * Correspondence: yuanhong70@zju.edu.cn 06 5 2022 5 2022 27 9 298131 3 2022 03 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Colorectal cancer (CRC) is a usual digestive tract malignancy and the third main cause of cancer death around the world, with a high occurrence rate and mortality rate. Conventional therapies for CRC have certain side effects and restrictions. However, the exciting thing is that with the rapid development of nanotechnology, nanoparticles have gradually become more valuable drug delivery systems than traditional therapies because of their capacity to control drug release and target CRC. This also promotes the application of nano-drug targeted delivery systems in the therapy of CRC. Moreover, to make nanoparticles have a better colon targeting effect, many approaches have been used, including nanoparticles targeting CRC and in response to environmental signals. In this review, we focus on various targeting mechanisms of CRC-targeted nanoparticles and their latest research progress in the last three years, hoping to give researchers some inspiration on the design of CRC-targeted nanoparticles. targeted therapy colorectal cancer nanoparticles drug delivery systems National Natural Science Foundation of China81773644 81773648 This research was funded by National Natural Science Foundation of China grant number 81773644 and 81773648. ==== Body pmc1. Introduction Colorectal cancer (CRC), the third most common malignancy worldwide, forms in the colon or rectum. It is also the third most common cancer in men and the second in women separately. It is estimated that, by 2035, the number of CRC patients will reach 2.4 million and CRC will cause 1.3 million deaths worldwide [1,2]. According to the histopathological features, CRC can be divided into four distinct stages, which are also the basis of European Society for Medical Oncology (ESMO) guidelines to take appropriate treatments. In stage 0 of CRC, polyps are formed in the lining of colonic mucosa. Then, in stage I, polyps deteriorate into tumors and begin to migrate into the mucosa. The surgical resection of local tumors can achieve a good therapeutic effect without additional treatment in the above periods. There are many methods of operation, such as excision of polyps on the intestinal wall, excision of intestinal segments of tumors, or standardized excision, and finally reconnecting intestinal segments for ileostomy or colostomy. In stage II, cancer can spread beyond the colon, but it will not metastasize through lymph nodes at this stage. In stage III, the cancer cells spread to the colon wall and surrounding lymph nodes, but not to nearby organs. During these periods, the patients must be treated with both radiotherapy and chemotherapy. Fluoropyrimidine or combined with other chemotherapy drug is usually used to treat CRC. In addition, microsatellite instability status can be used to determine the survival time of patients with localized tumors, and patients with high microsatellite instability will have longer survival time. However, fluorouracil-based chemotherapy is not suitable for them and may even have adverse effects [3]. In stage IV, cancer accelerates to spread to other organs of the body [4]. At this time, it is not only necessary to remove the tumor through surgery, but also to kill the tumor cells through systemic chemotherapy or the combination of chemotherapy and biological targeted therapy. The biological targeted therapy of CRC uses bevacizumab, which can combine with the vascular endothelial growth factor produced by CRC cells to block tumor angiogenesis, cetuximab as well as panitumumab, which can bind to epidermal growth factor receptor (EGFR) overexpressed in CRC to prevent the growth of tumor cells. As can be seen from the above, the survival rate of CRC patients can be increased if they are diagnosed as early as possible and treated accordingly, especially before CRC metastasis. Because the first step in the development of CRC is the formation of neoplastic polyps in the colonic mucosa, which are not very dangerous. However, because the specificity of early symptoms of CRC is not obvious and it is difficult to make a differential diagnosis, CRC patients are generally diagnosed in the advanced stage or when metastasis occurs [5]. Disappointingly, traditional therapies in this stage have certain side effects and limitations. For chemotherapy, due to poor physical and chemical properties, low bioavailability, and poor tissue selectivity of chemotherapeutic drugs, patients may suffer from many serious side effects [6,7,8]. Radiotherapy can cause serious DNA damage in patients sensitive to radiation damage, leading to the further development of tumors [9]. The side effect caused by receptor inhibitors is skin irritation [10]. Immunotherapy drugs cost a lot and have many side effects [11,12]. So, it is necessary to optimize the drug delivery to a specific target and minimize the side effects [13]. It is encouraging that, with the development of nanotechnology, there are many nano-drug delivery systems targeting CRC with high-efficiency anticancer, such as lipid nanoparticles, polymer nanoparticles, and liposomes, which can realize the delivery of chemotherapeutic drugs, genes, and vaccines. Various nano-drug delivery systems targeting CRC are presented in Figure 1. They can significantly improve the stability of drug in vivo, reduce the systemic toxicity of drug, and overcome drug resistance. Figure 2 shows some representative targeting mechanisms. This paper reviews the latest targeting mechanisms and prospects of nano-drug delivery systems for the treatment of CRC, with emphasis on the nanoparticles that can target the colon to further clarify the direction of future research and provide ideas for the design of nanoparticles for the better therapy of CRC. 2. Nano-Drug Delivery Systems Targeting CRC 2.1. Passive Targeting Nanoparticles It is generally believed that the anatomy and pathophysiology of solid tumors differ from that of normal tissue. For example, solid tumors have a high blood vessel density to meet the nutrition and oxygen required by tumor cells growth. Furthermore, the tumor lacks functional lymphatic vessels and the gap between tumor endothelial cells is relatively large, which can extravasate or retain macromolecular drugs. The phenomenon that makes nanoparticles accumulate in tumor cells is called the enhanced permeability and retention (EPR) effect [14,15,16]. The discovery of this phenomenon has promoted the emergence of passive targeting nanoparticles, which can target tumors through the EPR effect. In research, oxaliplatin was encapsulated into N, O-carboxymethyl chitosan nanoparticles by ionic crosslinking, and resveratrol was encapsulated into them by emulsification crosslinking. The particle size of the former was about 190.0 nm and that of the latter was about 164.2 nm. The nanoparticles enhanced the solubility, stability, and EPR effect of oxaliplatin and resveratrol, showing stronger anti-CRC activity in subcutaneous tumor-bearing mice compared to the free drug [17]. However, according to research findings, nanoparticles smaller than 10 nm are filtered out by the kidneys, and particles larger than 100 nm are captured by the liver, so the ideal nanoparticles should be between 10 nm and 100 nm in size, which can circulate for a long time to increase the possibility of uptake by tumors [18]. Moreover, Anitha et al. [19] mentioned that the degree of vascularization in CRC is low. It can be seen that there are controversies about the EPR effect in CRC. In recent years, scientists have gradually realized that the tumor-targeting mediated by EPR is highly heterogeneous. Hence, it is necessary to enhance the targeting ability of nanoparticles based on EPR by combining them with other targeting mechanisms [20,21]. 2.2. Active Targeting Nanoparticles Ligand-modified nanoparticles can actively accumulate in the tumor through a ligand–receptor binding mechanism, thus delivering the drug to the target. These nanoparticles are called active targeting nanoparticles [22]. Cancer cells overexpress some types of receptors and secrete biomolecules that can promote cell proliferation of cancer and surrounding tissues through paracrine or autocrine pathways [23]. In studies, there are two main methods to combine ligands with nanoparticles: one is to modify them chemically during the synthesis of nanoparticles, and the other is to bind ligands with polymers before the synthesis of nanoparticles [24,25]. Ligand-modified nanoparticles can accumulate in the tumor through passive targeting, and then enter tumor cells through active targeting, resulting in a more selective and enhanced therapeutic effect. In the recent three years, the active targeting design idea of nano-drug targeted delivery systems for CRC mainly adopt the receptor–ligand binding strategy, which involved many highly expressed receptors in CRC, such as folate receptor [26], EGFR [27], CD44, epithelial cell adhesion molecule (EpCAM) [28], CD133, αvβ3 integrin receptor, carcinoembryonic antigen [29], nucleolin [30], mannose receptor [31], hyaluronic acid receptor [32], N-acetyl-d-glucosamine [33], transferrin receptor [34], checkpoint kinase 2 [35], CXCR4+ [36], lipoprotein receptor-related protein-1 [37], MUC1 [38], NRP-1 [39], P-selectin [40], sigma-2 receptors [41], SSTRs [42], and glucocorticoid receptor [43]. The characteristics of these nanoparticles are shown in Table 1. Among them, the nanoparticles targeting EpCAM, folate receptor, epidermal growth factor, and CD44 were studied more. 2.2.1. Nanoparticles Targeting EpCAM Epithelial cell adhesion molecule (EpCAM) is a 314 amino acid and 39 kDa transmembrane glycoprotein that almost only exists on epithelial cells playing a role in adhering cells in normal cells [44,45]. As early as 1979, it has been reported as the main surface antigen of CRC [46], Later, scholars have found that overexpression of EpCAM also existed in human adenocarcinoma cells, and it could also carry out intercellular signal transduction [47]. In many kinds of research, to make nanoparticles target CRC after systemic administration, a DNA EpCAM aptamer (SYL3C) having a strong affinity with EpCAM protein and cancer cells expressing EpCAM was modified on nanoparticles. To reduce drug toxicity and improve therapeutic effect, Ge et al. [28] synthesized biological conjugates loaded with celastrol, which could be captured by CRC overexpressed with EpCAM. EpCAM aptamer, PEG, and dendrimers constituted the conjugates. The results showed that SW620 would undergo extensive apoptosis when exposed to biological conjugates. Moreover, in the xenograft mice and zebrafish models, the biological conjugate showed low toxicity. The dendrimer used in the study is composed of biocompatible components and has an excellent multivalent effect, which can improve the safety of chemotherapeutic drugs. In another study, the author prepared a cationic liposome to deliver miR-139-5p. The main lipids contained in it are HSPC, DOTAP, Chol, and DSPE-PEG2000-COOH. Moreover, to make the nanoparticles target CRC, the surface of it was modified with EpCAM aptamer. It is found that nanoparticles had inhibitory effects on HCT cells, consistent with the experimental results of animal pharmacodynamics experiments. They also restrained the tumor growth of CRC mice injected subcutaneously with HCT8 [48]. In addition to the specific targets of CRC, some targets of other tumors are also expressed in CRC, which can be used as the targets of nano-drug delivery systems in the targeted therapy of CRC. 2.2.2. Nanoparticles Targeting Folate Receptor The folate receptor is a cell membrane-anchored protein, which is slightly expressed or absent in normal cells, but overexpressed in various malignant tumor cells, especially in CRC [22]. Therefore, it can be used as a target. Natural polymers chitosan and chondroitin sulfate have the advantages of biocompatibility and biodegradability. Soe et al. [26] prepared self-assembled nanoparticles using these two materials to incorporate the hydrophobic drug, bortezomib, and modified the nanoparticles to actively target folate receptors. Folate was attached to DSPE-PEG and then coupled with nanoparticles. It was found that DSPE-PEG can not only promote blood circulation, but also improve core drug loading and stability. In the pharmacodynamic experiment of animal models, researchers found that nanoparticles had a strong inhibitory effect on the transplanted tumor and were almost non-toxic to other parts outside the tumor. 2.2.3. Nanoparticles Targeting EGFR The overexpression of epidermal growth factor receptor (EGFR) occurs in about one-third of epithelial malignancies. It can stimulate tumor growth, proliferation, promote angiogenesis, and promote cell invasion and metastasis. Therefore, the poor prognosis of CRC may have a great relationship with the overexpression of EGFR in CRC [49]. In order to solve this problem, targeted therapy using monoclonal antibodies to block biological pathways has been used to treat CRC. The commonly used monoclonal antibody is cetuximab, which is used for CRC under FDA approval [50]. It can selectively target the EGFR to block signal transduction in vivo [49]. This characteristic also makes cetuximab widely used to modify nano-drug delivery systems, actively target CRC, and increase the accumulation in CRC [51]. Sankha Bhattacharya [27] formulated PLGA-PEG-coated nanoparticles to deliver anti-epidermal growth factor receptor-5-fluorouracil (5-FU), which could improve the pharmacodynamics and distribution of the drug in vivo. The Anti-EGFR mAb was bound to the polymeric nanoparticles by using m-maleimidobenzoyl-N hydroxysuccinimide ester to increase target specificity. On the other hand, polymeric nanoparticles consisting of PLGA and PEG can block opsonic action through the reticular epithelial system. These nanoparticles have great clinical value because their preparation methods are solvent emulsification and evaporation, which are simple and rapid [52]. Another attractive carrier is gold nanoparticles, which are commercially available and easily functionalized. Hallal et al. [53] found that AuNPs loaded with cetuximab had better EGFR endocytosis and could inhibit downriver signaling pathways compared with cetuximab and gold nanoparticles alone, which could inhibit cell proliferation and accelerate cell apoptosis. This finding is of great help in solving the problem of drug resistance in the treatment of EGFR monoclonal antibodies. Moreover, many studies found that the therapeutic effect of monoclonal antibodies could be improved by combining McAb with chemotherapeutic drugs [54,55]. Chen et al. [56] constructed a hybrid nano delivery system to deliver 5-FU, including mesoporous silica nanoparticles (MSN), a supported lipid bilayer (SLB), and cetuximab. Furthermore, the SLB-MSN was functionalized with hydrophilic PEG, which made the nanoparticles more stable, circulate longer in the blood, and more likely to accumulate in the tumor through the EPR effect. Then, to enable the nanoparticles to actively target CRC, cetuximab was coupled to the PEG terminal in SLB-MSN. The SLB-MSN has the advantages of both MSN and liposome, showing significant biocompatibility. It can also encapsulate much drug, achieve controlled release, and its stability can be greatly improved. 2.2.4. Nanoparticles Targeting CD44 CD44 is a transmembraneous glycoprotein that exists as an adhesion molecule on the cell surface and takes part in intercellular and cytomatrix interactions, as well as cell adhesion and migration [57]. The overexpression of CD44 in CRC cells is closely related to tumor metastasis [58,59]. Therefore, it is also a biomarker for CRC as well as a target. Hyaluronic acid (HA) is the main ligand of CD44, consisting of units of D-glucuronic acid and N-acetyl-D-glucosamine alternately. Especially, it is very popular with researchers, because of its non-toxic, excellent biocompatibility, and biodegradability [60]. Wang et al. [58] synthesized HA-decorated polydopamine nanoparticles to deliver chlorin e6 to the CRC. By coupling the HA with nanoparticles, the nanoparticles were easily internalized via endocytosis under the guidance of CD44. Moreover, chlorin e6 [61], the second-generation PS approved by FDA, and polydopamine, the photothermal agent with excellent light-thermal conversion efficiency were conjugated to promote the synergistic effect of photodynamic therapy and photothermal therapy by avoiding the toxicity induced by mental-ion. The result showed that the tumor growth was inhibited strongly in HCT-116 tumor mice, which also confirmed that the antitumor efficiency could be enhanced by combining the two treatments. 2.2.5. Biomimetic Nano Delivery Systems In addition to the ligand–receptor binding strategy, the nanoparticles can also actively target the tumor by bionic technology, which mainly uses biofilms to coat nanoparticles. This strategy can not only prevent nanoparticles from being identified by the immune system, but also make use of membrane proteins, glycoproteins, and homologous adhesion to make nanoparticles specifically accumulate in the tumor. Recently, the bionic cell membranes used in the reported nanoparticles are mainly erythrocyte membrane [62], cancer cell membrane [63], leukocyte membrane [64], and so on. Furthermore, to make the nanoparticles endowed with more properties, it is better to camouflage the nanoparticles with hybrid cell membranes. Wang et al. [63] prepared Zn1.25Ga1.5Ge0.25O4: Cr3+, Yb3+, Er3+ persistent luminescence nanoparticles coated with mesoporous silica to deliver IR825 as well as irinotecan and encapsulated the nanoparticles with a cancer cell-macrophage hybrid membrane. The membrane made the nanoparticles actively target the tumor, then the nanoparticles continued to glow at the tumor, which could provide an accurate position for subsequent photothermal therapy. Moreover, after the CT26 tumor-bearing mice were treated with three months of nanoparticles and laser irradiation, the relative tumor volume of mice decreased significantly. In a study, hollow long persistence luminescence nanomaterials loaded with cisplatin were synthesized and coated with a hybrid cell membrane, which was mainly composed of a customized erythrocyte membrane that could prevent nanoparticles from being recognized by the immune system and biofilm expressing PD-1 that gave nanoparticles targeting ability. Then, this team evaluated their efficacy through in vitro anti-cancer experiments and concluded that these nanoplatforms, which played the role of immunotherapy and chemotherapy, were efficient in CT26 tumor-bearing mice [65]. These studies indicated that bionic technology may be an effective method for tumor-targeted therapy. 3. Nano-Drug Delivery Systems in Response to Environmental Signals 3.1. Nanoparticles Based on Stimulus Response Some nanoparticles can achieve precisely targeted therapy for CRC patients under the condition of stimulation, such as external magnetic field, temperature, reactive oxygen species (ROS), and near-infrared (NIR). In recent years, many pieces of research focused on magnetic hyperthermia. Magnetic nanoparticles were injected into the tumor and deposited under the action of an external magnetic field, and then give patients local radiofrequency hyperthermia [66]. In research, the complex of biological superparamagnetic chitosan-based nanocomposite was prepared to deliver SN-38 by combining with hydrophilic polymeric prodrug poly (l-glutamic acid)-SN-38, which showed obviously enhanced accumulation in CRC and more easily be internalized by cells with the assistance of a topical magnetic field. Moreover, magnetic nano complexes obtained a tumor inhibition ratio of up to 81% in the mice model of CRC xenografts [67]. More importantly, Dabaghi et al. [68] found that, compared with the use of magnetic hyperthermia individually or chemotherapy independently on the basis of magnetic nanoparticles loaded with 5-FU, the combination of the two was more effective in the treatment of CRC, and preferable therapeutic effects were also produced in mice model of CRC. It can be concluded that the magnetic nano complex system can boost tumor-targeted accumulation and improve the anti-colorectal cancer treatment effect. Apart from the injection, rectal administration can also be used to treat CRC based on the difference between body temperature and external temperature. Xing et al. [69] prepared topotecan-loaded solid lipid nanoparticles (SLNs) and then encapsulated them into a thermo-sensitive aqueous gel to realize controlled release and to be low toxic. Because nanoparticles could maintain free flow below 30 °C and convert into gel form under physiological conditions, they are very convenient for rectal administration. The construction of nanoparticles based on the response of the tumor environment is also the main strategy of targeted therapy. Considering that excessive ROS produced in the tumor promotes the transition from inflammation to cancer, Zhang et al. [70] synthesized a ROS-responsive and hydrogen peroxide-eliminating material by a cyclic polysaccharide, then used it to formulate functional nanoparticles loaded with irinotecan. Therefore, the nanoparticles can not only lighten oxidative stress and inflammatory response, but also release irinotecan under the simulation of the high level of ROS in the diseased colon. After oral administration of nanoparticles, the tumorigenesis and growth of colitis-induced CRC mice were expressively inhibited. In addition to inducing photodynamic therapy for CRC [71], NIR can also be used as an external stimulus to stimulate the release of drug from nanoparticles at the target. Yadav et al. [72] designed a compact shell-crosslinked micelle to deliver indocyanine green (ICG) and doxorubicin (DOX), which could release the drug when stimulated by NIR, because under NIR irradiation, the ICG generated ROS which broke the structure of micelles by destroying the diselenide bond in the micelles, and then a large amount of DOX would be released. Additionally, Anugrah et al. [73] prepared a hydrogel composed of alginate to encapsulate ICG and DOX, whose diselenide bonds could be decomposed by the ROS generated by the ICG and then released DOX by gel-sol transformation under the NIR light. Therefore, NIR light-responsive drug delivery system can also be applied in the targeted therapy of CRC. 3.2. Oral Colon-Targeted Nano Delivery Systems Nanomaterials will inevitably be recognized and swallowed by the immune system after intravenous injection, which will lead to adverse side reactions and reduce efficacy. In addition, the poor vascularization of CRC leads to a reduction in the number of nanoparticles by intravenous injection reaching the CRC and then limits the efficacy of nanoparticles [74]. Therefore, more and more scientists are committed to developing nano preparations that can be administered orally and still maintain the targeting of CRC. Furthermore, oral administration can improve patient compliance. To achieve this goal, the colon-targeted drug delivery systems are supposed to prevent the drug from gastrointestinal degradation before the drug enters the colon, hence increasing the drug concentration in the tumor. In many reports about colon-targeted therapy as shown in Table 2, pH [75], time [76], or enzyme-responsive [77] nanoplatforms are designed for CRC. 3.2.1. pH-Dependent Nano Platforms Among various methods, the most focused approach is to develop a pH-dependent system. The gastrointestinal tract is structurally divided into gastro, small bowels, and intestinum crassum. In addition to the difference in physiological function, the pH of each part is also different. The pH of the stomach is 1~3, that of the small intestine increases to 5.5~6.8, which is close to the neutral environment, and the pH of the colon is 6~8 [81]. Inspired by enteric coating tablets, the researchers began to develop a nano-drug delivery system containing enteric-coated materials to make drug released in the colon [82]. At low pH values, these nanoparticles are intact, but at high pH values, due to the dissolution of the coating materials, the nanoparticles swell and adhere to the colon to release the drug on a special area. The two most widely used coating materials are Eudragit and polysaccharides. Samprasit et al. [75] prepared nanoparticles that could adhere to the mucosa for the delivery of 𝛼-Mangostin to the colon against CRC, which could also stay in colon mucosa for a long time through adhesion. These nanoparticles were formed by chitosan and thiolated chitosan, the two were crossed by genipin, and the surface of nanoparticles was modified by Eudragit L100. There are many advantages of polymeric nanoparticles, such as encapsulating the drug, preventing drug degradation from various conditions, and improving the mucoadhesion and absorption within the GIT. However, they could not target the colon that was why the author modified the nanoparticles with Eudragit L100. Furthermore, it was mentioned that chitosan could not tolerate the acidity of the digestive tract after oral administration. The researcher in this article found that nanoparticles with chitosan as the main material and crosslinked with Eudragit S100 were inspiring vehicles that can specifically target the CRC and continuously release drug [83]. Pectin is a natural biopolymer extracted from plant polysaccharides, which acts as acid protection in the gastrointestinal environment [84]. Mohamed et al. [78] prepared SLNs coated with pectin and dry skim milk, which could release the encapsulated curcumin in the colon. Therefore, the oral bioavailability of curcumin was significantly enhanced. In addition to Eudragit and polysaccharides, another paper [79] used beta-lactoglobulin as a carrier to keep irinotecan from being destroyed in the stomach and the drug was released in the small bowel. Moreover, the MTT assay showed that these nanocarriers are more toxic to HT-29 cancer cell lines and AGS than the free drug. With the discovery of more and more enteric-coated materials, targeted nano-drug delivery systems based on pH will be more and more used in the targeted therapy of CRC. However, the efficacy of a pH-dependent system is still limited, mainly because the gastrointestinal tract varies greatly between different individuals. To deal with this problem, Taymouri et al. [76] prepared polymeric-coated capsules to deliver simvastatin (SIM) to the colon specifically, which was sensitive to both pH and time. Firstly, for the purpose of improving the solubility of the drug, the author used the anti-solvent crystallization technology to prepare the nanosuspension of SIM and studied whether the nanosuspension could produce a better anticancer effect against HT-29 compared with free drug. Then, a capsule composed of ethyl cellulose and Eudragit S100 was developed, in which ethylcellulose had a controlled release performance and could produce time-dependent release and Eudragit S100 had pH-dependent solubility. Finally, the optimized nanosuspension was mixed with sodium dodecyl sulfate to be freeze-dried and loaded into the capsule. The findings showed that SIM was not released in the stomach, but in the colon. In addition, compared with free SIM, SIM nanoparticles significantly enhanced the cytotoxicity of HT-29. There are some limitations of pH- or time-dependent nano-drug delivery systems. Firstly, pH-dependent nanoparticles may not fully target the colon because the colon (pH 6.8) is similar in pH to the small bowel (pH 7.4). Secondly, the uncertain time taken for gastrointestinal transit of the nanoplatforms makes time-dependent nanoparticles sometimes miss the targets [85]. Optimistically, designing nanoparticles that release drug dependent on colonic microbial degradation is another reliable strategy for the targeted therapy of CRC. 3.2.2. Enzyme-Triggered Nanoparticles There are more than 400 kinds of microbial flora in the colon, including two major categories of aerobic, such as Escherichia coli, and anaerobic, such as clostridium [86]. Moreover, some polysaccharides can only be degraded to smaller monosaccharides by the anaerobic microbiota of the colon, and then used by bacteria as an energy source, but cannot be digested by gastric and intestinal enzymes [87]. Hence, polysaccharides are of important function in the enzyme targeted therapy of CRC. They can not only control the location of drug release, but also are biodegradable and biocompatible natural polymers. In a study, colonic enzyme-responsive dextran-based oligoester crosslinked nanoparticles were fabricated to deliver 5-FU. It was found in vitro release studies that the nanoparticles released 75% of the drug within 12 h of incubation with glucanase, but there was no drug release under the pH conditions of the stomach and small intestine [77]. Tiryaki et al. [85] prepared nanoparticles containing organic and inorganic materials, that is, silica aerogels were coated with dextran and dextran aldehyde. By coated with dextran and dextran aldehyde, drug loaded in the silica aerogel particles were released merely in the colon because the dextran was degraded by dextranase. Apart from coating the nanoparticles, nanoparticles can also be incorporated into microparticles that are resistant to gastrointestinal enzymes. dos Santos et al. [80] synthesized chitosan nanoparticles loaded with 5-FU and then the nanoparticles were further encapsulated with microparticles made of degraded starch and pectin. Retrograde starch is a kind of modified starch that can resist enzyme degradation of the superior digestive tract. Additionally, their results showed that fewer nanoparticles were released from the microparticles than nanoparticles alone in the gastrointestinal lumen. It can be seen from these studies that polysaccharides are the key factor of enzyme-triggered colon targeted preparations. 4. Multifunctional Targeted Nanoparticles In order to further improve the targeting of nano preparations, more and more scholars have designed nanoparticles with multiple targeting mechanisms. Rajpoot et al. [88] designed dual-targeted nanoparticles, which contained folate-modified SLNs and enteric polymer-coated alginate microspheres that encapsulated the nanoparticles. Because of the pH-responsive of enteric polymer, the enteric-coated microbeads could release drug in the colon region after oral administration. Additionally, the combination of folate enabled nanoparticles to target CRC. In another study, trackable near-infrared persistent luminescence mesoporous silica nanoparticles coated with lactobacillus reuteri biofilm (LRM) were prepared that could protect drug from being digested and reaching the colon. The LRM played an important role in the nanoparticles. It could not only lengthen the release time of 5-FU and protect the 5-FU in the stomach, but also endow the nanoparticles with the ability to actively target the colon because LRM could recognize CRC through some of its biological components, such as adhesin [89]. It is also desirable to design drug delivery systems based on multiple environmental signals response. Ma et al. [90] prepared a pH- and enzyme-dependent nano preparation to deliver chemotherapeutic drug. In the study, indomethacin, 5-FU, and curcumin were encapsulated, respectively, into Eudragit RS nanoparticles by nanoprecipitation and then incorporated in biphasic microcapsules composed of chitosan and hydroxypropyl methyl cellulose by aerosolization. At last, coating the microcapsules with enteric soluble Eudragit S100 could protect the microcapsules from being degraded in the stomach. Furthermore, the microcapsules were released until they reached the colon since the chitosan was decomposed by bacterial enzymes, which could make the drug accumulate in CRC. 5. Challenges of Nano-Drug Delivery Systems in CRC Considering the high efficacy of nanoparticles in animal models of CRC, the clinical application of nanoparticles is promising. However, there are still some shortcomings of nano-drug delivery systems. The biggest challenge is how to produce nano-preparations on a large scale and evaluate their preclinical safety and effectiveness. At present, the composition and structure of nanoplatforms for targeted therapy of CRC are becoming more and more complex, which makes the preparation process of nanoplatforms more complicated and difficult to be synthesized repeatedly. In addition, the physical and chemical properties of nanoplatforms should be controlled in the production process, which brings higher requirements to manufacturing units and increases the difficulty and cost of scale-up production. Microfluidic technology attracts the attention of researchers. Valencia et al. [91] made the self-assembly of lipid–polymer and lipid–quantum dots (QDs) realized by microfluidic technology, which prepared stable and uniform nanoparticles. On the other hand, the complex structure of nanoparticles also brings potential toxicity risks to patients, so it is necessary to select a model that is more similar to the pathogenesis of human CRC to evaluate their toxicity in the preclinical study [92]. Additionally, in clinical trials, it is found that nanoplatforms tend to reduce the toxicity of drug rather than improve efficacy [93]. According to the analysis of many scholars, most nanoparticles accumulate at tumors based on EPR effect, even actively targeted nano preparations, but EPR effect is more consistent among animals, while there are differences in EPR effect for CRC patients, which will affect the efficacy of nano preparations [94,95]. Therefore, personalized therapy is needed, that is, nanoparticles are used in patients with strong EPR effect to achieve better efficacy [96], or to use preparations that can accumulate in the tumor independent of EPR effect, such as temperature response-based hydrogel for local treatment of CRC. Therefore, with the development of chemistry, biology, manufacturing, and other industries, nano-preparations will play a more important role in the targeted therapy of CRC. 6. Conclusions and Perspective In the treatment of CRC, colon-targeted nano-drug delivery systems can change the distribution and release of drug in humans by accumulating in CRC, which achieves a better therapeutic effect and reduces side effects in comparison with traditional treatment methods. Therefore, how to design nanoparticles with high targeting ability has great significance for CRC. In the review, we mainly summarize and classify the colon-targeted nanoparticles from the perspective of targeting power, demonstrating the richness and creativity of nanoparticles targeting CRC in academic research. Due to the complex pathophysiological changes of CRC, multifunctional targeted nanoparticles may have greater targeting ability. The goal of these studies is to provide patients with more effective preparations. However, because of the differences between preclinical animal models and clinical CRC patients as well as the stability of nano preparations, the translation of targeted nano preparations from research to clinic is difficult, which are also the directions that we need to consider and focus on when conducting research. Author Contributions Writing—original draft, K.W.; Writing—review and editing, K.W., R.S., T.M., F.H. and H.Y.; Funding acquisition, H.Y. and F.H. All authors have read and agreed to the published version of the manuscript. Conflicts of Interest The authors declare no conflict of interest. Abbreviations CRC colorectal cancer ESMO European Society for Medical Oncology EGFR epithelial growth factor EPR enhanced permeability and retention EpCAM epithelial cell adhesion molecule 5-FU 5-fluorouracil MSN mesoporous silica nanoparticles SLB supported lipid bilayer HA hyaluronic acid ROS reactive oxygen species NIR near infrared SLNs solid lipid nanoparticles ICG indocyanine green DOX doxorubicin SIM simvastatin LRM lactobacillus reuteri biofilm Figure 1 Types of nano-drug delivery systems targeting CRC. Figure 2 A schematic diagram of some representative targeting mechanism. molecules-27-02981-t001_Table 1 Table 1 Actively targeting nano delivery system for the treatment of colorectal cancer. Ligand Receptor Nano Delivery System Active Agents Size (nm) Charge (mv) Administration Route Ref. Folate Folate Receptor Self-assembled nanoparticles Bortezomib 196 28 IV [26] Anti-EGFR mAb Epidermal Growth Factor Receptor PLGA and PEG-based polymeric nanoparticles 5-Fluorouracil (5-FU) 252 −31 IV [27] EpCAM aptamer Epithelial cell adhesion molecule PAMAM dendrimers Celastrol 300 −6 IV [28] SS-Fc Carcinoembryonic antigen PEGylated hollow mesoporous ruthenium nanoparticles [Ru(bpy)2(tip)]2+, RBT 110 22 IV [29] AS1411 aptamer Nucleolin Silica nanoparticles coated with chitosan AntimiR-21, doxorubicin (DOX) 87 16 IV [30] Mannose Mannose receptor Cyclodextrin-based host–guest complexes Regorafenib 100 \ IV [31] Hyaluronic acid RHAMM, CD44 Hyaluronic Acid–Doxorubicin nanoparticles DOX 175 −5 IV [32] Wheat germ agglutinin N-acetyl-d-glucosamine, sialic acid PLGA nanoparticles 5-FU 156 −18 \ [33] Transferrin Transferrin receptor Polymersomes DOX 72 −2 IV [34] LRP-1 targeting peptide Lipoprotein receptor-related protein-1 Human serum albumin nanoparticles 5-FU 208 −13 IV [37] MUC1 aptamer MUC1 Mesenchymal-stem-cell-derived exosomes DOX 50 −80 IV [38] Tumor-homing peptide tLyp-1 NRP-1 Nanoparticles Paclitaxel, chlorin e6 107 −25 IV [39] Fucoidan P-selectin Nanoscale metal organic framework Talazoparib, temozolomide 84 −18 IV [40] Anisamide Sigma receptors Lipidic core nanocapsules Thymoquinone 217 −36 \ [41] Dexamethasone Glucocorticoid receptor Cationic liposomes ESC8, anti-Hsp90 plasmid 251 28 IV [43] molecules-27-02981-t002_Table 2 Table 2 Selected examples of oral colon-targeted formulations. Formulations Ingredients Targeting Strategy Ref. Polymeric nanoparticles Chitosan and thiolated chitosan, Eudragit L100, genipin pH responsive, mucoadhesiveness [75] Solid lipid nanoparticles Pectin, skimmed milk powder, lipid pH responsive [78] Beta-lactoglobulin nanoparticles Beta-lactoglobulin pH responsive [79] Polymeric coated capsule, nanosuspension Ethyl cellulose, Eudragit S100 pH responsive, time-dependent [76] Polymeric nanoparticles Dextran, bifunctional telechelic oligoester Enzyme responsive [77] Microparticle Chitosan, retrograded starch, pectin Enzyme responsive [80] Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Douaiher J. Ravipati A. Grams B. Chowdhury S. Alatise O. Are C. Colorectal cancer-global burden, trends, and geographical variations J. Surg. Oncol. 2017 115 619 630 10.1002/jso.24578 28194798 2. Mattiuzzi C. Sanchis-Gomar F. Lippi G. Concise update on colorectal cancer epidemiology Ann. Transl. Med. 2019 7 609 10.21037/atm.2019.07.91 32047770 3. Ribic C.M. Sargent D.J. Moore M.J. Thibodeau S.N. French A.J. Goldberg R.M. Hamilton S.R. Laurent-Puig P. Gryfe R. Shepherd L.E. Tumor microsatellite-instability status as a predictor of benefit from fluorouracil-based adjuvant chemotherapy for colon cancer N. Engl. J. Med. 2003 349 247 257 10.1056/NEJMoa022289 12867608 4. Hashiguchi Y. Muro K. Saito Y. Ito Y. Ajioka Y. Hamaguchi T. Hasegawa K. Hotta K. Ishida H. Ishiguro M. Japanese Society for Cancer of the Colon and Rectum (JSCCR) guidelines 2019 for the treatment of colorectal cancer Int. J. Clin. Oncol. 2020 25 1 42 10.1007/s10147-019-01485-z 31203527 5. De Rosa M. Pace U. Rega D. Costabile V. Duraturo F. Izzo P. Delrio P. Genetics, diagnosis and management of colorectal cancer (Review) Oncol. Rep. 2015 34 1087 1096 10.3892/or.2015.4108 26151224 6. Alomrani A. Badran M. Harisa G.I. Alshehry M. Alhariri M. Alshamsan A. Alkholief M. The use of chitosan-coated flexible liposomes as a remarkable carrier to enhance the antitumor efficacy of 5-fluorouracil against colorectal cancer Saudi Pharm. J. 2019 27 603 611 10.1016/j.jsps.2019.02.008 31297013 7. Goldberg R.M. Tabah-Fisch I. Bleiberg H. de Gramont A. Tournigand C. Andre T. Rothenberg M.L. Green E. Sargent D.J. Pooled analysis of safety and efficacy of oxaliplatin plus fluorouracil/leucovorin administered bimonthly in elderly patients with colorectal cancer J. Clin. Oncol. 2006 24 4085 4091 10.1200/JCO.2006.06.9039 16943526 8. Son H.S. Lee W.Y. Lee W.S. Yun S.H. Chun H.K. Compliance and effective management of the hand-foot syndrome in colon cancer patients receiving capecitabine as adjuvant chemotherapy Yonsei Med. J. 2009 50 796 802 10.3349/ymj.2009.50.6.796 20046420 9. Joye I. Haustermans K. Early and late toxicity of radiotherapy for rectal cancer Recent Results Cancer Res. 2014 203 189 201 10.1007/978-3-319-08060-4_13 25103006 10. Masago K. Imamichi F. Masuda Y. Ariga N. Fujitomi K. Fukumine Y. Hatakenaka K. Fujita S. Katakami N. Team Management of Skin Rash Associated with Use of Epidermal Growth Factor Receptor-Tyrosine Kinase Inhibitors Asia Pac. J. Oncol. Nurs. 2018 5 430 434 10.4103/apjon.apjon_33_18 30271827 11. Tang Y.Z. Szabados B. Leung C. Sahdev A. Adverse effects and radiological manifestations of new immunotherapy agents Br. J. Radiol. 2019 92 20180164 10.1259/bjr.20180164 30281331 12. Feldmann S.M. Cancer immunotherapy in routine cost-effective cancer care? EMBO Mol. Med. 2018 10 e9660 10.15252/emmm.201809660 30322840 13. Skalickova S. Loffelmann M. Gargulak M. Kepinska M. Docekalova M. Uhlirova D. Stankova M. Fernandez C. Milnerowicz H. Ruttkay-Nedecky B. Zinc-Modified Nanotransporter of Doxorubicin for Targeted Prostate Cancer Delivery Nanomaterials 2017 7 435 10.3390/nano7120435 14. Matsumura Y. Maeda H. A new concept for macromolecular therapeutics in cancer chemotherapy: Mechanism of tumoritropic accumulation of proteins and the antitumor agent smancs Cancer Res. 1986 46 6387 6392 2946403 15. Maeda H. Sawa T. Konno T. Mechanism of tumor-targeted delivery of macromolecular drugs, including the EPR effect in solid tumor and clinical overview of the prototype polymeric drug SMANCS J. Control. Release 2001 74 47 61 10.1016/S0168-3659(01)00309-1 11489482 16. Maeda H. Bharate G.Y. Daruwalla J. Polymeric drugs for efficient tumor-targeted drug delivery based on EPR-effect Eur. J. Pharm. Biopharm. 2009 71 409 419 10.1016/j.ejpb.2008.11.010 19070661 17. Wang Y. Ma J. Qiu T. Tang M. Zhang X. Dong W. In vitro and in vivo combinatorial anticancer effects of oxaliplatin- and resveratrol-loaded N,O-carboxymethyl chitosan nanoparticles against colorectal cancer Eur. J. Pharm. Sci. 2021 163 105864 10.1016/j.ejps.2021.105864 33965502 18. Acharya S. Sahoo S.K. PLGA nanoparticles containing various anticancer agents and tumour delivery by EPR effect Adv. Drug Deliv. Rev. 2011 63 170 183 10.1016/j.addr.2010.10.008 20965219 19. Anitha A. Maya S. Sivaram A.J. Mony U. Jayakumar R. Combinatorial nanomedicines for colon cancer therapy Wiley Interdiscip. Rev. Nanomed. Nanobiotechnol. 2016 8 151 159 10.1002/wnan.1353 26061225 20. Nichols J.W. Bae Y.H. EPR: Evidence and fallacy J. Control. Release 2014 190 451 464 10.1016/j.jconrel.2014.03.057 24794900 21. Golombek S.K. May J.N. Theek B. Appold L. Drude N. Kiessling F. Lammers T. Tumor targeting via EPR: Strategies to enhance patient responses Adv. Drug Deliv. Rev. 2018 130 17 38 10.1016/j.addr.2018.07.007 30009886 22. Hu Y. He Y. Ji J. Zheng S. Cheng Y. Tumor Targeted Curcumin Delivery by Folate-Modified MPEG-PCL Self-Assembly Micelles for Colorectal Cancer Therapy Int. J. Nanomed. 2020 15 1239 1252 10.2147/IJN.S232777 23. Dabkeviciene D. Jonusiene V. Zitkute V. Zalyte E. Grigaitis P. Kirveliene V. Sasnauskiene A. The role of interleukin-8 (CXCL8) and CXCR2 in acquired chemoresistance of human colorectal carcinoma cells HCT116 Med. Oncol. 2015 32 258 10.1007/s12032-015-0703-y 26519257 24. Yu D.H. Lu Q. Xie J. Fang C. Chen H.Z. Peptide-conjugated biodegradable nanoparticles as a carrier to target paclitaxel to tumor neovasculature Biomaterials 2010 31 2278 2292 10.1016/j.biomaterials.2009.11.047 20053444 25. Jain A.K. Massey A. Yusuf H. McDonald D.M. McCarthy H.O. Kett V.L. Development of polymeric-cationic peptide composite nanoparticles, a nanoparticle-in-nanoparticle system for controlled gene delivery Int. J. Nanomed. 2015 10 7183 7196 10.2147/IJN.S95245 26. Soe Z.C. Poudel B.K. Nguyen H.T. Thapa R.K. Ou W. Gautam M. Poudel K. Jin S.G. Jeong J.H. Ku S.K. Folate-targeted nanostructured chitosan/chondroitin sulfate complex carriers for enhanced delivery of bortezomib to colorectal cancer cells Asian J. Pharm. Sci. 2019 14 40 51 10.1016/j.ajps.2018.09.004 32104437 27. Bhattacharya S. Anti-EGFR-mAb and 5-Fluorouracil Conjugated Polymeric Nanoparticles for Colorectal Cancer Recent Pat. Anti-Cancer Drug Discov. 2021 16 84 100 10.2174/1574892815666201221121859 28. Ge P.J. Niu B.N. Wu Y.H. Xu W.X. Li M.Y. Sun H.S. Zhou H. Zhang X.K. Xie J.J. Enhanced cancer therapy of celastrol in vitro and in vivo by smart dendrimers delivery with specificity and biosafety Chem. Eng. J. 2020 383 123228 10.1016/j.cej.2019.123228 29. Xu M. Wen Y.Y. Liu Y.N. Tan X.J. Chen X. Zhu X.F. Wei C.F. Chen L.M. Wang Z. Liu J. Hollow mesoporous ruthenium nanoparticles conjugated bispecific antibody for targeted anti-colorectal cancer response of combination therapy Nanoscale 2019 11 9661 9678 10.1039/C9NR01904A 31065660 30. Khatami F. Matin M.M. Danesh N.M. Bahrami A.R. Abnous K. Taghdisi S.M. Targeted delivery system using silica nanoparticles coated with chitosan and AS1411 for combination therapy of doxorubicin and antimiR-21 Carbohydr. Polym. 2021 266 118111 10.1016/j.carbpol.2021.118111 34044928 31. Bai H.Z. Wang J.W. Phan C.U. Chen Q. Hu X.R. Shao G.Q. Zhou J. Lai L.H. Tang G.P. Cyclodextrin-based host-guest complexes loaded with regorafenib for colorectal cancer treatment Nat. Commun. 2021 12 759 10.1038/s41467-021-21071-0 33536421 32. Pan D.C. Krishnan V. Salinas A.K. Kim J. Sun T. Ravid S. Peng K. Wu D. Nurunnabi M. Nelson J.A. Hyaluronic acid-doxorubicin nanoparticles for targeted treatment of colorectal cancer Bioeng. Transl. Med. 2021 6 e10166 10.1002/btm2.10166 33532580 33. Pandey A.N. Rajpoot K. Jain S.K. Using 5-fluorouracil-encored plga nanoparticles for the treatment of colorectal cancer: The in-vitro characterization and cytotoxicity studies Nanomed. J. 2020 7 211 224 10.22038/nmj.2020.07.0005 34. Wei Y.H. Gu X.L. Sun Y.P. Meng F.H. Storm G. Zhong Z.Y. Transferrin-binding peptide functionalized polymersomes mediate targeted doxorubicin delivery to colorectal cancer in vivo J. Control. Release 2020 319 407 415 10.1016/j.jconrel.2020.01.012 31923538 35. Ahmad A. Ansari M.M. Verma R.K. Khan R. Aminocellulose-Grafted Polymeric Nanoparticles for Selective Targeting of CHEK2-Deficient Colorectal Cancer ACS Appl. Bio. Mater. 2021 4 5324 5335 10.1021/acsabm.1c00437 36. Serna N. Alamo P. Ramesh P. Vinokurova D. Sanchez-Garcia L. Unzueta U. Gallardo A. Cespedes M.V. Vazquez E. Villaverde A. Nanostructured toxins for the selective destruction of drug-resistant human CXCR4(+) colorectal cancer stem cells J. Control. Release 2020 320 96 104 10.1016/j.jconrel.2020.01.019 31931052 37. Lee K.J. Ko E.J. Park Y.Y. Park S.S. Ju E.J. Park J. Shin S.H. Suh Y.A. Hong S.M. Park I.J. A novel nanoparticle-based theranostic agent targeting LRP-1 enhances the efficacy of neoadjuvant radiotherapy in colorectal cancer Biomaterials 2020 255 120151 10.1016/j.biomaterials.2020.120151 32505033 38. Bagheri E. Abnous K. Farzad S.A. Taghdisi S.M. Ramezani M. Alibolandi M. Targeted doxorubicin-loaded mesenchymal stem cells-derived exosomes as a versatile platform for fighting against colorectal cancer Life Sci. 2020 261 118369 10.1016/j.lfs.2020.118369 32882265 39. Shi Y. Shan S.F. Li C.H. Song X.J. Zhang C.X. Chen J.Y. You J. Xiong J.Y. Application of the Tumor Site Recognizable and Dual-Responsive Nanoparticles for Combinational Treatment of the Drug-Resistant Colorectal Cancer Pharm. Res. 2020 37 72 10.1007/s11095-020-02791-2 32215748 40. DuRoss A.N. Landry M.R. Thomas C.R. Neufeld M.J. Sun C. Fucoidan-coated nanoparticles target radiation-induced P-selectin to enhance chemoradiotherapy in murine colorectal cancer Cancer Lett. 2021 500 208 219 10.1016/j.canlet.2020.11.021 33232787 41. Ramzy L. Metwally A.A. Nasr M. Awad G.A.S. Novel thymoquinone lipidic core nanocapsules with anisamide-polymethacrylate shell for colon cancer cells overexpressing sigma receptors Sci. Rep. 2020 10 10987 10.1038/s41598-020-67748-2 32620860 42. Abdellatif A.A.H. Ibrahim M.A. Amin M.A. Maswadeh H. Alwehaibi M.N. Al-Harbi S.N. Alharbi Z.A. Mohammed H.A. Mehany A.B.M. Saleem I. Cetuximab Conjugated with Octreotide and Entrapped Calcium Alginate-beads for Targeting Somatostatin Receptors Sci. Rep. 2020 10 4736 10.1038/s41598-020-61605-y 32170176 43. Jinka S. Rachamalla H.K. Bhattacharyya T. Sridharan K. Jaggarapu M. Yakati V. Banerjee R. Glucocorticoid receptor-targeted liposomal delivery system for delivering small molecule ESC8 and anti-miR-Hsp90 gene construct to combat colon cancer Biomed. Mater. 2021 16 024105 10.1088/1748-605X/abdb08 33434900 44. Trzpis M. McLaughlin P.M. de Leij L.M. Harmsen M.C. Epithelial cell adhesion molecule: More than a carcinoma marker and adhesion molecule Am. J. Pathol. 2007 171 386 395 10.2353/ajpath.2007.070152 17600130 45. Maetzel D. Denzel S. Mack B. Canis M. Went P. Benk M. Kieu C. Papior P. Baeuerle P.A. Munz M. Nuclear signalling by tumour-associated antigen EpCAM Nat. Cell Biol. 2009 11 162 171 10.1038/ncb1824 19136966 46. Herlyn M. Steplewski Z. Herlyn D. Koprowski H. Colorectal carcinoma-specific antigen: Detection by means of monoclonal antibodies Proc. Natl. Acad. Sci. USA 1979 76 1438 1442 10.1073/pnas.76.3.1438 286328 47. Chong J.M. Speicher D.W. Determination of disulfide bond assignments and N-glycosylation sites of the human gastrointestinal carcinoma antigen GA733-2 (CO17-1A, EGP, KS1-4, KSA, and Ep-CAM) J. Biol. Chem. 2001 276 5804 5813 10.1074/jbc.M008839200 11080501 48. Zhao Y.Y. Xu J.J. Le V.M. Gong Q.Y. Li S.Y. Gao F. Ni L. Liu J.W. Liang X. EpCAM Aptamer-Functionalized Cationic Liposome-Based Nanoparticles Loaded with miR-139-5p for Targeted Therapy in Colorectal Cancer Mol. Pharm. 2019 16 4696 4710 10.1021/acs.molpharmaceut.9b00867 31589818 49. Li C. Cai G. Song D. Gao R. Teng P. Zhou L. Ji Q. Sui H. Cai J. Li Q. Development of EGFR-targeted evodiamine nanoparticles for the treatment of colorectal cancer Biomater Sci. 2019 7 3627 3639 10.1039/C9BM00613C 31328737 50. Sabra R. Billa N. Roberts C.J. Cetuximab-conjugated chitosan-pectinate (modified) composite nanoparticles for targeting colon cancer Int. J. Pharm. 2019 572 118775 10.1016/j.ijpharm.2019.118775 31678385 51. Pedrosa P. Corvo M.L. Ferreira-Silva M. Martins P. Carvalheiro M.C. Costa P.M. Martins C. Martins L. Baptista P.V. Fernandes A.R. Targeting Cancer Resistance via Multifunctional Gold Nanoparticles Int. J. Mol. Sci. 2019 20 5510 10.3390/ijms20215510 31694227 52. Rafiei P. Haddadi A. Docetaxel-loaded PLGA and PLGA-PEG nanoparticles for intravenous application: Pharmacokinetics and biodistribution profile Int. J. Nanomed. 2017 12 935 947 10.2147/IJN.S121881 28184163 53. El Hallal R. Lyu N. Wang Y. Effect of Cetuximab-Conjugated Gold Nanoparticles on the Cytotoxicity and Phenotypic Evolution of Colorectal Cancer Cells Molecules 2021 26 567 10.3390/molecules26030567 33499047 54. Prewett M. Bassi R. Carrick F. Finnerty B. Ellis L.M. Witte L. Tonra J.R. Hicklin D.J. Fan F. Cetuximab (ERBITUX (R)) enhances clinical efficacy of oxaliplatin in human colon carcinoma xenografts and reverses oxaliplatin resistance Clin. Cancer Res. 2005 11 9048S 55. Prewett M.C. Hooper A.T. Bassi R. Ellis L.M. Waksal H.W. Hicklin D.J. Enhanced antitumor activity of anti-epidermal growth factor receptor monoclonal antibody IMC-C225 in combination with irinotecan (CPT-11) against human colorectal tumor xenografts Clin. Cancer Res. 2002 8 994 1003 12006511 56. Chen R. Huang Y. Wang L. Zhou J. Tan Y. Peng C. Yang P. Peng W. Li J. Gu Q. Cetuximab functionalization strategy for combining active targeting and antimigration capacities of a hybrid composite nanoplatform applied to deliver 5-fluorouracil: Toward colorectal cancer treatment Biomater. Sci. 2021 9 2279 2294 10.1039/D0BM01904F 33538278 57. Medrano-González P.A. Rivera-Ramírez O. Montaño L.F. Rendón-Huerta E.P. Proteolytic Processing of CD44 and Its Implications in Cancer Stem Cells Int. 2021 2021 6667735 10.1155/2021/6667735 33505471 58. Wang X.L. Ouyang X.M. Chen J.L. Hu Y. Sun X.Y. Yu Z.W. Nanoparticulate photosensitizer decorated with hyaluronic acid for photodynamic/photothermal cancer targeting therapy Nanomedicine 2019 14 151 167 10.2217/nnm-2018-0204 30511886 59. Lokeshwar B.L. Lokeshwar V.B. Block N.L. Expression of CD44 in prostate cancer cells: Association with cell proliferation and invasive potential Anticancer Res. 1995 15 1191 1198 7544562 60. Huang J. Zhang H. Yu Y. Chen Y. Wang D. Zhang G. Zhou G. Liu J. Sun Z. Sun D. Biodegradable self-assembled nanoparticles of poly (D,L-lactide-co-glycolide)/hyaluronic acid block copolymers for target delivery of docetaxel to breast cancer Biomaterials 2014 35 550 566 10.1016/j.biomaterials.2013.09.089 24135268 61. Gao S. Wang J. Tian R. Wang G. Zhang L. Li Y. Li L. Ma Q. Zhu L. Construction and Evaluation of a Targeted Hyaluronic Acid Nanoparticle/Photosensitizer Complex for Cancer Photodynamic Therapy ACS Appl. Mater. Interfaces 2017 9 32509 32519 10.1021/acsami.7b09331 28875691 62. Liu J.M. Zhang D.D. Fang G.Z. Wang S. Erythrocyte membrane bioinspired near-infrared persistent luminescence nanocarriers for in vivo long-circulating bioimaging and drug delivery Biomaterials 2018 165 39 47 10.1016/j.biomaterials.2018.02.042 29501968 63. Wang Z.H. Liu J.M. Zhao N. Li C.Y. Lv S.W. Hu Y.Z. Lv H. Wang D. Wang S. Cancer Cell Macrophage Membrane Camouflaged Persistent Luminescent Nanoparticles for Imaging-Guided Photothermal Therapy of Colorectal Cancer ACS Appl. Nano Mater. 2020 3 7105 7118 10.1021/acsanm.0c01433 64. Zhang Q. Dehaini D. Zhang Y. Zhou J. Chen X. Zhang L. Fang R.H. Gao W. Zhang L. Neutrophil membrane-coated nanoparticles inhibit synovial inflammation and alleviate joint damage in inflammatory arthritis Nat. Nanotechnol. 2018 13 1182 1190 10.1038/s41565-018-0254-4 30177807 65. Wang Z.H. Liu J.M. Yang F.E. Hu Y.Z. Lv H. Wang S. Tailor-Made Cell-Based Biomimetic Nanoprobes for Fluorescence Imaging Guided Colorectal Cancer Chemo-immunotherapy ACS Appl. Bio Mater. 2021 4 1920 1931 10.1021/acsabm.0c01553 66. Yusefi M. Lee-Kiun M.S. Shameli K. Teow S.Y. Ali R.R. Siew K.K. Chan H.Y. Wong M.M.T. Lim W.L. Kuca K. 5-Fluorouracil loaded magnetic cellulose bionanocomposites for potential colorectal cancer treatment Carbohydr. Polym. 2021 273 118523 10.1016/j.carbpol.2021.118523 34560940 67. Wu D. Li Y. Zhu L.X. Zhang W.Y. Xu S.M. Yang Y. Yan Q.Y. Yang G.S. A biocompatible superparamagnetic chitosan-based nanoplatform enabling targeted SN-38 delivery for colorectal cancer therapy Carbohydr. Polym. 2021 274 118641 10.1016/j.carbpol.2021.118641 34702462 68. Dabaghi M. Quaas R. Hilger I. The Treatment of Heterotopic Human Colon Xenograft Tumors in Mice with 5-Fluorouracil Attached to Magnetic Nanoparticles in Combination with Magnetic Hyperthermia Is More Efficient than Either Therapy Alone Cancers 2020 12 2562 10.3390/cancers12092562 69. Xing R. Mustapha O. Ali T. Rehman M. Zaidi S.S. Baseer A. Batool S. Mukhtiar M. Shafique S. Malik M. Development, Characterization, and Evaluation of SLN-Loaded Thermoresponsive Hydrogel System of Topotecan as Biological Macromolecule for Colorectal Delivery Biomed Res. Int. 2021 2021 9968602 10.1155/2021/9968602 34285920 70. Zhang Q.X. Zhang F.Z. Li S.S. Liu R.F. Jin T.T. Dou Y. Zhou Z.H. Zhang J.X. A Multifunctional Nanotherapy for Targeted Treatment of Colon Cancer by Simultaneously Regulating Tumor Microenvironment Theranostics 2019 9 3732 3753 10.7150/thno.34377 31281510 71. Kumar B. Murali A. Bharath A.B. Giri S. Guar gum modified upconversion nanocomposites for colorectal cancer treatment through enzyme-responsive drug release and NIR-triggered photodynamic therapy Nanotechnology 2019 30 315102 10.1088/1361-6528/ab116e 30893650 72. Yadav S. Ramesh K. Kumar P. Jo S.-H. Yoo S.I. Gal Y.-S. Park S.-H. Lim K.T. Near-Infrared Light-Responsive Shell-Crosslinked Micelles of Poly(d,l-lactide)-b-poly((furfuryl methacrylate)-co-(N-acryloylmorpholine)) Prepared by Diels–Alder Reaction for the Triggered Release of Doxorubicin Materials 2021 14 7913 10.3390/ma14247913 34947507 73. Anugrah D.S.B. Ramesh K. Kim M. Hyun K. Lim K.T. Near-infrared light-responsive alginate hydrogels based on diselenide-containing cross-linkage for on demand degradation and drug release Carbohydr. Polym. 2019 223 115070 10.1016/j.carbpol.2019.115070 31427031 74. Kumar B. Kulanthaivel S. Mondal A. Mishra S. Banerjee B. Bhaumik A. Banerjee I. Giri S. Mesoporous silica nanoparticle based enzyme responsive system for colon specific drug delivery through guar gum capping Colloids Surf. B Biointerfaces 2017 150 352 361 10.1016/j.colsurfb.2016.10.049 27847225 75. Samprasit W. Opanasopit P. Chamsai B. Mucoadhesive chitosan and thiolated chitosan nanoparticles containing alpha mangostin for possible Colon-targeted delivery Pharm. Dev. Technol. 2021 26 362 372 10.1080/10837450.2021.1873370 33423571 76. Taymouri S. Ahmadi Z. Mirian M. Tavakoli N. Simvastatin nanosuspensions prepared using a combination of pH-sensitive and timed-release approaches for potential treatment of colorectal cancer Pharm. Dev. Technol. 2021 26 335 348 10.1080/10837450.2021.1872086 33430677 77. Abid M. Naveed M. Azeem I. Faisal A. Nazar M.F. Yameen B. Colon specific enzyme responsive oligoester crosslinked dextran nanoparticles for controlled release of 5-fluorouracil Int. J. Pharm. 2020 586 119605 10.1016/j.ijpharm.2020.119605 32650112 78. Mohamed J.M. Alqahtani A. Ahmad F. Krishnaraju V. Kalpana K. Pectin co-functionalized dual layered solid lipid nanoparticle made by soluble curcumin for the targeted potential treatment of colorectal cancer Carbohydr. Polym. 2021 252 117180 10.1016/j.carbpol.2020.117180 33183627 79. Bijari N. Ghobadi S. Derakhshandeh K. -lactoglobulin-irinotecan inclusion complex as a new targeted nanocarrier for colorectal cancer cells Res. Pharm. Sci. 2019 14 216 227 10.4103/1735-5362.258488 31160899 80. Dos Santos A.M. Meneguin A.B. Akhter D.T. Fletcher N. Houston Z.H. Bell C. Thurecht K.J. Gremia M.P.D. Understanding the role of colon-specific microparticles based on retrograded starch/pectin in the delivery of chitosan nanoparticles along the gastrointestinal tract Eur. J. Pharm. Biopharm. 2021 158 371 378 10.1016/j.ejpb.2020.12.004 33309846 81. Koziolek M. Grimm M. Becker D. Iordanov V. Zou H. Shimizu J. Wanke C. Garbacz G. Weitschies W. Investigation of pH and Temperature Profiles in the GI Tract of Fasted Human Subjects Using the Intellicap((R)) System J. Pharm. Sci. 2015 104 2855 2863 10.1002/jps.24274 25411065 82. Fattal E. Youssef M. Couvreur P. Andremont A. Treatment of experimental salmonellosis in mice with ampicillin-bound nanoparticles Antimicrob. Agents Chemother. 1989 33 1540 1543 10.1128/AAC.33.9.1540 2684009 83. Sood A. Dev A. Mohanbhai S.J. Shrimali N. Kapasiya M. Kushwaha A.C. Choudhury S.R. Guchhait P. Karmakar S. Disulfide-Bridged Chitosan-Eudragit S-100 Nanoparticles for Colorectal Cancer ACS Appl. Nano Mater. 2019 2 6409 6417 10.1021/acsanm.9b01377 84. Wang T.R. Ma X.Y. Lei Y. Luo Y.C. Solid lipid nanoparticles coated with cross-linked polymeric double layer for oral delivery of curcumin Colloids Surf. B-Biointerfaces 2016 148 1 11 10.1016/j.colsurfb.2016.08.047 27588376 85. Tiryaki E. Elalmis Y.B. Ikizler B.K. Yucel S. Novel organic/inorganic hybrid nanoparticles as enzyme-triggered drug delivery systems: Dextran and Dextran aldehyde coated silica aerogels J. Drug Deliv. Sci. Technol. 2020 56 101517 10.1016/j.jddst.2020.101517 86. Gibson S.A. McFarlan C. Hay S. MacFarlane G.T. Significance of microflora in proteolysis in the colon Appl. Env. Microbiol. 1989 55 679 683 10.1128/aem.55.3.679-683.1989 2648991 87. Wang M. Hao W.S. Zhang L. Zhu Y.N. Chen K. Ma S. Cheng X.F. Zhao J.S. Lipid-polymer nano core-shell type hybrid system for colon specific drug delivery J. Drug Deliv. Sci. Technol. 2021 63 102540 10.1016/j.jddst.2021.102540 88. Rajpoot K. Jain S.K. Oral delivery of pH-responsive alginate microbeads incorporating folic acid-grafted solid lipid nanoparticles exhibits enhanced targeting effect against colorectal cancer: A dual-targeted approach Int. J. Biol. Macromol. 2020 151 830 844 10.1016/j.ijbiomac.2020.02.132 32061847 89. Wang Z.H. Liu J.M. Li C.Y. Wang D. Lv H. Lv S.W. Zhao N. Ma H. Wang S. Bacterial Biofilm Bioinspired Persistent Luminescence Nanoparticles with Gut-Oriented Drug Delivery for Colorectal Cancer Imaging and Chemotherapy ACS Appl. Mater. Interfaces 2019 11 36409 36419 10.1021/acsami.9b12853 31525949 90. Ma Y.M. Thurecht K.J. Coombes A.G.A. Development of enteric-coated, biphasic chitosan/HPMC microcapsules for colon-targeted delivery of anticancer drug-loaded nanoparticles Int. J. Pharm. 2021 607 121026 10.1016/j.ijpharm.2021.121026 34418474 91. Valencia P.M. Basto P.A. Zhang L.F. Rhee M. Langer R. Farokhzad O.C. Karnik R. Single-Step Assembly of Homogenous Lipid—Polymeric and Lipid—Quantum Dot Nanoparticles Enabled by Microfluidic Rapid Mixing ACS Nano 2010 4 1671 1679 10.1021/nn901433u 20166699 92. Choi S.Y.C. Lin D. Gout P.W. Collins C.C. Xu Y. Wang Y.Z. Lessons from patient-derived xenografts for better in vitro modeling of human cancer Adv. Drug Deliv. Rev. 2014 79–80 222 237 10.1016/j.addr.2014.09.009 93. Min Y.Z. Caster J.M. Eblan M.J. Wang A.Z. Clinical Translation of Nanomedicine Chem. Rev. 2015 115 11147 11190 10.1021/acs.chemrev.5b00116 26088284 94. Bertrand N. Wu J. Xu X.Y. Kamaly N. Farokhzad O.C. Cancer nanotechnology: The impact of passive and active targeting in the era of modern cancer biology Adv. Drug Deliv. Rev. 2014 66 2 25 10.1016/j.addr.2013.11.009 24270007 95. Prabhakar U. Maeda H. Jain R.K. Sevick-Muraca E.M. Zamboni W. Farokhzad O.C. Barry S.T. Gabizon A. Grodzinski P. Blakey D.C. Challenges and Key Considerations of the Enhanced Permeability and Retention Effect for Nanomedicine Drug Delivery in Oncology Cancer Res. 2013 73 2412 2417 10.1158/0008-5472.CAN-12-4561 23423979 96. Kim T.H. Lee S. Chen X.Y. Nanotheranostics for personalized medicine Expert Rev. Mol. Diagn. 2013 13 257 269 10.1586/erm.13.15 23570404

PMC009xxxxxx/PMC9099629.txt

==== Front Int J Mol Sci Int J Mol Sci ijms International Journal of Molecular Sciences 1422-0067 MDPI 10.3390/ijms23094768 ijms-23-04768 Review Role of Short-Chain Fatty Acids Produced by Gut Microbiota in Innate Lung Immunity and Pathogenesis of the Heterogeneous Course of Chronic Obstructive Pulmonary Disease https://orcid.org/0000-0002-7083-2692 Kotlyarov Stanislav Krick Stefanie Academic Editor Department of Nursing, Ryazan State Medical University, 390026 Ryazan, Russia; SKMR1@yandex.ru 26 4 2022 5 2022 23 9 476828 2 2022 22 4 2022 © 2022 by the author. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Chronic obstructive pulmonary disease (COPD) is a widespread socially significant disease. The development of COPD involves the innate immune system. Interestingly, the regulation of the innate lung immune system is related to the gut microbiota. This connection is due to the production by gut microorganisms of short-chain fatty acids (SCFAs) such as acetate, propionate, and butyrate. Nutritional disturbances and changes in the structure of the intestinal microbiota lead to a decrease in SCFAs production and their effect on pulmonary immunity. The presence of a metabolic and immune axis linking the lungs and gut plays an important role in the pathogenesis of COPD. In addition, the nature of nutrition and SCFAs may participate in the development of the clinically heterogeneous course of COPD. COPD innate immune system gut microbiota short-chain fatty acids nutritional support phenotype This research received no external funding. ==== Body pmc1. Introduction Chronic obstructive pulmonary disease (COPD) is one of the most common respiratory diseases of social significance [1]. It is an important cause of health care seeking, hospitalizations, and disability [2,3]. Epidemiological studies show unfavorable trends in the prevalence of COPD in many countries [4,5]. The urgency of the problem is also increased by the fact that COPD is among the leading causes of death in the world. Moreover, treatment and rehabilitation of patients remain an unsolved problem in many respects, which requires new more detailed data on pathophysiological mechanisms of disease development and on those parts of pathogenesis, which can be effectively influenced for therapeutic purposes. The development and progression of the disease is associated with exposure to tobacco smoke components and exogenous aeropollutants [6,7,8]. Inflammation in the bronchial tree is known to underlie the pathogenesis of COPD and is characterised by the involvement of many cells, including macrophages and neutrophils [9]. Prolonged persistent inflammation is associated with impaired immune mechanisms that provide a balance between maintaining and resolving inflammation [10]. It is known that COPD is a disease characterized by both pulmonary and extrapulmonary clinical heterogeneity [11,12]. Moreover, some features of the clinical picture and phenotypes of COPD are related to nutrition. Body mass index (BMI) is considered as one of the important markers in assessing the prognosis of COPD [13]. Lower body weight is associated with a greater risk of adverse outcome, whereas excess body weight has an even better prognosis [14]. This phenomenon has been called the “obesity paradox” [15,16,17]. The course of COPD, in which higher energy requirements lead to loss of fat and muscle tissue, is thought to be prognostically unfavorable [18,19]. The increased energy requirement may be related both to respiration and to the maintenance of inflammation in the bronchi. Dietary modification is considered an important tool that can improve the clinical course of COPD [20]. It is suggested that the nature of nutrition may be related to the pathophysiological mechanisms of COPD not only through the influence on energy and metabolic processes, but also through the regulation of some immune mechanisms [21]. A growing body of evidence supports a metabolic and immune axis linking the gut and lungs [22,23]. These links are bidirectional, and the gut microbiome is one of the central links in this interaction. The gut is the site of localization of much of the commensal bacterial mass [24,25]. This microbiome is metabolically active, participating in the production of many substances, such as, short-chain fatty acids (SCFAs) [26]. Accumulating evidence suggests that SCFAs can be considered as a leading link in the metabolic and immune axis between the gut and lungs. Indeed, SCFAs exhibit a variety of functions in immune defence, making them a potentially important target for clinical and experimental research. Thus, the aim of this review is to discuss the putative role of SCFAs produced by intestinal microbiota in innate lung immune defence and the pathogenesis of the heterogeneous course of COPD. 2. Short-Chain Fatty Acids 2.1. Production of SCFAs SCFAs are straight- or branched-chain fatty acids with less than six carbon atoms. The most common SCFAs are acetate, propionate, and butyrate, which are produced by the intestinal microbiota through anaerobic fermentation of non-digestible polysaccharides. The rate of production, amount, and ratio of SCFAs depend on several factors, such as the species structure and quantitative composition of the microbiota in the colon, the nature of the substrate, and the time of its passage through the gut [27]. Substrates for SCFA formation are dietary fibers, including resistant starch, cellulose, and pectin [28]. Dietary fibers that are not enzymatically processed by humans become a substrate for the microbiota of the large intestine. The colon microbiota hydrolyze undigested carbohydrates into oligosaccharides and then into monosaccharides, followed by fermentation through glycolysis and the pentose phosphate pathway [29]. Acetate is the major SCFA in the colon, which is due to the widespread occurrence of acetate production pathways among bacteria [30]. Acetate is formed from pyruvate via acetyl-CoA or via the Wood-Ljungdahl pathway [29]. Butyrate is produced by condensation of two acetyl-CoA molecules to form acetoacetyl-CoA followed by reduction to butyryl-CoA [31]. In this process, butyrate production is mediated by butyrate kinase. The other pathway of butyrate formation, which uses acetate of exogenous origin, involves butyryl-CoA:acetate CoA-transferase [29,32]. This variant involves coordination between acetate-producing and butyrate-producing bacteria [33]. Propionate production can occur via the succinate pathway or the acrylate pathway, in which lactate is reduced to propionate, and via the propanediol pathway from deoxy sugars such as fucose and rhamnose (Figure 1) [29,31]. In addition to carbohydrates, SCFA formation in the gut can also occur through amino acid metabolism [31]. Protein fermentation is associated with the formation of branched-chain SCFAs. For example, isobutyrate, 2-methylbutyrate, and isovalerate can be formed from branched-chain amino acids such as valine, isoleucine, and leucine [34]. In addition to fatty acid formation, amino acid fermentation produces potentially harmful metabolites such as phenolic and indole compounds, amines, and ammonia [31,34,35]. 2.2. SCFAs Transport SCFAs are found in the largest concentrations in the large intestine, ranging from about 70 to 130 mmol/kg [36,37]. Acetate, propionate, and butyrate are found in the colon in a molar ratio of approximately 57:22:21 [29,37]. SCFAs enter the systemic bloodstream via passive diffusion or involving specific transporters, such as the monocarboxylate transporter 1 (MCT1) and sodium-linked monocarboxylate transporter 1 (SMCT1) [28]. SCFAs are transported through the colonocytes, where most of the butyrate produced is used as the main energy source [38]. Colonocytes can obtain up to 60–70% of their energy from SCFA by β-oxidation and metabolism in the tricarboxylic acid cycle (TCA cycle) [28,29,39,40]. The remaining SCFA fractions enter the portal bloodstream through the basolateral membrane. In the portal vein, the ratio of acetate, propionate, and butyrate is approximately 69:23:8 [29]. In turn, propionate, passing through the portal vein, is metabolized in the liver, where it can be used for gluconeogenesis [41,42]. Because of these processes, acetate is the most common SCFA, as most of it enters the systemic bloodstream. As a result, plasma concentrations of acetate, propionate, and butyrate are approximately 25–250 μmol/l, 1.4–13.4 μmol/l, and 0.5–14.2 μmol/l, respectively [37,43]. The higher plasma levels of acetate are also related to the fact that acetate may not only be of intestinal origin. It can be produced by fatty acid oxidation and amino acid metabolism, by ketogenesis in hepatocytes, and by ethanol oxidation by microsomal cytochrome P450 enzymes [43,44,45,46]. 2.3. SCFAs Signal Transduction SCFAs are thought to exert their action by interacting with the G-protein-coupled receptors GPR43 and GPR41, also known as free fatty acid receptor (FFA)2 and FFA3, respectively [47,48,49,50]. SCFAs differ in their selectivity in activating FFA2 and FFA3 receptors. It has been shown that FFA2 is activated to a greater extent by fatty acids with a shorter chain, whereas FFA3 is characterized by an inverse relationship [51]. In particular, this dependence on the number of carbon atoms for FFA2 can be represented as C2 = C3 > C4 > C5 = C1, and for FFA3 as follows: C3 = C4 = C5 > C2 = C1 [52]. GPR43 is expressed in immune cells, including neutrophils, monocytes, and lymphocytes [47,48,49,53]. Other receptors for SCFAs are GPR109a, which is also known as HCA2 and olfactory receptor 78 (Olfr78) [54,55,56]. The GPR109A receptor is expressed predominantly on adipocytes as well as in immune cells such as neutrophils and macrophages [57,58]. Butyrate slightly activates GPR109A, whereas propionate and acetate do not activate the receptor [59]. In addition to binding to the described receptors, SCFAs exert their action through the inhibition of histone deacetylase (HDAC). Histone acetylation is an important mechanism for controlling gene transcription. During this process, acetyl groups are added to histone tails by histone acetyltransferases (HATs) and removed by (HDACs). HDACs are a class of enzymes that inhibit transcription through the removal of acetyl groups from chromatin [47,60]. Butyrate, being the strongest HDAC inhibitor, alters the expression of many genes with different functions [61,62]. Through HDAC inhibition, SCFAs are involved in the regulation of many cellular functions, including migration and survival [47,63,64,65,66]. 2.4. Participation of SCFAs in the Regulation of Metabolic and Immune processes A growing body of evidence strengthens the understanding of the importance of SCFAs in the regulation of some metabolic and immune processes (Figure 2). Through HDAC3 inhibition, butyrate can induce a metabolic switch of macrophages toward an anti-inflammatory M2 phenotype [67,68]. These metabolic and immunological changes in macrophages are in many ways opposite to the well-known proinflammatory M1 activation of macrophages induced by lipopolysaccharide (LPS) stimulation [69]. In the M1 phenotype, macrophages are known to rely primarily on glycolysis as an energy source, which has similarities to the Warburg effect that is characteristic of tumor cells [70]. Although glycolysis is a less efficient energy source, its volume can be rapidly increased, which can rapidly provide energy for the cell. Macrophages differentiated in the presence of butyrate show decreased glycolysis, increased adenosine monophosphate (AMP), increased AMP kinase (AMPK) phosphorylation, and inhibition of the mammalian target of rapamycin (mTOR) [43]. mTOR is a known positive regulator of several enzymes involved in glycolysis, including hexokinase II, glyceraldehyde 3-phosphate dehydrogenase and lactate dehydrogenase-B [71,72]. Inhibition of mTOR may explain the decrease in glycolysis in macrophages that were differentiated in the presence of butyrate [73]. In addition, butyrate-treated macrophages showed an increase in ribulose 5-phosphate, which against the background of decreased intracellular glucose levels indicates an increased flux into the pentose-phosphate pathway, which may contribute to NADPH formation [73]. This effect on cellular metabolism is of particular interest. Thus, by altering metabolic programming in alveolar macrophages, SCFAs may be involved in the regulation of lung immune tone [74]. In addition, by influencing metabolic pathways such as glycolysis enhancement and OXPHOS activity, SCFAs can participate in CD8+ T-cell activation [75]. By influencing cellular metabolism, butyrate has also been shown to promote memory potential in activated CD8+ T cells [76]. These data indicate the involvement of SCFAs in the regulation of distinct parts of both innate and adaptive immunity. Interestingly, butyrate can act differentially in its effect on cellular metabolism. Butyrate can stimulate the proliferation of normal colonocytes while inhibiting the proliferation of cancerous colonocytes, which are characterized by the Warburg effect in their metabolism. This differential effect is known as the “butyrate paradox” [77]. This effect is due to the fact that butyrate can act as an HDAC inhibitor in cancer cells but as a HAT activator in normal cells. Macrophages differentiated in the presence of butyrate show increased antimicrobial activity. This is due to the inhibition of histone deacetylase 3 (HDAC3), a shift in macrophage metabolism, decreased mTOR kinase activity, increased LC3-associated host defense and antimicrobial peptide production [73]. The anti-inflammatory effects of butyrate have also been linked to inhibition of the NF-kB signaling pathway as well as production by mononuclear cells and neutrophils of anti-inflammatory cytokines such as IL-10 [62,68,78,79]. In addition to neutrophils, butyrate stimulates the production of IL-10 and inhibits the production of IL -12 and interferon-γ (IFN-γ) by dendritic cells [80]. Another HDAC-related effect of butyrate is the inhibition of nitric oxide (NO) production (via inducible nitric oxide synthase (iNOS)) and proinflammatory cytokines (interleukin (IL)-6, IL-12, IL-1β) induced by LPS [68,79,81]. At the same time, low (0.5–2.5 mM) SCFAs concentrations increased and high (25–50 mM) concentrations decreased iNOS expression in the experiment [82]. Propionate and butyrate were also shown to reduce tumor necrosis factor alpha (TNF-α), cytokine-induced neutrophil chemoattractant-2 (CINC-2αβ), and NO production by LPS-stimulated neutrophils [78]. In addition to cytokines, the anti-inflammatory ability of SCFAs may also be related to the regulation of prostaglandin E2 (PGE2) production [83]. SCFAs such as acetate, propionate, and butyrate increased neutrophil migration by increasing L-selectin expression on neutrophils and releasing CINC-2 alphabeta [84]. At the same time, it has been shown that SCFAs can reduce excessive airway infiltration by neutrophils by reducing levels of C-X-C Motif Chemokine Ligand 1 (CXCL1) produced by lung monocytes and macrophages [75]. SCFAs are involved in regulating the differentiation, recruitment, and activation of neutrophils, dendritic cells, macrophages, and monocytes, as well as T cells [43,54]. Butyrate can affect the differentiation of dendritic cells derived from human monocytes by keeping dendritic cells in a stable, immature stage [80]. SCFAs inhibit the maturation of monocytes, macrophages, and dendritic cells by altering their ability to capture antigens and reducing their ability to produce proinflammatory cytokines such as IL-12 and TNF-α [43,54,62]. SCFAs are also involved in the regulation of inflammation in endothelial cells. Butyrate inhibited TNF-α-induced activation of NF-kB and enhanced the expression of peroxisome proliferator-activated receptor alpha (PPARα) in HUVECs [85]. In addition, butyrate reduced TNF-α-induced expression of vascular cell adhesion molecule 1 (VCAM-1) and Inter-Cellular Adhesion Molecule 1 (ICAM-1) mRNA in these cells [85]. In turn, propionate could also reduce cytokine-induced VCAM-1 and ICAM-1 expression by inhibiting NF-kB activation [86]. In another study, butyrate was shown to increase ICAM-1 and E-selectin expression in vascular endothelium, with a time-dependent effect [87]. It should be noted that ICAM-1 may play an important role in the pathogenesis of emphysema [88]. ICAM-1 is involved in the transendothelial migration of neutrophils during inflammation [89,90]. In this case, higher soluble ICAM-1 was associated with a more rapid progression of emphysema [88]. Thus, SCFAs have anti-inflammatory and immunomodulatory effects [91]. Other evidence suggests that SCFAs can have not only anti-inflammatory but also pro-inflammatory effects on lung cells, which depends on the concentration of SCFAs and the type of cells studied [92]. Interestingly, however, elevated concentrations of SCFAs can induce the release of interleukin IL-8, IL-6, and IL-1β either alone or in combination with toll-like receptor ligands TLR2 and TLR7 [93]. Thus, the accumulated data to date allow SCFAs to be considered as either pro- or anti-inflammatory molecules, depending on cell type as well as on conditions [54]. These and other data have strengthened the understanding of the importance of nutrition in the pathogenesis of COPD, which is based on chronic inflammation characterized by an imbalance between pro- and anti-inflammatory mechanisms. 3. Relationship between Gut Microbiota, Diet and Short-Chain Fatty Acid Production The intestinal microbial community is characterized by considerable diversity, which depends on many external and internal factors, including age and dietary habits. The intestinal microbiota is involved in maintaining the tone of the immune system. Studies in mice, show that a lack of microbiota causes developmental defects in many body systems, including the immune system [94]. Bacteroidetes (such as Bacteroides and Prevotella) and Firmicutes (such as Lactobacillus, Enterococcus, Clostridium and Bacillus) are thought to be common bacteria in the adult gut, whereas Actinobacteria (such as Bifidobacterium) and Proteobacteria (such as Escherichia) are lower [95]. Most Westerners are carriers of high levels of Bacteroides, whereas Prevotella is more common in non-Westerners who consume a plant-rich diet [96]. Bacteroidetes and Firmicutes are mainly localized in the proximal colon and are involved in SCFAs production [97,98,99]. And members of the Bacteroidetes type mainly produce acetate and propionate, while the Firmicutes type produces butyrate [29,100,101]. Given the emerging close relationship between the gut microbiome and lung function, there is increasing evidence of possible abnormalities in gut microflora structure in smoking and COPD [102,103,104,105]. Smokers have a higher proportion of Bacteroidetes and Prevotella and a lower proportion of Firmicutes and Proteobacteria in the gut microbiota compared to nonsmokers [106,107]. In another study, it was shown that stopping smoking led to an increase in microbial diversity and a significant change in gut microbial composition. At the phylum level, Firmicutes and Actinobacteria increased and the proportion of Bacteroidetes and Proteobacteria decreased [105]. In turn, chronic exposure to inhaled particulate matter, which is another important risk factor for COPD, causes gut dysbacteriosis and metabolic disorders in an experimental model in rats [108]. Chronic exposure to particulate matter for 24 weeks induced COPD-like pathological changes and lung inflammation in rats, which were accompanied by decreased abundance and diversity of the gut microbiota and decreased levels of SCFAs [108]. The species diversity of gut microflora and the number of Bacteroides and Bifidobacteria decreases in the elderly [109,110]. In addition, patients with chronic diseases have changes in gut microflora with an increase in a number of harmful bacteria [111]. Overweight and obesity were associated with a change in the ratio of individual SCFAs in favor of propionate [112]. This was consistent with a higher proportion of Bacteroidetes in the total gut bacterial structure in overweight and obese individuals than in lean volunteers [112]. At the same time, the total SCFAs concentration in fecal samples was more than 20% higher in obese than in lean volunteers [112]. In contrast, other studies have shown a reduction in the Bacteroidetes community in obese patients [113,114]. The gut microbiota is closely related to dietary habits. A comparative study of dietary patterns in African and European children showed that a diet high in fiber and low in animal protein was associated with a significant enrichment of Actinobacteria and Bacteroidetes and less Firmicutes and Proteobacteria. This corresponded to a greater production of short-chain fatty acids due to bacterial fermentation of plant fibers [115]. It should be noted that there are connections between the dietary precursors of SCFAs as well as the composition of plasma SCFAs [116,117,118]. This may be due to the fact that different compositions of the gut microflora may differentially utilize various sources of fermentable fiber. A diet rich in protein has a significant effect on SCFAs production. High-protein and moderate-carbohydrate and high-protein and low-carbohydrate diets have been shown to increase the proportion of branched-chain fatty acids and the concentration of phenylacetic acid and N-nitroso compounds after four weeks. The high-protein and low-carbohydrate diet also decreased the proportion of butyrate in short-chain fatty acid concentrations in feces [119]. This seems important given the need for nutritional support in COPD due to the skeletal muscle hypotrophy that accompanies this reduction in physical activity. A high-fat diet has been shown to reduce the number of Bacteroidetes [120,121,122]. The fecal microbiota of vegans was enriched for Verrucomicrobia phyla, while it was reduced for Proteobacteria phyla and lactic acid bacteria. And the vegan group had lower amounts of branched-chain fatty acids, such as iso-valerate and iso-butyrate, as well as acetate and propionate [97]. In turn, the consumption of vegetables, fruits, whole-grain cereals, and oily fish may help protect against worsening lung function in adults, especially in male smokers and COPD patients [123]. Among men, high fiber intake has been shown to be inversely related to the incidence of COPD in both current and former smokers [124]. In addition, the decrease in COPD incidence among men, both smokers and ex-smokers, was associated with high fruit and vegetable intake [125]. Interestingly, the reduced risk of COPD in women was associated with a prolonged consumption of fruit only, not vegetables [126]. It should be noted that these studies do not take into account the contribution of individual food components and micronutrients in providing these connections. Thus, the nature of nutrition may influence not only the structure of the gut microbiota but also the production of SCFAs, potentially affecting immune regulation, which is important in COPD. This reinforces the importance of additional studies that could improve the understanding of the role of SCFAs in the pathogenesis of COPD. In addition, an important focus of future research should be to investigate the causal relationship between diet, gut microbiome structure, SCFAs production, and COPD progression. 4. Clinical Significance of Short-Chain Fatty Acids in the Pathogenesis of COPD The presence of SCFAs in the sputum confirms the connection between the lungs and the intestine [82]. SCFAs are thought to contribute to the maintenance of lung immunometabolic tone [74]. As already noted, SCFAs can act on various parts of the innate immune defenses of the lungs. In addition to the described anti-inflammatory effects, the protective role of butyrate and propionate may include the restoration and maintenance of the barrier function of damaged airway epithelium by increasing the expression of ZO-1 dense contact proteins [127]. This seems important given the development of airway epithelial barrier dysfunction and impaired tight cell contacts in smoking and COPD [128]. In this regard, the regulation of barrier function under the influence of SCFAs may have some clinical significance [127]. The results showed differences in the composition of the gut microbiome in COPD and in healthy individuals. This is consistent with lower overall levels of SCFAs in the stage III-IV COPD group than in patients with stage I-II COPD and in healthy individuals [129]. Experimental data indicate that transplantation of fecal microbiota to mice led to inflammation in the lungs and increased IL-1β and TNF-α in plasma. At the same time, additional exposure to smoke from the biomass led to an accelerated decrease in lung function, emphysematous changes, airway remodeling, and mucus hypersecretion. These changes were accompanied by higher levels of claudin 1, α smooth-muscle actin (α-SMA), neutrophil elastase (NE) and matrix metalloproteinase 2 (MMP-2), and MUC5AC. In addition, gut microbiota obtained from stage III-IV COPD patients decreased body weight in recipient mice [129]. It should be noted that another recent study did not show an association of gut microbial diversity with the severity of COPD [130]. Thus, SCFAs are an important part of the immunological axis linking the gut microbiota and the lungs, which has implications in the pathogenesis of COPD (Figure 3). In another study, patients with COPD with a stable moderately severe course of the disease had increased the total concentration of SCFAs in exhaled breath condensate [131]. Another study involving 38 COPD patients showed a decrease in acetate and an increase in the relative content of propionate, butyrate and branched-chain SCFAs in sputum and feces [132]. Patients with asthma showed a significant decrease in the total content of SCFAs in the feces, changes in the absolute concentrations of individual acids: acetate, propionate, butyrate, and changes in the total content of short-chain branched-chain fatty acids were also detected [133]. Different variations in the composition of SCFAs were shown, but changes in the metabolic profile were independent of the disease phenotype [133]. These and other data suggest that SCFA levels may be related to the pathogenesis of COPD, but the details of these relationships require clarification. 5. Clinical Significance of Short-Chain Fatty Acids in the Pathogenesis of COPD COPD is a chronic respiratory disease characterized by progressive airflow limitation [134]. The rate of decline in lung function is individual and can vary widely from person to person. It should be noted that individual trajectory of COPD progression includes not only airflow limitation [135,136,137]. It is of great interest to phenotype patients according to the peculiarities of the clinical picture [138,139]. It should be noted that, to date, discussions about COPD phenotypes have not led to a generally accepted understanding of clinical variants, the separation of which would help to improve approaches to the management of patients [140,141]. The concept of COPD phenotypes has deep historical roots and goes back to the recognition of the two main components of the disease, such as emphysema and chronic bronchitis. They were described long before the term COPD itself appeared and were well known to clinicians. 5.1. Role of Short-Chain Fatty Acids in the Development of Emphysema Emphysema is an important clinical phenotype of COPD. Numerous studies are devoted to the analysis of pathogenetic mechanisms of its development, among which of interest are the works studying the relationship of emphysema with nutrition. It is important to know about the development of emphysema during prolonged starvation, which was shown in the study of Warsaw ghetto prisoners, as well as observations in patients with anorexia nervosa [142,143]. These data, reinforced by experimental animal models, confirmed the link between nutrient deficiency and alveolar tissue destruction that leads to emphysema [144]. There are various explanations that justify these connections. Interestingly, anorexia nervosa is characterized by a decrease in intestinal microbial diversity, which is associated with the production of SCFAs, including butyrate and propionate [145,146]. A high-fiber diet has been shown to help prevent the progression of emphysema caused by cigarette smoke exposure. This is because a high-fiber diet changes the composition of the gut microbial community, resulting in increased production of SCFAs (acetate, propionate, and butyrate). SCFAs attenuated the pathological changes associated with the progression of emphysema and influenced the inflammatory response caused by cigarette smoke exposure [147]. Mice with emphysema that received fermentable fiber (pectin) were shown to have less inflammation than mice with emphysema that received nonfermentable fiber such as cellulose. This corresponded to lower concentrations of acetate, propionate, and butyrate in the emphysema group compared with the other groups. At the same time, macrophage and neutrophil counts were lower in the high fiber (cellulose and pectin) group than in the emphysema group [147]. In addition, inflammatory cytokine levels were lower in the high-fiber (cellulose and pectin) diet. In another study, butyrate was shown to reduce hypoxia-induced accumulation of alveolar (mainly CD68+) and interstitial (CD68+ and CD163+) macrophages in rat lungs [148]. The key events leading to emphysema are considered to be interalveolar septal avascularization, which is associated with the apoptosis of endothelial cells and alveolar epithelial cells [149]. In this regard, the information about butyrate participation in endothelial cell function is of interest. The anti-inflammatory properties of butyrate are in part related to its effect on the activation of NF-kB and PPARalpha and the associated expression of VCAM-1 and ICAM-1 [85,148]. In addition, butyrate inhibited the activation of endothelial NRLP3 inflammasome in endothelial cells [150]. Butyrate has also been shown to be involved in the inhibition of angiogenesis through HIF-1α as well as through the inhibition of vascular endothelial growth factor (VEGF) and cyclooxygenase-2 (COX-2) [148,151,152]. Butyrate has also been shown to enhance the expression of tight junction proteins in pulmonary microvascular endothelial cells, affecting endothelial barrier function [148]. The effects of butyrate in endothelial cells are also related to NO production [153]. It has been shown that butyrate and acetate, can improve AngII-induced endothelial dysfunction through increased NO bioavailability [154]. The reduction of reactive oxygen species (ROS) levels in the vascular wall and the subsequent prevention of NO inactivation represent a key mechanism through which SCFAs act on endothelial function [154]. These findings are of interest given the frequent association of COPD with atherosclerosis. Thus, the production of SCFAs by the intestinal microbiota may be one possible mechanism in the prevention of emphysema. It was found that a serum peptide-based enteral diet can suppress elastase-induced emphysema in mice by altering SCFAs levels in the cecum [155]. Modulation of the gut microbiota by prebiotics and transplantation of fecal microbiota from a high-fiber diet altered the composition of the gut microbiota, attenuating smoking-induced emphysema [156]. This was associated with a decrease in local and systemic inflammation through production of SCFAs, which protect against alveolar destruction and cellular apoptosis. The levels of IL-6 and IFN-γ in bronchoalveolar lavage fluid (BALF) were lowest in the high-fiber diet group, and the local concentration of SCFAs was markedly higher in mice with emphysema following fecal microbiota transplantation and the high-fiber diet than in mice with emphysema [156]. And the high-fiber diet had a more protective effect against emphysema than the high-protein diet [156]. Experimental data showed that Firmicutes species were most dominant in samples from the feces of emphysematous mice. Transplantation of fecal microbiota from a high-fiber diet resulted in increased abundance of Bacteroides phylum, decreasing the Firmicutes/Bacteroides ratio [156]. Thus, the development of emphysema, a key phenotype of COPD, may to some extent be related to the role of SCFAs in the maintenance of lung immune function. It should be noted that the pathophysiology of emphysema includes many known mechanisms unrelated to SCFAs, many of whose links are currently the subject of research. That said, the possible role of SCFAs in the pathophysiology of emphysema in COPD is a promising area for future research that could provide answers to some questions related to the clinical efficacy of nutritional support for patients. 5.2. Role of Short-Chain Fatty Acids in COPD Exacerbations An important characteristic of the natural course of COPD is the frequency and severity of exacerbations. COPD exacerbations make a significant contribution to the clinical picture of COPD. The results of numerous studies suggest that the frequency of exacerbations is associated with a more rapid decrease in FEV1 and an unfavorable prognosis [157,158,159]. Given these facts, some authors propose to consider high exacerbation frequency as a separate phenotype [158,160,161]. Infectious exacerbations of COPD are associated with disturbances in the structure of microbiota in the bronchi [162]. Colonization of the bronchi by microorganisms is necessary to maintain the immunological tone of the lungs. The available data suggest certain links between the intestinal and lung microbiome [163]. The respiratory tract microbiome can be supplemented with microorganisms from the gastrointestinal tract, which is important. It is important to note that diet can affect not only the gut microflora, but also the respiratory tract microbiota [163,164]. SCFAs can have a direct effect on microorganisms as well as affecting their virulence [165]. Interestingly, high concentrations of SCFAs caused significant inhibition of Pseudomonas aeruginosa growth, which was enhanced at lower pH. At the same time, low concentrations of SCFAs resulted in enhanced bacterial growth [82]. Meanwhile, the administration of prebiotics in the form of oligosaccharides can modulate the immune and inflammatory response and outcome of pulmonary Pseudomonas aeruginosa infection in C57BL/6 mice through effects on the gut microbiota [166]. In addition, the structure of the gut and lung microbiota was shown to be dynamic, with changes associated with exacerbations. [167]. COPD exacerbations were characterized by a decrease in the relative content of Firmicutes and Actinobacteria, and an increase in Bacteroidetes and Proteobacteria compared with stable COPD and nonsmokers [168]. It should be noted that antibiotic therapy used to treat COPD exacerbations can have a serious impact on the structure of the gut microbiome and have long-term implications for metabolic and immunologic health [169]. In addition, macrolide use in childhood has also been shown to be associated with long-term changes in gut microbiota composition and function. These changes are associated with metabolic disease and obesity, and may also affect the development of the immune system, leading to respiratory hypersensitivity [169,170]. 5.3. Role of Short-Chain Fatty Acids in the Decline of Lung Function The rate at which lung function decreases is important in assessing the prognosis of patients with COPD. Given that bronchial obstruction is irreversible, progressive decline in lung function is associated with multiple systemic effects and the increased severity of comorbid conditions. Rapid progressive decline in lung function in COPD is suggested by some authors to be a separate phenotype, given its relationship with the prognosis of the disease. Moreover, the rate of decline in lung function is associated with many factors, including the frequency of exacerbations. Interestingly, nutrition can also influence lung function. The high intake of sweets, oils, fat, and coffee has been shown to be negatively associated with lung function, including FEV1/FVC, and has been associated with an increased prevalence of COPD in men [20,171,172,173,174]. In contrast, high intake of fruits and vegetables, fatty fish, and low-fat foods was negatively associated with a diagnosis of COPD [175,176,177]. Interestingly, however, the rate of decline in lung function may be related to impaired biodiversity of the gut microflora. It has been shown that a rate of decrease in FEV1 of more than 40 mL/year corresponded to a greater decrease in bacterial diversity in the gut [178]. In a one-year follow-up, it was shown that at the phylum level, Firmicutes were more abundant in the group with decreased lung function, whereas Bacteroidetes were more abundant when lung function was virtually unreduced [178]. This corresponded to a more unstable gut microbiota in the reduced group than in the control group. The relative abundance of Prevotella was also shown to decrease after 1 year in the reduction group, which may indicate a subtype of COPD in which Prevotella abundance may be related to the rate of decline in lung function [178]. It has been previously noted that Prevotella may be associated with inflammatory disease [96]. Another study showed a link between differences in the bacterial community structure of the gut and sensitization to aeroallergens and lung function (FEV1) in asthma [179]. These data are of interest given the presence of a Asthma-COPD overlap syndrome (ACOS), which is important because it has some diagnostic difficulties in real clinical practice. Consumption of a single dose of soluble fiber (3.5 g of inulin) in patients with asthma was shown to result in a significant reduction in airway inflammation and improvement in lung function (FEV1 and FEV1/FVC). This effect was characterized by a decrease in the total number of sputum cells, including neutrophils, macrophages, and lymphocytes, and a decrease in IL-8 in sputum and exhaled NO. This was consistent with the significantly increased expression of GPR41 and GPR43 genes in sputum cells [180]. Increased circulating levels of SCFAs in mice fed a high-fiber diet protected against allergic inflammation in the lungs. In contrast, a low-fiber diet increased allergic airway disease by reducing SCFAs levels. In addition, elevated levels of circulating SCFAs were found to protect against allergic inflammation in the lungs. Propionate, by affecting the maturation of lung dendritic cells, promoted their high phagocytic capacity but impaired their ability to stimulate the effector function of T helper type 2 (TH2) cells [164]. 5.4. Short-Chain Fatty Acids, Body Weight, and Physical Frailty Phenotype The role of diet as an important factor modifying the course of COPD has been the subject of numerous studies. Many of these studies have focused on the role of polyunsaturated fatty acids (PUFAs), especially ω-3 fatty acids [181,182]. Their influence on disease progression and prognosis has been analyzed, which is related to the involvement of both PUFAs themselves and their metabolites in the regulation of inflammation and resolution of inflammation. Analysis of the role of dietary fiber is another important area that has been shown to be clearly related to the course of COPD. A population-based prospective cohort of 35 339 Swedish women evaluated the association between baseline and long-term dietary fiber intake and COPD risk. The results of this study showed that high fiber intake is an important modifiable factor that may reduce the risk of COPD primarily in current and former smokers [183]. The relationship between SCFAs production and body weight is of interest, given its relevance to the pattern of COPD course. Fecal SCFAs concentrations were shown to be higher in overweight people than in lean people (80±6 vs 56±6 mmol/kg, p = 0.02). It was also found that overweight individuals may absorb more SCFA from the large intestine [184]. Another study showed that overweight or obese individuals had higher levels of fecal acetate, propionate, butyrate, and valerate compared to lean subjects [185]. GPR43 is known to play an important role in white adipose tissue (WAT) [186]. In an experiment in mice, SCFAs have been shown to suppress insulin-mediated fat accumulation through GPR43 activation. In mice attempting a high-fat diet, SCFAs levels in feces and plasma acetate were decreased, while GPR43 expression in the WAT was markedly higher compared to mice receiving a normal diet [186]. This may be part of the mechanism of energy balance regulation, which includes suppression of excess energy accumulation and increased fat consumption. In this mechanism, GPR43 may function as an energy sensor that promotes the use of excess energy in other tissues rather than to store it as fat in adipose tissue [186]. The physical frailty phenotype is less well known, but its clinical and prognostic significance is beyond doubt [141,187,188]. Physical frailty is a complex syndrome characterized by loss of physiological and cognitive reserve [189,190]. Impaired muscle metabolism and depletion of muscle mass is an important clinical characteristic of COPD, especially in older individuals. Reduced physical activity due to physical frailty is suggested to be considered as a COPD phenotype associated with unfavorable prognosis. Of considerable interest is the information that the species diversity of the fecal microbiota of the elderly is inversely related to physical performance and the Rockwood clinical frailty scale [191,192]. The composition of the gut microbiota has been shown to be closely related to the development of physical frailty in older adults [191]. A key link that may link gut microbiota structure to skeletal muscle function is SCFAs [193,194]. SCFAs can act as ligands for FFAR2 and FFAR3 in skeletal muscle cells, regulating metabolic pathways related to glucose uptake and metabolism as well as modulating mitochondrial biogenesis [195]. Through the inhibition of histone deacetylase, butyrate can lead to the prevention of apoptosis in muscle. Butyrate can affect muscle metabolism, including improving glucose metabolism and increasing enzymes involved in oxidative metabolism, thus preventing age-related muscle atrophy in mice [196]. Butyrate has been shown to increase the muscle fiber cross-sectional area and reduce intramuscular fat accumulation in older mice [196]. Thus, SCFAs may be involved in numerous extrapulmonary effects related to effects on metabolism in tissues, which is of great research and clinical interest. 5.5. Short-Chain Fatty Acids, the Central Nervous System, and the COPD Emotional Fragility Phenotype Emotional lability is increasingly recognized as a phenotype of COPD because of its significant impact on treatment efficacy. Anxiety and depression are known to be associated with decreased quality of life, as well as increased hospitalizations and mortality. The data accumulated to date have strengthened our understanding of the links between COPD and these disorders. Of great interest is the evidence of the influence of gut microbiota on the central nervous system through SCFAS, due to their several neuroactive properties. The exact mechanisms of these connections are still a subject for research. SCFAs have been shown to affect several neurological and mental diseases and behavioral processes. Their involvement in neuronal development, microglia maturation and the release of synaptic neurotransmitters is also known [43,197,198,199,200]. The involvement of SCFAs in neuroimmune processes in neurodegenerative diseases is important [200]. Short-chain fatty acids are involved in the onset of depression, which has been shown in macaques [201]. And it was found that in addition to plasma concentrations, some SCFAs (acetic acid, propanedioic acid, and butyric acid) are also impaired in macaque cerebrospinal fluid in a natural depression model. Butyrate levels differ significantly in both serum and liquor samples from these macaques [201]. These results are consistent with evidence of lower fecal SCFAs concentrations in depressed patients than in controls [202]. Another study showed that acetate production by the rodent intestinal microbiota on a high-fat diet leads to effects on the central nervous system. This is due to activation of the parasympathetic nervous system, resulting in the increased secretion of ghrelin, which promotes hyperphagia, and increased energy deposition in the form of fat due to increased glucose-stimulated insulin secretion [203]. At the same time, FFAR3, which is activated mainly by propionate and butyrate, regulates sympathetic activity by sensing nutritional status, thereby maintaining the body’s energy homeostasis [38]. In addition, butyrate has an antidepressant-like effect in mouse models and also improves cognitive abilities in rats [204,205,206]. These data are of clinical interest and are an important topic for further research. Thus, in accordance with modern concepts, COPD is considered to be not only as a lung disease, but also from the position of its numerous systemic effects. Moreover, SCFAs may be involved in the regulation of many links related to the development and progression of COPD. In this regard, the complex influence of SCFAs on various pulmonary and extrapulmonary effects of COPD can be assumed. 6. Conclusion COPD is a clinically heterogeneous disease characterized by the development and progressive restriction of airflow due to chronic inflammation in the bronchi. Despite the known etiological factor, cigarette smoking, many aspects of COPD pathogenesis are still unknown. The causes of heterogeneity of the COPD course, which is associated with individual trajectories of progression and prognosis, remain a subject of discussion. Nutrition has a marked effect on the course of the disease. Indeed, patients who are severely underweight are at greatest risk for adverse COPD outcomes. The development of emphysema is also related to dietary patterns. Non-digestible carbohydrates, such as fiber, have been shown to favorably influence prognosis. Although these fibers are inaccessible to human digestive enzymes, they are actively metabolized by intestinal microflora. The products of enzymatic activity are SCFAs, which are an important link in the gut-lung immune axis. Short-chain fatty acids demonstrate a variety of functions in the regulation of inflammation and may play an important role in the clinical picture of COPD. SCFA production depends on the nature of the diet and the structure of the microflora. In this regard, dietary modification to include non-digestible fibers in the diet is seen as an important therapeutic tool that can affect not only the course of the disease, but also its outcome. It should be noted that many of the effects of a diet containing dietary fiber may be related not only to the microbial fermentation of this fiber in the gut and SCFAs production, but also to various other dietary components, including micronutrients and vitamins. These findings are reflected in numerous studies that emphasize the importance of individual nutritional components [207,208,209]. Given the contribution of other nutritional components, it would not be very correct to link the clinical features of the course of COPD solely with SCFAs. It can be stated that the problems of diet in the natural history of COPD are far from being solved and require new research. In addition, individual trajectories of the natural history of COPD are shaped by many external and internal factors, a simplified understanding of which will not contribute to the interpretation of research results and improve approaches to the management of patients. Indeed, many questions concerning both the pathophysiology of COPD and the involvement of SCFAs in these processes remain unanswered to date. Importantly, COPD patients are not a clinically homogeneous group, which requires a differentiated approach in the evaluation of research findings. The molecular mechanisms exhibited by SCFAs require new experimental and clinical confirmations. In this regard, investigation of the role of SCFAs in the clinically heterogeneous course of COPD may be a promising area for future research. These data will help to broaden the understanding of the pathophysiological mechanisms and their impairments that are associated with COPD phenotypes. A better understanding of these mechanisms will enhance the development of more effective therapeutic intervention programs that will be better adapted to individual disease course trajectories. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare that they have no conflicts of interest. Figure 1 Short-chain fatty acids formation pathways. Figure 2 Scheme of the involvement of short-chain fatty acids in cell function and inflammation. Figure 3 The participation of short-chain fatty acids in the pathogenesis of the clinically heterogeneous course of COPD. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Chronic Obstructive Pulmonary Disease (COPD) Available online: https://www.who.int/news-room/fact-sheets/detail/chronic-obstructive-pulmonary-disease-(copd) (accessed on 25 February 2022) 2. Soriano J.B. Kendrick P.J. Paulson K.R. Gupta V. Abrams E.M. Adedoyin R.A. Adhikari T.B. Advani S.M. Agrawal A. Ahmadian E. Prevalence and attributable health burden of chronic respiratory diseases, 1990–2017: A systematic analysis for the Global Burden of Disease Study 2017 Lancet Respir. Med. 2020 8 585 596 10.1016/S2213-2600(20)30105-3 32526187 3. Çolak Y. Afzal S. Nordestgaard B.G. Vestbo J. Lange P. Prevalence, Characteristics, and Prognosis of Early Chronic Obstructive Pulmonary Disease. The Copenhagen General Population Study Am. J. Respir. Crit. Care Med. 2020 201 671 680 10.1164/rccm.201908-1644OC 31770495 4. Hashemi S.Y. Momenabadi V. Faramarzi A. Kiani A. Trends in burden of chronic obstructive pulmonary disease in Iran, 1995–2015: Findings from the global burden of disease study Arch. Public Health 2020 78 45 10.1186/s13690-020-00426-x 32509302 5. Blanco I. Diego I. Bueno P. Casas-Maldonado F. Miravitlles M. Geographic distribution of COPD prevalence in the world displayed by Geographic Information System maps Eur. Respir. J. 2019 54 1900610 10.1183/13993003.00610-2019 31000678 6. Kotaki K. Ikeda H. Fukuda T. Yuhei K. Yuki F. Kawasaki M. Wakamatsu K. Sugahara K. Trends in the prevalence of COPD in elderly individuals in an air-polluted city in Japan: A cross-sectional study Int. J. Chronic Obstr. Pulm. Dis. 2019 14 791 798 10.2147/COPD.S189372 7. Ni L. Chuang C.-C. Zuo L. Fine particulate matter in acute exacerbation of COPD Front. Physiol. 2015 6 10.3389/fphys.2015.00294 8. Ding Q. Li J. Xu S. Gao Y. Guo Y. Xie B. Li H. Wei X. Different Smoking Statuses on Survival and Emphysema in Patients with Acute Exacerbation of Chronic Obstructive Pulmonary Disease Int. J. Chron. Obstruct. Pulmon. Dis. 2022 17 505 515 10.2147/COPD.S346456 35281478 9. Wang Y. Xu J. Meng Y. Adcock I.M. Yao X. Role of inflammatory cells in airway remodeling in COPD Int. J. Chron. Obstruct. Pulmon. Dis. 2018 13 3341 3348 10.2147/COPD.S176122 30349237 10. Yang A. Wu Y. Yu G. Wang H. Role of specialized pro-resolving lipid mediators in pulmonary inflammation diseases: Mechanisms and development Respir. Res. 2021 22 204 10.1186/s12931-021-01792-y 34261470 11. Agustí A. Vogelmeier C. Faner R. COPD 2020: Changes and challenges Am. J. Physiol.-Lung Cell. Mol. Physiol. 2020 319 L879 L883 10.1152/ajplung.00429.2020 32964724 12. Agusti A. Calverley P.M. Celli B. Coxson H.O. Edwards L.D. Lomas D.A. MacNee W. Miller B.E. Rennard S. Silverman E.K. Characterisation of COPD heterogeneity in the ECLIPSE cohort Respir. Res. 2010 11 122 10.1186/1465-9921-11-122 20831787 13. Kim E.K. Singh D. Park J.H. Park Y.B. Kim S.I. Park B. Park J. Kim J. Kim M.A. Lee J.H. Impact of Body Mass Index Change on the Prognosis of Chronic Obstructive Pulmonary Disease Respiration 2020 99 943 953 10.1159/000511022 33264797 14. Guo Y. Zhang T. Wang Z. Yu F. Xu Q. Guo W. Wu C. He J. Body mass index and mortality in chronic obstructive pulmonary disease: A dose-response meta-analysis Medicine 2016 95 e4225 10.1097/MD.0000000000004225 27428228 15. Iyer A.S. Dransfield M.T. The “Obesity Paradox” in Chronic Obstructive Pulmonary Disease: Can It Be Resolved? Ann. Am. Thorac. Soc. 2018 15 158 159 10.1513/AnnalsATS.201711-901ED 29388820 16. Spelta F. Fratta Pasini A.M. Cazzoletti L. Ferrari M. Body weight and mortality in COPD: Focus on the obesity paradox Eat. Weight. Disord.-Stud. Anorex. Bulim. Obes. 2018 23 15 22 10.1007/s40519-017-0456-z 17. Collins P.F. Stratton R.J. Kurukulaaratchy R. Elia M. S163 The ‘Obesity Paradox’ in chronic obstructive pulmonary disease Thorax 2010 65 A73 A74 10.1136/thx.2010.150953.14 18. Barreiro E. Jaitovich A. Muscle atrophy in chronic obstructive pulmonary disease: Molecular basis and potential therapeutic targets J. Thorac. Dis. 2018 10 S1415 S1424 10.21037/jtd.2018.04.168 29928523 19. Wüst R.C.I. Degens H. Factors contributing to muscle wasting and dysfunction in COPD patients Int. J. Chronic Obstr. Pulm. Dis. 2007 2 289 300 20. Scoditti E. Massaro M. Garbarino S. Toraldo D.M. Role of Diet in Chronic Obstructive Pulmonary Disease Prevention and Treatment Nutrients 2019 11 1357 10.3390/nu11061357 21. Kotlyarov S. Kotlyarova A. Anti-Inflammatory Function of Fatty Acids and Involvement of Their Metabolites in the Resolution of Inflammation in Chronic Obstructive Pulmonary Disease Int. J. Mol. Sci. 2021 22 12803 10.3390/ijms222312803 34884621 22. Enaud R. Prevel R. Ciarlo E. Beaufils F. Wieërs G. Guery B. Delhaes L. The Gut-Lung Axis in Health and Respiratory Diseases: A Place for Inter-Organ and Inter-Kingdom Crosstalks Front. Cell. Infect. Microbiol. 2020 10 9 10.3389/fcimb.2020.00009 32140452 23. Anand S. Mande S.S. Diet, Microbiota and Gut-Lung Connection Front. Microbiol. 2018 9 2147 10.3389/fmicb.2018.02147 30283410 24. Wang B. Yao M. Lv L. Ling Z. Li L. The Human Microbiota in Health and Disease Engineering 2017 3 71 82 10.1016/J.ENG.2017.01.008 25. Vaughan A. Frazer Z.A. Hansbro P.M. Yang I.A. COPD and the gut-lung axis: The therapeutic potential of fibre J. Thorac. Dis. 2019 11 S2173 S2180 10.21037/jtd.2019.10.40 31737344 26. Blaak E.E. Canfora E.E. Theis S. Frost G. Groen A.K. Mithieux G. Nauta A. Scott K. Stahl B. van Harsselaar J. Short chain fatty acids in human gut and metabolic health Benef. Microbes 2020 11 411 455 10.3920/BM2020.0057 32865024 27. Wong J.M.W. de Souza R. Kendall C.W.C. Emam A. Jenkins D.J.A. Colonic Health: Fermentation and Short Chain Fatty Acids J. Clin. Gastroenterol. 2006 40 235 243 10.1097/00004836-200603000-00015 16633129 28. Parada Venegas D. De la Fuente M.K. Landskron G. González M.J. Quera R. Dijkstra G. Harmsen H.J.M. Faber K.N. Hermoso M.A. Short Chain Fatty Acids (SCFAs)-Mediated Gut Epithelial and Immune Regulation and Its Relevance for Inflammatory Bowel Diseases Front. Immunol. 2019 10 277 10.3389/fimmu.2019.00277 30915065 29. den Besten G. van Eunen K. Groen A.K. Venema K. Reijngoud D.-J. Bakker B.M. The role of short-chain fatty acids in the interplay between diet, gut microbiota, and host energy metabolism J. Lipid Res. 2013 54 2325 2340 10.1194/jlr.R036012 23821742 30. Morrison D.J. Preston T. Formation of short chain fatty acids by the gut microbiota and their impact on human metabolism Gut Microbes 2016 7 189 200 10.1080/19490976.2015.1134082 26963409 31. Louis P. Flint H.J. Formation of propionate and butyrate by the human colonic microbiota Environ. Microbiol. 2017 19 29 41 10.1111/1462-2920.13589 27928878 32. Deleu S. Machiels K. Raes J. Verbeke K. Vermeire S. Short chain fatty acids and its producing organisms: An overlooked therapy for IBD? eBioMedicine 2021 66 103293 10.1016/j.ebiom.2021.103293 33813134 33. Duncan S.H. Holtrop G. Lobley G.E. Calder A.G. Stewart C.S. Flint H.J. Contribution of acetate to butyrate formation by human faecal bacteria Br. J. Nutr. 2004 91 915 923 10.1079/BJN20041150 15182395 34. Smith E.A. Macfarlane G.T. Dissimilatory amino Acid metabolism in human colonic bacteria Anaerobe 1997 3 327 337 10.1006/anae.1997.0121 16887608 35. Dai Z.L. Wu G. Zhu W.Y. Amino acid metabolism in intestinal bacteria: Links between gut ecology and host health Front. Biosci. 2011 16 1768 1786 10.2741/3820 36. Mortensen P.B. Clausen M.R. Short-chain fatty acids in the human colon: Relation to gastrointestinal health and disease Scand. J. Gastroenterol. Suppl. 1996 216 132 148 10.3109/00365529609094568 8726286 37. Cummings J.H. Pomare E.W. Branch W.J. Naylor C.P. Macfarlane G.T. Short chain fatty acids in human large intestine, portal, hepatic and venous blood Gut 1987 28 1221 1227 10.1136/gut.28.10.1221 3678950 38. Kasubuchi M. Hasegawa S. Hiramatsu T. Ichimura A. Kimura I. Dietary gut microbial metabolites, short-chain fatty acids, and host metabolic regulation Nutrients 2015 7 2839 2849 10.3390/nu7042839 25875123 39. Dong J.P. Zheng Y. Wu T. He Q. Teng G.G. Wang H.H. Protective effect of Saccharomyces boulardii on intestinal mucosal barrier of dextran sodium sulfate-induced colitis in mice Chin. Med. J. 2019 132 1951 1958 10.1097/CM9.0000000000000364 31335471 40. Roediger W.E. Utilization of nutrients by isolated epithelial cells of the rat colon Gastroenterology 1982 83 424 429 10.1016/S0016-5085(82)80339-9 7084619 41. Roy C.C. Kien C.L. Bouthillier L. Levy E. Short-chain fatty acids: Ready for prime time? Nutr. Clin. Pract. 2006 21 351 366 10.1177/0115426506021004351 16870803 42. Wiltrout D.W. Satter L.D. Contribution of propionate to glucose synthesis in the lactating and nonlactating cow J. Dairy Sci. 1972 55 307 317 10.3168/jds.S0022-0302(72)85487-0 5062250 43. Dalile B. Van Oudenhove L. Vervliet B. Verbeke K. The role of short-chain fatty acids in microbiota–gut–brain communication Nat. Rev. Gastroenterol. Hepatol. 2019 16 461 478 10.1038/s41575-019-0157-3 31123355 44. Layden B.T. Angueira A.R. Brodsky M. Durai V. Lowe W.L. Jr. Short chain fatty acids and their receptors: New metabolic targets Transl. Res. 2013 161 131 140 10.1016/j.trsl.2012.10.007 23146568 45. Yamashita H. Kaneyuki T. Tagawa K. Production of acetate in the liver and its utilization in peripheral tissues Biochim. Et Biophys. Acta (BBA)-Mol. Cell Biol. Lipids 2001 1532 79 87 10.1016/S1388-1981(01)00117-2 46. Bell-Parikh L.C. Guengerich F.P. Kinetics of cytochrome P450 2E1-catalyzed oxidation of ethanol to acetic acid via acetaldehyde J. Biol. Chem. 1999 274 23833 23840 10.1074/jbc.274.34.23833 10446146 47. Sturm E.M. Knuplez E. Marsche G. Role of Short Chain Fatty Acids and Apolipoproteins in the Regulation of Eosinophilia-Associated Diseases Int. J. Mol. Sci. 2021 22 4377 10.3390/ijms22094377 33922158 48. Le Poul E. Loison C. Struyf S. Springael J.-Y. Lannoy V. Decobecq M.-E. Brezillon S. Dupriez V. Vassart G. Van Damme J. Functional characterization of human receptors for short chain fatty acids and their role in polymorphonuclear cell activation J. Biol. Chem. 2003 278 25481 25489 10.1074/jbc.M301403200 12711604 49. Nilsson N.E. Kotarsky K. Owman C. Olde B. Identification of a free fatty acid receptor, FFA2R, expressed on leukocytes and activated by short-chain fatty acids Biochem. Biophys. Res. Commun. 2003 303 1047 1052 10.1016/S0006-291X(03)00488-1 12684041 50. Ulven T. Short-chain free fatty acid receptors FFA2/GPR43 and FFA3/GPR41 as new potential therapeutic targets Front. Endocrinol. 2012 3 111 10.3389/fendo.2012.00111 51. Bolognini D. Tobin A.B. Milligan G. Moss C.E. The Pharmacology and Function of Receptors for Short-Chain Fatty Acids Mol. Pharmacol. 2016 89 388 398 10.1124/mol.115.102301 26719580 52. Milligan G. Stoddart L.A. Smith N.J. Agonism and allosterism: The pharmacology of the free fatty acid receptors FFA2 and FFA3 Br. J. Pharmacol. 2009 158 146 153 10.1111/j.1476-5381.2009.00421.x 19719777 53. Maslowski K.M. Vieira A.T. Ng A. Kranich J. Sierro F. Yu D. Schilter H.C. Rolph M.S. Mackay F. Artis D. Regulation of inflammatory responses by gut microbiota and chemoattractant receptor GPR43 Nature 2009 461 1282 1286 10.1038/nature08530 19865172 54. Corrêa-Oliveira R. Fachi J.L. Vieira A. Sato F.T. Vinolo M.A.R. Regulation of immune cell function by short-chain fatty acids Clin. Transl. Immunol. 2016 5 e73 10.1038/cti.2016.17 27195116 55. Pluznick J. A novel SCFA receptor, the microbiota, and blood pressure regulation Gut Microbes 2014 5 202 207 10.4161/gmic.27492 24429443 56. Thangaraju M. Cresci G.A. Liu K. Ananth S. Gnanaprakasam J.P. Browning D.D. Mellinger J.D. Smith S.B. Digby G.J. Lambert N.A. GPR109A is a G-protein-coupled receptor for the bacterial fermentation product butyrate and functions as a tumor suppressor in colon Cancer Res. 2009 69 2826 2832 10.1158/0008-5472.CAN-08-4466 19276343 57. Kostylina G. Simon D.M. Fey M.F. Yousefi S. Simon H.-U. Neutrophil apoptosis mediated by nicotinic acid receptors (GPR109A) Cell Death Differ. 2007 15 134 142 10.1038/sj.cdd.4402238 17932499 58. Li M. Van Esch B.C.A.M. Wagenaar G.T.M. Garssen J. Folkerts G. Henricks P.A.J. Pro- and anti-inflammatory effects of short chain fatty acids on immune and endothelial cells Eur. J. Pharmacol. 2018 831 52 59 10.1016/j.ejphar.2018.05.003 29750914 59. Carretta M.D. Quiroga J. López R. Hidalgo M.A. Burgos R.A. Participation of Short-Chain Fatty Acids and Their Receptors in Gut Inflammation and Colon Cancer Front. Physiol. 2021 12 662739 10.3389/fphys.2021.662739 33897470 60. Bolden J.E. Peart M.J. Johnstone R.W. Anticancer activities of histone deacetylase inhibitors Nat. Rev. Drug Discov. 2006 5 769 784 10.1038/nrd2133 16955068 61. Park J. Kim M. Kang S.G. Jannasch A.H. Cooper B. Patterson J. Kim C.H. Short-chain fatty acids induce both effector and regulatory T cells by suppression of histone deacetylases and regulation of the mTOR–S6K pathway Mucosal. Immunol. 2015 8 80 93 10.1038/mi.2014.44 24917457 62. Chang P.V. Hao L. Offermanns S. Medzhitov R. The microbial metabolite butyrate regulates intestinal macrophage function via histone deacetylase inhibition Proc. Natl. Acad. Sci. USA 2014 111 2247 2252 10.1073/pnas.1322269111 24390544 63. Lin K.T. Wang Y.W. Chen C.T. Ho C.M. Su W.H. Jou Y.S. HDAC inhibitors augmented cell migration and metastasis through induction of PKCs leading to identification of low toxicity modalities for combination cancer therapy Clin. Cancer Res. 2012 18 4691 4701 10.1158/1078-0432.CCR-12-0633 22811583 64. Xu Z. Tao J. Chen P. Chen L. Sharma S. Wang G. Dong Q. Sodium butyrate inhibits colorectal cancer cell migration by downregulating Bmi-1 through enhanced miR-200c expression Mol. Nutr. Food Res. 2018 62 1700844 10.1002/mnfr.201700844 29418071 65. Kankaanranta H. Janka-Junttila M. Ilmarinen-Salo P. Ito K. Jalonen U. Ito M. Adcock I.M. Moilanen E. Zhang X. Histone deacetylase inhibitors induce apoptosis in human eosinophils and neutrophils J. Inflamm. 2010 7 1 15 10.1186/1476-9255-7-9 66. Aoyama M. Kotani J. Usami M. Butyrate and propionate induced activated or non-activated neutrophil apoptosis via HDAC inhibitor activity but without activating GPR-41/GPR-43 pathways Nutrition 2010 26 653 661 10.1016/j.nut.2009.07.006 20004081 67. Zhang Z. Tang H. Chen P. Xie H. Tao Y. Demystifying the manipulation of host immunity, metabolism, and extraintestinal tumors by the gut microbiome Signal. Transduct. Target. Ther. 2019 4 41 10.1038/s41392-019-0074-5 31637019 68. Jardou M. Lawson R. Supportive therapy during COVID-19: The proposed mechanism of short-chain fatty acids to prevent cytokine storm and multi-organ failure Med. Hypotheses 2021 154 110661 10.1016/j.mehy.2021.110661 34385045 69. Viola A. Munari F. Sánchez-Rodríguez R. Scolaro T. Castegna A. The Metabolic Signature of Macrophage Responses Front. Immunol. 2019 10 1462 10.3389/fimmu.2019.01462 31333642 70. Nonnenmacher Y. Hiller K. Biochemistry of proinflammatory macrophage activation Cell. Mol. Life Sci. 2018 75 2093 2109 10.1007/s00018-018-2784-1 29502308 71. Fan H. Wu Y. Yu S. Li X. Wang A. Wang S. Chen W. Lu Y. Critical role of mTOR in regulating aerobic glycolysis in carcinogenesis (Review) Int. J. Oncol. 2021 58 9 19 10.3892/ijo.2020.5152 72. Sun Q. Chen X. Ma J. Peng H. Wang F. Zha X. Wang Y. Jing Y. Yang H. Chen R. Mammalian target of rapamycin up-regulation of pyruvate kinase isoenzyme type M2 is critical for aerobic glycolysis and tumor growth Proc. Natl. Acad. Sci. USA 2011 108 4129 4134 10.1073/pnas.1014769108 21325052 73. Schulthess J. Pandey S. Capitani M. Rue-Albrecht K.C. Arnold I. Franchini F. Chomka A. Ilott N.E. Johnston D.G.W. Pires E. The Short Chain Fatty Acid Butyrate Imprints an Antimicrobial Program in Macrophages Immunity 2019 50 432 445 10.1016/j.immuni.2018.12.018 30683619 74. Liu Q. Tian X. Maruyama D. Arjomandi M. Prakash A. Lung immune tone via gut-lung axis: Gut-derived LPS and short-chain fatty acids’ immunometabolic regulation of lung IL-1β, FFAR2, and FFAR3 expression Am. J. Physiol.-Lung Cell. Mol. Physiol. 2021 321 L65 L78 10.1152/ajplung.00421.2020 33851870 75. Trompette A. Gollwitzer E.S. Pattaroni C. Lopez-Mejia I.C. Riva E. Pernot J. Ubags N. Fajas L. Nicod L.P. Marsland B.J. Dietary Fiber Confers Protection against Flu by Shaping Ly6c- Patrolling Monocyte Hematopoiesis and CD8+ T Cell Metabolism Immunity 2018 48 992 1005.e1008 10.1016/j.immuni.2018.04.022 29768180 76. Bachem A. Makhlouf C. Binger K.J. de Souza D.P. Tull D. Hochheiser K. Whitney P.G. Fernandez-Ruiz D. Dähling S. Kastenmüller W. Microbiota-Derived Short-Chain Fatty Acids Promote the Memory Potential of Antigen-Activated CD8+ T Cells Immunity 2019 51 285 297 e285 10.1016/j.immuni.2019.06.002 31272808 77. Donohoe D.R. Collins L.B. Wali A. Bigler R. Sun W. Bultman S.J. The Warburg effect dictates the mechanism of butyrate-mediated histone acetylation and cell proliferation Mol. Cell 2012 48 612 626 10.1016/j.molcel.2012.08.033 23063526 78. Vinolo M.A. Rodrigues H.G. Hatanaka E. Sato F.T. Sampaio S.C. Curi R. Suppressive effect of short-chain fatty acids on production of proinflammatory mediators by neutrophils J. Nutr. Biochem. 2011 22 849 855 10.1016/j.jnutbio.2010.07.009 21167700 79. Liu T. Li J. Liu Y. Xiao N. Suo H. Xie K. Yang C. Wu C. Short-chain fatty acids suppress lipopolysaccharide-induced production of nitric oxide and proinflammatory cytokines through inhibition of NF-κB pathway in RAW264.7 cells Inflammation 2012 35 1676 1684 10.1007/s10753-012-9484-z 22669487 80. Liu L. Li L. Min J. Wang J. Wu H. Zeng Y. Chen S. Chu Z. Butyrate interferes with the differentiation and function of human monocyte-derived dendritic cells Cell Immunol. 2012 277 66 73 10.1016/j.cellimm.2012.05.011 22698927 81. He J. Zhang P. Shen L. Niu L. Tan Y. Chen L. Zhao Y. Bai L. Hao X. Li X. Short-Chain Fatty Acids and Their Association with Signalling Pathways in Inflammation, Glucose and Lipid Metabolism Int. J. Mol. Sci. 2020 21 6365 10.3390/ijms21176356 82. Ghorbani P. Santhakumar P. Hu Q. Djiadeu P. Wolever T.M. Palaniyar N. Grasemann H. Short-chain fatty acids affect cystic fibrosis airway inflammation and bacterial growth Eur. Respir. J. 2015 46 1033 1045 10.1183/09031936.00143614 26022954 83. Cox M.A. Jackson J. Stanton M. Rojas-Triana A. Bober L. Laverty M. Yang X. Zhu F. Liu J. Wang S. Short-chain fatty acids act as antiinflammatory mediators by regulating prostaglandin E2 and cytokines World J. Gastroenterol. 2009 15 5549 5557 10.3748/wjg.15.5549 19938193 84. Vinolo M.A. Rodrigues H.G. Hatanaka E. Hebeda C.B. Farsky S.H. Curi R. Short-chain fatty acids stimulate the migration of neutrophils to inflammatory sites Clin. Sci. 2009 117 331 338 10.1042/CS20080642 19335337 85. Zapolska-Downar D. Siennicka A. Kaczmarczyk M. Kołodziej B. Naruszewicz M. Butyrate inhibits cytokine-induced VCAM-1 and ICAM-1 expression in cultured endothelial cells: The role of NF-kappaB and PPARalpha J. Nutr. Biochem. 2004 15 220 228 10.1016/j.jnutbio.2003.11.008 15068815 86. Zapolska-Downar D. Naruszewicz M. Propionate reduces the cytokine-induced VCAM-1 and ICAM-1 expression by inhibiting nuclear factor-kappa B (NF-kappaB) activation J. Physiol. Pharmacol. 2009 60 123 131 19617655 87. Miller S.J. Zaloga G.P. Hoggatt A.M. Labarrere C. Faulk W.P. Short-chain fatty acids modulate gene expression for vascular endothelial cell adhesion molecules Nutrition 2005 21 740 748 10.1016/j.nut.2004.11.011 15925300 88. Aaron C.P. Schwartz J.E. Bielinski S.J. Hoffman E.A. Austin J.H.M. Oelsner E.C. Donohue K.M. Kalhan R. Berardi C. Kaufman J.D. Intercellular adhesion molecule 1 and progression of percent emphysema: The MESA Lung Study Respir. Med. 2015 109 255 264 10.1016/j.rmed.2014.10.004 25457724 89. Blidberg K. Palmberg L. James A. Billing B. Henriksson E. Lantz A.-S. Larsson K. Dahlén B. Adhesion molecules in subjects with COPD and healthy non-smokers: A cross sectional parallel group study Respir. Res. 2013 14 47 10.1186/1465-9921-14-47 23635004 90. Bochner B.S. Luscinskas F.W. Gimbrone M.A. Jr. Newman W. Sterbinsky S.A. Derse-Anthony C.P. Klunk D. Schleimer R.P. Adhesion of human basophils, eosinophils, and neutrophils to interleukin 1-activated human vascular endothelial cells: Contributions of endothelial cell adhesion molecules J. Exp. Med. 1991 173 1553 1557 10.1084/jem.173.6.1553 1709678 91. Yoon H.J. Park M.K. Lee H. Park T.S. Park D.W. Moon J.-Y. Kim T.H. Sohn J.W. Kim S.-H. Yoon H.-R. Effects of respiratory short-chain fatty acids on bronchial inflammation in asthma World Allergy Organ. J. 2020 13 10.1016/j.waojou.2020.100204 92. Rutting S. Xenaki D. Malouf M. Horvat J.C. Wood L.G. Hansbro P.M. Oliver B.G. Short-chain fatty acids increase TNFα-induced inflammation in primary human lung mesenchymal cells through the activation of p38 MAPK Am. J. Physiol.-Lung Cell. Mol. Physiol. 2019 316 L157 L174 10.1152/ajplung.00306.2018 30407866 93. Mirmonsef P. Zariffard M.R. Gilbert D. Makinde H. Landay A.L. Spear G.T. Short-chain fatty acids induce pro-inflammatory cytokine production alone and in combination with toll-like receptor ligands Am. J. Reprod. Immunol. 2012 67 391 400 10.1111/j.1600-0897.2011.01089.x 22059850 94. Erturk-Hasdemir D. Kasper D.L. Resident commensals shaping immunity Curr. Opin. Immunol. 2013 25 450 455 10.1016/j.coi.2013.06.001 23830047 95. Ananya F.N. Ahammed M.R. Fahem M.M. Kafle S. Viswanathan M. Desai D. Akku R. Khan F. Hernandez T.E. Bala S.K. Association of Intestinal Microbial Dysbiosis With Chronic Obstructive Pulmonary Disease Cureus 2021 13 e19343 10.7759/cureus.19343 34909304 96. Ley R.E. Gut microbiota in 2015: Prevotella in the gut: Choose carefully Nat. Rev. Gastroenterol. Hepatol. 2016 13 69 70 10.1038/nrgastro.2016.4 26828918 97. Reiss A. Jacobi M. Rusch K. Andreas S. Association of dietary type with fecal microbiota and short chain fatty acids in vegans and omnivores J. Int. Soc. Microbiota 2016 1 1 9 10.18143/jism_v1i1.782 98. Eckburg P.B. Bik E.M. Bernstein C.N. Purdom E. Dethlefsen L. Sargent M. Gill S.R. Nelson K.E. Relman D.A. Diversity of the human intestinal microbial flora Science 2005 308 1635 1638 10.1126/science.1110591 15831718 99. Alonso V.R. Guarner F. Intestinal microbiota composition in adults Probiotic Bacteria and Their Effect on Human Health and Well-Being Karger Publishers Basel, Switzerland 2013 Volume 107 17 24 100. Macfarlane S. Macfarlane G.T. Regulation of short-chain fatty acid production Proc. Nutr. Soc. 2003 62 67 72 10.1079/PNS2002207 12740060 101. Feng W. Ao H. Peng C. Gut Microbiota, Short-Chain Fatty Acids, and Herbal Medicines Front. Pharmacol. 2018 9 1354 10.3389/fphar.2018.01354 30532706 102. Sze M.A. Hogg J.C. Sin D.D. Bacterial microbiome of lungs in COPD Int. J. Chronic Obstr. Pulm. Dis. 2014 9 229 238 10.2147/COPD.S38932 103. Ekbom A. Brandt L. Granath F. Löfdahl C.G. Egesten A. Increased risk of both ulcerative colitis and Crohn’s disease in a population suffering from COPD Lung 2008 186 167 172 10.1007/s00408-008-9080-z 18330638 104. Jess T. Loftus E.V. Jr. Harmsen W.S. Zinsmeister A.R. Tremaine W.J. Melton L.J. 3rd Munkholm P. Sandborn W.J. Survival and cause specific mortality in patients with inflammatory bowel disease: A long term outcome study in Olmsted County, Minnesota, 1940–2004 Gut 2006 55 1248 1254 10.1136/gut.2005.079350 16423890 105. Biedermann L. Zeitz J. Mwinyi J. Sutter-Minder E. Rehman A. Ott S.J. Steurer-Stey C. Frei A. Frei P. Scharl M. Smoking cessation induces profound changes in the composition of the intestinal microbiota in humans PLoS ONE 2013 8 e59260 10.1371/journal.pone.0059260 23516617 106. Benjamin J.L. Hedin C.R. Koutsoumpas A. Ng S.C. McCarthy N.E. Prescott N.J. Pessoa-Lopes P. Mathew C.G. Sanderson J. Hart A.L. Smokers with active Crohn’s disease have a clinically relevant dysbiosis of the gastrointestinal microbiota Inflamm. Bowel. Dis. 2012 18 1092 1100 10.1002/ibd.21864 22102318 107. Lee S.H. Yun Y. Kim S.J. Lee E.-J. Chang Y. Ryu S. Shin H. Kim H.-L. Kim H.-N. Lee J.H. Association between Cigarette Smoking Status and Composition of Gut Microbiota: Population-Based Cross-Sectional Study J. Clin. Med. 2018 7 282 10.3390/jcm7090282 108. Li N. Yang Z. Liao B. Pan T. Pu J. Hao B. Fu Z. Cao W. Zhou Y. He F. Chronic exposure to ambient particulate matter induces gut microbial dysbiosis in a rat COPD model Respir. Res. 2020 21 271 10.1186/s12931-020-01529-3 33076910 109. Woodmansey E.J. Intestinal bacteria and ageing J. Appl. Microbiol. 2007 102 1178 1186 10.1111/j.1365-2672.2007.03400.x 17448153 110. Woodmansey E.J. McMurdo M.E. Macfarlane G.T. Macfarlane S. Comparison of compositions and metabolic activities of fecal microbiotas in young adults and in antibiotic-treated and non-antibiotic-treated elderly subjects Appl. Environ. Microbiol. 2004 70 6113 6122 10.1128/AEM.70.10.6113-6122.2004 15466557 111. Marsland B.J. Trompette A. Gollwitzer E.S. The Gut-Lung Axis in Respiratory Disease Ann. Am. Thorac. Soc. 2015 12 (Suppl. 2) S150 S156 10.1513/AnnalsATS.201503-133AW 26595731 112. Schwiertz A. Taras D. Schäfer K. Beijer S. Bos N.A. Donus C. Hardt P.D. Microbiota and SCFA in lean and overweight healthy subjects Obesity 2010 18 190 195 10.1038/oby.2009.167 19498350 113. Armougom F. Henry M. Vialettes B. Raccah D. Raoult D. Monitoring bacterial community of human gut microbiota reveals an increase in Lactobacillus in obese patients and Methanogens in anorexic patients PLoS ONE 2009 4 e7125 10.1371/journal.pone.0007125 19774074 114. Ley R.E. Turnbaugh P.J. Klein S. Gordon J.I. Human gut microbes associated with obesity Nature 2006 444 1022 1023 10.1038/4441022a 17183309 115. De Filippo C. Cavalieri D. Di Paola M. Ramazzotti M. Poullet J.B. Massart S. Collini S. Pieraccini G. Lionetti P. Impact of diet in shaping gut microbiota revealed by a comparative study in children from Europe and rural Africa Proc. Natl. Acad. Sci. USA 2010 107 14691 14696 10.1073/pnas.1005963107 20679230 116. Wang Y. Wang H. Howard A.G. Meyer K.A. Tsilimigras M.C.B. Avery C.L. Sha W. Sun S. Zhang J. Su C. Circulating Short-Chain Fatty Acids Are Positively Associated with Adiposity Measures in Chinese Adults Nutrients 2020 12 2127 10.3390/nu12072127 32708978 117. Yang J. Keshavarzian A. Rose D.J. Impact of dietary fiber fermentation from cereal grains on metabolite production by the fecal microbiota from normal weight and obese individuals J. Med. Food 2013 16 862 867 10.1089/jmf.2012.0292 24044495 118. Baxter N.T. Schmidt A.W. Venkataraman A. Kim K.S. Waldron C. Schmidt T.M. Dynamics of Human Gut Microbiota and Short-Chain Fatty Acids in Response to Dietary Interventions with Three Fermentable Fibers MBio 2019 10 e02566-18 10.1128/mBio.02566-18 30696735 119. Russell W.R. Gratz S.W. Duncan S.H. Holtrop G. Ince J. Scobbie L. Duncan G. Johnstone A.M. Lobley G.E. Wallace R.J. High-protein, reduced-carbohydrate weight-loss diets promote metabolite profiles likely to be detrimental to colonic health Am. J. Clin. Nutr. 2011 93 1062 1072 10.3945/ajcn.110.002188 21389180 120. Murphy E.F. Cotter P.D. Healy S. Marques T.M. O’Sullivan O. Fouhy F. Clarke S.F. O’Toole P.W. Quigley E.M. Stanton C. Composition and energy harvesting capacity of the gut microbiota: Relationship to diet, obesity and time in mouse models Gut 2010 59 1635 1642 10.1136/gut.2010.215665 20926643 121. Hildebrandt M.A. Hoffmann C. Sherrill-Mix S.A. Keilbaugh S.A. Hamady M. Chen Y.Y. Knight R. Ahima R.S. Bushman F. Wu G.D. High-fat diet determines the composition of the murine gut microbiome independently of obesity Gastroenterology 2009 137 1716 1724.e1712 10.1053/j.gastro.2009.08.042 19706296 122. Senghor B. Sokhna C. Ruimy R. Lagier J.-C. Gut microbiota diversity according to dietary habits and geographical provenance Hum. Microbiome J. 2018 7–8 1 9 10.1016/j.humic.2018.01.001 123. Shaheen S.O. Jameson K.A. Syddall H.E. Aihie Sayer A. Dennison E.M. Cooper C. Robinson S.M. Group T.H.C.S. The relationship of dietary patterns with adult lung function and COPD Eur. Respir. J. 2010 36 277 284 10.1183/09031936.00114709 20075056 124. Kaluza J. Harris H. Wallin A. Linden A. Wolk A. Dietary Fiber Intake and Risk of Chronic Obstructive Pulmonary Disease: A Prospective Cohort Study of Men Epidemiology 2018 29 254 260 10.1097/EDE.0000000000000750 28901975 125. Kaluza J. Larsson S.C. Orsini N. Linden A. Wolk A. Fruit and vegetable consumption and risk of COPD: A prospective cohort study of men Thorax 2017 72 500 509 10.1136/thoraxjnl-2015-207851 28228486 126. Kaluza J. Harris H.R. Linden A. Wolk A. Long-term consumption of fruits and vegetables and risk of chronic obstructive pulmonary disease: A prospective cohort study of women Int. J. Epidemiol. 2018 47 1897 1909 10.1093/ije/dyy178 30239739 127. Richards L.B. Li M. Folkerts G. Henricks P.A.J. Garssen J. van Esch B.C.A.M. Butyrate and Propionate Restore the Cytokine and House Dust Mite Compromised Barrier Function of Human Bronchial Airway Epithelial Cells Int. J. Mol. Sci. 2020 22 65 10.3390/ijms22010065 128. Tatsuta M. Kan-o K. Ishii Y. Yamamoto N. Ogawa T. Fukuyama S. Ogawa A. Fujita A. Nakanishi Y. Matsumoto K. Effects of cigarette smoke on barrier function and tight junction proteins in the bronchial epithelium: Protective role of cathelicidin LL-37 Respir. Res. 2019 20 251 10.1186/s12931-019-1226-4 31706310 129. Li N. Dai Z. Wang Z. Deng Z. Zhang J. Pu J. Cao W. Pan T. Zhou Y. Yang Z. Gut microbiota dysbiosis contributes to the development of chronic obstructive pulmonary disease Respir. Res. 2021 22 274 10.1186/s12931-021-01872-z 34696775 130. Chiu Y.C. Lee S.W. Liu C.W. Lin R.C. Huang Y.C. Lan T.Y. Wu L.S. Comprehensive profiling of the gut microbiota in patients with chronic obstructive pulmonary disease of varying severity PLoS ONE 2021 16 e0249944 10.1371/journal.pone.0249944 33836012 131. Karavaeva T.M. Maksimenya M.V. Tereshkov P.P. Gaimolenko I.N. Medvedeva T.A. Parshina A.A. Long-chain fatty acids and short-chain fatty acids in exhaled breath condensate of patients with chronic obstructive pulmonary disease Biomed. Khim. 2021 67 169 174 10.18097/pbmc20216702169 33860775 132. Ardatskaya M.D. Ponomareva E.V. Shevtsov V.V. Evdokimova S.A. Odintsov S.V. Diagnostic and tactical importance of studying short chain fatty acids in different biological substrates took place in patients with chronic obstructive pulmonary disease, lung cancer and community-acquired pneumonia developed after anticancer therapy Exp. Clin. Gastroenterol. 2016 11 17 25 133. Zolnikova O.Y. Kokina N.I. Trukhmanov A.S. Ivashkin V.T. Intestinal Short-Chain Fatty Acids in Patients with Bronchial Asthma Russ. J. Gastroenterol. Hepatol. Coloproctol. 2019 29 53 59 10.22416/1382-4376-2019-29-2-53-59 134. Agarwal A.K. Raja A. Brown B.D. Chronic Obstructive Pulmonary Disease StatPearls Publishing Treasure Island, FL, USA 2022 135. Roche N. Devillier P. Berger P. Bourdin A. Dusser D. Muir J.-F. Martinat Y. Terrioux P. Housset B. Individual trajectory-based care for COPD: Getting closer, but not there yet ERJ Open Res. 2021 7 00451 02021 10.1183/23120541.00451-2021 34912881 136. Marott J.L. Ingebrigtsen T.S. Çolak Y. Vestbo J. Lange P. Lung Function Trajectories Leading to Chronic Obstructive Pulmonary Disease as Predictors of Exacerbations and Mortality Am. J. Respir. Crit. Care Med. 2020 202 210 218 10.1164/rccm.201911-2115OC 32289231 137. Agustí A. Celli B. Natural history of COPD: Gaps and opportunities ERJ Open Res. 2017 3 00117 02017 10.1183/23120541.00117-2017 29255718 138. Mannino D.M. Watt G. Hole D. Gillis C. Hart C. McConnachie A. Davey Smith G. Upton M. Hawthorne V. Sin D.D. The natural history of chronic obstructive pulmonary disease Eur. Respir. J. 2006 27 627 643 10.1183/09031936.06.00024605 16507865 139. Lange P. Ahmed E. Lahmar Z.M. Martinez F.J. Bourdin A. Natural history and mechanisms of COPD Respirology 2021 26 298 321 10.1111/resp.14007 33506971 140. Corlateanu A. Mendez Y. Wang Y. Garnica R.d.J.A. Botnaru V. Siafakas N. Chronic obstructive pulmonary disease and phenotypes: A state-of-the-art Pulmonology 2020 26 95 100 10.1016/j.pulmoe.2019.10.006 31740261 141. Mirza S. Benzo R. Chronic Obstructive Pulmonary Disease Phenotypes: Implications for Care Mayo Clin. Proc. 2017 92 1104 1112 10.1016/j.mayocp.2017.03.020 28688465 142. Coxson H.O. Chan I.H.T. Mayo J.R. Hlynsky J. Nakano Y. Birmingham C.L. Early Emphysema in Patients with Anorexia Nervosa Am. J. Respir. Crit. Care Med. 2004 170 748 752 10.1164/rccm.200405-651OC 15256394 143. Fliederbaum J. Heller A. Zweibaun K. Zarchi J. Clinical aspects of hunger disease in adults Curr. Concepts Nutr. 1979 7 11 43 535413 144. Wright J.L. Cosio M. Churg A. Animal models of chronic obstructive pulmonary disease Am. J. Physiol. Lung Cell. Mol. Physiol. 2008 295 L1 L15 10.1152/ajplung.90200.2008 18456796 145. Reed K.K. Abbaspour A. Bulik C.M. Carroll I.M. The intestinal microbiota and anorexia nervosa: Cause or consequence of nutrient deprivation Curr. Opin. Endocr. Metab. Res. 2021 19 46 51 10.1016/j.coemr.2021.06.003 34458645 146. Mack I. Cuntz U. Grämer C. Niedermaier S. Pohl C. Schwiertz A. Zimmermann K. Zipfel S. Enck P. Penders J. Weight gain in anorexia nervosa does not ameliorate the faecal microbiota, branched chain fatty acid profiles and gastrointestinal complaints Sci. Rep. 2016 6 26752 10.1038/srep26752 27229737 147. Jang Y.O. Kim O.-H. Kim S.J. Lee S.H. Yun S. Lim S.E. Yoo H.J. Shin Y. Lee S.W. High-fiber diets attenuate emphysema development via modulation of gut microbiota and metabolism Sci. Rep. 2021 11 7008 10.1038/s41598-021-86404-x 33772084 148. Karoor V. Strassheim D. Sullivan T. Verin A. Umapathy N.S. Dempsey E.C. Frank D.N. Stenmark K.R. Gerasimovskaya E. The Short-Chain Fatty Acid Butyrate Attenuates Pulmonary Vascular Remodeling and Inflammation in Hypoxia-Induced Pulmonary Hypertension Int. J. Mol. Sci. 2021 22 9916 10.3390/ijms22189916 34576081 149. Petrache I. Petrusca D.N. The involvement of sphingolipids in chronic obstructive pulmonary diseases Handb. Exp. Pharmacol. 2013 216 247 264 10.1007/978-3-7091-1511-4_12 150. Yuan X. Wang L. Bhat O.M. Lohner H. Li P.-L. Differential effects of short chain fatty acids on endothelial Nlrp3 inflammasome activation and neointima formation: Antioxidant action of butyrate Redox Biol. 2018 16 21 31 10.1016/j.redox.2018.02.007 29475132 151. Ogawa H. Rafiee P. Fisher P.J. Johnson N.A. Otterson M.F. Binion D.G. Sodium butyrate inhibits angiogenesis of human intestinal microvascular endothelial cells through COX-2 inhibition FEBS Lett. 2003 554 88 94 10.1016/S0014-5793(03)01110-4 14596920 152. Kim S.H. Kim K.W. Jeong J.W. Inhibition of hypoxia-induced angiogenesis by sodium butyrate, a histone deacetylase inhibitor, through hypoxia-inducible factor-1alpha suppression Oncol. Rep. 2007 17 793 797 17342317 153. Morikawa A. Sugiyama T. Koide N. Mori I. Mu M.M. Yoshida T. Hassan F. Islam S. Yokochi T. Butyrate enhances the production of nitric oxide in mouse vascular endothelial cells in response to gamma interferon J. Endotoxin. Res. 2004 10 32 38 10.1179/096805104225003852 15025822 154. Robles-Vera I. Toral M. de la Visitación N. Aguilera-Sánchez N. Redondo J.M. Duarte J. Protective Effects of Short-Chain Fatty Acids on Endothelial Dysfunction Induced by Angiotensin II Front. Physiol. 2020 11 277 10.3389/fphys.2020.00277 32372967 155. Tomoda K. Kubo K. Dairiki K. Yamaji T. Yamamoto Y. Nishii Y. Nakamura A. Yoshikawa M. Hamada K. Kimura H. Whey peptide-based enteral diet attenuated elastase-induced emphysema with increase in short chain fatty acids in mice BMC Pulm. Med. 2015 15 64 10.1186/s12890-015-0059-2 26059026 156. Jang Y.O. Lee S.H. Choi J.J. Kim D.-H. Choi J.-M. Kang M.-J. Oh Y.-M. Park Y.-J. Shin Y. Lee S.W. Fecal microbial transplantation and a high fiber diet attenuates emphysema development by suppressing inflammation and apoptosis Exp. Mol. Med. 2020 52 1128 1139 10.1038/s12276-020-0469-y 32681029 157. Dransfield M.T. Kunisaki K.M. Strand M.J. Anzueto A. Bhatt S.P. Bowler R.P. Criner G.J. Curtis J.L. Hanania N.A. Nath H. Acute Exacerbations and Lung Function Loss in Smokers with and without Chronic Obstructive Pulmonary Disease Am. J. Respir. Crit. Care Med. 2017 195 324 330 10.1164/rccm.201605-1014OC 27556408 158. Hurst J.R. Vestbo J. Anzueto A. Locantore N. Müllerova H. Tal-Singer R. Miller B. Lomas D.A. Agusti A. MacNee W. Susceptibility to Exacerbation in Chronic Obstructive Pulmonary Disease N. Engl. J. Med. 2010 363 1128 1138 10.1056/NEJMoa0909883 20843247 159. Sadatsafavi M. Xie H. Etminan M. Johnson K. FitzGerald J.M. Network C.R.R. The association between previous and future severe exacerbations of chronic obstructive pulmonary disease: Updating the literature using robust statistical methodology PLoS ONE 2018 13 e0191243 10.1371/journal.pone.0191243 29351331 160. Barnes P.J. Inflammatory endotypes in COPD Allergy 2019 74 1249 1256 10.1111/all.13760 30834543 161. Han M.K. Quibrera P.M. Carretta E.E. Barr R.G. Bleecker E.R. Bowler R.P. Cooper C.B. Comellas A. Couper D.J. Curtis J.L. Frequency of exacerbations in patients with chronic obstructive pulmonary disease: An analysis of the SPIROMICS cohort Lancet Respir. Med. 2017 5 619 626 10.1016/S2213-2600(17)30207-2 28668356 162. Kotlyarov S. Kotlyarova A. Molecular Mechanisms of Lipid Metabolism Disorders in Infectious Exacerbations of Chronic Obstructive Pulmonary Disease Int. J. Mol. Sci. 2021 22 7634 10.3390/ijms22147634 34299266 163. Madan J.C. Koestler D.C. Stanton B.A. Davidson L. Moulton L.A. Housman M.L. Moore J.H. Guill M.F. Morrison H.G. Sogin M.L. Serial Analysis of the Gut and Respiratory Microbiome in Cystic Fibrosis in Infancy: Interaction between Intestinal and Respiratory Tracts and Impact of Nutritional Exposures MBio 2012 3 e00251-12 10.1128/mBio.00251-12 164. Trompette A. Gollwitzer E.S. Yadava K. Sichelstiel A.K. Sprenger N. Ngom-Bru C. Blanchard C. Junt T. Nicod L.P. Harris N.L. Gut microbiota metabolism of dietary fiber influences allergic airway disease and hematopoiesis Nat. Med. 2014 20 159 166 10.1038/nm.3444 24390308 165. Machado M.G. Sencio V. Trottein F. Bäumler A.J. Short-Chain Fatty Acids as a Potential Treatment for Infections: A Closer Look at the Lungs Infect. Immun. 2021 89 e00188-21 10.1128/IAI.00188-21 34097474 166. Bernard H. Desseyn J.L. Gottrand F. Stahl B. Bartke N. Husson M.O. Pectin-Derived Acidic Oligosaccharides Improve the Outcome of Pseudomonas aeruginosa Lung Infection in C57BL/6 Mice PLoS ONE 2015 10 e0139686 10.1371/journal.pone.0139686 26599638 167. Sun Z. Zhu Q.L. Shen Y. Yan T. Zhou X. Dynamic changes of gut and lung microorganisms during chronic obstructive pulmonary disease exacerbations Kaohsiung J. Med. Sci. 2020 36 107 113 10.1002/kjm2.12147 31782610 168. Wu Y. Luo Z. Liu C. Variations in fecal microbial profiles of acute exacerbations and stable chronic obstructive pulmonary disease Life Sci. 2021 265 118738 10.1016/j.lfs.2020.118738 33181175 169. Korpela K. Salonen A. Virta L.J. Kekkonen R.A. Forslund K. Bork P. de Vos W.M. Intestinal microbiome is related to lifetime antibiotic use in Finnish pre-school children Nat. Commun. 2016 7 10410 10.1038/ncomms10410 26811868 170. Russell S.L. Gold M.J. Hartmann M. Willing B.P. Thorson L. Wlodarska M. Gill N. Blanchet M.-R. Mohn W.W. McNagny K.M. Early life antibiotic-driven changes in microbiota enhance susceptibility to allergic asthma EMBO Rep. 2012 13 440 447 10.1038/embor.2012.32 22422004 171. Shin M.-K. Kwak S.H. Park Y. Jung J.Y. Kim Y.S. Kang Y.A. Association between Dietary Patterns and Chronic Obstructive Pulmonary Disease in Korean Adults: The Korean Genome and Epidemiology Study Nutrients 2021 13 4348 10.3390/nu13124348 34959900 172. De Filippis F. Pellegrini N. Vannini L. Jeffery I.B. La Storia A. Laghi L. Serrazanetti D.I. Di Cagno R. Ferrocino I. Lazzi C. High-level adherence to a Mediterranean diet beneficially impacts the gut microbiota and associated metabolome Gut 2016 65 1812 1821 10.1136/gutjnl-2015-309957 26416813 173. Wood L.G. Attia J. McElduff P. McEvoy M. Gibson P.G. Assessment of dietary fat intake and innate immune activation as risk factors for impaired lung function Eur. J. Clin. Nutr. 2010 64 818 825 10.1038/ejcn.2010.68 20531437 174. Brigham E.P. Steffen L.M. London S.J. Boyce D. Diette G.B. Hansel N.N. Rice J. McCormack M.C. Diet Pattern and Respiratory Morbidity in the Atherosclerosis Risk in Communities Study Ann. Am. Thorac. Soc. 2018 15 675 682 10.1513/AnnalsATS.201707-571OC 29446981 175. Varraso R. Fung T.T. Hu F.B. Willett W. Camargo C.A. Prospective study of dietary patterns and chronic obstructive pulmonary disease among US men Thorax 2007 62 786 791 10.1136/thx.2006.074534 17504819 176. Varraso R. Fung T.T. Barr R.G. Hu F.B. Willett W. Camargo C.A. Jr. Prospective study of dietary patterns and chronic obstructive pulmonary disease among US women Am. J. Clin. Nutr. 2007 86 488 495 10.1093/ajcn/86.2.488 17684223 177. Zheng P.F. Shu L. Si C.J. Zhang X.Y. Yu X.L. Gao W. Dietary Patterns and Chronic Obstructive Pulmonary Disease: A Meta-analysis COPD: J. Chronic Obstr. Pulm. Dis. 2016 13 515 522 10.3109/15412555.2015.1098606 26678388 178. Chiu Y.-C. Lee S.-W. Liu C.-W. Lan T.-Y. Wu L.S.-H. Relationship between gut microbiota and lung function decline in patients with chronic obstructive pulmonary disease: A 1-year follow-up study Respir. Res. 2022 23 10 10.1186/s12931-022-01928-8 35033061 179. Begley L. Madapoosi S. Opron K. Ndum O. Baptist A. Rysso K. Erb-Downward J.R. Huang Y.J. Gut microbiota relationships to lung function and adult asthma phenotype: A pilot study BMJ Open Respir. Res. 2018 5 e000324 10.1136/bmjresp-2018-000324 180. Halnes I. Baines K.J. Berthon B.S. MacDonald-Wicks L.K. Gibson P.G. Wood L.G. Soluble Fibre Meal Challenge Reduces Airway Inflammation and Expression of GPR43 and GPR41 in Asthma Nutrients 2017 9 57 10.3390/nu9010057 181. Wood L.G. Omega-3 polyunsaturated fatty acids and chronic obstructive pulmonary disease Curr. Opin. Clin. Nutr. Metab. Care 2015 18 128 132 10.1097/MCO.0000000000000142 25581034 182. Dirjayanto V.J. Evidence on the Efficacy of Omega-3 Polyunsaturated Fatty Acids as an Adjunct Therapy for Chronic Obstructive Pulmonary Disease J. Asian Med. Stud. Assoc. 2021 9 1 10.52629/jamsa.v9i1.238 183. Szmidt M.K. Kaluza J. Harris H.R. Linden A. Wolk A. Long-term dietary fiber intake and risk of chronic obstructive pulmonary disease: A prospective cohort study of women Eur. J. Nutr. 2020 59 1869 1879 10.1007/s00394-019-02038-w 31280344 184. Rozenbloom S.R. Fernandes J. Wolever T.M. Short chain fatty acid absorption in lean vs obese humans FASEB J. 2012 26 lb307 10.1096/fasebj.26.1_supplement.lb307 185. Fernandes J. Su W. Rahat-Rozenbloom S. Wolever T.M.S. Comelli E.M. Adiposity, gut microbiota and faecal short chain fatty acids are linked in adult humans Nutr. Diabetes 2014 4 e121 10.1038/nutd.2014.23 24979150 186. Kimura I. Ozawa K. Inoue D. Imamura T. Kimura K. Maeda T. Terasawa K. Kashihara D. Hirano K. Tani T. The gut microbiota suppresses insulin-mediated fat accumulation via the short-chain fatty acid receptor GPR43 Nat. Commun. 2013 4 1829 10.1038/ncomms2852 23652017 187. Gephine S. Mucci P. Grosbois J.-M. Maltais F. Saey D. Physical Frailty in COPD Patients with Chronic Respiratory Failure Int. J. Chronic Obstr. Pulm. Dis. 2021 16 1381 1392 10.2147/COPD.S295885 188. Lee S.Y. Nyunt M.S.Z. Gao Q. Gwee X. Chua D.Q.L. Yap K.B. Wee S.L. Ng T.P. Co-occurrence of Physical Frailty and COPD and Association With Disability and Mortality: Singapore Longitudinal Ageing Study Chest 2021 S0012-3692(21)05081-9 10.1016/j.chest.2021.12.633 189. Chen X. Mao G. Leng S.X. Frailty syndrome: An overview Clin. Interv. Aging 2014 9 433 441 10.2147/CIA.S45300 24672230 190. Lauretani F. Longobucco Y. Ferrari Pellegrini F. De Iorio A.M. Fazio C. Federici R. Gallini E. La Porta U. Ravazzoni G. Roberti M.F. Comprehensive Model for Physical and Cognitive Frailty: Current Organization and Unmet Needs Front. Psychol. 2020 11 569629 10.3389/fpsyg.2020.569629 33324282 191. Claesson M.J. Jeffery I.B. Conde S. Power S.E. O’Connor E.M. Cusack S. Harris H.M. Coakley M. Lakshminarayanan B. O’Sullivan O. Gut microbiota composition correlates with diet and health in the elderly Nature 2012 488 178 184 10.1038/nature11319 22797518 192. Jackson M.A. Jeffery I.B. Beaumont M. Bell J.T. Clark A.G. Ley R.E. O’Toole P.W. Spector T.D. Steves C.J. Signatures of early frailty in the gut microbiota Genome Med. 2016 8 8 10.1186/s13073-016-0262-7 26822992 193. Clark A. Mach N. The Crosstalk between the Gut Microbiota and Mitochondria during Exercise Front. Physiol. 2017 8 319 10.3389/fphys.2017.00319 28579962 194. Frampton J. Murphy K.G. Frost G. Chambers E.S. Short-chain fatty acids as potential regulators of skeletal muscle metabolism and function Nat. Metab. 2020 2 840 848 10.1038/s42255-020-0188-7 32694821 195. Ticinesi A. Lauretani F. Milani C. Nouvenne A. Tana C. Del Rio D. Maggio M. Ventura M. Meschi T. Aging Gut Microbiota at the Cross-Road between Nutrition, Physical Frailty, and Sarcopenia: Is There a Gut-Muscle Axis? Nutrients 2017 9 1303 10.3390/nu9121303 29189738 196. Walsh M.E. Bhattacharya A. Sataranatarajan K. Qaisar R. Sloane L. Rahman M.M. Kinter M. Van Remmen H. The histone deacetylase inhibitor butyrate improves metabolism and reduces muscle atrophy during aging Aging Cell 2015 14 957 970 10.1111/acel.12387 26290460 197. SSzczesniak O. Hestad K.A. Hanssen J.F. Rudi K. Isovaleric acid in stool correlates with human depression Nutr. Neurosci. 2016 19 279 283 10.1179/1476830515Y.0000000007 25710209 198. Dinan T.G. Cryan J.F. Gut instincts: Microbiota as a key regulator of brain development, ageing and neurodegeneration J. Physiol. 2017 595 489 503 10.1113/JP273106 27641441 199. Kelly J.R. Minuto C. Cryan J.F. Clarke G. Dinan T.G. Cross Talk: The Microbiota and Neurodevelopmental Disorders Front. Neurosci. 2017 11 490 10.3389/fnins.2017.00490 28966571 200. Sharon G. Sampson T.R. Geschwind D.H. Mazmanian S.K. The Central Nervous System and the Gut Microbiome Cell 2016 167 915 932 10.1016/j.cell.2016.10.027 27814521 201. Deng F.-L. Pan J.-X. Zheng P. Xia J.-J. Yin B.-M. Liang W.-W. Li Y.-F. Wu J. Xu F. Wu Q.-Y. Metabonomics reveals peripheral and central short-chain fatty acid and amino acid dysfunction in a naturally occurring depressive model of macaques Neuropsychiatr. Dis. Treat. 2019 15 1077 1088 10.2147/NDT.S186071 31118641 202. Skonieczna-Żydecka K. Grochans E. Maciejewska D. Szkup M. Schneider-Matyka D. Jurczak A. Łoniewski I. Kaczmarczyk M. Marlicz W. Czerwińska-Rogowska M. Faecal Short Chain Fatty Acids Profile is Changed in Polish Depressive Women Nutrients 2018 10 1939 10.3390/nu10121939 203. Perry R.J. Peng L. Barry N.A. Cline G.W. Zhang D. Cardone R.L. Petersen K.F. Kibbey R.G. Goodman A.L. Shulman G.I. Acetate mediates a microbiome-brain-β-cell axis to promote metabolic syndrome Nature 2016 534 213 217 10.1038/nature18309 27279214 204. Valvassori S.S. Resende W.R. Budni J. Dal-Pont G.C. Bavaresco D.V. Réus G.Z. Carvalho A.F. Gonçalves C.L. Furlanetto C.B. Streck E.L. Sodium Butyrate, a Histone Deacetylase Inhibitor, Reverses Behavioral and Mitochondrial Alterations in Animal Models of Depression Induced by Early- or Late-life Stress Curr. Neurovasc. Res. 2015 12 312 320 10.2174/1567202612666150728121121 26216027 205. Valvassori S.S. Varela R.B. Arent C.O. Dal-Pont G.C. Bobsin T.S. Budni J. Reus G.Z. Quevedo J. Sodium butyrate functions as an antidepressant and improves cognition with enhanced neurotrophic expression in models of maternal deprivation and chronic mild stress Curr. Neurovasc. Res. 2014 11 359 366 10.2174/1567202611666140829162158 25233278 206. Silva Y.P. Bernardi A. Frozza R.L. The Role of Short-Chain Fatty Acids From Gut Microbiota in Gut-Brain Communication Front. Endocrinol. 2020 11 25 10.3389/fendo.2020.00025 32082260 207. Zhai T. Li S. Hu W. Li D. Leng S. Potential Micronutrients and Phytochemicals against the Pathogenesis of Chronic Obstructive Pulmonary Disease and Lung Cancer Nutrients 2018 10 813 10.3390/nu10070813 29941777 208. Pearson P. Britton J. McKeever T. Lewis S.A. Weiss S. Pavord I. Fogarty A. Lung function and blood levels of copper, selenium, vitamin C and vitamin E in the general population Eur. J. Clin. Nutr. 2005 59 1043 1048 10.1038/sj.ejcn.1602209 16015272 209. Mekal D. Czerw A. Deptala A. Dietary Behaviour and Nutrition in Patients with COPD Treated with Long-Term Oxygen Therapy Int. J. Environ. Res. Public Health 2021 18 12793 10.3390/ijerph182312793 34886519

PMC009xxxxxx/PMC9099630.txt

==== Front J Clin Med J Clin Med jcm Journal of Clinical Medicine 2077-0383 MDPI 10.3390/jcm11092503 jcm-11-02503 Article Acceptability of Human Papilloma Virus Self-Sampling for Cervical Cancer Screening in a Cohort of Patients from Romania (Stage 2) Grigore Mihaela 1 https://orcid.org/0000-0002-4718-392X Vasilache Ingrid-Andrada 1* Cianga Petru 2 Constantinescu Daniela 2 Duma Odetta 3 Matasariu Roxana Daniela 1 Scripcariu Ioana-Sadiye 1 Jordanova Ekaterina S. Academic Editor 1 Department of Obstetrics and Gynecology, ‘Grigore T. Popa’ University of Medicine and Pharmacy, 700115 Iasi, Romania; mihaela.grigore@umfiasi.ro (M.G.); roxana.matasariu@umfiasi.ro (R.D.M.); isscripcariu@gmail.com (I.-S.S.) 2 Department of Immunology, ‘Grigore T. Popa’ University of Medicine and Pharmacy, 700115 Iasi, Romania; petru.cianga@umfiasi.ro (P.C.); daniela.constantinescu@umfiasi.ro (D.C.) 3 Department of Epidemiology, ‘Grigore T. Popa’ University of Medicine and Pharmacy, 700115 Iasi, Romania; odetta.duma@umfiasi.ro * Correspondence: tanasaingrid@yahoo.com 29 4 2022 5 2022 11 9 250308 4 2022 28 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). (1) Background: Low patient’s adherence to conventional cervical cancer screening methods determined the need to take into consideration alternative approaches, and vaginal HPV self-sampling is one of them. We aimed to evaluate, using an online survey, the Romanian women’s acceptability of vaginal HPV self-sampling. (2) Methods: A 13-questions online survey was distributed on three Facebook groups, and the results were summarized. (3) Results: Despite of good educational background, 10.8% (n = 60) of the respondents did not know what a Pap smear is, and 33% (n = 183) were not informed about the free national cervical cancer screening program. Multivariate analysis revealed an increased likelihood of vaginal self-sampling acceptance among respondents who did not know about Pap test (OR: 7.80; 95%CI: 1.062–57.431; p = 0.021), national cervical cancer screening program (OR: 1.96; 95%CI: 1.010–3.806; p = 0.02), HPV infection (OR: 7.35; 95%CI: 3.099–17.449; p< 0.001) or HPV test (OR: 1.67; 95%CI: 0.950–2.948; p = 0.03). Moreover, women who did not previously undergo a cervical cancer screening program were more likely to accept the new screening method (OR: 1.62; 95%CI: 0.878–3.015; p = 0.04). (4) Conclusions: Our results showed high acceptability rates of vaginal HPV self-sampling among participants. self-sampling cervical cancer screening acceptability HPV detection University of Medicine and Pharmacy ‘Grigore T. Popa’IDEI No. 10309/29.06.2020 This research was funded by the University of Medicine and Pharmacy ‘Grigore T. Popa’, Iasi, Romania (grant IDEI No. 10309/29.06.2020). ==== Body pmc1. Introduction With an estimated 604,000 new cases and 342,000 deaths worldwide in 2020, cervical cancer is the fourth most commonly diagnosed cancer and the fourth major cause of cancer death in women [1]. Each year in Romania approximately 1800 women die from cervical cancer, and almost 3400 receive this diagnosis [2,3]. The high incidence and mortality rates associated with cervical cancer place this country in the top of the European Union statistics [4]. Human papillomavirus (HPV) infection, some sexually transmittable infections (HIV and Chlamydia trachomatis), smoking, high parity, and long-term use of oral contraceptives are among the risk factors for developing cervical cancer [5,6]. HPV oncogenic types 16 and 18, responsible for almost 70% of the cervical neoplasia cases [7], are the main target of public health strategies for eradicating cervical cancer. Because of the highly efficient primary (HPV vaccine) and secondary (screening) prevention strategies, cervical cancer is regarded as a preventable disease. However, low patients’ adherence to the national screening programs, mainly due to the lack of information and understanding, represented a barrier for achieving consistent results [8]. In recent years, vaginal HPV self-sampling emerged as a promising alternative to the conventional screening strategies based on Pap smears and/or HPV DNA testing. HPV self-sampling is a screening procedure that involves using a kit to collect cervico-vaginal samples, which are then sent to a laboratory for genotyping high-risk oncogenic HPV strains. It is easy to use by patients at home, is cost-effective, and has good overall accuracy in detecting HPV strains. Although recent studies have indicated that provider-collected cervical samples had the highest HPV-DNA sensitivity (84–100%), self-collected vaginal HPV-DNA tests appeared to have good performance, with a sensitivity ranging from 66 to 88% [9,10,11]. If the patient receives a positive result for the HPV test, she will establish an appointment with the gynecologist for deciding the best therapeutic plan. Furthermore, self-sampling was found to be superior to HPV DNA testing performed by a clinician in terms of acceptance and preference [12,13]. The main benefits cited for vaginal self-sampling were less pain or physical discomfort, convenience, ability to perform the test in private, and less embarrassment or anxiety [14,15,16,17,18]. The challenges associated with the implementation of vaginal self-sampling include: difficulties regarding explaining the self-sampling procedure to participating women, specimen transportation and laboratory processing, as well as follow-up of positive women [19]. The provision of clear instructions accompanied by illustrations in suitable language would be a feasible approach for facilitating the self-sampling procedure [20,21]. On the other hand, the SARS-CoV-2 pandemic has accelerated the introduction of HPV self-sampling as a measure to increase the addressability to cervical cancer screening program, and self-sampling is included in the World Health Organization’s (WHO) recently issued guidelines on self-intervention for health, as well as the cervical cancer screening guidelines [22]. A recent study assessed the effectiveness of HPV self-sampling for cervical cancer screening during the SARS-CoV-2 pandemic, and found a high concordance for HPV detection between self-sampled and clinician-sampled specimens (90.2%), as well as a high willingness to repeat the procedure under the same conditions (89.2% of the participants) [23]. The main determinant of this study was the lack of literature data on the acceptability of doing self-sampling in the local population. We aimed to evaluate, using an online survey, the Romanian women’s acceptability of vaginal HPV self-sampling. 2. Materials and Methods Between 15 February and 16 March 2022, data were collected using a Qualtrics form prepared by the investigator and titled: ‘A survey for analyzing the Romanian women’s thoughts regarding vaginal HPV self-sampling’, which was posted on three Facebook groups (Mothers at first pregnancy, About kids, and Mothers from Iasi). The survey consisted of 13 questions that addressed maternal characteristics (age, living environment, and level of education), cervical cancer screening topic (6 closed questions that evaluated the women’s knowledge and experience with cervical cancer screening), as well as the women’s acceptability of vaginal HPV self-sampling. The participants were also shown an illustration of a vaginal self-sampling device along with its instructions for use so that they could complete the questionnaire fully informed. The questionnaire was posted once group administrators approved it, and participants were notified that their anonymous responses would be published in this study. Before beginning the survey, all participants were instructed that it was completely voluntary and anonymous, and that they could skip any questions they did not feel comfortable answering or exit at any time. Each participant was instructed to read and declare if she agreed or disagreed with the survey’s questions. No names or medical identifying numbers were recorded to safeguard the participants’ privacy. All procedures were followed in compliance with the applicable norms and legislation. The Institutional Ethics Committee of the University of Medicine and Pharmacy ‘Grigore T. Popa’ gave its approval to this investigation (No. 96/23.06.2021). Data collected were translated from the Romanian language into standard English. Incomplete surveys were excluded from the study. The total number of questionnaires distributed allowed statistical analysis with a cut-off for the absolute error and the type one error of 5%. Acceptability of vaginal self-sampling was defined as the target variable. Categorical parameters were expressed as numbers and percentages, and statistical comparisons were made using the Chi-square test. We used a multivariate logistic regressions analysis to explore the associations between demographic characteristics, knowledge about the cervical cancer screening, previous screening attendance, and the target variable. The statistical analysis was performed using SPSS software (version 28.0.1, IBM Corporation, Armonk, NY, USA). A p value of less than 0.05 was considered statistically significant. 3. Results Five hundred and eighty-two questionnaires were obtained and analyzed. Only 556 were included in the study due to incomplete data. Table 1 and Table 2 describe the acceptability of vaginal self-sampling for cervical screening among our respondents. The majority of the respondents belonged to the 30–40 years age group (n = 339; 60.9%%), and only a minority of respondents had more than 51 years (n = 28; 5%). Most participants lived in urban areas (n = 324; 58.2%), and had a bachelor degree (n = 264; 47.4%). Despite of good educational background, 10.9% (n = 61) of the respondents did not know what a Pap smear was, and 33% (n = 184) were not informed about the free national cervical cancer screening program. The results from the univariate analysis showed a significant association between age, medium, level of education, and the acceptability of vaginal HPV self-sampling (p < 0.001). The odds ratios (OR) and 95% confidence intervals (CI) from the multivariate logistic analysis are shown in Table 3. The women from the 20–30 years age group were the most likely to accept the new method of cervical cancer screening (OR: 15.78; 95% CI: 2.161–115.304; p = 0.003). There was also a good likelihood for accepting this screening method for women included in the 40–50 years age group (OR: 3.34; 95% CI: 1.020–10.960; p = 0.02). Women living in rural areas (OR: 2.87; 95% CI: 1.483–5.563; p < 0.001), as well as those who possessed a high-school diploma (OR: 2.73; 95% CI: 1.478–5.057; p < 0.001), were more likely to accept the screening method. On the other hand, it appeared that respondents from urban areas (OR: 0.34; 95% CI: 0.180–0.674; p < 0.001), as well as those with a bachelor degree (OR: 0.28; 95% CI: 0.155–0.532; p < 0.001), were less likely to embrace the method. An alarming number of respondents did not know what an HPV infection (n = 238; 42.8%) or HPV test (n = 267; 48%) was. Moreover, more than one third of the participants (n = 198; 35.6%) did not do a Pap test or an HPV test for cervical cancer screening during their lifetime. On the other hand, more than half of the women (n = 358; 64.3%) underwent cervical screening, and Pap smear was the most used screening method (n = 291; 81.2%). Univariate analysis indicated a significant association between knowledge about Pap test (p = 0.006), national cervical screening program (p = 0.043), HPV infection (p < 0.001), previous participation to the screening program (p < 0.001) and the acceptability of vaginal HPV self-sampling. Multivariate analysis revealed an increased likelihood of vaginal self-sampling acceptance among respondents who did not know about Pap test (OR: 7.80; 95%CI: 1.062–57.431; p = 0.021), national cervical cancer screening program (OR: 1.96; 95%CI: 1.010–3.806; p = 0.02), HPV infection (OR: 7.35; 95%CI: 3.099–17.449; p < 0.001) or HPV test (OR: 1.67; 95%CI: 0.950–2.948; p = 0.03). Moreover, women who did not previously undergo a cervical cancer screening program were more likely to accept the new screening method (OR: 1.62; 95%CI: 0.878–3.015; p = 0.04). The majority of the respondents (n = 523; 94%) considered self-sampling a good alternative to Pap test for women who do not consult their physicians. However, 87% (n = 484) of the subjects agreed that cervical sampling performed by a doctor is better than vaginal self-sampling (p = 0.02). 64.7% (n = 360) of the participants agreed with the statement that most women will choose self-sampling over a visit to a doctor, while 81.6% of them considered HPV vaginal self-sampling a good alternative to the conventional screening method. The answers to the last two questions indicated the women’s preference of self-sampling over the doctors’ appointments, with a significant level of acceptability among the participants (p < 0.001). 4. Discussion This observational retrospective study, based on a self-administered online survey, showed high acceptability rates of vaginal HPV self-sampling among participants, including under-screened women. Moreover, our study also confirmed the women’s lack of knowledge about cervical cancer screening, even though the majority of the respondents had higher education. In terms of acceptability, our multivariate analysis revealed that the women from the 20–30 years age group, followed by those in the 40–50 years age group, were most likely to accept the new method of cervical cancer screening, while those women in the 30–40 years age group were significantly less likely to embrace the method (OR: 0.07; 95%CI: 0.023–0.241; p < 0.001). These findings could be explained by the fact that very young women are more open to new technologies, while those middle-aged, who are less used to regular gynecological appointments, are in favor of self-sampling in the comfort of their own home. Similar results were reported in various studies that outlined increased acceptability of vaginal self-sampling among younger generations of women due to decreased embarrassment, time/effort investment and a ‘do-it- yourself’ attitude, as well as older women, due to previous bad gynecological experiences or the need for privacy [10,24,25,26]. Our results also indicated that women living in rural areas, as well as those who possessed a high-school diploma, were more likely to accept the screening method, while those from urban areas and with a bachelor degree were less likely to accept it. Several studies demonstrated that women living in rural areas do not benefit as much as women from urban areas in terms of cervical cancer screening [27,28]. Limited accessibility to Pap smear and/or HPV testing, rural practice configuration or scarce financial resources were cited as factors that negatively influence the women’s addressability to cervical cancer screening programs [29,30]. On the other hand, developed sanitary infrastructure along with higher consumption of healthcare services in the urban areas could be considered reasons for the women’s preference to traditional cervical cancer screening [31,32]. Although the literature data are conflicting regarding the influence of the educational background over the HPV self-sampling acceptability [33,34,35], we hypothesize that women with at least a bachelor degree tend to have a higher compliance to conventional cervical cancer screening. In Romania, there is an active national cervical cancer screening program developed by the Public Health Ministry in partnership with regional hospitals that offers free Pap testing for women aged between 25 and 64 years [36]. The lack of knowledge among the participants about the cervical cancer screening in Romania was a key finding of the study. An alarming number of respondents did not know what a Pap smear (n = 60; 10.8%), HPV infection (n = 237; 42.7%) or HPV test (n = 267; 48.1%) was despite a good educational background. These data are comparable to that outlined by various studies in developing or underdeveloped countries [37,38,39,40]. Our multivariate analysis revealed an increased likelihood of vaginal self-sampling acceptance among respondents who did not know about Pap tests, the national cervical cancer screening program, or HPV infection. Moreover, women who did not previously undergo a cervical cancer screening program appeared more likely to accept the new screening method. Our results are in line with those reported in an observational study by Lancrajan et al., which outlined the limited knowledge of Romanian women about cervical cancer screening [41]. Furthermore, previous studies indicated that self-sampling acceptability was higher among women who had never undergone cervical screening before, in both rich and underprivileged societies [42,43]. We hypothesize that women with scarce information regarding this topic would feel more comfortable with a self-administered test that does not require a visit to the gynecologist. This emphasizes the importance of a thorough education about the feasibility, benefits, and accuracy of self-sampling in order to increase screening participation. In this study, vaginal self-sampling was considered a good alternative to Pap test for women who do not consult their physicians, with an acceptance rate of 89.7%. The main reasons for low addressability to the cervical cancer screening program in Romania were explored by Todor et al., and included the lack of national coverage and the penetration in the rural areas, mass-media promotion campaigns, funding, involvement of general practitioners, and bureaucracy, as well as program monitoring [44]. Our results showed that 33% (n = 183) of the respondents were not informed about the free national cervical cancer screening program. The present study confirmed the low adherence of Romanian respondents to the cervical cancer screening program, despite the high acceptability rates of HPV self-sampling among respondents. Although the majority of respondents considered that cervical sampling performed by a doctor is better than vaginal self-sampling (86.9%), they have also expressed their preference to self-sampling over the doctors’ appointments (81.5% vs. 15.6%). Analysis of the patient’s preferences over these methods was not the purpose of this study, but the comfort, privacy, ergonomics, and accessibility of self-sampling were cited as some of the main reasons for choosing this alternative over a physician’s appointment [42,45,46,47]. A recent systematic review by Devarapalli et al., which evaluated the most important barriers associated with cervical cancer screening in low and middle-income countries, concluded that the following elements are important obstacles to fulfilling the screening’s objectives: psychological, structural, sociocultural, and religious barriers, as well as lack of knowledge and awareness [48]. All these aspects must be considered when establishing public health programs in order to improve the campaigns’ outcomes. Several studies have shown that women’s awareness of cervical cancer, HPV, and its vaccine can effectively improve the mortality rates and reduce incidences of this disease [49,50,51]. Therefore, effective measures to improve women’s level of knowledge should include all new means of information (mass-media, social media platforms, news websites, etc.) that allow dissemination of relevant data on a regular basis. Moreover, telemedicine has been shown to be a useful tool for providing medical advice during the SARS-CoV-2 pandemic, and could be further added to the screening programs [52,53]. The participants’ profile, which may not be representative of under-screened women targeted by vaginal self-sampling, is one main weakness in our study. The majority of the subjects had a high educational level and a high rate of cervical cancer screening compliance. This could be explained by the way women who had access to social media platforms were recruited. We were not able to generalize the findings of this study since our participants were limited to respondents attending social media groups. Other limitations of our study include: small number of participants included, short time-frame for data collection, and restricted information concerning the epidemiological characteristics of the respondents. We chose to design this survey with only a limited number of questions that allowed a reasonable completion rate, and a 3 min response time. Similar studies, that used web-based surveys disseminated through Facebook groups, acknowledged the difficulties regarding the maintenance of an active interest for the survey and representativeness of epidemiological data [54,55]. Further research is needed to evaluate the acceptability of vaginal HPV self-sampling in various cohorts of patients, especially those who manifest low compliance to conventional screening methods. 5. Conclusions This study demonstrated high acceptability rates of vaginal HPV self-sampling among respondents, indicating this as a feasible method for increasing the women’s compliance to cervical cancer screening. However, an alarming number of women did not have knowledge about cervical cancer and its screening possibilities. Therefore, more education and public health interventions are needed to raise awareness about this topic. Future studies could further evaluate the values and preferences of Romanian women regarding the secondary prevention of cervical cancer. Author Contributions Conceptualization, M.G., I.-S.S. and I.-A.V.; methodology, M.G., I.-S.S. and I.-A.V.; validation, P.C., D.C. and O.D.; formal analysis, M.G. and R.D.M.; investigation, I.-S.S. and I.-A.V.; data curation, P.C., D.C. and O.D.; writing—original draft preparation, M.G., I.-S.S. and I.-A.V.; writing—review and editing, M.G. and I.-A.V.; supervision, M.G.; project administration, M.G.; funding acquisition, M.G. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Ethics Committee of University of Medicine and Pharmacy ‘Grigore T. Popa’ (No. 96/23.06.2021). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The data presented in this study are available on request from the corresponding author. The data are not publicly available due to local policies. Conflicts of Interest The authors declare no conflict of interest. jcm-11-02503-t001_Table 1 Table 1 The acceptability of vaginal HPV self-sampling among the respondents, considering the demographic characteristics and cervical cancer screening knowledge. Variable Self-Sampling Acceptance (n/%) Total Number of Responses (n/%) p Value Yes No Age 20–30 years 108 (19.4%) 1 (0.1%) 109 (19.6%) <0.001 30–40 years 285 (51.2%) 54 (9.7%) 339 (69.9%) 40–50 years 78 (14%) 3 (0.5%) 81(14.5%) >51 years 27 (4.8%) 1 (0.1%) 28 (5%) Medium Urban 282 (50.7%) 42 (7.5%) 324 (58.3%) <0.001 Rural 216 (38.8%) 16 (2.8%) 232 (41.7%) Level of education Primary school (≤4 years of study) 3 (0.5%) 1 (0.1%) 4 (0.7%) <0.001 Pre-high school (5–8 years of study) 26 (4.6%) 1 (0.1%) 27 (4.8%) High-school (9–12 years of study) 246 (44.2%) 15 (2.6%) 261 (46.9%) ≥Bachelor degree 222 (39.9%) 42 (7.5%) 264 (47.4%) Do you know what Pap test is? Yes 438 (78.7%) 57 (10.2%) 495 (89%) 0.006 No 60 (10.7%) 1 (0.1%) 62 (11%) Do you know that in Romania there is a free national program for Pap testing? Yes 327 (58.8%) 45 (8.1%) 372 (66.9%) 0.043 No 171 (30.8%) 13 (2.3%) 184 (33.1%) Do you know what Human papilloma virus (HPV) infection is? Yes 267 (48%) 51 (9.1%) 318 (57.1%) <0.001 No 231 (41.5%) 7 (1.2%) 238 (42.9%) Do you know what HPV test is? Yes 252 (45.3%) 37 (6.6%) 289 (51.9%) 0.072 No 246 (44.2%) 21 (3.7%) 267 (48.1%) Did you previously do a Pap test or HPV test for cervical cancer screening? Yes 315 (56.6%) 43 (7.7%) 358 (64.3%) <0.001 No 183 (32.1%) 15 (2.6%) 198 (35.7%) If your previous answer was yes, which test did you take? Pap test 264 (47.4%) 27 (4.8%) 291 (81.2%) <0.001 HPV test 3 (0.5%) 1 (0.1%) 4 (1.1%) Both 45 (8%) 18 (3.2%) 63 (17.7%) jcm-11-02503-t002_Table 2 Table 2 The acceptability of vaginal HPV self-sampling among the respondents, considering their opinions regarding cervical cancer screening. Variable Self-Sampling Acceptance (n/%) Total Number of Responses (n/%) p Value Yes No Do you consider self-sampling a good alternative to Pap test for women who do not consult their physicians? Yes 498 (89.5%) 25 (4.4%) 523 (94.1%) <0.001 No 0 (0%) 30 (5.4%) 30 (5.4%) I have no opinion 0 (0%) 3 (0.5%) 3 (0.5%) I consider that cervical sampling performed by a doctor is better than self-sampling Yes 430 (77.3%) 54 (9.7%) 484 (87%) 0.02 No 39 (7%) 0 (0%) 39 (7%) I have no opinion 30 (5.4%) 3 (0.5%) 33 (6%) I consider that most women will choose self-sampling over a visit to a doctor Yes 339 (61%) 21 (3.7%) 360 (64.7%) <0.001 No 109 (19.6%) 27 (4.8%) 136 (24.4%) I have no opinion 51 (9.1%) 9 (1.6%) 60 (10.7%) I consider self-sampling a good alternative and I would use it instead of going to the doctor Yes 442 (79.4%) 12 (2.1%) 454 (81.6%) <0.001 No 45 (8.1%) 42 (7.5%) 87 (15.6%) I have no opinion 12 (2.1%) 3 (0.5%) 15 (21.7%) jcm-11-02503-t003_Table 3 Table 3 Multivariate logistic regression analysis to identify factors associated with the acceptability of HPV self-sampling. Variable Self-Sampling Acceptance p Value Odds ratio Lower Bound CI Upper Bound CI Age 20–30 years 15.78 2.161 115.304 0.003 30–40 years 0.07 0.023 0.241 <0.001 40–50 years 3.34 1.020 10.960 0.02 >51 years 3.26 0.436 24.503 0.12 Medium Urban 0.34 0.180 0.674 <0.001 Rural 2.87 1.483 5.563 <0.001 Level of education Primary school (≤4 years of study) 0.34 0.035 3.376 0.18 Pre-high school (5–8 years of study) 3.26 0.436 24.503 0.12 High-school (9–12 years of study) 2.73 1.478 5.057 <0.001 ≥Bachelor degree 0.28 0.155 0.532 <0.001 Did women know about Pap test? Yes 0.12 0.017 0.942 0.021 No 7.80 1.062 57.431 0.021 Did women know about the free national screening program? Yes 0.50 0.263 0.990 0.02 No 1.96 1.010 3.806 0.02 Did women know about Human papilloma virus (HPV) infection? Yes 0.13 0.057 0.323 <0.001 No 7.35 3.099 17.449 <0.001 Did women know about HPV test? Yes 0.59 0.339 1.053 0.03 No 1.67 0.950 2.948 0.03 Were women previously screened for cervical cancer? Yes 0.61 0.332 1.139 0.04 No 1.62 0.878 3.015 0.04 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Sung H. Ferlay J. Siegel R.L. Laversanne M. Soerjomataram I. Jemal A. Bray F. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries CA Cancer J. Clin. 2021 71 209 249 10.3322/caac.21660 33538338 2. International Agency for Research on Cancer Globocan 2018: Romania Available online: https://gco.iarc.fr/today/data/factsheets/populations/642-romania-fact-sheets.pdf (accessed on 15 March 2022) 3. Arbyn M. Weiderpass E. Bruni L. de Sanjosé S. Saraiya M. Ferlay J. Bray F. Estimates of incidence and mortality of cervical cancer in 2018: A worldwide analysis Lancet Glob. Health 2020 8 e191 e203 10.1016/S2214-109X(19)30482-6 31812369 4. Arbyn M. Raifu A.O. Weiderpass E. Bray F. Anttila A. Trends of cervical cancer mortality in the member states of the European Union Eur. J. Cancer 2009 45 2640 2648 10.1016/j.ejca.2009.07.018 19695864 5. Human papillomaviruses. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans International Agency for Research on Cancer Lyon, France 2007 Volume 90 1 636 6. Thun M. Linet M.S. Cerhan J.R. Haiman C.A. Schottenfeld D. Cancer Epidemiology and Prevention Oxford University Press Oxford, UK 2017 7. Whitlock E.P. Vesco K.K. Eder M. Lin J.S. Senger C.A. Burda B.U. Liquid-Based Cytology and Human Papillomavirus Testing to Screen for Cervical Cancer: A Systematic Review for the U.S. Preventive Services Task Force Ann. Intern. Med. 2011 155 687 697 10.7326/0003-4819-155-10-201111150-00376 22006930 8. Buskwofie A. David-West G. Clare C.A. A Review of Cervical Cancer: Incidence and Disparities J. Natl. Med. Assoc. 2020 112 229 232 10.1016/j.jnma.2020.03.002 32278478 9. Aranda Flores C.E. Gomez Gutierrez G. Ortiz Leon J.M. Cruz Rodriguez D. Sørbye S.W. Self-collected versus clinician-collected cervical samples for the detection of HPV infections by 14-type DNA and 7-type mRNA tests BMC Infect. Dis. 2021 21 504 10.1186/s12879-021-06189-2 34058992 10. Wong E.L.Y. Cheung A.W.L. Huang F. Chor J.S.Y. Can Human Papillomavirus DNA Self-sampling be an Acceptable and Reliable Option for Cervical Cancer Screening in Female Sex Workers? Cancer Nurs. 2018 41 45 52 10.1097/NCC.0000000000000462 28114260 11. Inturrisi F. Aitken C.A. Melchers W.J. Brule A.J.V.D. Molijn A. Hinrichs J.W. Niesters H.G. Siebers A.G. Schuurman R. Heideman D.A. Clinical performance of high-risk HPV testing on self-samples versus clinician samples in routine primary HPV screening in the Netherlands: An observational study Lancet Reg. Health Eur. 2021 11 100235 10.1016/j.lanepe.2021.100235 34918001 12. Huynh J. Howard M. Lytwyn A. Self-collection for vaginal human papillomavirus testing: Systematic review of studies asking women their perceptions J. Low. Genit. Tract. Dis. 2010 14 356 362 10.1097/LGT.0b013e3181dc115b 20885165 13. Tiro J.A. Betts A.C. Kimbel K. Buist D.S. Mao C. Gao H. Shulman L. Malone C. Beatty T. Lin J. Understanding Patients’ Perspectives and Information Needs Following a Positive Home Human Papillomavirus Self-Sampling Kit Result J. Women’s Health 2019 28 384 392 10.1089/jwh.2018.7070 14. Ilangovan K. Kobetz E. Koru-Sengul T. Marcus E.N. Rodriguez B. Alonzo Y. Carrasquillo O. Acceptability and Feasibility of Human Papilloma Virus Self-Sampling for Cervical Cancer Screening J. Women’s Health 2016 25 944 951 10.1089/jwh.2015.5469 26890012 15. Kilfoyle K.A. Marais A.C.D. Ngo M.A. Romocki L. Richman A.R. Barclay L. Brewer N.T. Rahangdale L. Smith J.S. Preference for Human Papillomavirus Self-Collection and Papanicolaou: Survey of Underscreened Women in North Carolina J. Low. Genit. Tract Dis. 2018 22 302 310 10.1097/LGT.0000000000000430 30179994 16. Gottschlich A. Rivera-Andrade A. Grajeda E. Alvarez C. Montano C.M. Meza R. Acceptability of Human Papillomavirus Self-Sampling for Cervical Cancer Screening in an Indigenous Community in Guatemala J. Glob. Oncol. 2017 3 444 454 10.1200/JGO.2016.005629 29094082 17. Crofts V. Flahault E. Tebeu P.M. Untiet S. Boulvain M. Vassilakos P. Petignat P. Flore G.K.-F.G. Education efforts may contribute to wider acceptance of human papillomavirus self-sampling Int. J. Women’s Health 2015 7 149 154 10.2147/IJWH.S56307 25674016 18. Ortiz A.P. Alejandro N. Pérez C.M. Otero Y. Soto-Salgado M. Palefsky J.M. Tortolero-Luna G. Romaguera J. Acceptability of cervical and anal HPV self-sampling in a sample of Hispanic women in Puerto Rico Puerto Rico Health Sci. J. 2012 31 205 212 19. Hawkes D. Keung M.H.T. Huang Y. McDermott T.L. Romano J. Saville M. Brotherton J.M.L. Self-Collection for Cervical Screening Programs: From Research to Reality Cancers 2020 12 1053 10.3390/cancers12041053 20. Podolak I. Kisia C. Omosa-Manyonyi G. Cosby J. Using a multimethod approach to develop implementation strategies for a cervical self-sampling program in Kenya BMC Health Serv. Res. 2017 17 222 10.1186/s12913-017-2160-0 28320374 21. Hanley S.J. Fujita H. Yokoyama S. Kunisawa S. Tamakoshi A. Dong P. Kobayashi N. Watari H. Kudo M. Sakuragi N. HPV self-sampling in Japanese women: A feasibility study in a population with limited experience of tampon use J. Med. Screen. 2016 23 164 170 10.1177/0969141315625702 26919866 22. World Health Organization (WHO) WHO Guideline on Self-Care Interventions for Health and Well-Being Available online: https://wwwwhoint/publications-detail-redirect/9789240030909 (accessed on 17 March 2022) 23. Ngu S.-F. Lau L.S.K. Li J. Wong G.C.Y. Cheung A.N.Y. Ngan H.Y.S. Chan K.K.L. Human Papillomavirus Self-Sampling for Primary Cervical Cancer Screening in Under-Screened Women in Hong Kong during the COVID-19 Pandemic Int. J. Environ. Res. Public Health 2022 19 2610 10.3390/ijerph19052610 35270303 24. Bosgraaf R.P. Ketelaars P.J. Verhoef V.M. Massuger L.F. Meijer C.J. Melchers W.J. Bekkers R.L. Reasons for non-attendance to cervical screening and preferences for HPV self-sampling in Dutch women Prev. Med. 2014 64 108 113 10.1016/j.ypmed.2014.04.011 24736093 25. Chen S.L. Hsieh P.C. Chou C.H. Tzeng Y.L. Determinants of women’s likelihood of vaginal self-sampling for human papillomavirus to screen for cervical cancer in Taiwan: A cross-sectional study BMC Women’s Health 2014 14 139 10.1186/s12905-014-0139-0 25420580 26. Fargnoli V. Petignat P. Burton-Jeangros C. To what extent will women accept HPV self-sampling for cervical cancer screening? A qualitative study conducted in Switzerland Int. J. Women’s Health 2015 7 883 888 10.2147/IJWH.S90772 26604830 27. Di J. Rutherford S. Chu C. Review of the Cervical Cancer Burden and Population-Based Cervical Cancer Screening in China Asian Pac. J. Cancer Prev. 2015 16 7401 7407 10.7314/APJCP.2015.16.17.7401 26625735 28. Sripan P. Chitapanarux I. Tharavichitkul E. Muangwong P. Pongnikorn D. Waisri N. Hanpragopsuk C. Maneesai P. Tansiri P. Poungsombat M. Disparities in the change of cervical cancer mortality rate between urban and rural Chiang Mai in the era of universal health care and the Thai national screening program Int. J. Equity Health 2021 20 175 10.1186/s12939-021-01515-1 34325727 29. Locklar L.R.B. Do D.P. Rural-urban differences in HPV testing for cervical cancer screening J. Rural Health 2022 38 409 415 10.1111/jrh.12615 34506669 30. Allen-Leigh B. Uribe-Zúñiga P. León-Maldonado L. Brown B.J. Lörincz A. Salmeron J. Lazcano-Ponce E. Barriers to HPV self-sampling and cytology among low-income indigenous women in rural areas of a middle-income setting: A qualitative study BMC Cancer 2017 17 734 10.1186/s12885-017-3723-5 29121873 31. Yeh P.T. E Kennedy C. De Vuyst H. Narasimhan M. Self-sampling for human papillomavirus (HPV) testing: A systematic review and meta-analysis BMJ Glob. Health 2019 4 e001351 10.1136/bmjgh-2018-001351 32. Shin M.B. Garcia P.J. Saldarriaga E.M. Fiestas J.L. Ásbjörnsdóttir K.H. Iribarren S.J. Barnabas R.V. Gimbel S. Cost of community-based human papillomavirus self-sampling in Peru: A micro-costing study Lancet Reg. Health 2021 8 100160 10.1016/S2214-109X(21)00120-0 33. Morgan K. Azzani M. Khaing S.L. Wong Y.-L. Su T.T. Acceptability of Women Self-Sampling versus Clinician-Collected Samples for HPV DNA Testing: A Systematic Review J. Low. Genit. Tract Dis. 2019 23 193 199 10.1097/LGT.0000000000000476 30933030 34. Ma’Som M. Bhoo-Pathy N. Nasir N.H. Bellinson J. Subramaniam S. Ma Y. Yap S.-H. Goh P.-P. Gravitt P. Woo Y.L. Attitudes and factors affecting acceptability of self-administered cervicovaginal sampling for human papillomavirus (HPV) genotyping as an alternative to Pap testing among multiethnic Malaysian women BMJ Open 2016 6 e011022 10.1136/bmjopen-2015-011022 35. Eche M.T. Vermaak K. Knowledge, attitude and practice of female university students regarding human papillomavirus and self-sampling in KwaZulu-Natal, South Africa: A cross-sectional survey BMC Women’s Health 2022 22 58 10.1186/s12905-022-01634-z 35246111 36. Romanian Public Health Ministry Planul National de Combatere a Cancerului (The National Plan for Fighting against Cancer) Available online: https://www.ms.ro/wp-content/uploads/2022/01/Plan-National-de-Combatere-a-Cancerului.pdf (accessed on 14 March 2022) 37. Dahiya N. Aggarwal K. Singh M.C. Garg S. Kumar R. Knowledge, attitude, and practice regarding the screening of cervical cancer among women in New Delhi, India Tzu Chi Med. J. 2019 31 240 243 10.4103/tcmj.tcmj_145_18 31867252 38. Maree J.E. Kampinda-Banda M. Knowledge and Practices of Cervical Cancer and Its Prevention Among Malawian Women J. Cancer Educ. 2020 35 86 92 10.1007/s13187-018-1443-4 30415315 39. Niyonsenga G. Gishoma D. Sego R. Uwayezu M.G. Nikuze B. Fitch M. Igiraneza P.C. Knowledge, utilization and barriers of cervical cancer screening among women attending selected district hospitals in Kigali-Rwanda Can. Oncol. Nurs. J. 2021 31 266 274 10.5737/23688076313266274 34395829 40. Romli R. Shahabudin S. Saddki N. Mokhtar N. Cervical cancer and pap smear screening: Knowledge, attitude and practice among working women in northern state of Malaysia Med. J. Malays. 2019 74 8 14 41. Lancrajan I.C. Lisencu I.C. Ignat L.F. Trisca R. Topor I.D. Coman M.A. Mocean F. Evaluation of Attitudes and Knowledge on Cervical Cancer among Women from Rural and Urban Areas of Romania J. Women’s Health 2017 6 5 10.4172/2325-9795.1000284 42. Piana L. Leandri F.X. Le Retraite L. Heid P. Tamalet C. Sancho-Garnier H. HPV-Hr detection by home self sampling in women not compliant with pap test for cervical cancer screening. Results of a pilot programme in Bouches-du-Rhône Bull. Cancer 2011 98 723 731 10.1684/bdc.2011.1388 21700548 43. Rosenbaum A. Gage J.C. Alfaro K.M. Ditzian L.R. Maza M. Scarinci I.C. Felix J.C. Castle P.E. Villalta S. Miranda E. Acceptability of self-collected versus provider-collected sampling for HPV DNA testing among women in rural El Salvador Int. J. Gynecol. Obstet. 2014 126 156 160 10.1016/j.ijgo.2014.02.026 44. Todor R.D. Bratucu G. Moga M.A. Candrea A.N. Marceanu L.G. Anastasiu C.V. Challenges in the Prevention of Cervical Cancer in Romania Int. J. Environ. Res. Public Health 2021 18 1721 10.3390/ijerph18041721 33578977 45. Bertucci M. Bonnet E. Satger L. Kreiche A. Chappert J.-L. Loy-Morel S. Segondy M. Daurès J.P. Boulle N. Acceptability of vaginal self-sampling with high-risk human papillomavirus testing for cervical cancer screening: A French questionnaire-based study Women Health 2021 61 83 94 10.1080/03630242.2020.1831683 33106125 46. Montealegre J.R. Landgren R.M. Anderson M.L. Hoxhaj S. Williams S. Robinson D.J. Scheurer M.E. Ramondetta L.M. Acceptability of self-sample human papillomavirus testing among medically underserved women visiting the emergency department Gynecol. Oncol. 2015 138 317 322 10.1016/j.ygyno.2015.05.028 26026733 47. Verdoodt F. Jentschke M. Hillemanns P. Racey C. Snijders P. Arbyn M. Reaching women who do not participate in the regular cervical cancer screening programme by offering self-sampling kits: A systematic review and meta-analysis of randomised trials Eur. J. Cancer 2015 51 2375 2385 10.1016/j.ejca.2015.07.006 26296294 48. Asthana S. Devarapalli P. Labani S. Nagarjuna N. Panchal P. Barriers affecting uptake of cervical cancer screening in low and middle income countries: A systematic review Indian J. Cancer 2018 55 318 326 10.4103/ijc.IJC_253_18 30829264 49. Strohl A.E. Mendoza G. Ghant M.S. Cameron K.A. Simon M.A. Schink J.C. Marsh E.E. Barriers to prevention: Knowledge of HPV, cervical cancer, and HPV vaccinations among African American women Am. J. Obstet. Gynecol. 2014 212 65.e1 65.e5 10.1016/j.ajog.2014.06.059 24983684 50. Wong L.P. Sam I.C. Ethnically diverse female university students’ knowledge and attitudes toward human papillomavirus (HPV), HPV vaccination and cervical cancer Eur. J. Obstet. Gynecol. Reprod. Biol. 2010 148 90 95 10.1016/j.ejogrb.2009.10.002 19910102 51. Lin W. Wang Y. Liu Z. Chen B. Yuan S. Wu B. Gong L. Awareness and attitude towards human papillomavirus and its vaccine among females with and without daughter(s) who participated in cervical cancer screening in Shenzhen, China Trop. Med. Int. Health 2019 24 1054 1063 10.1111/tmi.13283 31264319 52. Masson H. Cervical pap smears and pandemics: The effect of COVID-19 on screening uptake & opportunities to improve Women’s Health 2021 17 3 4 10.1177/17455065211017070 53. Hincapié M.A. Gallego J.C. Gempeler A. Piñeros J.A. Nasner D. Escobar M.F. Implementation and Usefulness of Telemedicine During the COVID-19 Pandemic: A Scoping Review J. Prim. Care Community Health 2020 11 2150132720980612 10.1177/2150132720980612 33300414 54. Kebede M.M. Schuett C. Pischke C.R. The Role of Continuous Glucose Monitoring, Diabetes Smartphone Applications, and Self-Care Behavior in Glycemic Control: Results of a Multi-National Online Survey J. Clin. Med. 2019 8 109 10.3390/jcm8010109 55. Grow A. Perrotta D. Del Fava E. Cimentada J. Rampazzo F. Gil-Clavel S. Zagheni E. Addressing Public Health Emergencies via Facebook Surveys: Advantages, Challenges, and Practical Considerations J. Med. Internet Res. 2020 22 e20653 10.2196/20653 33284782

PMC009xxxxxx/PMC9099631.txt

==== Front Cancers (Basel) Cancers (Basel) cancers Cancers 2072-6694 MDPI 10.3390/cancers14092196 cancers-14-02196 Review Structure–Activity Relationship of Benzofuran Derivatives with Potential Anticancer Activity Farhat Joviana 1 https://orcid.org/0000-0001-9307-3644 Alzyoud Lara 23 Alwahsh Mohammad 456 https://orcid.org/0000-0002-7804-8050 Al-Omari Basem 17* Nakanishi Takeo Academic Editor Sinha Birandra K. Academic Editor 1 Department of Epidemiology and Population Health, College of Medicine and Health Sciences, Khalifa University, Abu Dhabi P.O. Box 127788, United Arab Emirates; joviana.farhat@ku.ac.ae 2 College of Pharmacy, Al Ain University, Abu Dhabi P.O. Box 64141, United Arab Emirates; lara.alzyoud@aau.ac.ae 3 Health and Biomedical Research Center, Al Ain University, Abu Dhabi P.O. Box 64141, United Arab Emirates 4 Leibniz-Institut Für Analytische Wissenschaften-ISAS e.V., 44139 Dortmund, Germany; m.alwahsh@zuj.edu.jo 5 Institute of Pathology and Medical Research Center (ZMF), University Medical Center Mannheim, Heid Elberg University, 68167 Mannheim, Germany 6 Department of Pharmacy, Faculty of Pharmacy, AlZaytoonah University of Jordan, P.O. Box 130, Amman 11733, Jordan 7 KU Research and Data Intelligence Support Center (RDISC) AW 8474000331, Khalifa University of Science and Technology, Abu Dhabi P.O. Box 127788, United Arab Emirates * Correspondence: basem.alomari@ku.ac.ae 28 4 2022 5 2022 14 9 219615 3 2022 25 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Simple Summary Cancer is the leading cause of death worldwide and responsible for killing approximately 10 million people per year. Fused heterocyclic ring systems such as benzofuran have emerged as important scaffolds with many biological properties. Furthermore, derivatives of benzofurans demonstrate a wide range of biological and pharmacological activities, including anticancer properties. The main aim of this review is to highlight and discuss the contribution of benzofuran derivatives as anticancer agents by considering and discussing the chemical structure of 20 different compounds. Evaluating the chemical structure of these compounds will guide future medicinal chemists in designing new drugs for cancer therapy that might give excellent results in in vivo/in vitro applications. Abstract Benzofuran is a heterocyclic compound found naturally in plants and it can also be obtained through synthetic reactions. Multiple physicochemical characteristics and versatile features distinguish benzofuran, and its chemical structure is composed of fused benzene and furan rings. Benzofuran derivatives are essential compounds that hold vital biological activities to design novel therapies with enhanced efficacy compared to conventional treatments. Therefore, medicinal chemists used its core to synthesize new derivatives that can be applied to a variety of disorders. Benzofuran exhibited potential effectiveness in chronic diseases such as hypertension, neurodegenerative and oxidative conditions, and dyslipidemia. In acute infections, benzofuran revealed anti-infective properties against microorganisms like viruses, bacteria, and parasites. In recent years, the complex nature and the number of acquired or resistant cancer cases have been largely increasing. Benzofuran derivatives revealed potential anticancer activity with lower incidence or severity of adverse events normally encountered during chemotherapeutic treatments. This review discusses the structure–activity relationship (SAR) of several benzofuran derivatives in order to elucidate the possible substitution alternatives and structural requirements for a highly potent and selective anticancer activity. benzofuran SAR hybrid benzofurans anticancer activity anticancer potency anticancer selectivity This research received no external funding. ==== Body pmc1. Introduction Several heterocyclic compounds are found in many medications and have formed an essential base for medicinal chemistry research. This is mainly due to heterocyclic compounds’ versatility and distinctive physicochemical features [1]. Among these discovered heterocyclic compounds is benzofuran [2], known as a natural compound originating from plants such as Asteraceae, Rutaceae, Liliaceae, and Cyperaceae [3]. Benzofurans can also emerge from non-natural sources through the dehydrogenation of 2-ethylphenol [4,5]. Structurally, benzofuran is characterized by a distinctive motif consisting of fused benzene and furan rings, as illustrated in Figure 1 [6]. It is suggested that introducing substituents at specified positions within the benzofuran’s core [2] results in new derivatives with unique structural characteristics that may possess an excellent therapeutic value [7]. Therefore, in recent years, derivatives of benzofurans have been frequently used in the development of new drugs [8]. These derivatives exhibited a promising anti-infective activity against bacteria, viruses, and parasites [9,10,11]. For example, in treating neurodegenerative disorders, derivatives of benzofurans revealed potential efficacy in slowing down the progression of Alzheimer’s [12] as well as minimizing Parkinson’s severity [13] and presented potential neuroprotective functions in brain disorders [14]. Furthermore, derivatives of benzofurans have the ability to achieve anti-dyslipidemic and antioxidative effects [15]. Some researchers extended the use of benzofurans’ derivatives to design an effective class of benzofuran-based vasodilators to treat some cardiovascular conditions [16]. In practice, the synthetic derivatives of benzofurans are represented by Amiodarone, which is used in the treatment of ventricular and supraventricular arrhythmias [17], and by Bufuralol as a non-specific β-adrenergic blocker with an affinity for β1- and β2-adrenergic receptors [18,19]. Despite the major progress that has been achieved in research, there are still barriers limiting the effective improvement of therapy, especially in cancer [20]. Nowadays, cancer is known to be the leading cause of death worldwide, accounting for approximately 10 million deaths in 2020 [20]. As cancer cases are constantly increasing, oncology research is investing significant efforts to identify novel, safe and effective therapies to minimize critical side effects caused by conventional treatments [19]. Fused heterocyclic ring systems have emerged as important scaffolds with many biological properties [1,21]. Accordingly, the peculiar structural motif of oxygen-containing heterocycles demonstrates a wide range of biological and pharmacological activities, including anticancer properties [2,18,22]. Earlier structure–activity relationship (SAR) studies of benzofurans’ derivatives found that ester or heterocyclic ring substitutions at the C-2 position were crucial for the compounds’ cytotoxic activity [18]. These modifications have a significant role in influencing the selectivity of these compounds toward cancer cells, which have significant importance given the damage to normal cells caused by the cytotoxic side effects of anticancer therapy. Therefore, this review will discuss the SAR of several anticancer derivatives of benzofurans to determine the critical substitution patterns and structural requirements useful to gain potent and selective anticancer activity. 2. Materials and Methods The aim of this review is to highlight and discuss the contribution of benzofurans’ derivatives as anticancer agents. This review will discuss how the SAR of benzofuran can be used to predict their biological activity and better understand their applications in cancer treatment. A comprehensive electronic literature search of PubMed (MEDLINE), EMBASE, and Web of Science without language or date restrictions was conducted. The keywords related to “benzofuran” OR “derivatives” OR “compounds” OR “agent” OR “class” OR “anti-proliferative” OR “anti-tumor” OR “anti-cancer” OR “anti-neoplastic” OR “novel” OR “new” OR “active” OR “activity” OR “efficacy” OR “agent” OR “potent” OR “cytotoxic” OR “scaffolds” OR “heterocyclic” OR “modeling” OR “experimental” OR “computational” OR “potent” OR “selective” OR “drug design” OR “docking” OR “synthesis” OR “in vitro” were used to search the literature. All figures in this paper were produced by the authors using ACD/ChemSketch, which is a free molecular modeling software used to create images of chemical structures. 3. Benzofuran Derivatives as Anticancer Agents 3.1. Halogenated Derivatives of Benzofuran Some halogen additions into the benzofuran ring, such as bromine, chlorine, or fluorine atoms, have consistently resulted in a significant increase in anticancer activities [23,24,25,26,27,28,29,30]. This is most likely due to the ability of halogens to form a “halogen bond”; an attractive interaction between the electrophilic halogen and a molecule’s nucleophilic sites, which substantially improves the binding affinity [31,32]. For example, a set of seven derivatives (1,1′-(5,6-dimethoxy-3-methyl-1-benzofuran-2,7-diyl) diethanone) were synthesized via standard bromination reaction and condensation with aryl/heteroarylpiperazine [28]. Consequently, those novel halogen derivatives underwent 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) assays against three cancer cell lines (human chronic (K562), and acute (HL60) leukemia cells, human cervical cancer cells (HeLa)), and one normal endothelial cancer cell (HUVEC). Compound 1 (see Figure 2) has a bromine atom attached to the methyl group at the 3-position of the benzofuran ring; and was found to possess remarkable cytotoxic activity against K562, and HL60 leukemia cells with an inhibitory concentration (IC50) value of 5 μM and 0.1 μM (see Table 1), without cytotoxicity towards normal cells. This means that the position of the halogen in the benzofuran ring is a critical determinant of its biological activity [28]. In most cases, the halogen atom is attached to alkyl or acetyl chains rather than directly onto the benzofuran ring [28,33]. This placement does not deter the compound’s cytotoxic activity, as evidenced by electron-rich bromomethyl- or bromophenacyl-substituted benzofuran, which produced pronounced cytotoxic activity in both normal and cancer cells [34,35,36,37]. Selective Polo-like kinase 1 Polo-Box Domain (PLK1 PBD) inhibitor MCC1019 (compound 2) (see Figure 2) is a bromomethyl-substituted benzofuran developed by Abdelfatah and colleagues for the treatment of lung cancer and evaluated via in silico, in vitro, and in vivo models [33]. There are two key interactions between MCC1019 and PLK1 at residues Tryptophan 414 (Trp414) and Histidine 538 (His538), which must be maintained for optimum activity. In vitro testing against lung adenocarcinoma cells (A549) showed that MCC1019 successfully inactivated the serine-threonine kinase (AKT) signaling pathway and inhibited cancerous cell replication, causing a mitotic catastrophe [33,38]. This resulted in achieving selective inhibition of PLK1 PBD with an IC50 of 16.4 μM (see Table 1) [33]. Further in vivo testing using a murine lung cancer model demonstrated a significant anticancer activity by reducing the growth of metastatic lesions in the lung without affecting body weight or vital organ size [33]. The substitution of the N-phenyl ring of the benzofuran with halogen is considered beneficial due to their hydrophobic and electron-donating nature, which enhances benzofuran’ cytotoxic properties [27]. Several studies in the literature have emphasized the influence of the position of the halogen atom on the cytotoxic activity [39,40]. So far, the maximum activities have been recorded when a halogen atom is placed at the para position of the N-phenyl ring [41]. A series of fourteen apoptotic anticancer derivatives were developed using the Allosteric cannabinoid receptor type 1 (CB1) modulator 5-chlorobenzofuran-2-carboxamides [42]. Each compound was then tested to evaluate its antiproliferative activity against the human mammary gland epithelial cell line (MCF-10A) via cell viability assays [43]. Although multiple compounds exhibited excellent antiproliferative activity against tumor cells, compound 3 stood out as the most active derivative (see Figure 2). According to the SAR analysis, the presence of the N-phenethyl carboxamide significantly enhances its antiproliferative activity. This activity was further enhanced by morpholinyl substitution at the para position of the N-phenethyl ring [42]. This explains why compound 3 exhibits similar antiproliferative activity to oral anticancer drug doxorubicin (IC50 of 1.136 μM) [42], (see Figure 2 and Table 1). Interestingly, regardless of the halogen used (e.g., Cl, Br, F, etc.), all the aforementioned halogen-substituted compounds exhibit significant cytotoxicity [28,33,42]. This suggests that, while the nature of the halogen does not impact the cytotoxic activity of the compound, the position of the halogen is of great importance [28]. 3.2. Hybrid Benzofuran as Anticancer Agents Recent studies have highlighted novel classes of hybrid benzofurans’ derivatives like chalcone, triazole, piperazine, and imidazole substituted benzofuran, which have emerged as potent cytotoxic agents [44,45,46,47,48,49]. Utilizing the synergetic cytotoxic effect of heteroatom-substituted benzofuran presents a promising approach for the development of potent anticancer drugs with activities against malignant tumors. 3.2.1. Benzene-Sulfonamide-Based Benzofuran Derivatives Benzene-sulfonamide has become a biologically significant scaffold, with several of its derivatives being used as anticancer and antitumor agents [50]. Benzene-sulfonamide-based benzofuran derivative (5-[benzyl-(4-chlorophenyl)sulfonylamino]-n-[2-(dimethylamino)ethyl]-3-methyl-1-benzofuran-2-carboxamide) represented in compound 4 (see Figure 3) was designed and synthesized to inhibit the hypoxia-inducible factor (HIF-1) pathway [51], which is involved in the carcinogenesis of tumor protein (p53)-independent malignant cancers [51,52,53]. In vitro testing of compound 4 against HCT116 and HCT116−/− p53-null cell lines showed the inhibition of both p53-null cells and p53-mutated cells (see Figure 3 and Table 2). Thus, the addition of a chlorine atom at the para position along with the replacement of the ester group by N containing alkyl chains were major determinants for the antiproliferative activity [51]. 3.2.2. 6-Substituted Hexamethylene Amiloride (6-HMA)-Based Benzofuran Derivatives The urokinase-type plasminogen activator (uPA) system mediates cancer invasion and metastasis through the uPA and its receptor (uPAR) [54]. Targeting uPA is one of the key strategies for combating metastasis in malignant cancers including triple-negative breast cancer [55,56]. In recent in vitro and in vivo studies, high doses of amiloride, a potassium channel blocker, have been found to inhibit uPA proteolytic activity, prompting the search for novel amiloride analogs as uPA inhibitors [57,58,59]. In order to investigate the potential of amiloride-benzofuran derivatives as uPA inhibitors, a series of 6-HMA, 6-N,N-(hexamethylene) amiloride derivatives were synthesized via the Suzuki–Miyaura coupling reactions as potential uPA inhibitors [60]. The addition of a benzofuran group to 6-HMA yields a compound with higher potency, and cytotoxicity (Ki = 183 nM) [61]. In compound 5, the addition of fluorine atom at position 4 of 2-benzofuranyl resulted in a 2-fold increase in potency and inhibitory activity (Ki = 88 nM; IC50 = 0.43 μM) [61] (see Figure 3 and Table 2). Such halogen substitutions at the para position of benzofuran are more likely to form favorable hydrophobic interactions, and therefore are more potent [62,63]. 3.2.3. Quinazolinone- and Imidazolium-Based Benzofuran Derivatives Quinazolinone is an aromatic heterocyclic ring that contains a quinazoline with a carbonyl group [64]. Quinazolinone, like imidazole, is regarded by many as a privileged scaffold with significant anticancer properties [65]. Two of its derivatives, gefitinib, and erlotinib, were introduced to the market as anticancer agents [66]. One study reported the synthesis of a small library of benzofuran derivatives fused to two prominent scaffolds, imidazole and quinazolinone, to create a molecule with a desirable drug-like profile and cytotoxicity [47]. Accordingly, the cell viability and proliferation rates of nine hybrid derivatives (1-[[(1-(benzofuran-2-yl)-2-(quinazolin-4(3H)-one-3-yl)]ethyl-1-yl]-3-methylimidazol-1-ium chloride) or (compounds 6a–i) were tested via MTT assays against human breast cancer (MCF-7) cells [47] (see Figure 3). All derivatives successfully inhibited the growth of cancer cells except compound 6e. Analysis of its structural features suggested that the presence of two halogen-substituted rings coupled with the lack of methoxy substituent on the heterocyclic ring was detrimental to its activity, resulting in no cytotoxicity, as shown in Table 2. This is expected, as the addition of halogen-substituted rings is usually resulting compounds with little to no cytotoxic activity [47]. 3.2.4. Carbohydrazide- and Substituted Benzaldehydes-Based Benzofuran Derivatives The condensation of 3-methyl-2-benzofuran carbohydrazide with various substituted benzaldehydes yielded a set of new benzofuran derivatives, compounds 7a–k (see Figure 3). The eleven benzofuran analogues were screened for potential anticancer activity using the triphenyl blue dye exclusion technique on Erlich ascites carcinoma (EAC) cells [67,68,69]. Out of these eleven benzofuran analogues, derivatives 7a, 7c, 7d, 7f, 7i, and 7j demonstrated the greatest anticancer activity, as evidenced by their high cytotoxic concentration scores (CTC50) shown in Table 3. The SAR results have shown that the presence of the CONH group is necessary for anticancer activity [70]. The addition of phenol and chlorine groups in compounds 7c, 7d, and 7i increased the number of binding interactions formed with the target, resulting in improved anticancer activity (see Figure 3). As for the nitro group in compound 7a, it significantly boosted activity by reducing the melting temperature of DNA in EAC cells [71] (see Figure 3). Interestingly, the phenolic hydroxy group of benzofuran was found to be crucial for modulating anticancer activity. The presence of a hydrogen donating group promotes the formation of favorable interactions with the target, hence inducing its cytotoxic properties [17,44,70,72]. 3.2.5. Trimethoxyacetophenone-Based Benzofuran Derivatives Combretastatin A-4 (CA-A4), which is a naturally occurring chemical, isolated from the roots of Combretum Caffrum, has recently attracted considerable attention for its antitumor and antimitotic activity [2,73]. The CA-A4 analogue (compound 8) consists of trimethoxy acetophenone and a benzofuran core, and it has an IC5O of 0.43 μM (see Figure 3). Subsequently, Flynn and colleagues used compound 8 as the lead compound for the SAR-guided design of novel tubulin polymerization inhibitors [74]. The results demonstrated that the introduction of C7-OH and a C2-substituent, as seen in compound 8a (BNC105), improved its anticancer activity with a tubulin inhibition IC50 of 0.8 μM [74] (see Figure 3 and Table 2). Notably, the observed antimitotic activity is approximately tenfold stronger than that of the lead compound. The presence of a hydrogen bond donor (hydroxyl) at C7 adds to the pharmacophore’s interactions; as for the C-2 substituent, it maintains conformational bias, ensuring that the compound remains in the cis-conformation. Further efforts to enhance the activity were made by formulating a prodrug, disoduimphosphase ester derivative compound 8b (BNC105P), which is rapidly cleaved in vivo to return to its active state compound 8a [74] (see Figure 3). When tested in vitro, the prodrug produced tenfold stronger antitumor activity, eightyfold better selectivity, and a fivefold longer half-life than the free drug [74]. This means that adjusting the formulation is equally important to modifying the substituents on the compound in terms of increasing antitumor activity. 3.2.6. N-Methylpiperidine-Based Benzofuran Derivatives The hallmark of many cancers is the activation and dysregulation of the AKT/mammalian target of the rapamycin (mTOR) pathway, making it promising for drug discovery [75,76,77]. A series of mammalian targets of the rapamycin complex 1 (mTORC1) protein complex inhibitors were synthesized by performing different isosteric replacements on the lead compound ChemBridge 5219657, which was identified through high-throughput screening (HTS) [78,79]. Derivative 1-((2-(2-(benzyloxy) phenyl)-5-methoxybenzofuran-4-yl) methyl)-n, n-dimethylpiperidin-4-amine (compound 9) was found to exhibit the greatest cytotoxic activity against head and neck (SQ20B) cancer cell line with an IC50 value of 0.46 μM (see Figure 3 and Table 2). The replacement of the phenolic hydroxyl group with another H-bond donor like triflylamide conserved the cytotoxicity of the compound, whereas replacement with an H-bond acceptor altered its activity [78,79]. Whilst the introduction of triflate ester (a group that cannot donate or accept an H-bond) was well tolerated, the absolute removal of the phenolic hydroxy diminished the cytotoxicity of the compounds. Furthermore, substituting the dimethylamine and benzyl groups, with bulkier amine-containing groups such as 4-piperidino-piperidine, enhanced the cytotoxicity of the compounds [79]. Hypoxic microenvironments accelerate tumor metastasis and progression in solid tumor cancers, including pancreatic ductal adenocarcinoma (PADC) [80,81]. With the HIF-1 pathway being a target of interest, a small library of thirty-two benzofuran-derived HIF-1 inhibitors based on compound 10 were developed [82] (see Figure 3). MTT assays have found that derivatives 10a and 10b exhibit similar activity, but derivative 5-(4-bromo-N-(4-bromobenzyl) phenylsulfonamido)-3-methyl-N-(1-methylpiperidin-4-yl) benzofuran-2-carboxamide (compound 10b) emerged as the most promising candidate due to its significant antiproliferative activity and selective inhibition of HIF-1 pathway [83] (see Figure 3 and Table 2). The inclusion of hydrophilic heteroatom-containing groups, like piperidine, on the benzofuran ring significantly improved the compound’s physicochemical properties [82]. Additionally, the para-substituted halogen on the phenylsulfonyl- and N-containing alkyl chains contributed to the resultant antiproliferative activity [83]. cancers-14-02196-t002_Table 2 Table 2 In vitro cytotoxicity of hybrid benzofuran derivatives 4–20 against multiple cancer cell lines. Compound Cell Line IC50 (μM) References 4 HCT116 (p53-null) 2.91 [51] MDA-MB-435s (p53-mutated) 4.71 5 uPA 0.43 [61] 6a MCF-7 7.70 [47] 6b MCF-7 9.14 6c MCF-7 1.00 6d MCF-7 20.58 6e MCF-7 inactive 6f MCF-7 73.26 6g MCF-7 1.00 6h MCF-7 100 6i MCF-7 0.57 8 Tubulin 0.43 [74] 8a Tubulin 0.76 8b Tubulin ND 9 SQ20B 0.46 [79] 10 ND ND [82] 10a PANC-1 1.52 BxPC3 1.08 HCT116 2.39 HCT116(p53−/−) 1.66 MCF-7 2.84 A549 2.98 MDA-MB-231 3.73 10b PANC-1 1.07 BxPC3 0.65 HCT116 1.81 HCT116(p53−/−) 1.61 MCF-7 2.39 A549 2.68 MDA-MB-231 1.90 11a A549 0.12 [84] Hela 26.32 SGC7901 2.75 11b A549 6.25 Hela 18.71 SGC7901 36.23 11c A549 8.11 Hela 28.74 SGC7901 >40 11d A549 34.13 Hela 12.68 SGC7901 7.45 12 HT-1080 8.86 [85] 13a HL60 2.34 [86] SMMC-7721 2.63 A549 4.5 MCF-7 3.24 SW480 3.61 13b HL60 0.64 SMMC-7721 2.10 A549 3.34 MCF-7 4.78 SW480 5.56 13c HL60 0.61 [86] SMMC-7721 2.30 A549 5.35 MCF-7 3.03 SW480 3.14 13d HL60 0.08 SMMC-7721 0.52 A549 0.55 MCF-7 0.51 SW480 0.47 14a ND ND [87] 14b ND ND 14c ND ND 14d ND ND 15a MCF-7 1.90 [88] A549 2.38 Colo-205 2.11 A2780 1.05 15b MCF-7 3.90 A549 4.17 Colo-205 ND A2780 ND 15c MCF-7 0.011 A549 0.073 Colo-205 0.10 A2780 0.034 15d MCF-7 7.23 A549 6.91 Colo-205 2.84 A2780 10.2 15e MCF-7 12.5 A549 5.34 Colo-205 ND A2780 9.55 15f MCF-7 3.16 A549 ND Colo-205 7.10 A2780 8.64 15g MCF-7 10.76 A549 19.42 Colo-205 ND A2780 ND 15h MCF-7 1.55 A549 1.93 Colo-205 1.28 A2780 2.13 15i MCF-7 0.21 A549 0.43 Colo-205 0.17 A2780 1.84 15j MCF-7 0.14 A549 0.25 Colo-205 0.12 A2780 0.33 16 K562 ND [89] 17a K562 ND 17b K562 ND 18 A549 9 [90] MCF-7 2 PC-3 10 19 A549 6.3 [49] 20a A549 10.9 20b A549 Inactive The definitions of all abbreviations are provided in a list at the end of the manuscript. cancers-14-02196-t003_Table 3 Table 3 In vitro cytotoxicity inhibition of hybrid benzofuran derivatives 7a–k against EAC cancer cell lines. Compound CTC50 (μM/mL) Reference 7a 35.5 [71] 7b 472 7c 33.5 7d 33.75 7e 255 7f 43 7g 280 7h 365 7i 34 7j 49 7k 478 The definitions of all abbreviations are provided in a list at the end of the manuscript. 3.2.7. Piperazine-Based Benzofuran Derivatives Piperazine is a six-membered ring containing two nitrogen atoms at opposite positions [91]. In vitro and/or in vivo studies have shown that several piperazine compounds revealed significant activities against a variety of cancers cell lines [92]. Given this, a hybrid of 2-benzoyl benzofuran with N-aryl piperazine linker is considered to be more biologically active than unsubstituted benzofuran [18,40,84]. Benzofuran piperazine hybrids were designed, synthesized, and tested via MTT assays against lung cancer (A549), human cervical carcinoma (Hela), and colonic cancer (SGC7901) cell lines [84]. Derivatives bearing keto-substituent on the piperazine ring (compounds 11a–d) exhibited the most cytotoxic activity against cancer cells [84] (see Figure 4). Similarly, the addition of an electron-withdrawing group or halide such as fluoro-, chloro-, and cyano- at the para position of benzene in compounds 11b, 11c, and 11d was beneficial for anticancer activity [84] (see Figure 4). Furthermore, compound 11a showed promising activity and selectivity to lung (A549) and colonic cancer (SGC7901) cell lines with IC50 values of 0.12 μM and 2.75 μM, respectively [84] (see Figure 4 and Table 2). 3.2.8. Neolignans-Based Benzofuran Derivatives Naturally occurring dihydrobenzofuran neolignans are often found in high concentrations in aerial parts of plants like Mappianthus iodoies, Dorstenia kameruniana, and Aristolochia fordiana [45,93,94,95,96,97]. Many neolignans have shown considerable activity against a variety of cancers cell lines [98]. Neolignan-based benzofurans’ derivatives are expected to benefit from the synergistic cytotoxic effect of both molecules. Thus, eight dihydro benzofuran neolignans analogs were isolated from the seeds of crataegus pinnatifida [85]. In vitro testing recognized 7R,8S-balanophonin (compound 12) as the most potent analogue, with stronger inhibitory activity against HT-1080 cancer cells than positive control 5-fluorouracil (5FU) (IC50 = 35.62 μM) (see Figure 4 and Table 2). The SAR studies of hybrid dihydrobenzofuran neolignans revealed that the presence of a double bond at C-7′/C-8′ next to the aromatic ring is vital for cytotoxicity and that the reduction of the double bond can reduce the activity by tenfold or greater [85]. 3.2.9. Imidazole-Based Benzofuran Derivatives Imidazoles are five-membered, nitrogen-containing heterocycles with significant anticancer activity against a variety of biological targets [99]. However, there is no consensus surrounding the cytotoxic activity of benzofuran-imidazole derivatives [48,100,101]. It has been reported that the addition of an imidazole ring to the benzofuran produced compounds with weak cytotoxic properties [36]. Therefore, to yield optimal benzofuran imidazole hybrids, some modifications must be implemented. The 2-benzylbenzofuran ring is altered to 2-alkylbenzofuran to improve both the steric effect and charge distribution of the compound [100,101]. Then, electron-rich groups like 2-bromophenacyl, phenacyl, and napthylacyl- are substituted onto the imidazole ring, preferably into the 3-positon [102]. These alterations are crucial to ensure a maximal cytotoxic activity against cancer cells. Similar findings were observed in 2-phenyl-3-alkylbenzofuran imidazole/triazole hybrids (compounds 13a–d) (see Figure 4) [86]. These highly potent anticancer derivatives often include a 2-ethyl-imidazole or benzimidazole ring with a 2-bromobenzyl or napthylacyl substituent at the 3-position of the imidazole ring, all of which are important groups in modulating antitumor activity [103]. Among these compounds, compound 13d has shown the strongest inhibitory activity and selectivity towards breast cancer (MCF-7) and colon cancer (SW480) cells, with IC50 values ranging from 0.08 to 0.55 μM [86] (see Figure 4 and Table 2). 3.2.10. Pyrazole-Based Benzofuran Derivatives The non-receptor tyrosine kinase (c-Src) has been identified as a promising target for cancer treatment, sparking the interest of researchers [104,105,106]. Pyrazole is a five-membered aromatic heterocyclic ring containing two neighboring nitrogen atoms [107]. Pyrazole derivatives have previously demonstrated antitumor activity against numerous types of cancer [108]. In an effort to discover novel potent c-Src inhibitors as anticancer agents, a set of benzofuran-pyrazoles hybrids containing chalcones, pyrazoline, isoxazole, and thiopyrimidine substituents were in vitro-synthesized and tested for their anticancer activity [87]. Compounds 14c and 14d, which consist of 3-furano-N-acetylpyrazoline and 3-furano-isoxazole rings, respectively, exhibited remarkable and broad-spectrum anticancer activity (see Figure 4). Incorporating acetyl, an electron-withdrawing group, into the N-1 of the pyrazoline ring appears to be essential for antiproliferative activity. Hence, the derivatives lacking the acetyl group such as compound 14a exhibited weak anticancer activity in-vitro [87] (see Figure 4). Increasing the size of the hetero-ring systems attached to the parent core resulted in weak-to-moderate antiproliferative potency [21]. Among all derivatives, compound 14b containing 3-pyrrolo-N-acetylpyrazoline demonstrated significant antiproliferative and anticancer activity against leukemia, lung cancer, colon cancer, central nervous system (CNS) cancer, melanoma, ovarian cancer, breast cancer, and renal cancer cells [87] (see Figure 4). Enzyme assays of compound 14c detected significant inhibition of Src and zeta-chain-associated protein (ZAP-70) kinases [87] (see Figure 4). Overall, the potent antitumor activity and favorable absorption, distribution, metabolism, and excretion (ADME) characteristics of compound 14b make it a viable candidate worthy of further investigation and modifications (Figure 4). 3.2.11. Imidazopyridine-Based Benzofuran Derivatives Imidazopyridine is fused bicyclic heterocycles that are synthesized by several strategies such as condensation, oxidative coupling, tandem reaction, etc. [109]. A series of imidazopyridine-substituted benzofurans (compounds 15a–j) were derived from sulfonamides, and subsequently underwent MTT assays to evaluate their in vitro cytotoxicity against human cancer cells [110] (see Figure 4). Moreover, compounds 15a, 15c, 15h, 15i, and 15j were found to produce considerable anticancer activity against tested cell lines [88] (see Figure 4). Among these, derivative 15c, with the greatest cytotoxicity, significantly inhibited the growth of breast (MCF-7), lung (A549), colon (Colo-205), and ovarian (A2780) cancer cell lines with IC50 values of 0.011, 0.073, 0.10, and 0.034 μM [88], respectively (see Figure 4 and Table 2). The SAR has shown that the addition of electron-positive groups at the para position on the phenyl group significantly improved anticancer activity, regardless if it’s a strong group like 4-methoxy (compound 15c) or a weak group like 4-methyl (compound 15h). On the other hand, substitution with electron-withdrawing groups like chloro (15d), bromo (15e), nitro (15f), and cyano (15g) resulted in significant drop-in activity (Figure 4 and Table 2). Interestingly, compound 15a lacked any phenyl ring substituents but still maintained good anticancer activity [88] (see Figure 4). Furthermore, replacing the aryl ring with a hetero-aromatic ring such as 2,6-dimethylpyridine (compound 15i) or 4,5-dimethylthiophene (compound 15j) rings was more beneficial for anticancer activity than keeping the aryl ring (compound 15a and 15h) [88] (see Figure 4). 3.2.12. Aurones-Chromone- and -Coumarin-Based Benzofuran Derivatives Flavonoids, aurones, chromones, and coumarins are abundantly found in plants, fungi, and bacteria [109]. These natural products are capable of modulating a wide range of biological pathways and achieving selective anticancer activity with few side effects [109,111,112]. Yet, only a limited number of hybrids with aurone-chromone, -coumarin fused heterocycles have been reported. Therefore, a series of 26 hybrid compounds between benzofuran core of aurones-chromone and -coumarin were designed [89]. This combination takes advantage of the potential synergistic anticancer effect of these flavonoids [113,114,115]. These derivatives were then evaluated for their anticancer activity against a panel of human leukemia cells (K562) at different concentrations. In particular, compounds 16, 17a, and 17b were able to induce around 24% apoptosis [89] (see Figure 4). Interestingly, the potency of the compounds is unaffected by different substitutions of the chromone [29]. Furthermore, it appears that exchanging the benzofuranone or methylbenzofuranone moieties with napthofuranone induces a stronger apoptotic effect [89]. In order to understand the pro-apoptotic properties of these benzofuran–coumarin derivatives, (Z)-7-methoxy-4-[(6-methyl-3-oxobenzofuran-2(3H)-ylidene) methyl]-2H-coumarine (compound 17a) was compared to 7-methoxy-coumarin-4-aldehyde and (Z)-2-(4-methoxybenzylidene)-6-methylbenzofuran-3(2H)-one by testing them in K526 cells at doses ranging from 5 to 100 μM [89]. The results demonstrated that compound 17a produced the strongest apoptosis induction at higher doses, outperforming both unsubstituted benzofuran and coumarins [89]. These findings imply that coupling aurone-like benzofuran with a chromone or coumarin can yield novel compounds with more potent pro-apoptotic properties compared to unconjugated benzofuran. 3.2.13. Chalcone-Based Benzofuran Derivatives Many naturally occurring compounds are derived from plants, including the simple chalcone scaffold [116]. These structures are simple to synthesize, allowing for the chalcones to be incorporated into several derivatives with a wide range of biological activities [117]. Moreover, chalcones have been recognized as a valuable scaffold with potent anticancer activity [118]. Thus, a synergistic cytotoxic effect could be observed after the hybridization of chalcones and benzofuran, yielding compounds that are used to treat malignant tumors [18,29,44]. Encouraged by the anticancer potential of chalcones, a set of 1-(7-ethoxy-1-benzofuran-2-yl) substituted chalcone derivatives via the base-catalyzed Claisen-Schmidt reaction was synthesized [46,90]. All derivatives were then tested by sulforhodamine B (SRB) and adenosine 5′-triphosphate (ATP) cell viability assays, against breast (MCF-7), non-small-cell lung (A549), and prostate (PC-3) cancer cell lines [43]. The best cytotoxic activity was observed in chalcone derivative compound 18, with IC50 values ranging from 2 to 10 μM [90] (see Figure 4 and Table 2). Interestingly, compound 18 showed selective cytotoxicity toward human breast cancer cell line (MC-7), while being non-toxic towards normal breast cancer cells (MRC5). Furthermore, compound 18 was successful in inducing apoptosis in cancer cells while maintaining a promising safety profile, indicating that hybrid benzofuran chalcones have greater cytotoxic activity compared to unsubstituted benzofuran [90]. 3.2.14. Oxadiazole- and Triazole-Based Benzofuran Derivatives Oxadiazoles and triazoles are nitrogen-oxygen and nitrogen-containing five-membered heterocyclic aromatic rings commonly hybridized with other anticancer scaffolds, such as benzofuran [119,120,121]. These hybrid derivatives have shown substantial anticancer potential and play essential roles in cancer management [122,123]. Hence, ultrasound- and microwave-assisted green synthetic protocols were implemented for synthesizing a set of 15 benzofurans–oxadiazole and –triazole. Then, those compounds were evaluated for the anticancer activity against the lung cancer cell line (A549) [49]. Compound 19, benzofuran-oxadiazole hybrid, was reported as the most potent anticancer derivative, with cell viability of 27.49 μM and IC50 of 6.3 μM, outperforming reference drugs crizotinib and cisplatin, which had IC50 of 8.54 and 3.88 μM, respectively [49]. The enhanced anticancer activity is believed to be due to meta methoxy or para ethoxy substitutions on the phenyl ring of N-(substituted-phenyl)-acetamide (see Figure 4). Although benzofuran triazole derivatives 20a and b exhibit excellent thrombolysis activity and minimal toxicity, they did not demonstrate strong anticancer activity against A549 cancer cells (see Figure 4 and Table 2). The presence of two adjacent electron-withdrawing chloro groups at the ortho and para positions of the phenyl ring in compound 20a was detrimental to its anticancer activity [49]. Similarly, in compound 20b, the addition of two adjacent methyl groups on the ortho and para positions of the phenyl ring yielded an inactive compound [49]. These SAR studies highlight the possible positive and negative impacts of structural modifications to oxadiazole- and triazole-benzofuran derivatives as anticancer drug candidates. 3.3. Cytotoxicity of Benzofurans’ Derivatives against Selected Cancer Cell Lines Many of the compounds presented have been tested against the same cancer cell lines, and while they show high cytotoxicity, it is notable how different substitutions can influence the compound’s cytotoxicity against selected cancer cell lines. Imidazopyridine-benzofuran analogs bearing electron-positive groups at the 4-position on the phenyl group, for example, have significantly improved anticancer activity against various cancer cell lines (A549, MCF-7, HL-60, SW480, A2780, and Colo-205) [88]. The most effective modifications to the cytotoxicity of MCF-7 cell lines were quinazolinone and Imidazolium, Imidazole, and Chalcone-based benzofuran compounds [47,86,90]. A halogen atom attached to the methyl group at the 3-position of the benzofuran ring promotes cytotoxicity toward both A549 and HL60 cell lines [28,33]. More specifically, the presence of oxadiazole and triazole-benzofuran hybrids further boosts the anticancer activity of A549 cells [49]. The majority of the novel hybrids demonstrated potential anticancer agents against specific cancer cell lines, while maintaining a remarkable safety profile against normal cells. Most of the novel hybrids demonstrated potential anticancer agents against specific cancer cell lines while maintaining a remarkable safety profile against normal cells. Hence, benzofuran derivatives have the potential to be developed as novel therapeutic agents given their recent experimental findings and documented selectivity against cancer cells. 4. Conclusions This review suggests benzofuran as a versatile scaffold with significant anticancer activity on various human cells such as breast, lung, and prostate cancer. Understanding the SAR of benzofurans’ derivatives facilitates the design and development of novel, safe, and potent in vitro therapeutic options in cancer. Therefore, this may provide a more robust assessment of anticancer activities before considering in vivo studies. The anticancer activity of benzofuran scaffolds is dependent on the type of substituent present and is frequently multifactorial. Furthermore, hybrid structures bearing benzofuran moiety stand out as highly potent anticancer agents. They utilize the functionalization or structural configuration of the conjugate molecule. This review recommends that studying the chemical structure of these compounds will result in anticancer agents that limit tumor progression with minimal adverse effects. Therefore, this could potentially have an impact on improving patients’ adherence to medication and subsequently disease prognosis. Author Contributions Conceptualization, J.F., L.A. and M.A.; methodology, J.F., L.A., M.A. and B.A.-O.; validation, J.F., L.A., M.A. and B.A.-O.; formal analysis, J.F., L.A. and M.A.; investigation, J.F., L.A., M.A. and B.A.-O.; resources, J.F., L.A. and M.A.; data curation, J.F., L.A. and M.A.; writing—original draft preparation, J.F., L.A. and M.A.; writing—review and editing, B.A.-O.; visualization, L.A.; supervision, B.A.-O.; project administration, J.F. All authors have read and agreed to the published version of the manuscript. Conflicts of Interest The authors declare no conflict of interest. Abbreviations 5FU 5-fluorouracil 6-HMA 6-substituted hexamethylene amiloride 6-HMA 6- N, N-hexamethylene ADME Absorption, distribution, metabolism, and excretion AKT signaling Serine–threonine kinase signaling ATP Adenosine 5′-triphosphate A549 Hypotriploid alveolar basal epithelial cell lines A2780 Ovarian cancer cell line BNC105P Disoduimphosphase ester derivative compound 8b BxPC3 Human pancreatic cancer cell lines CB1 5-chlorobenzofuran-2-carboxamides CNS Central nervous system CB1 modulator Cannabinoid receptor type 1 modulator CA-A4 Combretastatin A-4 CTC50 cytotoxic concentration scores Colo-205 Colon cancer cell lines EAC Erlich ascites carcinoma cells HT-29 Human colorectal adenocarcinoma cell lines HCT116 Human colorectal carcinoma cell lines HIF pathway Hypoxia-inducible factor pathway HUVEC normal endothelial cancer cell lines HT-1080 Fibrosarcoma cell lines HL60 Human acute leukemia cells HeLa human cervical cancer cells HTS High-Throughput Screening IC50 half-maximal inhibitory concentration Ki Dissociation constant K562 Human leukemia cell lines MCF-7 human breast cancer cells MTT 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide mTOR pathway mammalian target of the rapamycin mTORC1 Mammalian target of rapamycin complex 1 MDA-MB-23 Metastatic adenocarcinoma cell lines MCF-10A human mammary gland epithelial cell line MRC5 normal breast cancer cells NA Not Applicable Panc-1 Human pancreatic cancer cell lines P53 Tumor protein PADC pancreatic ductal adenocarcinoma cell lines PC-3 Prostate cancer cell lines PLK1 PBD inhibitor Polo-like kinase 1 Polo-Box Domain inhibitor SAR Structure–activity relationship SRB sulforhodamine B SQ20B head and neck cancer cell lines SMMC-7721 Hepatocellular carcinoma cell lines SGC7901 Colonic cancer cell lines Src Non-receptor tyrosine kinase protein SW480 Colon cancer cell lines TNBC Triple-negative breast cancer cell lines uPA urokinase-type plasminogen activator uPAR urokinase-type plasminogen activator receptor ZAP-70 kinases Zeta-chain-associated protein kinase 70 Figure 1 Chemical structure of benzofuran. Figure 2 Chemical structure of halogenated derivatives of benzofuran 1–3. Figure 3 Chemical structures of anticancer hybrid benzofuran 4–10. Figure 4 Chemical structures of anticancer hybrid benzofuran 11–20. cancers-14-02196-t001_Table 1 Table 1 In vitro inhibitory activities of halogenated derivatives of benzofuran against multiple cancer cell lines. Compound Cell Line IC50, μM GI50, μM Reference 1 K562 5 ND [28] HL60 0.1 ND 2 PLK1 PBD 16.4 ND [33] 3 A-549 ND 1.8 [42] MCF-7 ND 0.7 Panc-1 ND 1.3 HT-29 ND 1.6 The definitions of all abbreviations are provided in a list at the end of the manuscript. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Chand K. Hiremathad A. Singh M. Santos M.A. Keri R.S. A Review on Antioxidant Potential of Bioactive Heterocycle Benzofuran: Natural and Synthetic Derivatives Pharmacol. Rep. 2017 69 281 295 10.1016/j.pharep.2016.11.007 28171830 2. Radadiya A. Shah A. Bioactive Benzofuran Derivatives: An Insight on Lead Developments, Radioligands and Advances of the Last Decade Eur. J. Med. Chem. 2015 97 356 376 10.1016/j.ejmech.2015.01.021 25703339 3. Proksch P. Rodriguez E. Chromenes and Benzofurans of the Asteraceae, Their Chemistry and Biological Significance Phytochemistry 1983 22 2335 2348 10.1016/0031-9422(83)80118-6 4. Heravi M.M. Zadsirjan V. Hamidi H. Tabar Amiri P.H. Total Synthesis of Natural Products Containing Benzofuran Rings RSC Adv. 2017 7 24470 24521 10.1039/C7RA03551A 5. Collin G. Höke H. Benzofurans Ullmann’s Encyclopedia of Industrial Chemistry Wiley Hoboken, NJ, USA 2007 10.1002/14356007.L03_L01 6. Jung J.W. Park J.H. Seo K.H. Oh E.J. Lee D.Y. Lim D.W. Han D. Song M.C. Baek N.I. New Hydroxy Fatty Acid from the Root Bark of Morus alba L. J. Korean Soc. Appl. Biol. Chem. 2015 58 541 543 10.1007/s13765-015-0071-5 7. Modell A.E. Blosser S.L. Arora P.S. Systematic Targeting of Protein–Protein Interactions Trends Pharmacol. Sci. 2016 37 702 713 10.1016/j.tips.2016.05.008 27267699 8. Xu Z. Zhao S. Lv Z. Feng L. Wang Y. Zhang F. Bai L. Deng J. Benzofuran Derivatives and Their Anti-Tubercular, Anti-Bacterial Activities Eur. J. Med. Chem. 2019 162 266 276 10.1016/j.ejmech.2018.11.025 30448416 9. Nevagi R.J. Dighe S.N. Dighe S.N. Biological and Medicinal Significance of Benzofuran Eur. J. Med. Chem. 2015 97 561 581 10.1016/j.ejmech.2014.10.085 26015069 10. Hiremathad A. Patil M.R. Chethana K.R. Chand K. Santos M.A. Keri R.S. Benzofuran: An Emerging Scaffold for Antimicrobial Agents RSC Adv. 2015 5 96809 96828 10.1039/C5RA20658H 11. Zhong M. Peng E. Huang N. Huang Q. Huq A. Lau M. Colonno R. Li L. Discovery of Novel Potent HCV NS5B Polymerase Non-Nucleoside Inhibitors Bearing a Fused Benzofuran Scaffold Bioorg. Med. Chem. Lett. 2018 28 963 968 10.1016/j.bmcl.2018.01.029 29422387 12. Goyal D. Kaur A. Goyal B. Benzofuran and Indole: Promising Scaffolds for Drug Development in Alzheimer’s Disease ChemMedChem 2018 13 1275 1299 10.1002/cmdc.201800156 29742314 13. Subair T.I. Soremekun O.S. Olotu F.A. Soliman M.E.S. Therapeutic Path to Double Knockout: Investigating the Selective Dual-Inhibitory Mechanisms of Adenosine Receptors A1 and A2 by a Novel Methoxy-Substituted Benzofuran Derivative in the Treatment of Parkinson’s Disease Cell Biochem. Biophys. 2021 79 25 36 10.1007/s12013-020-00957-8 33222095 14. Wakabayashi T. Tokunaga N. Tokumaru K. Ohra T. Koyama N. Hayashi S. Yamada R. Shirasaki M. Inui Y. Tsukamoto T. Discovery of Benzofuran Derivatives That Collaborate with Insulin-Like Growth Factor 1 (IGF-1) to Promote Neuroprotection J. Med. Chem. 2016 59 5109 5114 10.1021/acs.jmedchem.6b00191 27163512 15. Sashidhara K.V. Modukuri R.K. Sonkar R. Rao K.B. Bhatia G. Hybrid Benzofuran-Bisindole Derivatives: New Prototypes with Promising Anti-Hyperlipidemic Activities Eur. J. Med. Chem. 2013 68 38 46 10.1016/j.ejmech.2013.07.009 23954239 16. Khalifa N.M. Srour A.M. El-Karim S.S.A. Saleh D.O. Al-Omar M.A. Synthesis and 2D-QSAR Study of Active Benzofuran-Based Vasodilators Molecules 2017 22 1820 10.3390/molecules22111820 17. Dawood K.M. An Update on Benzofuran Inhibitors: A Patent Review Expert Opin. Ther. Pat. 2019 29 841 870 10.1080/13543776.2019.1673727 31560232 18. Miao Y.H. Hu Y.H. Yang J. Liu T. Sun J. Wang X.J. Natural Source, Bioactivity and Synthesis of Benzofuran Derivatives RSC Adv. 2019 9 27510 27540 10.1039/C9RA04917G 35529241 19. Pucci C. Martinelli C. Ciofani G. Innovative Approaches for Cancer Treatment: Current Perspectives and New Challenges Ecancermedicalscience 2019 13 961 10.3332/ecancer.2019.961 31537986 20. Ferlay J. Ervik M. Lam F. Colombet M. Mery L. Piñeros M. Znaor A. Soerjomataram I. Bray F. Global Cancer Observatory: Cancer Today Available online: https://gco.iarc.fr/today (accessed on 11 December 2021) 21. Martins P. Jesus J. Santos S. Raposo L.R. Roma-Rodrigues C. Baptista P.V. Fernandes A.R. Heterocyclic Anticancer Compounds: Recent Advances and the Paradigm Shift towards the Use of Nanomedicine’s Tool Box Molecules 2015 20 16852 16891 10.3390/molecules200916852 26389876 22. Khodarahmi G. Asadi P. Hassanzadeh F. Khodarahmi E. Benzofuran as a Promising Scaffold for the Synthesis of Antimicrobial and Antibreast Cancer Agents: A Review J. Res. Med. Sci. 2015 20 1094 1104 10.4103/1735-1995.172835 26941815 23. Gao H. Zhang X. Pu X.J. Zheng X. Liu B. Rao G.X. Wan C.P. Mao Z.W. 2-Benzoylbenzofuran Derivatives Possessing Piperazine Linker as Anticancer Agents Bioorg. Med. Chem. Lett. 2019 29 806 810 10.1016/j.bmcl.2019.01.025 30709651 24. Negalurmath V.S. Boda S.K. Kotresh O. Anantha Lakshmi P.V. Basanagouda M. Benzofuran-Oxadiazole Hybrids: Design, Synthesis, Antitubercular Activity and Molecular Docking Studies Chem. Data Collect. 2019 19 100178 10.1016/j.cdc.2019.100178 25. Li W. Yang Z.H. Hu A.X. Yan X.W. Ding N. Ye J. Design, Synthesis, and Antitumor Activity of (E,Z)-1-(Dihydrobenzofuran-5-Yl)-3-Phenyl-2-(1,2,4-Triazol-1-Yl)-2-Propen-1-Ones Chem. Biol. Drug Des. 2015 86 1339 1350 10.1111/cbdd.12601 26041467 26. Jin L.P. Xie Q. Huang E.F. Wang L. Zhang B.Q. Hu J.S. Wan D.C.C. Jin Z. Hu C. Design, Synthesis, and Biological Activity of a Novel Series of Benzofuran Derivatives against Oestrogen Receptor-Dependent Breast Cancer Cell Lines Bioorg. Chem. 2020 95 103566 10.1016/j.bioorg.2020.103566 31935604 27. Lu B. Shen X. Zhang L. Liu D. Zhang C. Cao J. Shen R. Zhang J. Wang D. Wan H. Discovery of EBI-2511: A Highly Potent and Orally Active EZH2 Inhibitor for the Treatment of Non-Hodgkin’s Lymphoma ACS Med. Chem. Lett. 2018 9 98 102 10.1021/acsmedchemlett.7b00437 29456795 28. Krawiecka M. Kuran B. Kossakowski J. Cieslak M. Kazmierczak- Baranska J. Krolewska K. Nawrot B. Synthesis and Cytotoxic Properties of Halogen and Aryl-/Heteroarylpiperazinyl Derivatives of Benzofurans Anticancer Agents Med. Chem. 2014 15 115 121 10.2174/187152061501141204124709 29. Amin K.M. Syam Y.M. Anwar M.M. Ali H.I. Abdel-Ghani T.M. Serry A.M. Synthesis and Molecular Docking Study of New Benzofuran and Furo[3,2-g]Chromone-Based Cytotoxic Agents against Breast Cancer and P38α MAP Kinase Inhibitors Bioorg. Chem. 2018 76 487 500 10.1016/j.bioorg.2017.12.029 29310080 30. Kang T.S. Wang W. Zhong H.J. Dong Z.Z. Huang Q. Mok S.W.F. Leung C.H. Wong V.K.W. Ma D.L. An Anti-Prostate Cancer Benzofuran-Conjugated Iridium(III) Complex as a Dual Inhibitor of STAT3 and NF-ΚB Cancer Lett. 2017 396 76 84 10.1016/j.canlet.2017.03.016 28323031 31. Ford M.C. Ho P.S. Computational Tools To Model Halogen Bonds in Medicinal Chemistry J. Med. Chem. 2015 59 1655 1670 10.1021/acs.jmedchem.5b00997 26465079 32. Wilcken R. Zimmermann M.O. Lange A. Joerger A.C. Boeckler F.M. Principles and Applications of Halogen Bonding in Medicinal Chemistry and Chemical Biology J. Med. Chem. 2013 56 1363 1388 10.1021/jm3012068 23145854 33. Abdelfatah S. Berg A. Huang Q. Yang L.J. Hamdoun S. Klinger A. Greten H.J. Fleischer E. Berg T. Wong V.K.W. MCC1019, a Selective Inhibitor of the Polo-Box Domain of Polo-like Kinase 1 as Novel, Potent Anticancer Candidate Acta Pharm. Sin. B 2019 9 1021 1034 10.1016/j.apsb.2019.02.001 31649851 34. Napiórkowska M. Cieślak M. Kaźmierczak-Barańska J. Królewska-Golińska K. Nawrot B. Synthesis of New Derivatives of Benzofuran as Potential Anticancer Agents Molecules 2019 24 1529 10.3390/molecules24081529 31003438 35. Othman D.I. Abdelal A.M.M. El-Sayed M.A. El Bialy S.A.A. Novel Benzofuran Derivatives: Synthesis and Antitumor Activity Heterocycl. Commun. 2013 19 29 35 10.1515/hc-2012-0119 36. Wan W.C. Chen W. Liu L.X. Li Y. Yang L.J. Deng X.Y. Zhang H.B. Yang X.D. Synthesis and Cytotoxic Activity of Novel Hybrid Compounds between 2-Alkylbenzofuran and Imidazole Med. Chem. Res. 2014 23 1599 1611 10.1007/s00044-013-0760-8 37. Yang X.D. Wan W.C. Deng X.Y. Li Y. Yang L.J. Li L. Zhang H.B. Design, Synthesis and Cytotoxic Activities of Novel Hybrid Compounds between 2-Phenylbenzofuran and Imidazole Bioorg. Med. Chem. Lett. 2012 22 2726 2729 10.1016/j.bmcl.2012.02.094 22440627 38. Chen Y. Law P.-Y. Loh H. Inhibition of PI3K/Akt Signaling: An Emerging Paradigm for Targeted Cancer Therapy Curr. Med. Chem. Agents 2005 5 575 589 10.2174/156801105774574649 16305480 39. Choi M. Jo H. Park H.J. Sateesh Kumar A. Lee J. Yun J. Kim Y. Han S.B. Jung J.K. Cho J. Design, Synthesis, and Biological Evaluation of Benzofuran- and 2,3-Dihydrobenzofuran-2-Carboxylic Acid N-(Substituted)Phenylamide Derivatives as Anticancer Agents and Inhibitors of NF-ΚB Bioorg. Med. Chem. Lett. 2015 25 2545 2549 10.1016/j.bmcl.2015.04.050 25953156 40. Mao Z.W. Zheng X. Lin Y.P. Hu C.Y. Wang X.L. Wan C.P. Rao G.X. Design, Synthesis and Anticancer Activity of Novel Hybrid Compounds between Benzofuran and N-Aryl Piperazine Bioorg. Med. Chem. Lett. 2016 26 3421 3424 10.1016/j.bmcl.2016.06.055 27371110 41. Wang F. Feng K.R. Zhao J.Y. Zhang J.W. Shi X.W. Zhou J. Gao D. Lin G.Q. Tian P. Identification of Novel STAT3 Inhibitors Bearing 2-Acetyl-7-Phenylamino Benzofuran Scaffold for Antitumour Study Bioorg. Med. Chem. 2020 28 115822 10.1016/j.bmc.2020.115822 33126089 42. Youssif B.G.M. Mohamed A.M. Osman E.E.A. Abou-Ghadir O.F. Elnaggar D.H. Abdelrahman M.H. Treamblu L. Gomaa H.A.M. 5-Chlorobenzofuran-2-Carboxamides: From Allosteric CB1 Modulators to Potential Apoptotic Antitumor Agents Eur. J. Med. Chem. 2019 177 1 11 10.1016/j.ejmech.2019.05.040 31128433 43. Riss T.L. Moravec R.A. Niles A.L. Duellman S. Benink H.A. Worzella T.J. Minor L. Cell Viability Assays Eli Lilly & Company and the National Center for Advancing Translational Sciences Bethesda, MD, USA 2004 44. Abdelhafez O.M. Amin K.M. Ali H.I. Abdalla M.M. Ahmed E.Y. Design, Synthesis and Anticancer Activity of Benzofuran Derivatives Targeting VEGFR-2 Tyrosine Kinase RSC Adv. 2014 4 11569 11579 10.1039/c4ra00943f 45. Adem F.A. Kuete V. Mbaveng A.T. Heydenreich M. Ndakala A. Irungu B. Efferth T. Yenesew A. Cytotoxic Benzylbenzofuran Derivatives from Dorstenia Kameruniana Fitoterapia 2018 128 26 30 10.1016/j.fitote.2018.04.019 29715541 46. Coşkun D. Tekin S. Sandal S. Coşkun M.F. Synthesis, Characterization, and Anticancer Activity of New Benzofuran Substituted Chalcones J. Chem. 2016 2016 7678486 10.1155/2016/7678486 47. Asadi P. Khodarahmi G. Jahanian-Najafabadi A. Saghaie L. Hassanzadeh F. Biologically Active Heterocyclic Hybrids Based on Quinazolinone, Benzofuran and Imidazolium Moieties: Synthesis, Characterization, Cytotoxic and Antibacterial Evaluation Chem. Biodivers. 2017 14 10.1002/cbdv.201600411 28002645 48. Etebari M. Khodarahmi G.A. Jafarian-Dehkordi A. Nokhodian Z. Genotoxic Effects of Some L-[(Benzofuran-2-Yl)-Phenylmethyl]-Imidazoles on MCF-7 Cell Line Res. Pharm. Sci. 2012 7 189 195 23181097 49. Irfan A. Faiz S. Rasul A. Zafar R. Zahoor A.F. Kotwica-Mojzych K. Mojzych M. Exploring the Synergistic Anticancer Potential of Benzofuran–Oxadiazoles and Triazoles: Improved Ultrasound-and Microwave-Assisted Synthesis, Molecular Docking, Hemolytic, Thrombolytic and Anticancer Evaluation of Furan-Based Molecules Molecules 2022 27 1023 10.3390/molecules27031023 35164286 50. Jia Y. Li M. Cao Y. Feng W. Li X. Xue W. Shi H. Discovery of a Novel Benzenesulfonamide Analogue That Inhibits Proliferation and Metastasis against Ovarian Cancer OVCAR-8 Cells Drug Des. Devel. Ther. 2020 14 207 216 10.2147/DDDT.S225201 32021105 51. Yang Y.R. Wei J.L. Mo X.F. Yuan Z.W. Wang J.L. Zhang C. Xie Y.Y. You Q.D. Sun H.P. Discovery and Optimization of New Benzofuran Derivatives against P53-Independent Malignant Cancer Cells through Inhibition of HIF-1 Pathway Bioorg. Med. Chem. Lett. 2016 26 2713 2718 10.1016/j.bmcl.2016.03.112 27101893 52. Semenza G.L. Defining the Role of Hypoxia-Inducible Factor 1 in Cancer Biology and Therapeutics Oncogene 2010 29 625 634 10.1038/onc.2009.441 19946328 53. Nakamura H. Yasui Y. Maruyama M. Minegishi H. Ban H.S. Sato S. Development of Hypoxia-Inducible Factor (HIF)-1α Inhibitors: Effect of Ortho-Carborane Substituents on HIF Transcriptional Activity under Hypoxia Bioorg. Med. Chem. Lett. 2013 23 806 810 10.1016/j.bmcl.2012.11.081 23267767 54. Mahmood N. Mihalcioiu C. Rabbani S.A. Multifaceted Role of the Urokinase-Type Plasminogen Activator (UPA) and Its Receptor (UPAR): Diagnostic, Prognostic, and Therapeutic Applications Front. Oncol. 2018 8 24 10.3389/fonc.2018.00024 29484286 55. Ulisse S. Baldini E. Sorrenti S. D’Armiento M. The Urokinase Plasminogen Activator System: A Target for Anti-Cancer Therapy Curr. Cancer Drug Targets 2009 9 32 71 10.2174/156800909787314002 19200050 56. Croucher D.R. Saunders D.N. Lobov S. Ranson M. Revisiting the Biological Roles of PAI2 (SERPINB2) in Cancer Nat. Rev. Cancer 2008 8 535 545 10.1038/nrc2400 18548086 57. Matthews H. Ranson M. Tyndall J.D.A. Kelso M.J. Synthesis and Preliminary Evaluation of Amiloride Analogs as Inhibitors of the Urokinase-Type Plasminogen Activator (UPA) Bioorg. Med. Chem. Lett. 2011 21 6760 6766 10.1016/j.bmcl.2011.09.044 21978672 58. Song C.W. Lyons J.C. Makepeace C.M. Griffin R.J. Cragoe E.J. Effect’s of Hma, an Analog of Amiloride, on the Thermosensitivity of Tumors in Vivo Int. J. Radiat. Oncol. Biol. Phys. 1994 30 133 139 10.1016/0360-3016(94)90528-2 8083106 59. Luo J. Tannock I.F. Inhibition of the Regulation of Intracellular PH: Potential of 5-(N,N-Hexamethylene) Amiloride in Tumour-Selective Therapy Br. J. Cancer 1994 70 617 624 10.1038/bjc.1994.360 7917906 60. Chemler S.R. Trauner D. Danishefsky S.J. The B -Alkyl Suzuki-Miyaura Cross-Coupling Reaction: A Versatile C-C Bond-Forming Tool REVIEWS The B -Alkyl Suzuki-Miyaura Cross-Coupling Reaction: Development, Mechanistic Study, and Applications in Natural Product Synthesis Angew. Chem. Int. Ed. 2001 40 4544 4568 10.1002/1521-3773(20011217)40:24<4544::AID-ANIE4544>3.0.CO;2-N 61. Buckley B.J. Aboelela A. Minaei E. Jiang L.X. Xu Z. Ali U. Fildes K. Cheung C.Y. Cook S.M. Johnson D.C. 6-Substituted Hexamethylene Amiloride (HMA) Derivatives as Potent and Selective Inhibitors of the Human Urokinase Plasminogen Activator for Use in Cancer J. Med. Chem. 2018 61 8299 8320 10.1021/acs.jmedchem.8b00838 30130401 62. Gao Y. Ma C. Feng X. Liu Y. Haimiti X. BF12, a Novel Benzofuran, Exhibits Antitumor Activity by Inhibiting Microtubules and the PI3K/Akt/MTOR Signaling Pathway in Human Cervical Cancer Cells Chem. Biodivers. 2020 17 10.1002/cbdv.201900622 63. Soleimani A. Asadi J. Rostami-Charati F. Gharaei R. High Cytotoxicity and Apoptotic Effects of Natural Bioactive Benzofuran Derivative on the MCF-7 Breast Cancer Cell Line Comb. Chem. High Throughput Screen 2015 18 505 513 10.2174/1386207318666150430114815 25924658 64. Mishra S. Quinazolinone and Quinazoline Derivatives: Synthesis and Biological Application Intechopen London, UK 2019 10.5772/INTECHOPEN.89203 65. Radwan A.A. Alanazi F.K. Biological Activity of Quinazolinones Intechopen London, UK 2020 10.5772/intechopen.90621 66. Jafari E. Khajouei M.R. Hassanzadeh F. Hakimelahi G.H. Khodarahmi G.A. Quinazolinone and Quinazoline Derivatives: Recent Structures with Potent Antimicrobial and Cytotoxic Activities Res. Pharm. Sci. 2016 11 1 14 27051427 67. Kuttan R. Bhanumathy P. Nirmala K. George M.C. Potential Anticancer Activity of Turmeric (Curcuma Longa) Cancer Lett. 1985 29 197 202 10.1016/0304-3835(85)90159-4 4075289 68. Kumar B.R.P. Sharma G.K. Srinath S. Noor M. Suresh B. Srinivasa B.R. Microwave-Assisted, Solvent-Free, Parallel Syntheses and Elucidation of Reaction Mechanism for the Formation of Some Novel Tetraaryl Imidazoles of Biological Interest J. Heterocycl. Chem. 2009 46 278 284 10.1002/jhet.68 69. Sharma G.K. Pathak D. Microwave-Assisted, Solvent-Free and Parallel Synthesis of Some Novel Substituted Imidazoles of Biological Interest Chem. Pharm. Bull. 2010 58 375 380 10.1248/cpb.58.375 70. Shankar B. Jalapathi P. Saikrishna B. Perugu S. Manga V. Synthesis, Anti-Microbial Activity, Cytotoxicity of Some Novel Substituted (5-(3-(1H-Benzo[d]Imidazol-2-Yl)-4-Hydroxybenzyl)Benzofuran-2-Yl)(Phenyl)Methanone Analogs Chem. Cent. J. 2018 12 1 10.1186/s13065-017-0364-3 29318401 71. Kumar Saxena R. Kumar P. Varshney S. Synthesis and Biological Evaluation of Newer Benzofuran Derivatives as Potential Anticancer and Antithelmintic Agents World J. Pharm. Res. 2015 3 821 72. Khanam H. Shamsuzzaman Bioactive Benzofuran Derivatives: A Review Eur. J. Med. Chem. 2015 97 483 504 10.1016/j.ejmech.2014.11.039 25482554 73. Nam N.-H. Combretastatin A-4 Analogues as Antimitotic Antitumor Agents Curr. Med. Chem. 2005 10 1697 1722 10.2174/0929867033457151 74. Flynn B.L. Gill G.S. Grobelny D.W. Chaplin J.H. Paul D. Leske A.F. Lavranos T.C. Chalmers D.K. Charman S.A. Kostewicz E. Discovery of 7-Hydroxy-6-Methoxy-2-Methyl-3-(3,4,5-Trimethoxybenzoyl)Benzo[ b ]Furan (BNC105), a Tubulin Polymerization Inhibitor with Potent Antiproliferative and Tumor Vascular Disrupting Properties J. Med. Chem. 2011 54 6014 6027 10.1021/jm200454y 21774499 75. Vignot S. Faivre S. Aguirre D. Raymond E. MTOR-Targeted Therapy of Cancer with Rapamycin Derivatives Ann. Oncol. 2005 16 525 537 10.1093/annonc/mdi113 15728109 76. Albert S. Serova M. Dreyer C. Sablin M.P. Faivre S. Raymond E. New Inhibitors of the Mammalian Target of Rapamycin Signaling Pathway for Cancer Expert Opin. Investig. Drugs 2010 19 919 930 10.1517/13543784.2010.499121 20569080 77. Faivre S. Kroemer G. Raymond E. Current Development of MTOR Inhibitors as Anticancer Agents Nat. Rev. Drug Discov. 2006 5 671 688 10.1038/nrd2062 16883305 78. Salomé C. Narbonne V. Ribeiro N. Thuaud F. Serova M. De Gramont A. Faivre S. Raymond E. Désaubry L. Benzofuran Derivatives as a Novel Class of Inhibitors of MTOR Signaling Eur. J. Med. Chem. 2014 74 41 49 10.1016/j.ejmech.2013.12.020 24445311 79. Salomé C. Ribeiro N. Chavagnan T. Thuaud F. Serova M. De Gramont A. Faivre S. Raymond E. Désaubry L. Benzofuran Derivatives as Anticancer Inhibitors of MTOR Signaling Eur. J. Med. Chem. 2014 81 181 191 10.1016/j.ejmech.2014.05.014 24836070 80. Garrido-Laguna I. Hidalgo M. Pancreatic Cancer: From State-of-the-Art Treatments to Promising Novel Therapies Nat. Rev. Clin. Oncol. 2015 12 319 334 10.1038/nrclinonc.2015.53 25824606 81. Spivak-Kroizman T.R. Hostetter G. Posner R. Aziz M. Hu C. Demeure M.J. Von Hoff D. Hingorani S.R. Palculict T.B. Izzo J. Hypoxia Triggers Hedgehog-Mediated Tumor-Stromal Interactions in Pancreatic Cancer Cancer Res. 2013 73 3235 3247 10.1158/0008-5472.CAN-11-1433 23633488 82. Xu X.-L. Yang Y.-R. Mo X.-F. Wei J.-L. Zhang X.-J. You Q.-D. Design, Synthesis, and Evaluation of Benzofuran Derivatives as Novel Anti-Pancreatic Carcinoma Agents via Interfering the Hypoxia Environment by Targeting HIF-1α Pathway Eur. J. Med. Chem. 2017 137 45 62 10.1016/j.ejmech.2017.05.042 28554092 83. Naik R. Harmalkar D.S. Xu X. Jang K. Lee K. Bioactive Benzofuran Derivatives: Moracins A-Z in Medicinal Chemistry Eur. J. Med. Chem. 2015 90 379 393 10.1016/j.ejmech.2014.11.047 25461329 84. Ma Y. Zheng X. Gao H. Wan C. Rao G. Mao Z. Design, Synthesis, and Biological Evaluation of Novel Benzofuran Derivatives Bearing N-Aryl Piperazine Moiety Molecules 2016 21 1684 10.3390/molecules21121684 85. Huang X.X. Zhou C.C. Li L.Z. Peng Y. Lou L.L. Liu S. Li D.M. Ikejima T. Song S.J. Cytotoxic and Antioxidant Dihydrobenzofuran Neolignans from the Seeds of Crataegus Pinnatifida Fitoterapia 2013 91 217 223 10.1016/j.fitote.2013.09.011 24060910 86. Chen W. Deng X.Y. Li Y. Yang L.J. Wan W.C. Wang X.Q. Zhang H.B. Yang X.D. Synthesis and Cytotoxic Activities of Novel Hybrid 2-Phenyl-3- Alkylbenzofuran and Imidazole/Triazole Compounds Bioorg. Med. Chem. Lett. 2013 23 4297 4302 10.1016/j.bmcl.2013.06.001 23800685 87. Abd El-Karim S.S. Anwar M.M. Mohamed N.A. Nasr T. Elseginy S.A. Design, Synthesis, Biological Evaluation and Molecular Docking Studies of Novel Benzofuran-Pyrazole Derivatives as Anticancer Agents Bioorg. Chem. 2015 63 1 12 10.1016/j.bioorg.2015.08.006 26368040 88. Murthy I.S. Sireesha R. Deepti K. Srinivasa Rao P. Ramesh Raju R. Design, Synthesis and Biological Evaluation of Sulphonamide Derivatives of Benzofuran-Imidazopyridines as Anticancer Agents Chem. Data Collect. 2021 31 100608 10.1016/j.cdc.2020.100608 89. Zwergel C. Valente S. Salvato A. Xu Z. Talhi O. Mai A. Silva A. Altucci L. Kirsch G. Novel Benzofuran-Chromone and -Coumarin Derivatives: Synthesis and Biological Activity in K562 Human Leukemia Cells Med. Chem. Commun. 2013 4 1571 1579 10.1039/c3md00241a 90. Coskun D. Erkisa M. Ulukaya E. Coskun M.F. Ari F. Novel 1-(7-Ethoxy-1-Benzofuran-2-Yl) Substituted Chalcone Derivatives: Synthesis, Characterization and Anticancer Activity Eur. J. Med. Chem. 2017 136 212 222 10.1016/j.ejmech.2017.05.017 28494257 91. Al-Ghorbani M. Begum B.A. Mamatha S.V. Ara Khanum S. Piperazine and Morpholine: Synthetic Preview and Pharmaceutical Applications J. Chem. Pharm. Res. 2015 7 281 301 10.5958/0974-360X.2015.00100.6 Available online: www.jocpr.com (accessed on 17 November 2021) 92. Rathi A.K. Syed R. Shin H.S. Patel R.V. Piperazine Derivatives for Therapeutic Use: A Patent Review (2010-Present) Expert Opin. Ther. Pat. 2016 26 777 797 10.1080/13543776.2016.1189902 27177234 93. Cuca L.E. David Coy E. Alarcón M.A. Fernández A. Aristizábal F.A. Cytotoxic Effect of Some Natural Compounds Isolated from Lauraceae Plants and Synthetic Derivatives Biomédica 2011 31 335 378 10.7705/biomedica.v31i3.347 22674310 94. Jiang Z.H. Liu Y.P. Huang Z.H. Wang T.T. Feng X.Y. Yue H. Guo W. Fu Y.H. Cytotoxic Dihydrobenzofuran Neolignans from Mappianthus Iodoies Bioorg. Chem. 2017 75 260 264 10.1016/j.bioorg.2017.10.003 29040842 95. Mansoor T.A. Borralho P.M. Luo X. Mulhovo S. Rodrigues C.M.P. Ferreira M.J.U. Apoptosis Inducing Activity of Benzophenanthridine-Type Alkaloids and 2-Arylbenzofuran Neolignans in HCT116 Colon Carcinoma Cells Phytomedicine 2013 20 923 929 10.1016/j.phymed.2013.03.026 23643093 96. Sivaraman A. Kim J.S. Harmalkar D.S. Min K.H. Park J.W. Choi Y. Kim K. Lee K. Synthesis and Cytotoxicity Studies of Bioactive Benzofurans from Lavandula Agustifolia and Modified Synthesis of Ailanthoidol, Homoegonol, and Egonol J. Nat. Prod. 2020 83 3354 3362 10.1021/acs.jnatprod.0c00697 33073572 97. Zhou Z.B. Luo J.G. Pan K. Shan S.M. Zhang W. Kong L.Y. Bioactive Benzofuran Neolignans from Aristolochia Fordiana Planta Med. 2013 79 1730 1735 10.1055/s-0033-1351100 24288294 98. Zálešák F. Bon D.J.Y.D. Pospíšil J. Lignans and Neolignans: Plant Secondary Metabolites as a Reservoir of Biologically Active Substances Pharmacol. Res. 2019 146 104284 10.1016/j.phrs.2019.104284 31136813 99. Sharma P. Larosa C. Antwi J. Govindarajan R. Werbovetz K.A. Imidazoles as Potential Anticancer Agents: An Update on Recent Studies Molecules 2021 26 4213 10.3390/molecules26144213 34299488 100. Hranjec M. Sović I. Ratkaj I. Pavlović G. Ilić N. Valjalo L. Pavelić K. Kraljević Pavelić S. Karminski-Zamola G. Antiproliferative Potency of Novel Benzofuran-2-Carboxamides on Tumour Cell Lines: Cell Death Mechanisms and Determination of Crystal Structure Eur. J. Med. Chem. 2013 59 111 119 10.1016/j.ejmech.2012.11.009 23220640 101. Song W.J. Yang X.D. Zeng X.H. Xu X.L. Zhang G.L. Zhang H.B. Synthesis and Cytotoxic Activities of Novel Hybrid Compounds of Imidazole Scaffold-Based 2-Substituted Benzofurans RSC Adv. 2012 2 4612 4615 10.1039/c2ra20376f 102. Chen W. Yang X.D. Li Y. Yang L.J. Wang X.Q. Zhang G.L. Zhang H.B. Design, Synthesis and Cytotoxic Activities of Novel Hybrid Compounds between Dihydrobenzofuran and Imidazole Org. Biomol. Chem. 2011 9 4250 4255 10.1039/c1ob05116d 21505704 103. Ali I. Lone M.N. Aboul-Enein H.Y. Imidazoles as Potential Anticancer Agents Medchemcomm 2017 8 1742 1773 10.1039/C7MD00067G 30108886 104. Berger D.M. Dutia M. Birnberg G. Powell D. Boschelli D.H. Wang Y.D. Ravi M. Yaczko D. Golas J. Lucas J. 4-Anilino-7,8-Dialkoxybenzo[g]Quinoline-3-Carbonitriles as Potent Src Kinase Inhibitors J. Med. Chem. 2005 48 5909 5920 10.1021/jm050512u 16161995 105. Tsao A.S. He D. Saigal B. Liu S. Lee J.J. Bakkannagari S. Ordonez N.G. Waun K.H. Wistuba I. Johnson F.M. Inhibition of C-Src Expression and Activation in Malignant Pleural Mesothelioma Tissues Leads to Apoptosis, Cell Cycle Arrest, and Decreased Migration and Invasion Mol. Cancer Ther. 2007 6 1962 1972 10.1158/1535-7163.MCT-07-0052 17620427 106. Summy J.M. Gallick G.E. Src Family Kinases in Tumor Progression and Metastasis Cancer Metastasis Rev. 2003 22 337 358 10.1023/A:1023772912750 12884910 107. Li M. Zhao B.X. Progress of the Synthesis of Condensed Pyrazole Derivatives (from 2010 to Mid-2013) Eur. J. Med. Chem. 2014 85 311 340 10.1016/j.ejmech.2014.07.102 25104650 108. Bennani F.E. Doudach L. Cherrah Y. Ramli Y. Karrouchi K. Ansar M. Faouzi M.E.A. Overview of Recent Developments of Pyrazole Derivatives as an Anticancer Agent in Different Cell Line Bioorg. Chem. 2020 97 10.1016/j.bioorg.2019.103470 32120072 109. Patil V.M. Masand N. Verma S. Masand V. Chromones: Privileged Scaffold in Anticancer Drug Discovery Chem. Biol. Drug Des. 2021 98 943 953 10.1111/cbdd.13951 34519163 110. Van Meerloo J. Kaspers G.J.L. Cloos J. Cell Sensitivity Assays: The MTT Assay Methods Mol. Biol. 2011 731 237 245 10.1007/978-1-61779-80-5_20 21516412 111. Alsayari A. Muhsinah A.B. Hassan M.Z. Ahsan M.J. Alshehri J.A. Begum N. Aurone: A Biologically Attractive Scaffold as Anticancer Agent Eur. J. Med. Chem. 2019 166 417 431 10.1016/j.ejmech.2019.01.078 30739824 112. Thakur A. Singla R. Jaitak V. Coumarins as Anticancer Agents: A Review on Synthetic Strategies, Mechanism of Action and SAR Studies Eur. J. Med. Chem. 2015 101 476 495 10.1016/j.ejmech.2015.07.010 26188907 113. Boumendjel A. Aurones: A Subclass of Flavones with Promising Biological Potential Curr. Med. Chem. 2003 10 2621 2630 10.2174/0929867033456468 14529476 114. Kontogiorgis C. Detsi A. Hadjipavlou-Litina D. Coumarin-Based Drugs: A Patent Review (2008-Present) Expert Opin. Ther. Pat. 2012 22 437 454 10.1517/13543776.2012.678835 22475457 115. Zwergel C. Gaascht F. Valente S. Diederich M. Bagrel D. Kirsch G. Aurones: Interesting Natural and Synthetic Compounds with Emerging Biological Potential Nat. Prod. Commun. 2012 7 389 394 10.1177/1934578X1200700322 22545415 116. Zhuang C. Zhang W. Sheng C. Zhang W. Xing C. Miao Z. Chalcone: A Privileged Structure in Medicinal Chemistry Chem. Rev. 2017 117 7762 7810 10.1021/acs.chemrev.7b00020 28488435 117. Zhou B. Xing C. Diverse Molecular Targets for Chalcones with Varied Bioactivities Med. Chem. 2015 5 388 404 10.4172/2161-0444.1000291 26798565 118. Karthikeyan C. Narayana Moorthy N.S.H. Ramasamy S. Vanam U. Manivannan E. Karunagaran D. Trivedi P. Advances in Chalcones with Anticancer Activities Recent Pat. Anticancer Drug Discov. 2015 10 97 115 10.2174/1574892809666140819153902 25138130 119. Kumar V. Kaur K. Triazole and Oxadiazole Containing Natural Products: A Review Nat. Prod. J. 2014 4 115 130 10.2174/221031550402141009100114 120. Nayak S. Gaonkar S.L. Musad E.A. Dawsar A.M.A. 1,3,4-Oxadiazole-Containing Hybrids as Potential Anticancer Agents: Recent Developments, Mechanism of Action and Structure-Activity Relationships J. Saudi Chem. Soc. 2021 25 101284 10.1016/j.jscs.2021.101284 121. Xu Z. Zhao S.J. Liu Y. 1,2,3-Triazole-Containing Hybrids as Potential Anticancer Agents: Current Developments, Action Mechanisms and Structure-Activity Relationships Eur. J. Med. Chem. 2019 183 111700 10.1016/j.ejmech.2019.111700 31546197 122. Alam M.M. 1,2,3-Triazole Hybrids as Anticancer Agents: A Review Arch. Pharm. 2022 355 e2100158 10.1002/ardp.202100158 123. Pervaram S. Ashok D. Sarasija M. Reddy C.V.R. Sridhar G. Synthesis and Anticancer Activity of 1,2,4-Oxadiazole Fused Benzofuran Derivatives Russ. J. Gen. Chem. 2018 88 1219 1223 10.1134/S1070363218060282

PMC009xxxxxx/PMC9099632.txt

==== Front Polymers (Basel) Polymers (Basel) polymers Polymers 2073-4360 MDPI 10.3390/polym14091659 polymers-14-01659 Article pH- and Thermo-Responsive Water-Soluble Smart Polyion Complex (PIC) Vesicle with Polyampholyte Shells Pham Thu Thao 1 https://orcid.org/0000-0002-9087-7417 Pham Tien Duc 2 https://orcid.org/0000-0002-2838-5200 Yusa Shin-ichi 1* Pispas Asterios (Stergios) Academic Editor 1 Department of Applied Chemistry, Graduate School of Engineering, University of Hyogo, 2167 Shosha, Himeji 671-2280, Hyogo, Japan; phamthuthao.hus@gmail.com 2 Faculty of Chemistry, University of Science, Vietnam National University, Hanoi, 19 Le Thanh Tong, Hoan Kiem, Hanoi 100000, Vietnam; tienduchphn@gmail.com * Correspondence: yusa@eng.u-hyogo.ac.jp; Tel.: +81-79-267-4954; Fax: +81-79-266-8868 20 4 2022 5 2022 14 9 165909 4 2022 18 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). A diblock copolymer (P(VBTAC/NaSS)17-b-PAPTAC50; P(VS)17A50) composed of amphoteric random copolymer, poly(vinylbenzyl trimethylammonium chloride-co-sodium p-styrensunfonate) (P(VBTAC/NaSS); P(VS)) and cationic poly(3-(acrylamidopropyl) trimethylammonium chloride) (PAPTAC; A) block, and poly(acrylic acid) (PAAc49) were prepared via a reversible addition−fragmentation chain transfer radical polymerization. Scrips V, S, and A represent VBTAC, NaSS, and PAPTAC blocks, respectively. Water-soluble polyion complex (PIC) vesicles were formed by mixing P(VS)17A50 and PAAc49 in water under basic conditions through electrostatic interactions between the cationic PAPTAC block and PAAc49 with the deprotonated pendant carboxylate anions. The PIC vesicle collapsed under an acidic medium because the pendant carboxylate anions in PAAc49 were protonated to delete the anionic charges. The PIC vesicle comprises an ionic PAPTAC/PAAc membrane coated with amphoteric random copolymer P(VS)17 shells. The PIC vesicle showed upper critical solution temperature (UCST) behavior in aqueous solutions because of the P(VS)17 shells. The pH- and thermo-responsive behavior of the PIC vesicle were studied using 1H NMR, static and dynamic light scattering, and percent transmittance measurements. When the ratio of the oppositely charged polymers in PAPTAC/PAAc was equal, the size and light scattering intensity of the PIC vesicle reached maximum values. The hydrophilic guest molecules can be encapsulated into the PIC vesicle at the base medium and released under acidic conditions. It is expected that the PIC vesicles will be applied as a smart drug delivery system. polyion complex electrostatic interaction oppositely charged polyelectrolyte pH-responsive UCST behavior polyampholyte ==== Body pmc1. Introduction Owing to electrostatic interactions between cationic and anionic units, oppositely charged polyelectrolytes in water form a polyion complex (PIC) or polyelectrolytes complex [1,2]. Furthermore, PIC micelles with a PIC core and hydrophilic noncharged shells can be prepared without using an organic solvent in water [3,4]. Smart PIC aggregates have single- or multi-responsive properties against external stimuli, such as pH, temperature, and light irradiation. They show substantial changes in their properties with a slight change in the surrounding environments [5]. The PIC aggregates have attracted huge interest from many researchers because of their advantages. The PIC aggregates are expected to meet some drug delivery system (DDS) assignments, such as the targeted and controlled release of the drug. The pH of the medium is an important factor in the controlled release of a DDS. The microenvironment of cancer tumors is acidic, whereas it is nearly neutral for normal cells [6]. The change in the pH of the environment results in an ideal pH-responsive DDS carrier, which can be applied in cancer treatments. Therefore, many PIC aggregate systems, which target the release of drugs in response to the acidity of the surroundings, have been reported [7,8,9,10,11,12]. Water-soluble pH-responsive PIC vesicles covered with biocompatible poly(2-(methacryloyloxy)ethyl phosphorylcholine) (PMPC) shells have also been reported [3]. The PIC vesicle was coassembled via electrostatic interactions between PMAPTAC and ionized PAaH under a basic medium by mixing oppositely charged diblock copolymers consisting of PMPC and cationic poly(3-(methacrylamidopropyl) trimethylammonium chloride) (PMAPTAC), or anionic poly(sodium 6-acrylamidohexanoate) (PAaH) blocks. The hydrophilic guest molecule, Texas red-labeled dextran, was encapsulated into the inner water phase of the PIC vesicle. Under the acidic medium, PAaH was protonated to become noncharged resulting in a dissociation of the PIC vesicle, where the guest molecule was released. Therefore, the PIC vesicle is expected to effectively target the drug in the cytoplasm. Similarly, the environment’s temperature is a widely used parameter in the controlled release of drugs [13,14,15,16,17]. Matsuoka et al. [13] reported on thermo-responsive PIC aggregates. Diblock copolymers composed of poly(sulfopropyl dimethylammonium propylacrylamide) (PSPP) and cationic PMAPTAC or anionic poly(sodium p-styrenesulfonate) (PNaSS) were mixed in water to form PIC micelles with a PMAPTAC/PNaSS core and PSPP shells. The PIC micelles showed an upper critical solution temperature (UCST) caused by the PSPP shells. They formed large aggregates of micelles below the phase transition temperature of the PSPP shells owing to the PSPP chain shrinkage. The UCST behavior of the PIC aggregates can be controlled using the polymer concentration and the degree of polymerization (DP) of the PSPP chains. Furthermore, some other PIC systems have been fabricated with the target to release a drug in response to environments containing glucose to deal with diabetes [18,19] or to use light irradiation to control the release of the drug [20,21]. These single stimulus-responsive PIC aggregates have been studied and applied. Besides stimuli response, PIC aggregates have shown other outstanding advantages [22,23,24,25,26,27]. For instance, dual pH- and thermo-responsive PIC micelles were prepared from a co-assembly of oppositely charged diblock copolymers, poly(N-methyl-2-vinyl pyridinium iodide)-block-poly(ethylene oxide) (P2MVP-b-PEO) and poly(acrylic acid)-block-poly(isopropyl acrylamide) (PAAc-b-PNIPAM) in water [27]. Under a basic medium at 25 °C, the PIC micelles were formed owing to the electrostatic interaction between the PAAc and P2MPV blocks with the PAAc/P2MVP core and hydrophilic mixed PNIPAM/PEG shells. However, under an acidic medium, the size of the aggregates increased, which can be attributed to the aggregation formed by hydrogen bonding between PAAc and PNIPAM/PEO. At high concentrations, the core–shell structure of PIC micelles switches to core–shell–corona (onion type) with a PNIPAM core, covered with a mixed P2MVP/PAAc shell, and a PEO outer corona at 60 °C. Polyampholytes or amphoteric random copolymers are polymers that consist of cationic and anionic units in their structure [28,29,30]. Amphoteric random copolymers have no charge, which is comparable to the balance between oppositely charged groups since they contain both pendant cationic and anionic groups [31,32]. Therefore, to avoid complex precipitation, the polyampholytes can be used as a hydrophilic nonionic block. Furthermore, some amphoteric polymers show thermo-responsive and protein antifouling properties. Yusa et al. [28] reported that amphoteric diblock copolymers from poly(2-acrylamido-2-methylpropanesulfonic acid sodium salt) (PAMPS) and poly(3-(acrylamido)propyl trimethylammonium chloride) (PAPTAC) synthesized via a reversible addition−fragmentation chain transfer (RAFT) radical polymerization. The polymer exhibited a lower critical solution temperature (LCST) behavior due to the hydrogen-bonding interactions of the pendant amide groups and water molecules. The thermo-responsive behavior of the polymers can be controlled by adjusting the NaCl concentration. Meanwhile, P(APMPS/APTAC), a random copolymer prepared by RAFT with the same content as APMPS and APTAC, showed an ability in protein antifouling [33]. The PIC micelle consists of amphoteric random copolymer P(AMPS/APTAC) shells that show no interaction with proteins. Furthermore, a UCST-type thermo-responsive amphoteric random copolymer (P(VBTAC/NaSS)) composed of the pendant quaternary ammonium group in vinylbenzyl trimethylammonium chloride (VBTAC) and pendant sulfonate in NaSS with the same content of each monomer, was prepared via RAFT polymerization [34]. P(VBTAC/NaSS) showed the UCST behavior caused by electrostatic interactions between PVBTAC and PNaSS. The UCST behavior was affected by DP and concentrations of P(VBTAC/NaSS). Based on prior studies, we designed pH- and thermos-responsive PIC micelles. The interaction of the PIC vesicle with the hydrophobic surface may be controlled by temperature if the PIC vesicle surface hydrophobicity is changed by heating. For example, if the PIC vesicle surface changes to be hydrophobic with cooling, the interactions increase between the PIC vesicle and cell. Furthermore, if the PIC vesicle is dissociated in response to pH, the encapsulated guest molecules are released by pH change. Particularly, cancer tissue is more acidic than healthy tissue, so designing a pH-responsive PIC vesicle is useful for cancer treatment. In this study, we have prepared a diblock copolymer (P(VS)17A50) consisting of polyampholyte (P(VBTAC/NaSS)17; P(VS)17) block and a cationic PAPTAC50 (A50) block (Scheme 1). Furthermore, the PIC aggregates were prepared by mixing P(VS)17A50 and polyacrylic acid (PAAc49) in a basic medium because of electrostatic interactions between the cationic PAPTAC block and deprotonated anionic PAAc49. The PIC aggregate had a vesicle structure, which confirmed that the PIC aggregate could encapsulate a hydrophilic guest molecule into the hollow core. PAAc indicated pH-responsive behavior caused by protonation and the pendant carboxyl groups [35,36,37]. The PAAc pendant carboxyl groups are protonated and neutralized under acidic conditions, leading to the dissociation of the PIC vesicles. The PIC vesicle was formed from the PAPTAC/PAAc membrane, covered with the amphoteric random copolymer P(VS)17 shells. In this study, to synthesize P(VS)17, equimolar amounts of styrene-type cationic VBTAC and anionic NaSS were randomly copolymerized via RAFT. Then, the cationic APTAC monomer was polymerized to synthesize P(SV)17A50. Furthermore, PAAc49 was also synthesized via RAFT. Polymers were characterized with NMR and gel-permeation chromatography (GPC) measurements. The pH- and thermo-responsive properties of the PIC vesicle were studied with percent transmittance, dynamic and static light scattering, zeta-potential, and NMR measurements. Additionally, since the P(SV)17 polyampholyte shells exhibited the UCST behavior, controlling the interaction with proteins and cells in response to temperature may be possible. Furthermore, the P(SV)17 polyampholyte can suppress protein adsorption. Therefore, the PIC vesicles are expected to modify the surfaces of medical devices. 2. Materials and Methods 2.1. Materials Vinylbenzyl trimethylammonium chloride (VBTAC, 99%) methoxy poly(ethylene glycol) labeled pyrene (PEG-Py, Mn = 2000 g/mol) was purchased from Sigma-Aldrich (St. Louis, MO, USA), sodium p-styrenesulfonate (NaSS, 98%) from Tokyo Chemical Industry (Tokyo, Japan), and 4,4-azobis(4-cyanopentanoic acid) (V-501, 98%) and 2,2-azobisisobutyronitrile (AIBN, 98%) from Wako Pure Chemical (Osaka, Japan) were used without further purification. Acrylic acid (AAc, 98%) from Wako Pure Chemical (Osaka, Japan) was distilled under reduced pressure. 4-Cyanopentanoic acid dithiobenzoate (CPD) was synthesized using the previously reported method [38]. The water was purified using an ion-exchange column system. 2.2. Synthesis of P(VS)17 Random Copolymer First, amphoteric random copolymer (P(VS)17) was prepared via RAFT polymerization in accordance with the previous report [34]. VTBAC (3.18 g, 15.0 mmol), NaSS (3.09 g, 15.0 mmol), CPD (419 mg, 1.50 mmol), and V-501 (210 mg, 0.749 mmol) were dissolved in 30 mL of mixed solvent of 1.2 M NaCl aqueous solution and MeOH (9/1, v/v) with the feed molar ratio of [VBTAC]/[NaSS]/[CPD]/[V-501] = 10/10/1/0.5. The mixture was then, degassed by purging with Ar gas for 30 min. Polymerization occurred at 70 °C for 5 h. After polymerization, 1H NMR spectra were used to calculate the polymerization conversion of VBTAC and NaSS (conversion = 84.5%). The reaction mixture was dialyzed against 1.2-M NaCl aqueous solution for two days and pure water for one day and collected using a freeze-drying method (4.69 g, 67.9%). The theoretical DP(theo)) and number-average molecular weight (Mn(theo)) were 17 and 3.83 × 103 g/mol. The VBTAC and NaSS contents were determined using 1H NMR spectra of P(VS)17 and found to be 51.9 and 48.1 mol%, respectively. 2.3. P(VBTAC/NaSS)17-b-PAPTAC50 The diblock copolymer was prepared via RAFT polymerization using P(VS)17 as a macro-CTA. P(VS)17 (539 mg, 0.141 mmol), APTAC (1.46 g, 7.06 mmol), and V-501 (15.7 mg, 0.06 mmol) with a feed molar ratio of [P(VS)17]/[APTAC]/[AIBN] = 1/50/0.4 were dissolved in water (7.0 mL). The mixture was degassed by purging with Ar gas for 30 min. Polymerization occurred at 70 °C for 24 h. The polymerization conversion was calculated from 1H NMR integral intensity ratio of the vinyl protons and phenyl protons to 5.6 and 6.4–7.6 ppm, respectively, (conversion = 98.4%). Then, the reaction mixture was dialyzed against pure water for 5 days and collected using a freeze-drying method (1.55 g, 77.5%). From 1H NMR measurement, the DP(NMR)) of the APTAC block was estimated to be 50. The number-average molecular weight (Mn(GPC)) and polydispersity index (Mw/Mn) of P(VS)17A50 were estimated to be 5.70 × 104 g/mol and 1.04, respectively, based on the GPC measurement. 2.4. Synthesis of PAAc49 RAFT polymerization was used to prepare poly(acrylic acid) (PAAc49). AAc (7.20 g, 100 mmol), CPD (280 mg, 0.99 mg), and AIBN (65.7 mg, 0.40 mmol) with a feed molar ratio of [AAc]/[CPD]/[AIBN] = 100/1/0.4 were dissolved in methanol (100 mL). The mixture was degassed by purging with Ar gas for 30 min. Polymerization occurred at 60 °C for 47 h. From the 1H NMR integral intensity ratio, the polymerization conversion of the vinyl proton signal at 5.8 ppm and terminal phenyl group protons at 7.3–7.9 ppm (conversion = 36.3%) was estimated. Then, the reaction mixture was dialyzed against pure water for one week and collected using a freeze-drying method (0.97 g, 12.9%). From 1H NMR measurement the DP(NMR) of PAAc49 was calculated to be 49. From GPC measurement, the Mn(GPC) and Mw/Mn of P(VS)17A50 were 1.46 × 104 g/mol and 1.07, respectively. 2.5. Preparation of PIC Vesicles P(VS)17A50 and PAAc49 were individually dissolved in water at Cp = 1.0 g/L as stock solutions to prepare the PIC vesicles. The PIC vesicles were fabricated by mixing the cationic P(VS)17A50 solution with the anionic PAAc49 solution while stirring for 5 min at room temperature. The pH of the mixed solution was adjusted using 0.1-M NaOH and 0.1-M HCl aqueous solutions, then left for one day to obtain an equilibrium state. The mole fraction of positively charged (f+ = [APTAC]/([APTAC] + [PAAc]) in the PIC vesicle solution corresponds to the mixing ratio of the opposite polymers 2.6. Encapsulation of PEG-Py PEG-Py (0.01 g/L) aqueous solution was prepared. Then, P(VS)17A50 (0.1 g/L) and PAAc49 (0.1 g/L) were separately dissolved in the PEG-Py aqueous solution. Subsequently, the P(VS)17A50 aqueous solution was added to the PAAc49 aqueous solution with f+ = 0.5 and stirred for 5 min at room temperature. The mixed solution was adjusted to pH 10 using NaOH solution and then dialyzed for 20 days against an aqueous solution at pH 10 using a Harvard Apparatus (Holliston, MA, USA) polycarbonate membrane with 10-nm pores. The solvent was replaced every 12 h to remove the unbound PEG-Py molecules. A PEG-Py aqueous solution (0.01 g/L) without a PIC vesicle was dialyzed with the same procedure. Fluorescence spectroscopy measurements for PEG-Py in the presence and absence of the PIC vesicles were performed to detect PEG-Py content in the solution after dialysis. The loading capacity (LC) and loading efficiency (LE) were calculated according to the following Equations [39]:(1) LE (%)=Weight of encapsulated PEG−PyWeight of total PEG−Py × 100, (2) LC (%)=Weight of encapsulated PEG−PyWeight of the PIC aggregate × 100.  2.7. Measurements 1H NMR measurements were performed to determine the complex characteristics of the polymer and polyion complex using a JEOL (Tokyo, Japan) JNM-ECZ 400 MHz NMR. The sample solutions were prepared in D2O and adjusted with NaOD or DCl solutions to desired pH. The standard pulse program, stebpgp1s19, used a stimulated echo, bipolar gradient pulse, and one spoil gradient with 3-9-19 pulse sequence (WATERGATE) solvent suppression of the water signal. The GPC measurements were used to determine the number-average molecular weight (Mn(GPC)) and polydispersity index (Mw/Mn). For cationic polymer, P(VS)17A50, a mixture of Na2SO4 (0.3 M) and CH3COOH (0.5 M) aqueous solution was used as a cationic eluent. P(VS)17A50 signal was detected using Jasco (Tokyo, Japan) RI-2031 Plus refractive index (RI) detector equipped with a PU-2080 Plus column operating at 40 °C. Using a calibrated curve of standard poly(2-vinylpyridine), the values of Mn(GPC) and Mw/Mn for P(VS)17A50 were estimated. For anionic homopolymer PAAc49, GPC spectra were obtained from Tosoh RI-8020 RI working at 40 °C with a Shodex (Tokyo, Japan) GF-7M column, Tosoh (Yamaguchi, Japan) DP-8020 pump and a mixture of phosphate buffer (50 mM) at pH 9 and acetonitrile (9/1, v/v) as an anionic eluent. Using a calibration made from standard PNaSS, the values of Mn and Mw/Mn for polymers were estimated. Fourier transform infrared (FTIR) spectra were obtained using a Jasco FT/IR-4200 spectrometer. The PIC vesicle structure was confirmed via transmission electron microscopy (TEM) observation using a JEOL (Tokyo, Japan) JEM-2100. The PIC vesicle aqueous solutions with f+ = 0.5 and Cp = 0.1 g/L at pH 10 were prepared. Then, a drop of the sample solution was put on a copper grid coated with thin films of Formvar. The oversupply sample solution was filtered using filter paper. The next step was to stain with sodium phosphotungstate and dry under a vacuum for one night. The hydrodynamic radius (Rh), light scattering intensity (LSI), and zeta-potential of polymers and PIC vesicles were calculated using a Malvern Nano ZS with He–Ne laser (4 mW at 632.8 nm) from Malvern (Kobe, Japan). Before taking the measurements, the sample solutions were filtered using a 0.45-μm pore size membrane filter. Furthermore, to estimate the weight-average molecular weight (Mw), a z-average radius of gyration (Rg), and the second viral coefficient (A2) values of the polymers and PIC, SLS measurements were performed using the Otsuka Electronics Photal (Osaka, Japan) SLS-6500 with a He–Ne laser (4 mW at 632.8 nm) as a light source working at 25 °C. An Otsuka Electronics Photal DRM-3000 (Osaka, Japan) differential refract meter was used to calculate dn/dCp values at 633 nm. Percentage transmittance (%T) was measured on a Jasco (Tokyo, Japan) V-730 UV-vis spectrophotometer equipped with a temperature control system. Fluorescence spectra for PEG-Py were measured using a Hitachi (Tokyo, Japan) F-2500 fluorescence spectrometer. The excitation and emission slit widths were kept constantly at 20 and 2.5 nm, respectively. 3. Results 3.1. Preparation of P(VS)17A50 and PAAc49 All polymers in this study were prepared via normal RAFT polymerization. An amphoteric random copolymer comprising cationic VBVTAC and anionic NaSS was prepared via RAFT radical polymerization. According to 1H NMR, the total conversion (p) of VBTAC and NaSS was estimated to be 84.5%. The theoretical DP(theo) and theoretical number-average molecular weight (Mn(theo)) of P(VS)17 were calculated using p and the following equation:(3) DP(theo)=[M]0[CTA]0 × p100   (4) Mn(theo)=DPtheo × Mm+MCTA where, [M]0 and [CTA]0 are the initial concentrations of the monomer and CTA, respectively, and Mm and MCTA are the molecular weights of the monomer and CTA, respectively. The DP(theo) and Mn(theo) for P(VS)17 were calculated to be 17 and 3.83 × 103 g/mol, respectively. The composition of P(VS)17 was estimated from 1H NMR in D2O containing 1.2 M NaCl at 70 °C (Figure S1). The main chain proton signals were observed at 1.5–2.1 ppm. The pendent methyl protons of the VTBAC unit were observed at 2.9 ppm. The pendant phenyl protons of the VBTAC and NaSS units are presented as 6.4–7.6 ppm. The VBTAC content in the P(VS)17 block was calculated to be 51.9 mol% based on the integral intensities between the pendant phenyl and VBTAC methyl protons. The DP(NMR) for the PAPTAC block in P(VS)17A50 was 50, as estimated from the integral intensity of the methylene proton from the PAPTAC unit at 1.6 ppm compared with that for the phenyl protons of VBTAC and NaSS at 6.4–7.6 ppm (Figure 1a). PAAc49 has a DP(NMR) of 49, estimated from the ratio of integral intensities between main chain protons at 1.5–2.3 ppm and terminal phenyl protons at 7.3–7.9 ppm, derived from the CTA fragment (Figure 1b). The Mn(NMR) and Mn(theo) values for P(VS)17A50 were close. The Mw/Mn values estimated from GPC for P(VS)17A50 and PAAc49 (Figure S2) were 1.04 and 1.07, respectively, indicating that the polymers showed well-controlled structures. The DP, Mn, and Mw/Mn values for all polymers are summarized in Table 1. P(VS)17 could not be measured using GPC because it cannot dissolve in the eluent. FTIR spectrum was measured for P(VS)17A50. The pendant carbonyl signal can be seen at 1700 cm−1 (Figure S3). 3.2. Preparation and Characterization of PIC Vesicles The PIC vesicle was prepared by mixing P(VS)17A50 and PAAc49 with f+ = 0.5 at Cp = 2.0 g/L in D2O at pH 10 for the 1H NMR measurement (Figure 1c). For DLS measurements, the PIC vesicle was prepared at 0.1 g/L; however, the Cp was too low for NMR measurements so the PIC vesicle was prepared at Cp = 2.0 g/L. After mixing, the signals from the PAPTAC50 block and PAAc49 disappeared. The VBTAC unit’s methylene proton signal remained at 2.9 ppm, showing that the PAPTAC block and PAAc49 interacted to create the vesicle structure’s membrane and P(VS)17 produced shells. TEM was conducted to confirm the PIC aggregate shape. The TEM image was obtained for the stoichiometrically charge neutralized mixture of P(VS)17A50 and PAAc49 at pH 10 (Figure 2). The particles had a spherical shape with a contrasting white center and black edge, indicating a vesicle structure in the PIC aggregates. The average radius obtained from the TEM observations was 86.2 ± 17.3 nm, close to the hydrodynamic radius (Rh) estimated from the dynamic light scattering measurements (Rh = 86.6 nm). During TEM observation, the sample shrank due to the high vacuum conditions. The PIC vesicles collapsed as they shrank and stuck to the TEM grid membrane. At that time, the PIC vesicles expanded. Therefore, the size obtained from the TEM observation did not decrease much compared to the Rh value. The mixing ratio of the oppositely charged polymers determines the development of the PIC vesicle [40,41]. The plots of Rh, LSI, and zeta-potential with f+ at pH 10 were prepared (Figure 3). The Cp values in the PIC vesicle solutions were kept at 0.1 g/L. However, samples at f+ = 0 and 1 attributed to anionic PAAc49 and cationic P(VS)17A50, respectively, were prepared at Cp = 2.0 g/L because Rh and zeta-potential values could not be obtained because of very low LSI at Cp = 0.1 g/L. The Rh and LSI values of the PIC vesicle reached maximum values at f+ = 0.5, and then the zeta-potential reached 0 mV, suggesting the neutralization of the oppositely charged PAPTAC50 block and PAAc49. Meanwhile, the zeta-potential values of PAAc49 and the PAPTAC50 block at f+ = 0 and 1 were −27.7 and 23.8 mV, respectively, indicating negative and positive charges. The critical micelle concentration (CMC) of the PIC vesicle was determined to certify the thermodynamic equilibrium properties. The LSI of the PIC vesicle aqueous solutions was measured as a function of Cp. The PIC vesicle aqueous solution was prepared by mixing P(VS)17A50 and PAAc49 in water at pH 10 with f+ = 0.5 at Cp = 0.1 g/L; the solution was diluted to the desired Cp using water at pH 10. The CMC for the PIC vesicle in water was determined from the relationship between the LSI ratio (I/I0) and Cp (Figure 4). I and I0 correspond to the LSIs of the PIC vesicle aqueous solution and water at pH 10, respectively. The crossing point of the linear portions in the low and high Cp regions was defined as the CMC value, which was 2.47 × 10−3 g/L. The defined CMC value implies that the PIC vesicle is in a dynamic equilibrium state. The PIC vesicle is stable above the CMC. The PIC vesicle, however, can disintegrate below the CMC into a small assembly of P(VS)17A50/PAAc49 because of strong electrostatic interactions between them [33,42]. We studied the concentration dependence of the PIC vesicle because the SLS should be measured at different Cp values to perform the SLS measurements. The P(VS)17A50 and PAAc49 aqueous solutions were separately prepared at Cp = 0.1, 0.08, and 0.05 g/L. Then, P(VS)17A50 and PAAc49 Cp were mixed to form the PIC vesicle. The pH value of the aqueous solution was adjusted to 10 and left for one day to obtain an equilibrium state. Subsequently, the PIC vesicle aqueous solutions were diluted by water at pH 10 to adjust the target Cp. The Rh and LSI for the PIC vesicle were plotted as a function of the Cp after mixing (Figure 5). The Rh values for the PIC vesicles at Cp = 0.1, 0.08, and 0.05 g/L before mixing were 86.6, 71.2, and 67.8 nm, respectively. This indicated that the size of the PIC vesicles increased with an increasing Cp before mixing since the aggregation number (Nagg) of the PIC vesicle depends on the Cp. When the Cp increases, the Nagg increases to increase the size of the PIC vesicle [42]. Furthermore, the increase in the Nagg was confirmed by of the increase in the LSI value from 1.55 to 3.25 Mcps at Cp = 0.05 and 0.1 g/L, respectively (Figure S4). When the PIC vesicle was diluted with water at pH 10, the Rh values remained nearly constant, independent of the Cp. These findings suggest that the size of the PIC vesicle can be controlled by adjusting the Cp before mixing, and after the PIC vesicleformation they were stable against dilution. SLS measurements at 25 °C were used to determine further properties of the polymers and PIC vesicle, such as evident weight-average molecular weight (Mw(SLS)), and radius of gyration (Rg) (Figure S5). The RI increment (dn/dCp) values for all samples were also separately estimated. Nagg is the number of polymer chains required to form a PIC vesicle. The PIC vesicle was prepared by mixing P(VS)17A50 and PAAc49 at Cp = 0.1 g/L with f+ = 0.5 at pH 10 then the solution was diluted to 0.025 g/L by water at pH 10. Table 2 indicates the value of Mw(SLS), Nagg, Rg, and Rh for all samples. The values of Mw(SLS) for P(VS)17A50 and PAAc49 were 1.55 × 104 g/mol and 0.33 × 104 g/mol, respectively, which were close to the Mn(NMR) values estimated from the 1H NMR measurements. P(VS)17A50 and PAAc49 had Mn(NMR) values of 1.42 × 104 and 0.35 × 104 g/mol, respectively. The Mw(SLS) for the PIC vesicle was 1.06 × 108 g/mol. The value of Nagg for the PIC vesicle was estimated by dividing the Mw(SLS) of the PIC vesicle by that of the unimers and was found to be 5640. The Rg/Rh value for the PIC vesicle was calculated as 1.17, indicating a spherical shape [3,43,44]. The following equation can be used to calculate the density (d) of the polymers and PIC vesicles:(5) d=Mw(SLS)43πNARh3 The d values for P(VS)17A50, PAAc49, and the PIC vesicle were calculated to be 0.033, 0.028, and 0.064 g/cm3, respectively. The d value for the PIC vesicle was larger than those for P(VS)17A50 and PAAc49, suggesting that the polymer chains in the PIC vesicle are more densely packed than those of the unimers. The pendant carboxylic groups in PAAc49 are deprotonated under basic conditions. Therefore, the pH of the solution is an important factor in the PIC vesicle formation. To understand the effect of the pH value on the PIC vesicle, the Rh and LSI were measured as a function of the pH values. Initially, the PIC vesicle aqueous solution was prepared in water at Cp = 0.1 g/L with f+ = 0.5, then adjusted to the appropriate pH using NaOH and HCl aqueous solutions. The Rh, LSI, and zeta-potential values were plotted against the pH. By increasing pH from 3 to 5, the Rh of the PIC vesicle increased from 60.9 to 85.5 nm, then remained at ca. 86 nm above pH 5 (Figure 6a). The LSI value of the PIC vesicle showed the same tendency as that of the Rh. The acid dissociation constant (pKa) of PAAc is about 4.5 [45,46,47]. At pH above 4.5 of the PAAc49 aqueous solutions, the pendant carboxyl groups were deprotonated to generate anionic charges. The anionic PAAc49 with pendant carboxylate anions can interact with the cationic PAPTAC block in P(VS)17A50 to form the PIC vesicle through electrostatic interactions. This was supported by the zeta-potential value of the PIC vesicle aqueous solutions. At pH < 5, the zeta-potential of the PIC vesicle showed a positive value which can be attributed to a charge of the cationic PAPTAC block in P(VS)17A50 with protonated PAAc49 (Figure 6b). After increasing the pH value, at 5 ≤ pH the zeta-potential reached nearly 0 mV. It implied that the PIC vesicle formation was affected by the pH value, which means that the PIC vesicle can be formed above pH 5 and collapse below pH 5. The pH influence on the association behavior of the PIC vesicle is attributed to the pH-responsive property of PAAc49. The conformation of PAAc in water was reported, and a reversible transition from the coil to the globule dependence of the pH was caused by deprotonation of the carboxylic group. At low pH, PAAc was a compact globular shape. When the media of the solution became baser, PAAc expanded into an open coil shape [47,48]. Furthermore, the Rh and LSI values for P(VS)17A50 were constant at ca. 6 nm and 70 kcps, respectively, at all pH values ranging from 3 to 10 (Figure S6), suggesting that P(VS)17A50 was not affected by pH. The small Rh and LSI values indicated that the polymer was in the unimer state. The zeta-potential values of P(VS)17A50 and PAAc49 were measured as a function of pH (Figure S7). The zeta-potential for P(VS)17A50 stayed constant at ca. 17 mV, indicating that the polymer always gives a positive charge independent of the solution’s pH. However, the zeta-potential of PAAc49 decreased from pH 3 to 5, then remained constant from pH 6. The pKa for PAAc49 was 4.5; therefore, from pH 5, the PAAc49 deprotonated and indicated a negative charge. The 1H NMR spectra for the PIC vesicle suggested the A50/PAAc49 PIC membrane was covered with P(VS)17 amphoteric random copolymer shells. P(VS)17 showed the UCST behavior in aqueous solutions (Figure S8). Therefore, the PIC vesicle was expected to exhibit the UCST behavior owing to the P(VS)17 shells. The percentage transmittance (%T) of the PIC vesicle was plotted against temperature upon the cooling process indicating the UCST behavior of the PIC vesicle. The temperature dependences investigated the effect of pH on the UCST behavior on %T at different pH conditions ranging from pH 3 to 10 (Figure 7). The PIC vesicle was disrupted below pH 5 owing to protonation of the pendant carboxyl groups in PAAc49 (Figure 6). However, the aqueous solutions showed the UCST behavior excepted at pH 3. The phase transition temperature (Tp) of the PIC vesicle decreased from pH 4 to 6, then remained at 15 °C from pH 6. The UCST behavior of the PIC vesicle was affected by pH. We have studied incorporating a hydrophilic noncharged guest molecule, PEG-Py, into the PIC vesicles. The hydrophilic drugs can be encapsulated inside the PIC vesicle thanks to the inner water phase. Herein, P(VS)17A50 and PAAc49 solutions containing PEG-Py were mixed to form the PIC vesicle. The PEG-Py molecules which could not be incorporated into the PIC vesicle were removed by dialysis against pure water at pH 10. In comparison, a blank experiment was conducted, and the PEG-Py aqueous solution in the absence of the PIC vesicle was dialyzed with the same procedure. After dialysis for 20 days, fluorescence spectroscopy for PEG-Py in the presence and absence of the PIC aggregate was measured (Figure 8). The maximum emission wavelengths at 376 and 397 nm for PEG-Py were observed in the PIC vesicle solution, suggesting that the PIC vesicles can encapsulate PEG-Py. Hydrophobic pyrene molecules associate to produce excimer emission with a peak at 480 nm. PEG-Py can dissolve in water without interpolymer association in this study because no excimer emission was observed from PEG-Py fluorescence. In contrast, no signal of PEG-Py was detected from the reference solution in the absence of the PIC vesicle owing to the removal of the small PEG-Py molecules after dialysis membrane with a pore size of 10 nm. These findings suggested the PIC aggregate might be a vesicle structure because it can encapsulate hydrophilic PEG-Py. From the calibration curve of PEG-Py (Figure S9) and the fluorescence intensity, the weight of PEG-Py encapsulated into a PIC vesicle was estimated to be 0.0012 g. The LE and LC were calculated to be 12% and 1.2%, respectively, from Equations (1) and (2). Using an HCl aqueous solution, the PIC vesicle encapsulating PEG-Py at pH 10 was adjusted to pH 3. A dialysis bag with a 10-nm pore size was used to dialyze the solution with an aqueous solution at pH 3. The solution sample inside the dialysis bag was taken out at varying times to measure the fluorescence intensities to calculate the release rate of the encapsulated PEG-Py from the PIC vesicle (Figure 9). After 24 h, the release amount of PEG-Py at pH 3 was 86.2%. Noncharged hydrophilic guest molecules can be encapsulated into the PIC vesicle at pH 10 and released at pH 3 owing to the PIC vesicle’s disruption. The release rate of the encapsulated guest molecule may be changed above and below the UCST of the P(VS)17 shell. 4. Conclusions P(VS)17A50, a cationic diblock copolymer, and an anionic homopolymer, PAAc49, were successfully synthesized via a RAFT polymerization with a well-controlled structure. The pH- and thermo-responsive P(VS)17A50/PAAc49 PIC vesicle was prepared by mixing P(VS)17A50 and PAAc49 in basic aqueous solutions at room temperature due to electrostatic interactions. The Rh and LSI values of the PIC vesicle showed maximum values at f+ = 0.5, whereas the zeta-potential was near 0 mV. At low pH, the PIC vesicle disintegrated. The spherical shape with a vesicle structure was confirmed by TEM. The Rg/Rh ratio obtained from SLS and DLS measurements showed a value close to 1, indicating the spherical shape of the PIC vesicle. The thermo-responsive P(VS)17 shells caused UCST behavior in the PIC vesicle. The thermo-responsive behavior of the PIC vesicle was affected by the pH of the solution. The PIC vesicle can encapsulate water-soluble nonionic guest molecules at a high pH and release them under acidic conditions, implying that it could be used as a DDS material. The interaction of the PIC vesicle with the cell can be controlled by temperature because the surface hydrophobicity increases with decreasing temperature. Therefore, it is expected that the PIC vesicle can be taken in the cell at a low temperature below the UCS of the P(VS)17 shells. It is known that the pH around the cancer tissue is low. The encapsulated hydrophilic anticancer drugs can be released under acidic conditions from the P(VS)17A50/PAAc49 PIC vesicle. Acknowledgments The authors would like to express their thanks to Thi Ngoc Anh Doan, Shota Fujii, Kazuo Sakurai for their suggestions. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/polym14091659/s1, Figure S1. 1H NMR spectrum for P(VBTAC/NaSS)17 in D2O containing 1.2-M NaCl at 70 °C; Figure S2. Gel-permeation chromatography (GPC) elution curves for (a) P(VS)17A50 and (b) PAAc49 monitored with refractive index (RI) detector working at 40 °C; Figure S3. FTIR spectrum for P(VS)17A50; Figure S4. Light scattering intensity (LSI) for the P(VS)17A50/PAAc49 PIC vesicles with f+ = 0.5 in water at pH 10 as a function of polymer concentration after mixing P(VS)17A50 and PAAc49. At 25 °C the PIC aggregates solution at 0.1 (triangle), 0.08 (circle), and 0.05 (diamond) g/L were diluted with pH 10 aqueous solution; Figure S5. Typical examples of Zimm plots for (a) P(VS)17, (b) PAAc49, and (c) the P(VS)17A50/PAAc49 PIC vesicle with f+ = 0.5 in aqueous solutions at pH 10; Figure S6. Hydrodynamic radius (Rh, diamond) and light scattering intensity (LSI, circle) for P(VS)17A50 at Cp = 2.0 g/L in water as a function of pH at 25 °C; Figure S7. Zeta-potential for P(VS)17A50 (circle) and PAAc49 (triangle) in water at Cp = 2.0 g/L as a function of pH at 25 °C; Figure S8. Percent transmittance (%T) at 700 nm for the 0.1-M NaCl aqueous P(VS)17 solution at Cp = 1.0 g/L as a function of temperature upon the heating and cooling processes; Figure S9. (a) Fluorescence spectra for PEG-Py in water at pH 10 excited at 334 nm at various PEG-Py concentrations ([PEG-Py]) and (b) a calibration curve of fluorescence intensity at 376 nm using [PEG-Py] in water at pH 10. Click here for additional data file. Author Contributions Conceptualization, T.T.P. and S.-i.Y.; methodology, T.T.P., T.D.P. and S.-i.Y.; software, T.T.P.; validation, T.T.P. and S.-i.Y.; formal analysis, T.T.P. and S.-i.Y.; investigation, T.T.P.; writing—original draft preparation, T.T.P. and S.-i.Y.; writing—review and editing, T.T.P., T.D.P. and S.-i.Y.; visualization, T.T.P. and S.-i.Y.; supervision, S.-i.Y.; project administration, S.-i.Y.; funding acquisition, S.-i.Y. All authors have read and agreed to the published version of the manuscript. Funding This research was partially supported by KAKENHI grants (21H02005, 21K19931, 21H05027, 21H05535) from the Japan Society for the Promotion of Science (JSPS), JSPS Bilateral Joint Research Projects (JPJSBP120203509), the Cooperative Research Program of “Network Joint Research Center for Materials and Devices (20214044)”, the International Collaborative Research Program of Institute for Chemical Research, Kyoto University (2022-121), and MEXT Promotion of Distinctive Joint Research Center Program (JPMXP 0621467946). Data Availability Statement The data used to support the findings of this study are available from the corresponding authors upon request. Conflicts of Interest The authors declare no conflict of interest. Figures, Scheme and Tables polymers-14-01659-sch001_Scheme 1 Scheme 1 (a) Chemical structures of P(VBTAC/NaSS)17-b-PAPTAC50 ((P(VS)17A50) and PAAc49, and (b) conceptual formation of the PIC vesicle consist of P(VS)17A50 and PAAc49. Figure 1 1H NMR with water suppression spectra for P(VS)17A50 (a), PAAc49 (b), and the P(VS)17A50/PAAc49 PIC vesicle with f+ = 0.5 (c) at Cp = 2.0 g/L in D2O at pH 10 at 25 °C. Figure 2 (a) TEM image and (b) histogram of particle size distribution for the P(VS)17A50/PAAc49 PIC vesicle with f+ = 0.5 at Cp = 0.1 g/L in water at pH 10. Figure 3 (a) Hydrodynamic radius (Rh, diamond) and light scattering intensity (LSI, circle) and (b) zeta-potential for the P(VS)17A50/PAAc49 PIC vesicle at Cp = 0.1 g/L in water at pH 10 as a function of f+ at 25 °C. Figure 4 Light scattering intensity ratio (I/I0) for the P(VS)17A50/PAAc49 PIC vesicle in water pH 10 as a function of polymer concentration (Cp). I and I0 are the light scattering intensities of the PIC vesicle solution and solvent, respectively. Figure 5 Hydrodynamic radius (Rh) light scattering intensity (LSI) for the P(VS)17A50/PAAc49 PIC vesicle with f+ = 0.5 in water at pH 10 as a function of polymer concentration after mixing P(VS)17A50 and PAAc49. The PIC vesicle solution at 0.1 (triangle), 0.08 (circle), and 0.05 (diamond) g/L were diluted with pH 10 aqueous solution at 25 °C. Figure 6 (a) Hydrodynamic radius (Rh, diamond) and LSI (circle), and (b) zeta-potential for the P(VS)17A50/PAAc49 PIC vesicle with f+ = 0.5 at Cp = 0.1 g/L in water as a function of pH at 25 °C. Figure 7 (a) Percent transmittance (%T) at 700 nm for the P(VS)17A50/PAAc49 PIC vesicle aqueous solution with f+ = 0.5 as a function of temperature upon the cooling process at different pH at Cp = 0.1 g/L, and (b) pH dependence of the phase transition temperature (Tp) for the PIC vesicle aqueous solution with f+ = 0.5 at Cp = 0.1 g/L. Figure 8 Fluorescence spectra for PEG-Py after dialysis against water at pH 10 for 10 days excited at 334 nm in the presence (—) and absence (—) of the PIC vesicle. Figure 9 Release profiles of the encapsulated PEG-Py from the PIC vesicle. polymers-14-01659-t001_Table 1 Table 1 Degree of polymerization (DP), number-average molecular weight (Mn), and polydispersity index (Mw/Mn). Sample DP (theo) DP (NMR) Mn(theo) 1 × 10−4 (g/mol) Mn(NMR) 2 × 10−4 (g/mol) Mn(GPC) 3 × 10−4 (g/mol) Mw/Mn P(VS)17 17 - 0.38 - - - P(VS)17A50 49 50 1.40 1.41 5.70 1.04 PAAc49 36 49 0.26 0.34 1.40 1.07 1 Calculated from Equation (4) 2 Obtained from 1H NMR 3 Obtained from GPC. polymers-14-01659-t002_Table 2 Table 2 Dynamic and static light scattering data for P(VS)17A50, PAAc49, and the P(VS)17A50/PAAc49 PIC vesicle. Sample Mw(SLS) × 10−4 (g/mol) N agg Rg (nm) Rh (nm) Rg/Rh dn/dCp d1 (g/cm3) P(VS)17A50 1.55 1 - 5.7 - 0.1323 0.033 PAAc49 0.33 1 - 5.3 - 0.2278 0.028 PIC vesicle 10,600 5640 101 86.6 1.17 0.0736 0.064 1 Calculated from Equation (5). Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Lankalapalli S. Kolapalli V.R.M. Polyelectrolyte complexes: A review of their applicability in drug delivery technology Indian J. Pharm. Sci. 2009 71 481 487 10.4103/0250-474X.58165 20502564 2. Insua I. Wilkinson A. Fernandez-Trillo F. Polyion complex (PIC) particles: Preparation and biomedical applications Eur. Polym. J. 2016 81 198 215 10.1016/j.eurpolymj.2016.06.003 27524831 3. Ohara Y. Nakai K. Ahmed S. Matsumura K. Ishihara K. Yusa S. pH-responsive polyion complex vesicle with polyphosphobetaine shells Langmuir 2019 35 1249 1256 10.1021/acs.langmuir.8b00632 29940726 4. Slyusarenko N.V. Vasilyeva N.Y. Kazachenko A.S. Gerasimova M.A. Romanchenko A.S. Slyusareva E.A. Synthesis and properties of interpolymer complexes based on chitosan and sulfated arabinogalactan Polym. Sci. B 2020 62 272 278 10.1134/S1560090420020050 5. Bhat A.H. Khan I. Bhawani S.A. Abdul Rahim M.K. Gazal U. Stimuli-Responsive Polymer Composites for Fabric Applications. Smart Polym. Nanocomposites Woodhead Publishing Sawston, UK 2021 Chapter Fifteen 6. Justus C.R. Dong L. Yang L.V. Acidic tumor microenvironment and pH-sensing G protein-coupled receptors Front. Physiol. 2013 4 354 10.3389/fphys.2013.00354 24367336 7. Ohno S. Ishihara K. Yusa S. Formation of polyion complex (PIC) micelles and vesicles with anionic pH-responsive unimer micelles and cationic diblock copolymers in water Langmuir 2016 32 3945 3953 10.1021/acs.langmuir.6b00637 27048989 8. Chen J. Ding J. Zhang Y. Xiao C. Zhuang X. Chen X. Polyion complex micelles with gradient pH-sensitivity for adjustable intracellular drug delivery Polym. Chem. 2015 6 397 405 10.1039/C4PY01149J 9. Oishi M. Sasaki S. Nagasaki Y. Kataoka K. pH-responsive oligodeoxynucleotide (ODN)-poly(ethylene glycol) conjugate through acid-labile β-thiopropionate linkage: Preparation and polyion complex micelle formation Biomacromolecules 2003 4 1426 1432 10.1021/bm034164u 12959615 10. Luo Y.L. Yuan J.F. Liu X.J. Xie H. Gao Q.Y. Self-assembled polyion complex micelles based on PVP-b-PAMPS and PVP-b-PDMAEMA for drug delivery J. Bioact. Compat. Polym. 2010 25 292 304 10.1177/0883911510362459 11. Tian S. Liu G. Wang X. Wu T. Yang J. Ye X. Zhang G. Hu J. Liu S. pH-regulated reversible transition between polyion complexes (PIC) and hydrogen-bonding complexes (HBC) with tunable aggregation-induced emission ACS Appl. Mater. Interfaces 2016 8 3693 3702 10.1021/acsami.5b08970 26584477 12. Oishi M. Nagatsugi F. Sasaki S. Nagasaki Y. Kataoka K. Smart polyion complex micelles for targeted intracellular delivery of PEGylated antisense oligonucleotides containing acid-labile linkages Chembiochem 2005 6 718 725 10.1002/cbic.200400334 15756696 13. Kim D. Matsuoka H. Saruwatari Y. Formation of sulfobetaine-containing entirely ionic PIC (polyion complex) micelles and their temperature responsivity Langmuir 2020 36 10130 10137 10.1021/acs.langmuir.0c01577 32787061 14. De Santis S. Ladogana R.D. Diociaiuti M. Masci G. Pegylated and thermosensitive polyion complex micelles by self-assembly of two oppositely and permanently charged diblock copolymers Macromolecules 2010 43 1992 2001 10.1021/ma9026542 15. Li G. Guo L. Meng Y. Zhang T. Self-assembled nanoparticles from thermo-sensitive polyion complex micelles for controlled drug release Chem. Eng. J. 2011 174 199 205 10.1016/j.cej.2011.08.079 16. Pua M.L. Yoshitomi T. Chonpathompikunlert P. Hirayama A. Nagasaki Y. Redox-active injectable gel using thermo-responsive nanoscale polyion complex flower micelle for noninvasive treatment of local inflammation J. Control. Release 2013 172 914 920 10.1016/j.jconrel.2013.10.009 24157475 17. Kim D. Matsuoka H. Yusa S. Saruwatari Y. Collapse behavior of polyion complex (PIC) micelles upon salt addition and reforming behavior by dialysis and its temperature responsivity Langmuir 2020 36 15485 15492 10.1021/acs.langmuir.0c02456 33325225 18. Dong S. Jiang Y. Qin G. Liu L. Zhao H. Methionine-based pH and oxidation dual-responsive block copolymer: Synthesis and fabrication of protein nanogels Biomacromolecules 2020 21 4063 4075 10.1021/acs.biomac.0c00879 32914964 19. Ma R. Yang H. Li Z. Liu G. Sun X. Liu X. An Y. Shi L. Phenylboronic acid-based complex micelles with enhanced glucose-responsiveness at physiological pH by complexation with glycopolymer Biomacromolecules 2012 13 3409 3417 10.1021/bm3012715 22957842 20. Jin Q. Cai T. Wang Y. Wang H. Ji J. Light-responsive polyion complex micelles with switchable surface charge for efficient protein delivery ACS Macro Lett. 2014 3 679 683 10.1021/mz500290s 21. Zhou W. Ding P. Zhang C. Wu B. Guo X. Stuart M.A.C. Zhuo X. Wang J. Regulated polyion complex vesicles for efficient photothermal therapy Adv. Funct. Mater. 2022 2022 2108729 10.1002/adfm.202108729 22. Chiang W.H. Hsu Y.H. Chen Y.W. Chern C.S. Chiu H.C. Thermoresponsive interpolymeric complex assemblies from co-association of linear PAAc homopolymers with PNIPAAm segments containing PAAc-based graft copolymer Macromol. Chem. Phys. 2011 212 1869 1878 10.1002/macp.201100124 23. Zhang D. Liu J. Xie H. Zhu A. Xu Y. Zeng B. Luo W. Dai L. Polyion complex micelles formed by azobenzene-based polymer with multi-responsive properties J. Appl. Polym. Sci. 2021 138 50580 10.1002/app.50580 24. Ishii S. Kaneko J. Nagasaki Y. Dual stimuli-responsive redox-active injectable gel by polyion complex based flower micelles for biomedical applications Macromolecules 2015 48 3088 3094 10.1021/acs.macromol.5b00305 25. Liu Y. Li C. Wang H.Y. Zhang X.Z. Zhuo R.X. Synthesis of thermo- and pH-sensitive polyion complex micelles for fluorescent imaging Chem. A Eur. J. 2012 18 2297 2304 10.1002/chem.201102704 26. Ren J. Zhang Y. Zhang J. Gao H. Liu G. Ma R. An Y. Kong D. Shi L. pH/sugar dual responsive core-cross-linked PIC micelles for enhanced intracellular protein delivery Biomacromolecules 2013 14 3434 3443 10.1021/bm4007387 24063314 27. Voets I.K. Moll P. Aqli A. Jerome C. Detrembleur C. Waard P. Keizet A. Stuart M. Temperature responsive complex coacervate core micelles with a PEO and PNIPAAm corona J. Phys. Chem. B 2008 35 10833 10840 10.1021/jp8014832 28. Kawata Y. Kozuka S. Yusa S. Thermo-responsive behavior of amphoteric diblock copolymers bearing sulfonate and quaternary amino pendant groups Langmuir 2019 35 1458 1464 10.1021/acs.langmuir.8b01684 30086641 29. Kazachenko A.S. Akman F. Malyar Y.N. Issaoui N. Vasilieva N.Y. Karacharov A.A. Synthesis optimization, DFT and physicochemical study of chitosan sulfates J. Mol. Struct. 2021 1245 131083 10.1016/j.molstruc.2021.131083 30. Kazachenko A. Akman F. Medimagh M. Issaoui N. Vasilieva N. Malyar Y.N. Sudakova I.G. Karacharov A. Miroshnikova A. Marzook Al-Dossary O.M. Sulfation of Diethylaminoethyl-cellulose: QTAIM topological analysis and experimental and DFT studies of the properties ACS Omega 2021 6 22603 22615 10.1021/acsomega.1c02570 34514232 31. Rabiee A. Ershad-Langroudi A. Jamshidi H. Polyacrylamide-based polyampholytes and their applications Rev. Chem. Eng. 2014 30 501 519 10.1515/revce-2014-0004 32. North S.M. Armes S.P. Synthesis of polyampholytic diblock copolymers via RAFT aqueous solution polymerization Polym. Chem. 2021 12 4846 4855 10.1039/D1PY01020D 33. Nakahata R. Yusa S. Preparation of water-soluble polyion complex (PIC) micelles covered with amphoteric random copolymer shells with pendant sulfonate and quaternary amino groups Polymers 2018 10 205 10.3390/polym10020205 34. Sharker K.K. Ohara Y. Shigeta Y. Ozoe S. Yusa S. Upper critical solution temperature (UCST) behavior of polystyrene-based polyampholytes in aqueous solution Polymers 2019 11 265 10.3390/polym11020265 35. Seki K. Tirrell D.A. pH-Dependent complexation of poly(acrylic acid) derivatives with phospholipid vesicle membranes Macromolecules 1984 17 1692 1698 10.1021/ma00139a009 36. Li R. Zhang C. Wang C. Cheng Y. Hu D. Study on the mechanism of the reversible color change of polyacrylic acid modified gold nanoparticles responding to pH Materials 2021 14 3679 10.3390/ma14133679 34279250 37. Fukumoto H. Ishihara K. Yusa S. Thermo-responsive behavior of mixed aqueous solution of hydrophilic polymer with pendant phosphorylcholine group and poly(acrylic acid) Polymers 2021 13 148 10.3390/polym13010148 38. Mitsukami Y. Donovan M.S. Lowe A.B. Mccormick C.L. Water-soluble polymers. 81. Direct synthesis of hydrophilic styrenic-based homopolymers and block copolymers in aqueous solution via RAFT Macromolecules 2001 34 2248 2256 10.1021/ma0018087 39. Kozuka S. Kuroda K. Ishihara K. Yusa S. Interpolymer association of amphiphilic diblock copolymers bearing pendant siloxane and phosphorylcholine groups J. Polym. Sci. Part A Polym. Chem. 2019 57 1500 1507 10.1002/pola.29407 40. Nakamura N. Mochida Y. Toh K. Fukushima S. Anraku Y. Cabral H. Effect of mixing ratio of oppositely charged block copolymers on polyion complex micelles for in vivo application Polymers 2020 13 5 10.3390/polym13010005 41. Anraku Y. Kishimura A. Oba M. Yamasaki Y. Kataoka K. Spontaneous formation of nanosized unilamellar polyion complex vesicles with tunable size and properties J. Am. Chem. Soc. 2010 132 1631 1636 10.1021/ja908350e 20078126 42. Pramod P.S. Takamura K. Chaphekar S. Balasubramanian N. Jayakannan M. Dextran vesicular carriers for dual encapsulation of hydrophilic and hydrophobic molecules and delivery into cells Biomacromolecules 2012 13 3627 3640 10.1021/bm301583s 23082727 43. Burchard W. Static and dynamic light scattering from branched polymers and biopolymers Springer Berlin/Heidelberg, Germany 1983 44. Geismann C. Tomicki F. Ulbricht M. Block copolymer photo-grafted poly(ethylene terephthalate) capillary pore membranes distinctly switchable by two different stimuli Sep. Sci. Technol. 2009 44 3312 3329 10.1080/01496390903212755 45. Michaels A.S. Morelos O. Polyelectrolyte adsorption by kaolinite Ind. Eng. Chem 1995 47 1801 1809 10.1021/ie50549a029 46. Paek K. Yang H. Lee J. Park J. Kim B.J. Efficient colorimetric pH sensor based on responsive polymer-quantum dot integrated graphene oxide ACS Nano 2014 8 2848 2856 10.1021/nn406657b 24548181 47. Katchalsky A. Eisenberg H. Molecular weight of polyacrylic and polymethacrylic acid J. Polym. Sci. 1951 6 145 154 10.1002/pol.1951.120060202 48. Lim P.F.C. Chee L.Y. Chen S.B. Study of interaction between cetyltrimethylammonium bromide and poly(acrylic acid) by rheological measurements J. Phys. Chem. B 2003 107 6491 6496 10.1021/jp027864m

PMC009xxxxxx/PMC9099633.txt

==== Front Polymers (Basel) Polymers (Basel) polymers Polymers 2073-4360 MDPI 10.3390/polym14091675 polymers-14-01675 Article Stress–Strain Model for Lightweight Aggregate Concrete Reinforced with Carbon–Polypropylene Hybrid Fibers Yang Xue https://orcid.org/0000-0003-1493-5205 Wu Tao * Liu Xi Yan Libo Academic Editor School of Civil Engineering, Chang’an University, Xi’an 710061, China; ms_yangxue@163.com (X.Y.); lliuxii@163.com (X.L.) * Correspondence: wutaochd0922@yahoo.com; Tel.: +86-139-9132-2194 20 4 2022 5 2022 14 9 167521 3 2022 19 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). This research aimed to investigate the hybrid effects of carbon and polypropylene fibers on the stress–strain behavior of lightweight aggregate concrete (LWAC). The considered test variables were two fiber volume fractions of 0.2% and 0.4% and two water/binder ratios of 0.27 and 0.30. Eighteen groups of prisms fabricated with fiber-reinforced LWAC were tested under axial compressive load. Experimental studies were carried out to analyze the influence of different fiber combinations on the complete stress–strain behavior. It was found that the carbon–polypropylene hybrid fibers led to toughness enhancement that was numerically more significant than the sum of individual fibers, indicating a positive synergistic effect between them. Finally, a mathematical expression of the stress–strain curve accounting for the fiber combinations was developed. Compared with existing stress–strain models, the proposed model shows better accuracy in predicting the effect of carbon and polypropylene fibers in both single and hybrid forms on the stress–strain curve of LWAC. compressive stress–strain relationship lightweight aggregate concrete carbon fiber polypropylene fiber hybrid fiber reinforcement ==== Body pmc1. Introduction Lightweight aggregate concrete (LWAC) has been successfully used in structural engineering due to its advantages over conventional concrete, including lower density, superior thermal insulation, and higher specific strength [1,2]. However, disadvantages such as higher brittle texture and lower mechanical properties have restricted the wide range of applications for LWAC [3]. Incorporating fibers into concrete as a single or hybrid form is confirmed as an effective way of compensating for the adverse effects of LWAC. As is known, the most beneficial characteristic of fiber is the crack-bridging mechanism, which can significantly increase the toughness and post-cracking ductility of concrete [4,5,6]. Different types of fibers, such as steel, glass, carbon, nylon, and polypropylene, have been used to produce fiber-reinforced LWAC (FLWAC) [7]. Adding steel fiber to LWAC leads to a significant improvement in the mechanical performance of concrete but increases the density of LWAC. Instead, the incorporation of non-metallic fibers (carbon, basalt, polypropylene, etc.) in concrete has become attractive because of the high performance and low density [8,9]. LWAC comprising two or more types of fiber was investigated to achieve a positive synergistic response. Therefore, combinations of high-strength carbon fiber and high-ductility polypropylene fiber are hoped to reinforce LWAC at multiple scales. The stress–strain relationship is crucial to the axial design of short columns and the flexural designing of slabs and beams [10,11]. Moreover, a thorough understanding of comprehensive compressive stress–strain behavior is essential for the nonlinear analysis of structures and the derivation of design behavior curves [10]. However, the effect of non-metallic fibers on the compressive stress–strain response of LWAC may be different from that on the properties of normal-weight concrete (NWC). To design and analyze the performance of the structural application of LWAC reinforced with carbon and polypropylene fibers, it is necessary to investigate relevant mechanical properties and propose a stress–strain model. Continuing efforts regarding experimental and theoretical research have been dedicated to generating compressive stress–strain curves, and several empirical models have been established in the last few decades, such as those proposed by Carreira and Chu [12], Guo [13], and Wee et al. [14]. It was reported that the stress–strain response of concrete depended on the concrete composition [15]. Nevertheless, the constituents of LWAC differ from those of NWC, and it is uncertain whether a model proposed based on NWC can provide good predictions for LWAC. Meanwhile, the effect of fibers is not taken into account in the parameters of the proposed models. Studies on the compressive behavior of fiber-reinforced NWC as well as LWAC have been carried out. Tasnimi [16] derived a stress–strain model applicable to both LWAC and NWC, but it did not reflect the properties of lightweight aggregate (LWA). Recently, a unified model was developed by Lim and Ozbakkaloglu [17] to describe the stress–strain behavior of both NWC and LWAC, but the proposed model is unsuitable for fiber-reinforced concrete. Wang et al. [18] proposed a model to predict the compressive stress–strain behavior of concrete reinforced with basalt–polypropylene hybrid fibers. Currently, the available evidence regarding the axial compressive behavior of FLWAC is still insufficient. The axial compressive behavior of LWAC reinforced with non-metallic fiber has yet to be investigated, and it is essential to develop a model for the prediction of this concrete. This study explores the effects of single and hybrid carbon and polypropylene fibers on the compressive stress–strain behavior and model of LWAC. From the experimental results, analytical expressions for compressive strength and critical strain accounting for the fiber combinations were proposed. Finally, to capture the complete stress–strain response of FLWAC, a stress–strain model suggesting the peak stress and critical strain was derived. 2. Experimental Details 2.1. Materials Lightweight Aggregate: Artificially expanded shale ceramic with crushed shape provided by Guangda Co., Ltd. (Yichang, China) was selected as coarse LWA. According to Chinese Specification GB/T 17431.2-2010 [19], its properties and particle size distribution are provided in Table 1. The chemical composition of LWA provided by the manufacturer is also listed in Table 1. Fibers: Two types of fibers, namely, carbon and polypropylene fibers, were used, provided by Anjie composite material factory (Haining, China) and Hansen Co., Ltd. (Wuhan, China). Their respective features and properties, as provided by the suppliers, are listed in Table 2. Binder: Ordinary Portland cement, classified as P.O 42.5, was used, complying with Chinese Specification GB 175-2007 [20]. Silica fume and class F fly ash in accordance with GB/T 18736-2017 [21] were added as active mineral admixtures. The fine aggregate used was medium sand sieved to 4 mm with a bulk density of 1530 kg/m3. A high-performance polycarboxylate-based superplasticizer was employed in all mixtures to improve fiber dispersion and adjust the fluidity in fresh LWAC mixtures. 2.2. Preparation and Details of Specimens The effective water/binder ratio (W/B) employed was 0.27 and 0.3 for Series 1 and Series 2, respectively. Nine concrete mixtures were prepared with the same binder, aggregate, water, and superplasticizer for each series. In addition, silica fume and fly ash were applied to replace 0.08 and 0.12 by mass of cement, respectively. The mix proportions of plain LWAC are provided in Table 3. Accordingly, as shown in Table 4, different combinations of carbon and polypropylene fibers (volume fraction = 0.2% and 0.4%) were incorporated into the plain LWAC mixtures to investigate both the individual effects and synergistic effect of fiber reinforcement. The concrete was prepared using a forced-action mixer under lab conditions. Before mixing, LWA was first pre-wetted with additional water to ensure an SSD condition. Firstly, powder-type ingredients such as cement, fly ash, silica fume, and medium sand were dry mixed for about 3 min. Next, polypropylene fibers were added and mixed in dry state for another 2 min, and after that, a half-quantity of pre-mixed water with superplasticizer was introduced to the dry mix. The mixture was mixed for a further 3 min, followed by the addition of pre-soaked LWA. To compensate for the absorbed water of carbon fiber, it is suggested to first blend carbon fiber with the remaining water to ensure sufficient dispersion. Subsequently, 3 min after the aggregate was added, the remaining water was gradually poured into the mixer. Mixing was performed for another 3~5 min until a consistent mixture was obtained. Simultaneously, fresh concrete mixtures were cast into standard molds. All specimens were compacted on a high-frequency vibrating table for 30 s and kept covered with plastic sheets for 24 h before demolding. The hardened specimens were cured in a standard curing room for 28 days and then placed at room temperature until testing time. For each concrete mixture, three 100 mm cubes were cast and cured under the same conditions as the specimens to determine the cubic compressive strength in accordance with the Chinese Specification GB/T 50081-2007 [22]. 2.3. Experimental Instrumentation and Methods Axial compressive tests were performed on 100 mm × 100 mm × 300 mm prismatic specimens in compliance with the Chinese Specification GB/T 50081-2007 [22]. The specimens were tested on a 1000 kN electro-hydraulic servo universal testing machine at a displacement control rate of 0.02 mm/min. To avoid the end of concrete crushing, steel collars were employed to confine the top and bottom of the specimen. The axial deformation was measured using four linear variable displacement transducers (LVDTs) that were mounted at 90° around the test section of the specimen (see Figure 1). Before testing, the specimen was preloaded from 0 to 5 MPa and then unloaded to 0.5 MPa, which formed one loading–unloading cycle. Thus, more than three cycles were applied to the specimens to guarantee axial loading and prevent the slackness of the system [23]. During the test, the applied load and corresponding axial deformation were continuously recorded at a sampling frequency of 10 Hz to describe the stress–strain curve of the specimen. Compressive stress was the applied load divided by the cross-sectional area of each specimen, and then compressive strain was calculated from the average value of four LVDTs. With the stress and corresponding strain, the experimental stress–strain curves can be plotted. Compressive strength (fc) and critical strain (εc), corresponding to the peak point, were obtained directly from the experimental stress–strain curve. The initial tangential elastic modulus (Ec) and peak secant elastic modulus (E0) were calculated using Equations (1) and (2), respectively:(1) Ec=σc2−σc1ε2−0.00005 (2) E0=σcεc where σc2 is the compressive stress at the point of fc/3; σc1 is the compressive stress at the point of the strain of 0.0005; and ε2 is the compressive strain at the point of the stress of σc2. 3. Results and Discussion 3.1. Failure Mechanism The effect of fiber reinforcement on LWAC can be directly characterized by the modification of the crack pattern and the failure mode. Specimens containing the same fiber combination but with different W/B show similar failure modes, such as Plain–a and Plain–b, so the failure mode of specimens with W/B of 0.3 was taken as an example to analyze the effect of fibers on the concrete failure mechanism. As shown in Figure 2, failure modes for carbon-fiber-reinforced LWAC (CFLWAC) and polypropylene-fiber-reinforced LWAC (PFLWAC) expressed significant differences from those of hybrid-fiber-reinforced LWAC (HFLWAC). In the specimen of plain LWAC, noticeable concrete crushing can be recognized from dense cracks forming after a compressive test. The failure mode of specimens of LWAC with only carbon fiber was similar to that of plain LWAC, whereas the degree of concrete spalling was relatively lighter for concrete containing carbon fiber. Cracks extended through the LWA were accompanied by the rupture of carbon fiber, which can be attributed to the low strength of LWA as well as the low ductility of carbon fiber. Moreover, fractured LWA on the crack surface can also explain the higher brittleness of LWAC as compared with NWC with the same strength grade. In the case of LWAC containing polypropylene fiber, the specimen maintained integrity during the testing. However, there is a beneficial modification on the failure pattern in that there is no apparent concrete spalling, but dense cracks can be observed. This phenomenon may be due to the bridging effect of polypropylene fiber that restrains the cracking and lateral deformation of concrete. In addition, the ruptured polypropylene fiber can be viewed on the crack surface because of the high ductility of the polypropylene fiber, which allows a large deformation when subjected to tensile stress. Additionally, the difference in compressive strength was taken as partly responsible for the difference in failure modes between CFLWAC and PFLWAC. The compressive strength of LWAC containing carbon fiber was much higher than that solely containing polypropylene fiber. 3.2. Compressive Stress–Strain Behavior Figure 3 shows the typical stress–strain curve of LWAC in compression. During the loading procedure, the specimen’s response to axial compression was recorded to illustrate crack patterns corresponding to different stages in the stress–strain curve. As shown in the figure, four cracking stages during the compressive tests can be identified in the stress–strain curve. In stage 1, the stress can be assumed to be linearly increasing from point O to point A, illustrated by the elastic deformation of the aggregate and the cement paste. This linearity is maintained until approaching 90% of the peak point and is considerably larger than that of NWC (30~45%) [24]. Microcracks that existed before loading would not extend until the end of this stage. Point A is recognized as the point where the stress–strain curve deflects from linearity. Meanwhile, the cracks tend to propagate from point A. During stage 2, the strain increases more quickly than before, manifesting as a continuous decrease in the curve’s gradient. As the stress steadily increases, vertical cracks appear on the specimen until one major crack reaches its critical length at point B. Point B is referred to as the peak point, from which the peak stress and critical strain can be obtained. In the post-peak region, the specimen enters the cracking instability stage, and cracks propagate automatically, though the stress is decreasing. The duration of stage 3 is relatively short, as the stress decreases rapidly to about 50% of the peak stress, suggesting the brittleness of LWAC. Point C is referred to as the inflection point in the descending branch. In the case of stage 4, the load capacity mainly consists of the frictional resistance and residual stress. Figure 4 and Figure 5 present the effects of different fiber combinations on the compressive stress–strain behavior of LWAC with W/B of 0.27 and 0.3, respectively. All experimental stress–strain curves show a similar shape in ascending and descending branches, but the critical points referred to in Figure 3 change with changes in fiber dosages and W/B. As shown in the figure, the initial linear stage of the stress–strain curve shows no noticeable difference with the addition of fibers. However, the nonlinear stage of the ascending branch is significantly affected. When adding fibers to the LWAC, the drop in the ascending branch is flatter, and the nonlinear stage of the ascending branch is wider than that of the plain LWAC, indicating a larger energy dissipation capacity for the FLWAC. For LWAC containing a single type of fiber, the gradient of the descending branch is close to that of the corresponding plain LWAC, as plotted in Figure 4a and Figure 5a. Instead, as presented in Figure 4b and Figure 5b, the descending parts in HFLWAC become much flatter than those of plain LWAC, suggesting an increase in energy dissipation capacity when referring to the post-peak portion. 3.3. Stress–Strain Characteristics Table 5 contains the characteristic values of the experimental stress–strain curves, including peak stress, critical strain, initial tangential elastic modulus, and peak secant elastic modulus. The cubic compressive strength (fcu) obtained from the tests is also listed in Table 5. As can be seen, both axial and cubic compressive strengths decreased with fiber addition, except with the addition of only carbon fiber. The probable reason for this may be the high elastic modulus and tensile strength of the carbon fiber. Thus, a load-bearing skeleton can be formed in concrete. However, the low elastic modulus of polypropylene fiber caused more defects in the compactness of concrete. As can also be seen, with the decrease in W/B, the compressive strength of concrete increases, which can be attributed to the increase in compactness of the hydrated cement paste [25]. As expected, when W/B decreases from 0.30 to 0.27, the axial compressive strength of plain concrete increases from 47.87 MPa to 50.38 MPa. Meanwhile, the compressive strength of FLWAC in Series 1 is higher than that of corresponding concrete in Series 2. The critical strain is defined as the strain at the peak point, from which the vertical deformability of the specimen can be evaluated. The critical strain of FLWAC was in the range of 0.00231~0.00394, with an average of 0.002939, which is slightly higher than that of the plain LWAC. Besides that, it can be concluded that the critical strain of LWAC is higher than the critical strain of 0.002 for NWC, as reported in the Chinese Specification GB 50010-2010 [26]. The elastic modulus is extensively used to evaluate the deformability and stiffness of concrete [27]. As provided in Table 5, the Ec and E0 for plain LWAC and FLWAC range from 18.6 GPa to 23.4 GPa and 14.2 GPa to 20.2 GPa, respectively, lower than the Ec of NWC at the comparable strength specified by CEB-FIP Model Code [28]. This phenomenon might be correlated with the porous structure of the LWA, thereby increasing the deformability of aggregate, which is directly responsible for the reduction in the elastic modulus [29]. Moreover, it is believed that the addition of fibers to LWAC reduces its elastic modulus as a result of the interference of fibers with the compactness of concrete. The ratio of Ec to E0 can reflect the curvature of the ascending branch in the stress–strain curve. In other words, the value of Ec/E0 demonstrates the characteristic of nonlinearity in the ascending branch. The larger the Ec/E0, the more significant the nonlinearity. As can be observed, the plain LWAC has an Ec/E0 of 1.157 and 1.186, indicating the apparent linearity of LWAC. Besides that, Ec/E0 increases with the addition of fibers, suggesting significant nonlinearity in the ascending branch of the FLWAC. 3.4. Toughness The toughness is calculated as the area under the stress–strain curve up to the specified strain, representing the concrete’s energy absorption capacity and ductility. The specified strain was set to 0.009 and 0.015, which is 3 and 5 times the ultimate strain following the ACI 318 standard [30], which is sufficient for evaluating the post-peak deformation following Fanella and Naaman [31]. The specific toughness is defined as the toughness ratio to peak strength since the toughness is affected by the compressive strength. Therefore, the specific toughness is considered a better measure to reflect the effect of fibers on the energy absorption capacity. The definitions of toughness and specific toughness are shown in Figure 6. The toughness and specific toughness are calculated using Equations (3) and (4), respectively:(3) TFi=∫0εiσ(ε)dε (4) TRi=∫0εiσ(ε)dεfcεi=TFifcεi where εi represents ε1 and ε2, corresponding to a specific strain of 0.009 and 0.015; TF1 and TF2 are the toughness corresponding to the specific strain of ε1 and ε2; TR1 and TR2 are the specific toughness corresponding to the specific strain ε1 and ε2. Table 6 shows the compressive toughness and specific toughness of LWAC, with which the effects of the fiber combination and W/B on the energy absorption capacity of LWAC can be exhibited. In the case of LWAC in Series 1, an initial increase followed by a decrease in toughness and specific toughness can be observed following the increase in polypropylene fiber content. The lower compressive strength of PF0.2a and PF0.4a may be the main reason for the decrease in toughness. The addition of carbon fiber is conducive to increasing the energy absorption capacity and the ductility of LWAC. In the case of LWAC in Series 2, an increase in both toughness and specific toughness can be observed with the increase in carbon fiber content. The toughness and specific toughness of PF0.2b and PF0.4b approach those of plain LWAC. It can be inferred that carbon fiber has a superior effect on the post-peak behavior of LWAC as compared to polypropylene fiber. Moreover, by comparing Plain 1 with Plain 2, the toughness and specific toughness for Series 2 are more significant than the corresponding specimens in Series 1. Therefore, the reduction in W/B (from 0.3 to 0.27) reduces the effect of fiber addition on the toughness of concrete. When LWAC contained carbon–polypropylene hybrid fibers, an increase in toughness and specific toughness was observed. The percentage of the specific toughness increase is shown in Table 6 with the corresponding plain concrete as control concrete. For instance, HFLWAC shows an increase in TR5 of 42.6~56.3% and 26.2~38.3% for W/B of 0.27 and 0.30, respectively. Besides that, CF0.4PF0.2 and CF0.4PF0.4 show a specific toughness larger than CF0.2, which shows that the carbon fiber has a better effect on the energy absorption capacity than polypropylene fiber does, similar to the results in single-fiber-reinforced LWAC. 3.5. Assessment of Synergy With an appropriate mixing range, two or more different fibers in the concrete can derive benefits from each of the individual fibers and exhibit a positive synergistic effect, enhancing the material properties of concrete so that they are far superior to a single fiber [32,33]. To assess the effect of hybridization of carbon and polypropylene fibers on toughness, the synergistic effect coefficients were evaluated following the method suggested by Wang et al. [18] as follows:(5) αx−1=βCF−PF+βmin(CF,PF)βmin(CF,PF)+βmax(CF,PF) (6) αx−2=βCF−PF+βmax(CF,PF)βmin(CF,PF)+βmax(CF,PF) where αx−1 and αx−2 are introduced to represent the synergistic effect coefficients, and x is one of the material properties that need to be identified. In this study, the effect of synergy on toughness was predicted; αt3−1 and αt3−2 represent the synergistic effect coefficients of TR3, and αt5−1 and αt5−2 are the synergistic effect coefficients of TR5. The parameter β is the toughness enhancement ratio of FLWAC to the corresponding plain LWAC, βCF−PF is the toughness enhancement ratio of HFLWAC, and βmin(CF,PF) and βmax(CF,PF) are the minimum and maximum enhancement ratios of individual carbon and polypropylene fibers that compose the hybrid fibers, respectively. The idea behind this method is that when αx−1 > 1, the combination of fibers produces a positive synergy; meanwhile, when αx−1 < 1, but αx−2 > 1, it is a positive synergistic effect, and when αx−2 < 1, it is a negative synergistic effect. The synergistic effect coefficients noted for toughness are presented in Table 7. As indicated in the table, all specimens exhibited positive synergistic effects, which represents that the hybridization of carbon and polypropylene fibers leads to toughness enhancement that is numerically more significant than the sum of the individual fibers. Polypropylene fiber did not add much to the toughness but shows effectiveness in contributing to the toughness when combined with carbon fiber. In addition, hybridization was less effective at higher fiber dosages in both Series 1 and Series 2. For instance, in Series 1, the best performance was obtained with carbon fiber at 0.2% and polypropylene fiber at 0.4%. Besides that, it can be seen from Table 7 that W/B shows little influence on the reinforcement of fiber in LWAC. 4. Compressive Stress–Strain Model 4.1. Modeling of Compressive Strength The relationship between fcu and fc was basically linear, and a formula is proposed with relevant test data, as follows:(7) fc=0.82fcu+2.01 A comparison of the calculated results with the test values is shown in Figure 7. As can be observed, the calculated values match the experimental values well. Then, Equation (7) can be used to calculate the fc of LWAC containing carbon and polypropylene fibers. 4.2. Modeling of Critical Strain at Peak Compressive Stress In order to analyze the influence of the two types of fibers on the critical strain, the critical strain ratio (εc/ε0) is taken to compensate for the effect of W/B, where εc/ε0 is the ratio of critical strain for FLWAC (εc) to the critical strain for the corresponding plain LWAC (ε0). As Section 3.2 stated, the FLWAC with the same mix design as the corresponding plain LWAC contains different fiber combinations. It can be observed in Table 5 that εc/ε0 increases with the increase in VCF but slightly decreases with the increase in VPF. An analytical equation for describing the relationship between the critical strain ratio and fiber volume fraction was obtained by fitting experimental data, as shown in Equation (8) (R2 = 0.741):(8) εc/ε0=1+0.1281(RI)CF−0.0613(RI)PF−27150(RI)CF(RI)PF where εc is the critical strain of FLWAC, and ε0 is the critical strain of the corresponding plain LWAC; for example, in Series 1, ε0 is the critical strain of specimen Plain-a, and (RI)CF and (RI)PF are the reinforcing indexes (calculated as the aspect ratio multiplied by the volume fraction) of carbon fiber and polypropylene fiber, respectively. The critical strain of plain LWAC can be modeled using the following analytical expression proposed by Lim and Ozbakkaloglu [17]:(9) ε0=fc0.225kd1000(152D)0.1(2DH)0.13 where kd is the parameter to allow for density; D is the diameter of the specimen (D = 100 mm); and H is the height of the specimen (H = 300 mm). When ε0 is known, εc for FLWAC can be calculated. The values of εc calculated by Equations (8) and (9) based on (RI)CF and (RI)PF in this study are compared with the test values of εc in Figure 8. As can be seen from the comparison, the calculated values are in good agreement with the test results. Then, Equations (8) and (9) can be used to model the critical strain of LWAC containing single and hybrid carbon and polypropylene fibers. 4.3. Modeling of the Stress–Strain Curve 4.3.1. Existing Models As presented in Table 8, five typical existing stress–strain models, including those developed by Carreira and Chu [12], Abbass et al. [34], Ou et al. [35], Oliveira Júnior et al. [36], and Wang et al. [18], were applied to predict the compressive behavior of FLWAC. After reviewing these models, it is recognized that the five existing models, except for the model of Wang et al. [18], were modified based on the Carreira and Chu model that was initially used to describe the behavior of plain concrete in compression. Abbass et al. [34] and Ou et al. [35] modified the parameter β with the fiber reinforcing index, with which the fiber characteristic can be considered in the models. In the model established by Oliveira Júnior et al. [36], the material parameter β contains the fiber volume fraction and compressive strength. Wang et al.’s model was derived for the axial compressive behavior of basalt–polypropylene hybrid fiber reinforced concrete. Since the reinforcing index of carbon fiber in this study is similar to that of basalt fibers in [17], this model was applied to the experimental data of LWAC containing carbon and polypropylene fibers. Comparisons of the predicted stress–strain curves with the experimental curves for Series 1 and Series 2 are shown in Figure 9 and Figure 10. As can be observed, the ascending branch in stress–strain curves predicted by the existing models is in close agreement with the experimental results. In the descending branch, Carreira and Chu’s model and Wang et al.’s model underestimate the post-peak region of the curves. Abbass et al.’s model can be used to effectively describe the behavior of PFLWAC but underestimates the behavior of CFLWAC. However, Abbass et al.’s model and Ou et al.’s model can only be used in single-fiber-reinforced concrete for just one variable considered in the parameter β. In Oliveira Junior et al.’s model, to express the stress–strain curves of HFLWAC, the volume fraction of fiber used in the parameter β was calculated by the sum of the two types of fibers. The curves for single-fiber-reinforced LWAC that Oliveira Júnior et al. [36] predicted were close to the experimental curves; however, the curves for HFLWAC were underestimated. 4.3.2. New Proposal for the Stress–Strain Curve of FLWAC As previously mentioned, Carreira and Chu’s model (1985) gives a relatively good representation of the ascending portion. Therefore, the stress–strain model generated by Carreira and Chu (1985) and later improved by Wee et al. (1986) was chosen to predict the stress–strain curve of LWAC. The parameters of this model were further modified to predict the stress–strain behavior of carbon- and/or polypropylene-fiber-reinforced LWAC. The model is represented by the following:(10) σ=fc(k1β(ε/εc)k1β−1+(ε/εc)k2β) (11) β=11−(fc/εcEc) where β is the material parameter, which can be obtained from Equation (11); the parameter k1 is applied to the numerator, and in the first term of the denominator, k2 is applied to the exponent of the last term of the denominator. According to the above analysis, the regression equations of parameters k1 and k2 are determined by using (RI)CF, (RI)PF, and β as variables, as follows:(12) k1=1.2524−0.0197(RI)CF−0.3224(RI)PF−0.0875β (13) k2=1.1476−0.0158(RI)CF−0.2512(RI)PF−0.1059β Here, k1 and k2 can be calculated by Equations (12) and (13), and then the fitted stress–strain curve can be determined by Equation (10). Table 9 shows the experimental and calculated parameters. The COVs are 0.587 and 0.808 for k1 and k2; the average ratio is 1.00 and 1.02 for k1 and k2, respectively. 4.4. Comparison between Experimental Results and the Proposed Model Comparisons of the experimental and analytical stress–strain curves of FLWAC are shown in Figure 11 and Figure 12 for W/B of 0.27 and 0.30, respectively. There is a reasonable adjustment between the analytical ascending branch of the predicted curve and the experimental curve. When referring to the descending branch, a slight difference can be found between experimental results and predicted curves, which can be attributed to the inconsistent manner in which cracks occurred. In both cases, the proposed model shows better accuracy relative to the experimental curve than the existing models. 5. Conclusions The compressive stress–strain relationship of LWAC containing different combinations of carbon and polypropylene fibers and different W/B was investigated experimentally through the compressive testing of prisms. Furthermore, regression formulas of compressive strength and critical strain were derived, and a stress–strain model for HFLWAC was established. Therefore, the following conclusions can be drawn: The failure modes were less affected by W/B but displayed significant modifications with the addition of fibers. Although CFLWAC and plain LWAC showed similar failure modes with localized concrete crushing, LWAC containing polypropylene fiber maintained integrity during the testing, and no apparent crushing was observed after the test. Hybrid carbon–polypropylene fibers have better reinforcing effects on LWAC than single fibers. All specimens show positive synergistic effects, demonstrating the positive hybridization effect of carbon and polypropylene fibers. The axial compressive strength of FLWAC is proportional to fcu, and a conversion relationship between them was established by the regression of experimental data through Equation (7). An empirical formula of the critical strain of FLWAC based on fiber reinforcing indexes ((RI)CF and (RI)PF) and the critical strain of plain LWAC was proposed through Equations (8) and (9). All stress–strain curves show a similar trend in their mathematical characteristics. One stress–strain model can be estimated by using the appropriate curve parameter. The relationship between the parameters (k1 and k2) and fiber reinforcing indexes ((RI)CF and (RI)PF) is estimated through Equations (10)–(13). Author Contributions Conceptualization, T.W.; methodology, X.Y. and T.W.; formal analysis, X.Y. and X.L.; investigation, T.W., X.Y. and X.L.; resources, T.W.; data curation, X.L. and X.Y.; writing—original draft preparation, X.Y. and T.W. All authors have read and agreed to the published version of the manuscript. Funding This work was supported by the National Natural Science Foundation of China, China (52078042, 51878054, 51908041), the Natural Science Foundation of Shaanxi Province, China (2020GY-248), and the Fundamental Research Funds for the Central Universities (300102280712). Institutional Review Board Statement The study did not require ethical approval. Informed Consent Statement Not applicable. Data Availability Statement The data presented in this study are available on request from the corresponding author. Conflicts of Interest The authors declare no conflict of interest. Abbreviations f cu Cubic compressive strength/MPa f c Compressive strength corresponding to the peak point of stress–strain curve/MPa ε c Critical strain corresponding to the peak point of the stress–strain curve ε 0 Critical strain for corresponding plain LWAC E c Initial tangential elastic modulus/GPa E 0 Peak secant elastic modulus/GPa TF i Toughness TR i Specific toughness α x−i Synergistic effect coefficients Figure 1 Scheme of axial compression test. Figure 2 Failure modes of specimens under axial compression (W/B = 0.3): (a) Plain LWAC; (b) PFLWAC; (c) CFLWAC; (d) HFLWAC. Figure 3 A typical example of the compressive stress–strain curve. Figure 4 Effects of CF and PF on the compressive stress–strain curves of LWAC (W/B = 0.27): (a) fibers in single form; (b) fibers in hybrid form. Figure 5 Effects of CF and PF on the compressive stress–strain curves of LWAC (W/B = 0.3): (a) fibers in single form; (b) fibers in hybrid form. Figure 6 Definitions of toughness and specific toughness. Figure 7 Comparison of fc between experimental and calculated values. Figure 8 Comparison of εc between experimental and calculated values. Figure 9 Application of existing stress–strain models to the data of: (a) Plain–a; (b) CF0.2a; (c) CF0.4a; (d) PF0.2a; (e) PF0.4a; (f) CF0.2PF0.2a; (g) CF0.2PF0.4a; (h) CF0.4PF0.2a; (i) CF0.4PF0.4a. Figure 10 Application of existing stress–strain models to the data of: (a) Plain–b; (b) CF0.2b; (c) CF0.4b; (d) PF0.2b; (e) PF0.4b; (f) CF0.2PF0.2b; (g) CF0.2PF0.4b; (h) CF0.4PF0.2b; (i) CF0.4PF0.4b. Figure 11 Comparison of calculated complete stress–strain curves for: (a) Plain–a; (b) CF0.2a; (c) CF0.4a; (d) PF0.2a; (e) PF0.4a; (f) CF0.2PF0.2a; (g) CF0.2PF0.4a; (h) CF0.4PF0.2a; (i) CF0.4PF0.4a. Figure 12 Comparison of calculated complete stress–strain curves for: (a) Plain–b; (b) CF0.2b; (c) CF0.4b; (d) PF0.2b; (e) PF0.4b; (f) CF0.2PF0.2b; (g) CF0.2PF0.4b; (h) CF0.4PF0.2b; (i) CF0.4PF0.4b. polymers-14-01675-t001_Table 1 Table 1 Properties of LWA. Bulk Density (kg/m3) Apparent Density (kg/m3) Crushing Strength (MPa) 1 h/24 h Water Absorption (%) Total Porosity (%) Particle Size Distribution (%) 2.36~5 mm 5~10 mm 10~16 mm 860 1512 6.9 2.2/2.6 43.12 11 68 21 Chemical Composition SiO2 (%) Al2O3 (%) Fe2O3 (%) TiO2 (%) CaO (%) MgO (%) SiO3 (%) Alkalis as Na2O (%) LOI (%) 65.4 15.9 4.2 0.7 2.4 3.7 0.23 3.8 3.67 polymers-14-01675-t002_Table 2 Table 2 Properties of fibers. Fiber Details Carbon Fiber Polypropylene Fiber View Fiber shape Straight, filaments Straight, fibrillated Cut length (Lf) (mm) 8~10 15~22 Diameter (Df) (μm) 7 80 Aspect ratio (Lf/Df) 1100 225 Specific gravity (g/cm3) 1.8 0.91 Elongation (%) 2.1 17 Tensile strength (MPa) 4000 >400 Elastic modulus (GPa) 240 22 Water absorption <1% by weight Nil polymers-14-01675-t003_Table 3 Table 3 The mix proportions of plain LWAC (kg/m3). Mixture W/B Cement Silica Fume Fly Ash LWA Sand Water Superplasticizer Plain–a 0.27 440 44 66 603 684 148.5 6.8 Plain–b 0.3 440 44 66 578 667 165 5.2 polymers-14-01675-t004_Table 4 Table 4 Fiber addition in LWAC mixtures. Fiber W/B Carbon Fiber/(%) Polypropylene Fiber/(%) Carbon–Polypropylene Hybrid Fibers/(%) 0.2 0.4 0.2 0.4 0.2/0.2 0.2/0.4 0.4/0.2 0.4/0.4 Mix code 0.27 CF0.2a CF0.4a PF0.2a PF0.4a CF0.2PF0.2a CF0.2PF0.4a CF0.4PF0.2a CF0.4PF0.4a 0.3 CF0.2b CF0.4b PF0.2b PF0.4b CF0.2PF0.2b CF0.2PF0.4b CF0.4PF0.2b CF0.4PF0.4b polymers-14-01675-t005_Table 5 Table 5 Characteristic values of compressive stress–strain curves of LWAC. Specimen W/B VCF (%) VPF (%) fcu (MPa) fc (MPa) εc (10−6) εc/ε0 Ec (GPa) E0 (GPa) Ec/E0 Plain–a 0.27 0 0 61.34 (0.041) 50.38 2492 1.000 23.4 20.2 1.157 CF0.2a 0.2 0 61.17 (0.038) 55.97 3807 1.528 22.9 14.7 1.558 CF0.4a 0.4 0 74.03 (0.054) 63.86 3940 1.581 23.1 16.2 1.425 PF0.2a 0 0.2 50.83 (0.045) 41.15 2547 1.022 20.1 16.2 1.244 PF0.4a 0 0.4 42.15 (0.036) 36.05 2307 0.926 19.4 15.6 1.241 CF0.2PF0.2a 0.2 0.2 48.15 (0.062) 40.71 2834 1.137 20.7 14.4 1.441 CF0.2PF0.4a 0.2 0.4 42.90 (0.049) 39.93 2685 1.077 20.8 14.9 1.399 CF0.4PF0.2a 0.4 0.2 47.98 (0.039) 43.25 2769 1.111 21.1 15.6 1.351 CF0.4PF0.4a 0.4 0.4 62.15 (0.061) 50.15 2847 1.142 22.2 17.6 1.260 Plain–b 0.3 0 0 58.24 (0.030) 47.87 2500 1.000 22.7 19.1 1.186 CF0.2b 0.2 0 57.99 (0.033) 52.29 3680 1.472 23.2 14.2 1.633 CF0.4b 0.4 0 67.11 (0.031) 56.06 3870 1.548 22.4 14.5 1.546 PF0.2b 0 0.2 49.64 (0.041) 42.07 2354 0.942 21.8 17.9 1.220 PF0.4b 0 0.4 40.99 (0.064) 34.16 2314 0.926 18.6 14.8 1.260 CF0.2PF0.2b 0.2 0.2 48.26 (0.019) 41.68 2790 1.116 22.0 14.9 1.473 CF0.2PF0.4b 0.2 0.4 50.65 (0.030) 44.19 2548 1.019 21.7 17.3 1.251 CF0.4PF0.2b 0.4 0.2 51.36 (0.047) 44.46 2764 1.106 22.5 16.1 1.399 CF0.4PF0.4b 0.4 0.4 56.88 (0.028) 49.18 2968 1.187 21.3 16.6 1.285 polymers-14-01675-t006_Table 6 Table 6 Toughness and specific toughness of LWAC. Specimen σ0.009 (MPa) σ0.015 (MPa) Toughness Specific Toughness (%) TF 3 TF 5 TR 3 TR 5 Plain-a 22.09 13.88 0.2933 0.3954 0.6468 (-) 0.5232 (-) CF0.2a 39.58 20.73 0.3882 0.5465 0.7703 0.6506 CF0.4a 37.83 18.58 0.3888 0.5419 0.6764 0.5657 PF0.2a 26.01 15.98 0.2801 0.4030 0.7563 0.6529 PF0.4a 17.23 13.19 0.2177 0.3082 0.6701 0.5699 CF0.2PF0.2a 29.32 21.71 0.3070 0.4591 0.8379 (29.5%) 0.7518 (43.7%) CF0.2PF0.4a 28.45 26.08 0.3018 0.4650 0.8399 (29.9%) 0.7764 (48.4%) CF0.4PF0.2a 34.57 27.78 0.3206 0.5049 0.8656 (33.8%) 0.8178 (56.3%) CF0.4PF0.4a 34.63 25.69 0.3664 0.5413 0.8007 (23.8%) 0.7097 (35.6%) Plain-b 25.15 16.97 0.2919 0.4145 0.6775 (-) 0.5772 (-) CF0.2b 31.79 21.14 0.3317 0.4822 0.7049 0.6148 CF0.4b 38.02 31.60 0.3651 0.5600 0.7236 0.6897 PF0.2b 20.28 9.18 0.2525 0.3371 0.6669 0.5342 PF0.4b 17.73 11.77 0.2159 0.3046 0.7022 0.5944 CF0.2PF0.2b 31.94 23.03 0.3116 0.4712 0.8322 (22.8%) 0.7551 (30.8%) CF0.2PF0.4b 30.25 23.96 0.3128 0.4827 0.7866 (16.1%) 0.7282 (26.2%) CF0.4PF0.2b 37.65 28.16 0.3280 0.5296 0.8198 (21.0%) 0.7942 (37.6%) CF0.4PF0.4b 42.12 30.89 0.3642 0.5871 0.8229 (21.5%) 0.7985 (38.3%) Note: the data in parentheses show the percentage of specific toughness increase over that of plain concrete. polymers-14-01675-t007_Table 7 Table 7 Specific toughness synergistic coefficient. Specimen α t3−1 α t5−1 CF0.2PF0.2a 1.053 1.076 CF0.2PF0.4a 1.087 1.103 CF0.4PF0.2a 1.132 1.207 CF0.4PF0.4a 1.092 1.127 CF0.2PF0.2b 1.093 1.122 CF0.2PF0.4b 1.058 1.094 CF0.4PF0.2b 1.069 1.085 CF0.4PF0.4b 1.070 1.084 polymers-14-01675-t008_Table 8 Table 8 Existing stress–strain models for concrete in compression. Models Fitting Expressions Crucial Parameters Carreira and Chu, 1985 σ=fc(β(ε/εc)β−1+(ε/εc)β) β=1/(1−(fc/εcEc)) Abbass et al., 2018 σ=fc(β(ε/εc)β−1+(ε/εc)β) β=1.401(RI)2−1.56(RI)+2.42 Ou et al., 2012 σ=fc(β(ε/εc)β−1+(ε/εc)β) β=0.71(RI)2−2(RI)+3.05 Júnior et al., 2010 σ=fc(β(ε/εc)β−1+(ε/εc)β) β=(0.0536−0.5754Vf)fc Wang et al., 2019 Ascending branch: σ=fc(a1(ε/εc)+(6−5a1)(ε/εc)5+(4a1−5)(ε/εc)6) Descending branch: σ=fc((ε/εc)α(ε/εc−1)2+ε/εc) a1=1.417+0.697VBF−6.699VPPF α=5.638+24.01VBF−468.34VPPF polymers-14-01675-t009_Table 9 Table 9 Comparison of calculated values with the experimental values. Specimen VCF (%) VPF (%) β k 1 k 2 Experimental Calculated Ratio Experimental Calculated Ratio Plain–a 0 0 7.3508 0.56 0.6092 0.92 0.3833 0.3691 1.04 CF0.2a 0.2 0 2.7933 0.9546 0.9646 0.99 0.8572 0.8169 1.05 CF0.4a 0.4 0 3.3518 1.071 0.8724 1.23 0.8793 0.7230 1.22 PF0.2a 0 0.2 5.0967 0.4895 0.6614 0.74 0.4066 0.4948 0.82 PF0.4a 0 0.4 5.1409 0.4429 0.5124 0.86 0.4139 0.3771 1.10 CF0.2PF0.2a 0.2 0.2 3.2675 0.549 0.7781 0.71 0.5332 0.6537 0.82 CF0.2PF0.4a 0.2 0.4 3.5085 0.5538 0.6119 0.90 0.5083 0.5151 0.99 CF0.4PF0.2a 0.4 0.2 3.8499 0.5331 0.6838 0.78 0.4493 0.5572 0.81 CF0.4PF0.4a 0.4 0.4 4.8419 0.4509 0.4519 1.00 0.397 0.3391 1.17 Plain–b 0 0 6.3908 0.6855 0.6932 0.99 0.4144 0.4708 0.88 CF0.2b 0.2 0 2.5804 1.194 0.9833 1.21 0.9684 0.8395 1.15 CF0.4b 0.4 0 2.8304 0.812 0.9181 0.88 0.7265 0.7782 0.93 PF0.2b 0 0.2 5.5495 0.7507 0.6218 1.21 0.4665 0.4469 1.04 PF0.4b 0 0.4 4.8467 0.8236 0.5382 1.53 0.5093 0.4083 1.25 CF0.2PF0.2b 0.2 0.2 3.1157 0.6874 0.7914 0.87 0.5717 0.6698 0.85 CF0.2PF0.4b 0.2 0.4 4.9805 0.4463 0.4831 0.92 0.3622 0.3592 1.01 CF0.4PF0.2b 0.4 0.2 3.5076 0.7719 0.7137 1.08 0.5258 0.5934 0.89 CF0.4PF0.4b 0.4 0.4 4.5032 0.5982 0.4816 1.24 0.4146 0.3749 1.11 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Kockal N.U. Ozturan T. Effects of lightweight fly ash aggregate properties on the behavior of lightweight concretes J. Hazard. Mater. 2010 179 954 965 10.1016/j.jhazmat.2010.03.098 20399557 2. Sohel K.M.A. Al-Jabri K. Zhang M.H. Liew J.Y.R. Flexural fatigue behavior of ultra-lightweight cement composite and high strength lightweight aggregate concrete Constr. Build. Mater. 2018 173 90 100 10.1016/j.conbuildmat.2018.03.276 3. Guler S. The effect of polyamide fibers on the strength and toughness properties of structural lightweight aggregate concrete Constr. Build. Mater. 2018 173 394 402 10.1016/j.conbuildmat.2018.03.212 4. Li J.J. Niu J.G. Wan C.J. Jin Z. Comparison of flexural property between high performance polypropylene fiber reinforced lightweight aggregate concrete and steel fiber reinforced lightweight aggregate concrete Constr. Build. Mater. 2017 157 729 736 10.1016/j.conbuildmat.2017.09.149 5. El-Mal H. Sherbini A.S. Sallam H. Mode II fracture toughness of Hybrid FRCs Int. J. Concr. Struct. Mater. 2015 9 475 486 10.1007/s40069-015-0117-4 6. Qian K. Geng S.Y. Liang S.L. Fu F. Yu J. Effects of loading regimes on the structural behavior of RC beam-column sub-assemblages against disproportionate collapse Eng. Struct. 2022 251 113470 10.1016/j.engstruct.2021.113470 7. Hassanpour M. Shafigh P. Mahmud H.B. Lightweight aggregate concrete fiber reinforcement—A review Constr. Build. Mater. 2012 37 452 461 10.1016/j.conbuildmat.2012.07.071 8. Erdem S. Dawson A.R. Thom N.H. Impact load-induced micro-structural damage and micro-structure associated mechanical response of concrete made with different surface roughness and porosity aggregates Cem. Concr. Res. 2012 42 291 305 10.1016/j.cemconres.2011.09.015 9. Yanga G.J. Parkb M. Parka S.J. Recent progresses of fabrication and characterization of fibers-reinforced composites: A review Compos. Commun. 2019 14 34 42 10.1016/j.coco.2019.05.004 10. Lu Z.H. Zhao Y.G. Empirical stress-strain model for unconfined high-strength concrete under uniaxial compression J. Mater. Civ. Eng. 2010 22 1181 1186 10.1061/(ASCE)MT.1943-5533.0000095 11. Weng Y.H. Qian K. Fu F. Fang Q. Numerical investigation on load redistribution capacity of flat slab substructures to resist progressive collapse J. Build. Eng. 2020 29 101109 10.1016/j.jobe.2019.101109 12. Carreira D.J. Chu K.H. Stress-strain relationship for plain concrete in compression ACI J. 1985 82 797 804 13. Guo Z.H. Concrete Strength and Deformation Tsinghua University Press Beijing, China 1997 127 131 14. Wee T.H. Chin M.S. Mansur M.A. Stress-strain relationship of high-strength concrete in compression J. Mater. Civ. Eng. 1986 8 70 76 10.1061/(ASCE)0899-1561(1996)8:2(70) 15. Desnerck I.P. Schutter G.D. Taerwe L. Stress-strain behaviour of self-compacting concretes containing limestone fillers Struct. Concr. 2012 13 95 101 10.1002/suco.201100056 16. Tasnimi A.A. Mathematical model for complete stress–strain curve prediction of normal, lightweight and high-strength concretes Mag. Concr. Res. 1994 56 23 34 10.1680/macr.2004.56.1.23 17. Lim J.C. Ozbakkaloglu T. Stress-strain model for normal-and lightweight concretes under uniaxial and triaxial compression Constr. Build. Mater. 2014 71 92 509 10.1016/j.conbuildmat.2014.08.050 18. Wang D.H. Ju Y.Z. Shen H. Jin Z. Mechanical properties of high performance concrete reinforced with basalt fiber and polypropylene fiber Constr. Build. Mater. 2019 197 464 473 10.1016/j.conbuildmat.2018.11.181 19. CS (Chinese Standard) GB/T 17431.2-2010 Lightweight Aggregates and Its Test Methods—Part 2: Test Methods for Lightweight Aggregates China Architecture and Building Press Beijing, China 2010 (In Chinese) 20. CS (Chinese Standard) GB 175-2007 Common Portland Cement China Architecture and Building Press Beijing, China 2007 (In Chinese) 21. CS (Chinese Standard) GB/T 18736-2017 Mineral Admixtures for High Strength and High Performance Concrete China Architecture and Building Press Beijing, China 2017 (In Chinese) 22. CS (Chinese Standard) GB/T 50081-2007 Standard for Test Method of Mechanical Properties on Ordinary Concrete China Architecture and Building Press Beijing, China 2007 (In Chinese) 23. Wei Y. Wu Y.F. Compressive behavior of concrete columns confined by high strength steel wire Constr. Build. Mater. 2014 54 443 453 10.1016/j.conbuildmat.2013.12.083 24. Domagała L. Modification of properties of structural lightweight concrete with steel fibres J. Civ. Eng. Manag. 2011 17 36 44 10.3846/13923730.2011.553923 25. Smarzewski P. Influence of basalt-polypropylene fibres on fracture properties of high performance concrete Compos. Struct. 2019 209 23 33 10.1016/j.compstruct.2018.10.070 26. CS (Chinese Standard) GB 50010-2010 Code for Design of Concrete Structures China Architecture and Building Press Beijing, China 2010 (In Chinese) 27. Gao D.Y. Zhang L.J. Nokken M. Compressive behavior of steel fiber reinforced recycled coarse aggregate concrete designed with equivalent cubic compressive strength Constr. Build. Mater. 2017 141 235 244 10.1016/j.conbuildmat.2017.02.136 28. CEB-FIP CEB-FIP Model Code 2010 Model Code for Concrete Structures CEB-FIP London, UK 2010 29. Han B. Xiang T.Y. Axial compressive stress-strain relation and poisson effect of structural lightweight aggregate concrete Constr. Build. Mater. 2017 14 338 343 10.1016/j.conbuildmat.2017.04.101 30. ACI Committee Building Code Requirements for Structural Concrete and Commentary ACI 318-14 ACI Farmington Hills, MI, USA 2018 31. Fanella D.A. Naaman A.E. Stress-strain properties of fiber reinforced mortar in compression ACI J. Proc. 1985 82 475 483 32. Banthia N. Majdzadeh F. Wu J. Bindiganavile V. Fiber synergy in hybrid fiber reinforced concrete (HyFRC) in flexure and direct shear Cem. Concr. Res. 2014 48 91 97 10.1016/j.cemconcomp.2013.10.018 33. Abadel A. Abbas H. Almusallam T. Al-Salloum Y. Siddiqui N. Mechanical properties of hybrid fibre-reinforced concrete–analytical modelling and experimental behavior Mag. Concr. Res. 2015 68 823 843 10.1680/jmacr.15.00276 34. Abbass M. Khan M.I. Mourad S. Evaluation of mechanical properties of steel fiber reinforced concrete with different strengths of concrete Constr. Build. Mater. 2018 168 556 569 10.1016/j.conbuildmat.2018.02.164 35. Ou Y.C. Tsai M.S. Liu K.Y. Chang K.C. Compressive behavior of steel fiber reinforced concrete with a high reinforcing index J. Mater. Civ. Eng. 2012 24 207 215 10.1061/(ASCE)MT.1943-5533.0000372 36. Oliveira Júnior L.A. Borges V.E.S. Danin A.R. Stress-strain curves for steel fiber-reinforced concrete in compression Matéria 2010 15 260 266 10.1590/S1517-70762010000200025

PMC009xxxxxx/PMC9099634.txt

==== Front Int J Environ Res Public Health Int J Environ Res Public Health ijerph International Journal of Environmental Research and Public Health 1661-7827 1660-4601 MDPI 10.3390/ijerph19095754 ijerph-19-05754 Article Folate Status as a Nutritional Indicator among People with Substance Use Disorder; A Prospective Cohort Study in Norway Bemanian Mitra 12* https://orcid.org/0000-0001-8701-7638 Vold Jørn Henrik 123 Chowdhury Ranadip 4 Aas Christer Frode 123 Gjestad Rolf 3 Johansson Kjell Arne 12 https://orcid.org/0000-0001-8757-2092 Fadnes Lars Thore 12 Müller-Stierlin Annabel Academic Editor Teasdale Scott B. Academic Editor 1 Bergen Addiction Research, Department of Addiction Medicine, Haukeland University Hospital, 5012 Bergen, Norway; jorn.henrik.vold@helse-bergen.no (J.H.V.); christer.frode.aas@helse-bergen.no (C.F.A.); kjell.johansson@uib.no (K.A.J.); lars.fadnes@uib.no (L.T.F.) 2 Department of Global Public Health and Primary Care, Faculty of Medicine, University of Bergen, 5009 Bergen, Norway 3 Division of Psychiatry, Haukeland University Hospital, 5036 Bergen, Norway; rolf.gjestad@uib.no 4 Centre for Health Research and Development, Society for Applied Studies, New Delhi 110016, India; ranadip.chowdhury@sas.org.in * Correspondence: mitra.bemanian@student.uib.no 09 5 2022 5 2022 19 9 575412 4 2022 06 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Substance use disorder (SUD) is associated with poor nutrition. Vitamin B9, or folate, is an important micronutrient for health. The aim of this prospective longitudinal cohort study was to assess serum folate levels among people with SUD and to investigate the impact of factors related to substance use severity on folate status. Participants were recruited from outpatient clinics for opioid agonist therapy (OAT) and municipal health-care clinics for SUD in Western Norway. They were assessed annually, including blood sampling for determination of micronutrient status. Overall, 663 participants with a total of 2236 serum folate measurements were included. A linear mixed model was applied, and measures are presented as β-coefficients with 95% confidence interval (CI). Forty-eight percent (CI: 44–51) of the population had low serum folate levels (s-folate < 10 nmol/L), and 23% (CI: 20–26) were deficient (s-folate < 6.8 nmol/L) at baseline. Sixty percent (CI: 53–65) sustained their poor folate status in at least one subsequent assessment. Except for weekly use of cannabis (mean difference in serum folate [nmol/L]: −1.8, CI: −3.3, −0.25) and alcohol (1.9, CI: 0.15, 3.6), weekly use of no other substance class was associated with baseline differences in serum folate when compared to less frequent or no use. Injecting substances was associated with a reduction in serum folate over time (−1.2, CI: −2.3, −0.14), as was higher dosages of OAT medication (−1.1, CI: −2.2, −0.024). Our findings emphasize the need of addressing nutrition among people with severe SUD. substance use disorder nutrition folate vitamin opioid agonist therapy The Norwegian Research Council269855 Western Norway Regional Health AuthorityThis work was supported by The Norwegian Research Council (BEHANDLING, contract no. 269855) and the Western Norway Regional Health Authority («Åpen prosjektstøtte» for INTRO-HCV and «Strategiske forskningsmidler» for ATLAS4LAR) with the Department of Addiction Medicine, Haukeland University Hospital, Bergen, Norway as the responsible institution. The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. Authors are funded by their respective affiliations. ==== Body pmc1. Introduction Folate deficiency is a known complication of alcohol use disorder, but an association of folate status with other substance use disorders (SUDs) is not well-established [1,2]. Under- and malnutrition is common among people with SUD, in part due to limited food accessibility and unfavourable dietary habits [3,4]. The diet of those heavily burdened by SUD tends to be dominated by foods with a low nutritional value, in particular simple carbohydrates and sugar-sweetened snack foods and beverages [5,6,7]. This type of diet increases the risk of many micronutrient deficiencies, including that of folate deficiency [1,6]. Although there is some evidence on overall low dietary intake [8,9] and poor blood status of folate [10,11,12] among people with SUD, there is little updated literature on folate status and risk factors for folate deficiency in this population. Folate, or vitamin B9, is an essential micronutrient that is particularly abundant in a range of fruits, legumes, and green leafy vegetables [1,13]. An adequate blood status of folate is dependent on a well-balanced diet, especially in countries where grain and cereal products are not routinely fortified with its synthetic equivalent folic acid [1,14]. On a molecular level, folate is a key cofactor in one-carbon reactions vital for cell proliferation and genomic integrity [13,15]. Folate deficiency is an established cause of megaloblastic anaemia, pregnancy complications and foetal developmental malformations [13,16]. Further, sustained low or deficient folate status is associated with an increased risk of cognitive decline and neurological disease and symptoms, depressive disorders as well as some cancers [17,18,19,20]. Opioid agonist therapy (OAT), with methadone or buprenorphine-based medication, is widely accepted as the gold standard for treatment of severe opioid dependence [21]. The OAT is a well-documented medical intervention for reducing morbidity and mortality, and improving the health-related quality of life, of individuals with opioid dependence [22,23,24]. Nevertheless, patients in OAT have a significant disease burden and increased morbidity and mortality compared to the general population [25,26,27]. The high mortality of patients in OAT is equally attributable to substance overdoses and comorbid somatic disease, including chronic infections such as hepatitis C virus, cardiovascular and pulmonary disease [28,29]. Additionally, long-term use of opioids, including opioids for OAT, is associated with adverse metabolic and endocrinologic manifestations including the development of overweight, dyslipidaemia, hypogonadism, and metabolic syndrome [30,31,32]. Efforts should be made to reduce the disease burden in this population, and this includes addressing complications related to adverse dietary habits, food accessibility and nutritional deficiencies [6]. The poor dietary and nutritional status of people with SUD, and particularly the tendency to replace nutritious foods with so-called “empty calories” in the form simple carbohydrates and dietary sugar, puts them at high risk of developing folate deficiency [6]. Little is known about the extent of folate deficiency in this population, and whether the substance use patterns, injecting substance use, or treatment with opioid agonist therapy are associated with the levels of folate. Thus, the objective of this longitudinal cohort study was to determine the serum folate status of a population with SUD visiting outpatient OAT clinics and municipality health-care clinics for SUD in Bergen and Stavanger, Norway. Further, we aimed to assess the impact of sociodemographic and clinical factors related to severity of substance use on serum folate status. 2. Materials and Methods 2.1. Study Characteristics; Design, Population, Data Collection and Study Sample This is a prospective longitudinal cohort study nested to the multicentre INTRO-HCV and ATLAS4LAR studies [33,34]. Participants were recruited from a population of outpatients visiting OAT clinics or municipality health-care clinics for SUD in Bergen and Stavanger, Norway. Participants were assessed yearly with a specialized research nurse-led and questionnaire-based interview focused on somatic and mental health, psychosocial aspects and substance use patterns. Data were collected using the software CheckWare® (CheckWare, Trondheim, Norway). Clinical data were obtained from the electronic medical record. Data collected between May 2016 and June 2020 are presented. A total of 2236 valid serum folate measurements from 663 participants were included. Each participant had a mean of 3.4 (Standard Deviation (SD): 3.6) serum folate measurements during the study period, and a total of 516 (78%) participants had at least two serum folate measurements. 2.2. Measuring Serum Folate; Laboratory Assays and Definitions Venous blood samples were collected according to standard protocol and sent to the Department of Medical Biochemistry and Pharmacology at Haukeland University Hospital in Bergen and the Department of Medical Biochemistry at Stavanger University Hospital in Stavanger for analysis of serum folate concentration (both accredited by ISO-standard 15189). The former laboratory assessed folate concentration in serum samples by means of the Electrochemiluminescence Immunoassay (10% analytical variation at concentration 8.4 nmol/L) [35], whereas the latter used the Chemiluminescence Microparticle Immunoassay (12% analytical variation at concentration 9 nmol/L) [36]. Data on serum folate concentration was obtained from the electronic medical record. The unit used was nanomoles per litre (nmol/L), and values between 1.5 and 45.3 nmol/L were specified. Two separate cut-offs were used when describing folate status: folate deficiency was defined as s-folate < 6.8 nmol/L. This cut-off is set based on the risk of developing megaloblastic anaemia, and is widely used in literature when describing folate deficiency [37]. Low folate status was defined as s-folate < 10 nmol/L. This cut-off is based on metabolic manifestations of negative folate balance, namely elevated plasma homocysteine concentration [37]. This is also the cut-off which, according to local guidelines based on recommendations from the World Health Organization (WHO), warrants intervention in the form of nutritional guidance or supplementation [37,38,39]. 2.3. Study Variables; Baseline, OAT, Clinical and Sociodemographic Factors The baseline was defined as the serum folate measurement performed in closest proximity in time to the first health assessment for the individual. The subsequent serum folate measurements were listed chronologically and included as follow-up, and time was defined as years from baseline. The OAT was defined as receiving methadone or buprenorphine-based medication. We calculated an OAT dose ratio corresponding to the prescribed daily dose of medication divided by the mean of WHO’s recommended dose interval (90 mg for methadone and 18 mg for buprenorphine) [40]. As for the clinical factors, injecting substances was defined as having injected any substance within the prior six months. Frequent substance use was defined as using any of the following substances on a minimum weekly basis during the 12 months leading up to the annual health assessment: alcohol, cannabis, benzodiazepines, stimulants (amphetamines and cocaine) and non-OAT opioids (e.g., heroin). Hepatitis C virus (HCV) status was determined by means of a quantitative polymerase chain reaction assay, and non-zero values were defined as HCV infection. Regarding sociodemographic factors, housing conditions for the last 30 days leading up to the annual health assessment were defined as stable (living in owned or rented home or at an institution) or unstable (being homeless, living at shelter or with friends and family). Highest completed education level was classed into the following groups: primary school (7 years), middle school (10 years), high school (13 years), ≤3 years of college or university level education and >3 years of college or university level education. Age was categorized into the following groups: <30 years, 30–39 years, 40–49 years, 50–59 years and ≥60 years. 2.4. Statistical Analyses The Stata/SE 16.0 software (Stata Corporation, College Station, TX, USA) was used for the generation of descriptive statistics and a linear mixed model (LMM). The SPSS version 26.0 software (IBM, Armonk, NY, USA) was used for expectation maximization imputation. The software R version 4.0.3 (R foundation for Statistical Computing, Vienna, Austria) with the package mgcv was used for the preparation of generalized additive models (GAMs) and their graphical presentations. The website Sankeymatic (http://www.sankeymatic.com/build, accessed on 1 September 2021) was used for the generation of a Sankey diagram. The threshold of statistical significance was set to p < 0.05 for all analyses. Nine percent of values were missing across the sociodemographic and clinical variables, including variables assessing substance use patterns. These were assumed to be missing at random and expectation maximization imputation was performed to replace them with estimated values [41]. Descriptive data are presented with total numbers and percentages, only including valid, non-imputed values. Median (Interquartile range (IQR)) serum folate levels for different subgroups are presented. The prevalence of low and deficient folate status within subgroups are presented as percentages with 95% confidence intervals (CI). The CIs for prevalence measures were estimated by means of the one proportion Z-test. The GAM models and plots were generated to visualize the nonlinear associations of substance use severity with serum folate, adjusted for gender, age and education level. In order to do this, a substance use severity score was generated based on the type, frequency and number of substances used (for details see Supplementary Table S1). A Sankey diagram was generated to display the flow between folate status categories (i.e., deficient, low and adequate folate status) between the first and second folate assessment for participants with at least two folate measurements (n = 516). A LMM was performed to estimate associations of clinical and sociodemographic factors with baseline serum folate concentration, and the impact of selected clinical factors on changes in serum folate concentration over time (including OAT dose ratio, injecting substances, and weekly use of substances). The model was random intercept fixed slope with the estimator set to restricted maximum likelihood. Time was defined as years from baseline, and predictor variables were kept constant to the value held at baseline. To estimate the contribution of individual clinical factors, partial adjusted models predicting associations of individual clinical factors with serum folate concentration (adjusted for gender and age) are presented in addition to the adjusted model including all variables. 3. Results The baseline characteristics of the study sample are presented in Table 1. The mean age at baseline was 44 years (SD: 11), and 70% were males. Eighty-eight percent were patients enrolled in OAT, and of these 60% were prescribed buprenorphine-based medication and 40% were prescribed methadone. Twelve percent lived under unstable housing conditions, 47% were infected with HCV and less than five participants were human immunodeficiency virus (HIV) positive at baseline. Weekly substance use was reported by 76%, and the most common substances used on a weekly basis were cannabis (49%) and benzodiazepines (38%), followed by stimulants (26%) and alcohol (25%). The median (IQR) serum folate concentration of the study population was 10 (11) nmol/L at baseline (Table 1). The distribution was leptokurtic (kurtosis: 4.9) and right-skewed (skewness: 1.5). Forty-eight percent (CI: 44–51) of the population had a low serum folate status and 23% (CI: 20–26) were deficient at baseline (Table 1). Low folate status was particularly prevalent among those under the age of 30 (59%, CI: 48–69), those injecting substances (55%, CI: 50–60) or using stimulants on a weekly basis (54%, CI: 46–62), and those prescribed methadone for OAT (54%, CI: 48–60). The prevalence of low folate status was similar between participants receiving buprenorphine for OAT (44%, CI: 39–49) compared to those not receiving OAT (45%, CI 34–57). Folate deficiency was however substantially more prevalent among those prescribed methadone (32%, CI: 26–38) compared to those prescribed buprenorphine (18%, CI: 14–22) or not receiving OAT (15%, CI: 9–25). Figure 1 shows that 63% (CI: 57–69) of participants with an adequate serum folate status, 60% (CI: 53–65) of those with low levels and 43% (CI: 34–51) of those deficient at baseline sustained this status in the subsequent assessment. Figure 1 displays the movement between folate status categories from the first (left) to the second (right) folate assessment for participants with at least two serum folate measurements (n = 516). It shows that 63% (CI: 57–69) of participants with an adequate serum folate status, 60% (CI: 53–65) of those with low levels and 43% (CI: 34–51) of those deficient at baseline sustained this status in the subsequent assessment. Definitions: Adequate (green) = s-folate > 10 nmol/L, Low (yellow) = s-folate 6.8–10 nmol/L, Deficient (red) = s-folate < 6.8 nmol/L. Figure 2 shows that serum folate concentration was inversely associated with the substance use severity score at low to mid-range scores at baseline (scores 0–12). There were few participants with the highest substance use severity scores and thus limited precision for the highest substance use severity (7.5% had scores > 15). In Figure 2, the graph to the left displays a GAM plot of the associations of serum folate concentrations with the substance use severity score the solid line depicts the association at various severity scores, whereas the shaded area represents the 95% confidence intervals of these associations. The plot to the right displays the distribution of substance use severity scores in the population. A substance use severity score of 0 equals no use of substances, whereas a score of 25 equals daily use of 5 substance classes (cannabis, stimulants, opioids, alcohol and benzodiazepines). Serum folate concentration was inversely associated with the substance use severity score at low to mid-range scores at baseline (scores 0–12). There were few participants with substance use severity scores >15 resulting in limited precision for these values. Table 2 presents the results of a LMM of serum folate predicted by sociodemographic and clinical factors. Higher serum folate concentrations were found for those over the age of 50 years (mean serum folate difference: 2.4, CI: 0.32, 4.5) and sixty years (4.5, CI: 1.5, 7.4) compared to those under the age of 30 years at baseline. Weekly use of alcohol was associated with higher serum folate concentrations at baseline compared to less or no alcohol use (1.9, CI: 0.15, 3.6), whereas weekly use of cannabis was associated with lower serum folate concentrations (−1.8, CI: −3.3, −0.25) when compared to less or no use. Statistically significant negative associations were found between injecting substances and serum folate concentration over time (−1.2, CI: −2.3, −0.14), and between higher OAT dose ratios and serum folate concentration over time (−1.1, CI: −2.2, −0.024). The β-coefficients were similar between partly adjusted and fully adjusted analyses. One exception was the effect estimate of the variable for injecting substances which was significantly associated with folate in the partly adjusted model (−1.6, CI: −3.1, −0.13), while non-significant in the fully adjusted model (−1.2, CI: −2.9, 0.41). This is likely due to correlation between the variable for injecting substances and the variable for weekly use of stimulants (correlation coefficient: 0.41) 4. Discussion Overall, the serum folate status of this Norwegian population of outpatients with SUD was poor. Nearly half of the population had serum folate levels below the cut-off that warrants intervention in the form of nutritional guidance or supplementation, and 23% were deficient. Additionally, poor folate status was to a large part sustained over time. Although we lack healthy controls and comparable data on the prevalence of folate deficiency in the general Norwegian population, studies from other high-income countries generally report the prevalence of folate deficiency to under five percent, in contrast to low-income regions where the prevalence is in the realms of twenty percent [42,43]. It is, however, important to note that many of the comparable high-income countries have implemented mandatory food fortification programs that include folic acid [44]. Norway has not implemented this, and the prevalence of folate deficiency is therefore feasibly much higher. Denmark is comparable to Norway in terms of demographics as well as fortification policies, and a study on a representative Danish sample showed that a third had poor serum status of folate and that lifestyle factors, in particular diet quality, were associated with this [45]. In regions where foods are not routinely fortified with folic acid, adequate intake of the folate is highly dependent on a sufficiently balanced diet rich in its primary sources including fruits, vegetables, and grain products [14,46]. The poor folate status presented in this study on people with SUD is likely indicative of an insufficient dietary intake, which in turn could reflect an overall poor dietary status—as is widely reported in literature [5,6,7,8,11]. In line with recommendations from WHO on tackling poor folate status in at-risk populations, we argue that any actions taken towards improving the folate status of people with SUD should, when feasible, be long-term and keep a broad scope on diet and nutrition [16]. Substance use was negatively associated with serum folate status. Higher severity of substance use, meaning more frequent use of multiple substances, was associated with lower serum folate concentration at baseline. Additionally, two factors related to SUD severity, namely injecting substances, and higher dosages of OAT medication, were associated with reduced serum folate concentration over time. This is consistent with other literature reporting higher rates of anthropometric and protein malnutrition among people with more intense [47] and prolonged substance use [48], and among those who inject substances or use heroin in combination with other substances [49]. Low folate levels were more frequent among those using methadone (a full agonist to opioid receptors) compared to buprenorphine (partial agonist) or no OAT. This could indicate an opioid related effect or interaction between opioids and eating behaviour but could also be due to selection of people with more severe substance use disorder more often using methadone compared to buprenorphine. As mentioned, the poor folate status in this population of outpatients with SUD is likely indicative of insufficient dietary intake, which again could reflect an overall poor nutritional status. A nutritional assessment of a comparable Norwegian population with severe substance use revealed largely monotonous and unfavourable diets high in dietary sugar and low in nutritious foods, including fruits and vegetables [6]. Similar tendencies have been reported in other studies on people with SUD [5,7,8,11]. Opioid users in particular, including patients enrolled in OAT, are often reported to experience a strong preference for sweet taste and sugar-sweetened foods and beverages [6,7,50]. This could feasibly predispose users of opioids to replace nutritious foods with high-sugar palatable foods that low in nutrients (so-called “empty calories”), and thereby lead to a low dietary intake of many essential micronutrients—including folate. Although we argue that insufficient intake from food is likely an important cause of poor serum folate status in this population, other factors could be of importance as well. Some pharmaceuticals, such as barbiturates and certain anti-convulsants, are known to impair folate status [1,13]. Due to the high prevalence of somatic and psychiatric health complaints in this population, the use of such medications could be of importance [25,51]. Unfortunately, we were not able to account for this in our study. Chronic HCV infection and other hepatic diseases are prevalent in this population and could, feasibly, impact folate homeostasis through impaired enterohepatic circulation and reduced decreased hepatic storage capacity [52,53]. We did, however, not find any support for this in our analyses. 5. Strengths and Limitations A major strength of this study is its relatively large sample size among a “hard-to-reach” population, and its longitudinal design. Each participant, on average, had over three serum folate measurements over the course of three years. This allowed us to follow patterns in serum folate and predictor variables over time, adding strength to the descriptive analyses and associations presented in this study. This is particularly important as serum folate is highly impacted by very recent intake of food folates, and one single measurement is often not representative true folate status. One important limitation of this study is related to the timing of health assessments and blood samplings; although health assessments were performed annually, the blood samples were not necessarily drawn in exact concurrence with these assessments. Another limitation of this study includes the lack of healthy controls and limited pre-existing knowledge on the serum folate status of the general Norwegian population. This restricts our ability to compare the folate status of this group to that of the general population. Future studies should aim to recruit healthy controls in a case-control manner, or cohort data including both the general population and people with substance use disorder. This will better allow for comparisons between groups. Moreover, most participants in our study cohort were enrolled in OAT for opioid dependence, and although many other substance classes were commonly used, our results may not be generalizable to different SUDs such as predominantly alcohol addictions. 6. Conclusions The folate status of this population of outpatients with SUD was poor. Almost half of the population had serum folate levels below the cut-off that warrants interventions such as nutritional guidance or folic acid supplementation, and a quarter were deficient. Poor folate status was particularly prevalent among those with a more frequent use of multiple substances, those injecting substances, and those prescribed higher dosages of OAT medication, or full agonist OAT. These novel findings shed light on one of the many nutritional challenges faced by people with SUD, especially by those with the most severe substance use. Efforts should be made to address poor micronutrient status, as well as other aspects of suboptimal diet and nutrition, among people with severe SUD. Acknowledgments This study was nested to the INTRO-HCV study. The INTRO-HCV study group is comprised by the following investigators (according to location and alphabetical order of surnames). Bergen: Christer Frode Aas, Vibeke Bråthen Buljovcic, Fatemeh Chalabianloo, Jan Tore Daltveit, Silvia Eiken Alpers, Lars T. Fadnes (principal investigator), Trude Fondenes Eriksen, Per Gundersen, Velinda Hille, Kristin Holmelid Håberg, Kjell Arne Johansson, Rafael Alexander Leiva, Siv-Elin Leirvåg Carlsen, Martine Lepsøy Bonnier, Lennart Lorås, Else-Marie Løberg, Mette Hegland Nordbotn, Cathrine Nygård, Maria Olsvold, Christian Ohldieck, Lillian Sivertsen, Hugo Torjussen, Jørn Henrik Vold, Jan-Magnus Økland. Stavanger: Tone Lise Eielsen, Nancy Laura Ortega Maldonado, Ewa Joanna Wilk. proLAR: Ronny Bjørnestad, Ole Jørgen Lygren, Marianne Cook Pierron. Oslo: Olav Dalgard, Håvard Midgard, Svetlana Skurtveit. Bristol: Aaron G. Lim, Peter Vickerman. Supplementary Materials The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/ijerph19095754/s1: Table S1: Scoring scheme of the “substance use severity score”. Click here for additional data file. Author Contributions Conceptualization, L.T.F., M.B., J.H.V. and K.A.J.; methodology, R.G., L.T.F., C.F.A. and J.H.V.; formal analysis, J.H.V., L.T.F., R.C., R.G. and M.B.; data curation, L.T.F., C.F.A. and J.H.V.; writing—original draft preparation, M.B., L.T.F. and J.H.V.; writing—review and editing, C.F.A., R.C., K.A.J. and R.G.; supervision, L.T.F. and J.H.V.; project administration, L.T.F. and K.A.J.; funding acquisition, L.T.F. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki and approved by the Regional Ethical Committee for Health Research, Norway (REK no: 155386 and REK Vest 2017/51). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The datasets generated and analysed during the current study are not publicly available due to privacy concerns, but are available from the corresponding author on reasonable request. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Movement between folate status categories from the first to the second serum folate assessment. Figure 2 Generalized additive model plot of the association of serum folate concentration with the substance use severity score. ijerph-19-05754-t001_Table 1 Table 1 Descriptive characteristics of the study population, median serum folate and the prevalence of inadequate folate status and folate deficiency at baseline. Characteristic N (%) S-Folate (nmol/L) Median (IQR 1) Low S-Folate % of Group (CI 2) Deficient S-Folate % of Group (CI 2) Gender Male 466 (70) 13 (9.6) 48 (43–52) 22 (20–32) Female 197 (30) 15 (11) 48 (41–55) 26 (18–25) Age group <30 years 78 (12) 9.1 (7.1) 59 (48–69) 24 (16–35) 30–39 years 188 (28) 10 (9.3) 50 (43–57) 27 (21–33) 40–49 years 202 (30) 10 (12) 48 (41–54) 22 (17–28) 50–59 years 155 (23) 12 (13) 45 (38–53) 21 (15–28) ≥60 years 40 (6) 14 (12) 25 (14–40) 15 (8–34) Education level Primary school 39 (6) 11 (13) 49 (34–64) 28 (17–44) Middle school 297 (45) 10 (10) 50 (44–54) 26 (21–31) High school 263 (40) 11 (10) 46 (40–52) 21 (16–26) ≤3 years higher education 51 (8) 13 (14) 43 (30–57) 20 (11–32) >3 years higher education 13 (2) 13 (19) 46 (23–71) 0 (0–23) Housing conditions 3 Unstable 81 (12) 10 (11) 49 (39–60) 20 (13–29) Stable 582 (88) 10 (8.3) 47 (43–51) 23 (20–27) HCV infection 4 311 (47) 11 (10) 45 (40–51) 23 (19–28) Injecting substances 5 318 (48) 9.2 (8.0) 55 (50–60) 24 (20–29) Opioid agonist therapy Buprenorphine 352 (53) 11 (11) 44 (39–49) 18 (14–22) Methadone 229 (35) 9.1 (11) 54 (48–60) 32 (26–38) Not in OAT 73 (11) 11 (11) 45 (34–57) 15 (9–25) Weekly substance use 6 Alcohol 148 (25) 13 (11) 32 (25–40) 12 (7–18) Cannabis 297 (49) 9.8 (10) 52 (46–57) 28 (23–33) Stimulants 7 157 (26) 9.4 (6.9) 54 (46–62) 21 (15–28) Benzodiazepines 230 (38) 9.6 (9.1) 52 (45–58) 26 (21–32) Non-OAT opioids 85 (14) 11 (10) 47 (37–58) 17 (10–26) No weekly substance use 144 (24) 11 (12) 44 (37–53) 23 (17–30) Overall 663 (100) 10 (11) 48 (44–51) 23 (20–26) 1 IQR, interquartile range, 2 CI, 95% confidence interval, 3 Stable housing included living in owned or rented housing or at an institution, unstable housing included homelessness, living at temporary camping sites or with friends or family, 4 Hepatitis C virus infection, defined as non-zero values on a quantitative HCV-RNA assay at baseline, 5 Self-reported injection of any substance during the 12 months prior to the first health assessment, 6 Self-reported substance use on a minimum weekly basis during the 12 months. 7 Amphetamine, methamphetamine and cocaine. ijerph-19-05754-t002_Table 2 Table 2 Linear mixed model of serum folate concentration (nmol/L) adjusted for sociodemographic and clinical factors, including substance use patterns. Fixed Effects Partly Adjusted 1 Adjusted Effect Estimate Time Trend (per Year) Effect Estimate Time Trend (per Year) Estimate (CI) Slope (CI) Estimate (CI) Slope (CI) s-folate 12.8 (10.3, 15.3) 2.1 (0.79, 3.4) Gender Male Reference (0.0) Female 0.85 (−0.49, 2.2) Age <30 Reference (0.0) 30–39 0.71 (−1.3, 2.6) 40–49 1.5 (−0.46, 3.6) 50–59 2.4 (0.32, 4.5) ≥60 4.5 (1.5, 7.4) OAT dose ratio 2 −0.018 (−1.5, 1.5) −0.71 (−1.7, 0.33) 0.06 (−1.5, 1.6) −1.1 (−2.2, −0.024) Injecting substances 3 −1.6 (−3.1, −0.13) −0.93 (−1.9, 0.014) −1.2 (−2.9, 0.41) −1.2 (−2.3, −0.14) Weekly substance use 4 Alcohol 1.9 (0.13, 3.6) −0.53 (−1.6, 0.57) 1.9 (0.15, 3.6) −0.60 (−1.7, 0.50) Cannabis −2.1 (−3.6, −0.63) 0.12 (−0.82, 1.1) −1.8 (−3.3, −0.25) 0.14 (−0.84, 1.1) Non-OAT opioids 0.32 (−1.9, 2.5) 0.16 (−1.3, 1.6) 0.80 (−1.5, 3.1) 0.22 (−1.3, 1.8) Stimulants 5 −0.93 (−2.6, 0.77) −0.54 (−1.7, 0.61) −0.29 (−2.2, 1.6) −0.38 (−1.7, 0.94) Benzodiazepines −1.1 (−2.6, 0.45) 0.47 (−0.50, 1.4) −0.50 (−2.1, 1.2) 0.84 (−0.24, 1.9) The table displays the results of a linear mixed model (restricted maximum likelihood regression) estimating associations of serum folate concentration (nmol/L) with sociodemographic and clinical predictor variables at baseline (effect estimates), as well as the impact of predictors on changes in serum folate concentrations over time (time trends per year). Significant results are shown in italics. CI, 95% confidence interval; 1 Adjusted for gender and age. In the partly adjusted model, age and gender were included as categorical independent variables in the model together with one of the clinical variables (substance use, opioid agonist therapy, and injecting behaviour) separately, as well as interaction between this variable and time (using identity and timepoints as hierarchical group variables). In the adjusted model, age and gender were also included as categorical independent variables together with all the clinical variables (substance use, opioid agonist therapy, and injecting behaviour). Substance use variables, opioid agonist therapy, and injecting use were also included as independent variables in the models with interactions between each of these variables and time (using identity and timepoints as hierarchical group variables); 2 The patients’ prescribed daily dose of opioid agonist divided by the WHO mean expected dose (90 mg for methadone, 18 mg for buprenorphine). In this variable, zero represents no prescribed OAT medication; 3 Self-reported injection of any substance during the 12 months prior to the first assessment (baseline); 4 Self-reported use of a substance at a minimum weekly basis during the 12 months prior to the first assessment; 5 Amphetamine, methamphetamine and cocaine. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Allen L.H. Causes of vitamin B12 and folate deficiency Food Nutr. Bull. 2008 29 S20 S35 S20–S34; discussion S27–S35 10.1177/15648265080292S105 18709879 2. Halsted C.H. Villanueva J.A. Devlin A.M. Chandler C.J. Metabolic interactions of alcohol and folate J. Nutr. 2002 132 2367s 2372s 10.1093/jn/132.8.2367S 12163694 3. Mahboub N. Rizk R. Karavetian M. de Vries N. Nutritional status and eating habits of people who use drugs and/or are undergoing treatment for recovery: A narrative review Nutr. Rev. 2020 79 627 635 10.1093/nutrit/nuaa095 32974658 4. Neale J. Nettleton S. Pickering L. Fischer J. Eating patterns among heroin users: A qualitative study with implications for nutritional interventions Addiction 2012 107 635 641 10.1111/j.1360-0443.2011.03660.x 21933297 5. Kheradmand A. Kheradmand A. Nutritional status in patients under methadone maintenance treatment J. Subst. Use 2020 25 173 176 10.1080/14659891.2019.1672817 6. Saeland M. Haugen M. Eriksen F.L. Wandel M. Smehaugen A. Böhmer T. Oshaug A. High sugar consumption and poor nutrient intake among drug addicts in Oslo, Norway Br. J. Nutr. 2011 105 618 624 10.1017/S0007114510003971 20880416 7. Zador D. Lyons Wall P.M. Webster I. High sugar intake in a group of women on methadone maintenance in south western Sydney, Australia Addiction 1996 91 1053 1061 10.1111/j.1360-0443.1996.tb03601.x 8688819 8. Shrestha S. Jimenez E. Garrison L. Pribis P. Raisch D.W. Stephen J.M. Bakhireva L.N. Dietary intake among opioid- and alcohol-using pregnant women Subst. Use Misuse 2018 53 260 269 10.1080/10826084.2016.1265563 28635522 9. Varela P. Marcos A. Ripoll S. Santacruz I. Requejo A.M. Effects of human immunodeficiency virus infection and detoxification time on anthropometric measurements and dietary intake of male drug addicts Am. J. Clin. Nutr. 1997 66 509s 514s 10.1093/ajcn/66.2.509S 9250140 10. El-Nakah A. Frank O. Louria D.B. Quinones M.A. Baker H. A vitamin profile of heroin addiction Am. J. Public Health 1979 69 1058 1060 10.2105/AJPH.69.10.1058 484761 11. Tomedi L.E. Bogen D.L. Hanusa B.H. Wisner K.L. Bodnar L.M. A pilot study of the nutritional status of opiate-using pregnant women on methadone maintenance therapy Subst. Use Misuse 2012 47 286 295 10.3109/10826084.2011.635324 22217127 12. Yazici A.B. Akcay Ciner O. Yazici E. Cilli A.S. Dogan B. Erol A. Comparison of vitamin B12, vitamin D and folic acid blood levels in patients with schizophrenia, drug addiction and controls J. Clin. Neurosci. 2019 65 11 16 10.1016/j.jocn.2019.04.031 31076249 13. Bailey L.B. Stover P.J. McNulty H. Fenech M.F. Gregory J.F. 3rd Mills J.L. Pfeiffer C.M. Fazili Z. Zhang M. Ueland P.M. Biomarkers of nutrition for development-folate review J. Nutr. 2015 145 1636s 1680s 10.3945/jn.114.206599 26451605 14. Brevik A. Vollset S.E. Tell G.S. Refsum H. Ueland P.M. Loeken E.B. Drevon C.A. Andersen L.F. Plasma concentration of folate as a biomarker for the intake of fruit and vegetables: The hordaland homocysteine study Am. J. Clin. Nutr. 2005 81 434 439 10.1093/ajcn.81.2.434 15699232 15. Crider K.S. Yang T.P. Berry R.J. Bailey L.B. Folate and DNA methylation: A review of molecular mechanisms and the evidence for folate’s role Adv. Nutr. 2012 3 21 38 10.3945/an.111.000992 22332098 16. De Benoist B. Conclusions of a WHO technical consultation on folate and vitamin B12 deficiencies Food Nutr. Bull. 2008 29 S238 S244 10.1177/15648265080292S129 18709899 17. Bender A. Hagan K.E. Kingston N. The association of folate and depression: A meta-analysis J. Psychiatr. Res. 2017 95 9 18 10.1016/j.jpsychires.2017.07.019 28759846 18. Jang S. Han J.W. Shin J. Kim T.H. Kwak K.P. Kim K. Kim B.J. Kim S.G. Kim J.L. Kim T.H. Normal-but-low serum folate levels and the risks for cognitive impairment Psychiatry Investig. 2019 16 532 538 10.30773/pi.2019.05.29 19. Pieroth R. Paver S. Day S. Lammersfeld C. Folate and its impact on cancer risk Curr. Nutr. Rep. 2018 7 70 84 10.1007/s13668-018-0237-y 30099693 20. Reynolds E.H. The neurology of folic acid deficiency Handb. Clin. Neurol. 2014 120 927 943 10.1016/b978-0-7020-4087-0.00061-9 24365361 21. Degenhardt L. Grebely J. Stone J. Hickman M. Vickerman P. Marshall B.D.L. Bruneau J. Altice F.L. Henderson G. Rahimi-Movaghar A. Global patterns of opioid use and dependence: Harms to populations, interventions, and future action Lancet 2019 394 1560 1579 10.1016/S0140-6736(19)32229-9 31657732 22. Connery H.S. Medication-assisted treatment of opioid use disorder: Review of the evidence and future directions Harv. Rev. Psychiatry 2015 23 63 75 10.1097/HRP.0000000000000075 25747920 23. Karow A. Verthein U. Pukrop R. Reimer J. Haasen C. Krausz M. Schäfer I. Quality of life profiles and changes in the course of maintenance treatment among 1015 patients with severe opioid dependence Subst. Use Misuse 2011 46 705 715 10.3109/10826084.2010.509854 21047149 24. Sordo L. Barrio G. Bravo M.J. Indave B.I. Degenhardt L. Wiessing L. Ferri M. Pastor-Barriuso R. Mortality risk during and after opioid substitution treatment: Systematic review and meta-analysis of cohort studies BMJ 2017 357 j1550 10.1136/bmj.j1550 28446428 25. Aas C.F. Vold J.H. Gjestad R. Skurtveit S. Lim A.G. Gjerde K.V. Løberg E.M. Johansson K.A. Fadnes L.T. Substance use and symptoms of mental health disorders: A prospective cohort of patients with severe substance use disorders in Norway Subst. Abus. Treat. Prev. Policy 2021 16 20 10.1186/s13011-021-00354-1 26. Rosca P. Haklai Z. Goldberger N. Zohar P. Margolis A. Ponizovsky A.M. Mortality and causes of death among users of methadone maintenance treatment in Israel, 1999–2008 Drug Alcohol Depend. 2012 125 160 163 10.1016/j.drugalcdep.2012.03.013 22483811 27. Vold J.H. Gjestad R. Aas C.F. Chalabianloo F. Skurtveit S. Løberg E.M. Johansson K.A. Fadnes L.T. Impact of clinical and sociodemographic factors on fatigue among patients with substance use disorder: A cohort study from Norway for the period 2016–2020 Subst. Abus. Treat. Prev. Policy 2020 15 93 10.1186/s13011-020-00334-x 33317568 28. Bech A.B. Clausen T. Waal H. Šaltytė Benth J. Skeie I. Mortality and causes of death among patients with opioid use disorder receiving opioid agonist treatment: A national register study BMC Health Serv. Res. 2019 19 440 10.1186/s12913-019-4282-z 31266495 29. Degenhardt L. Charlson F. Stanaway J. Larney S. Alexander L.T. Hickman M. Cowie B. Hall W.D. Strang J. Whiteford H. Estimating the burden of disease attributable to injecting drug use as a risk factor for HIV, hepatitis C, and hepatitis B: Findings from the Global Burden of Disease Study 2013 Lancet Infect. Dis. 2016 16 1385 1398 10.1016/S1473-3099(16)30325-5 27665254 30. Antony T. Alzaharani S.Y. El-Ghaiesh S.H. Opioid-induced hypogonadism: Pathophysiology, clinical and therapeutics review Clin. Exp. Pharm. Physiol. 2020 47 741 750 10.1111/1440-1681.13246 31. Elman I. Howard M. Borodovsky J.T. Mysels D. Rott D. Borsook D. Albanese M. Metabolic and addiction indices in patients on opioid agonist medication-assisted treatment: A comparison of buprenorphine and methadone Sci. Rep. 2020 10 5617 10.1038/s41598-020-62556-0 32221389 32. Vallecillo G. Robles M.J. Torrens M. Samos P. Roquer A. Martires P.K. Sanvisens A. Muga R. Pedro-Botet J. Metabolic syndrome among individuals with heroin use disorders on methadone therapy: Prevalence, characteristics, and related factors Subst. Abus. 2018 39 46 51 10.1080/08897077.2017.1363122 28771091 33. Helse Bergen ATLAS4LAR: Kartlegging og Behandling av Lungesykdom i Legemiddelassistert Behandling Available online: https://helse-bergen.no/avdelinger/rusmedisin/rusmedisin-seksjon-forsking/bar/atlas4lar-kartlegging-og-behandling-av-lungesykdom-i-legemiddelassistert-behandling (accessed on 10 May 2021) 34. Fadnes L.T. Aas C.F. Vold J.H. Ohldieck C. Leiva R.A. Chalabianloo F. Skurtveit S. Lygren O.J. Dalgård O. Vickerman P. Integrated treatment of hepatitis C virus infection among people who inject drugs: Study protocol for a randomised controlled trial (INTRO-HCV) BMC Infect. Dis. 2019 19 943 10.1186/s12879-019-4598-7 31703669 35. Analyseoversikten Folat Available online: https://analyseoversikten.no/analyse/110 (accessed on 17 March 2021) 36. Helse Stavanger Folater Available online: https://labhandbok.sus.no/docs/doc_17511/index.html (accessed on 17 March 2021) 37. World Health Organization Serum and Red Blood Cell Folate Concentrations for Assessing Folate Status in Populations Available online: http://apps.who.int/iris/bitstream/10665/162114/1/WHO_NMH_NHD_EPG_15.01.pdf?ua=1 (accessed on 17 March 2021) 38. Analyseoversikten s-Folat—Endring av Aksjonsgrense for Folatmangel Available online: https://analyseoversikten.no/labinfo/24 (accessed on 17 March 2021) 39. Nasjonal Brukerhåndbok i Medisinsk Biokjemi Folat, P Available online: https://www.prosedyrer.no/index.php?action=showtopic&topic=4ad7a834f5b503c9b9f6 (accessed on 17 March 2021) 40. World Health Organization Guidelines for the Psychosocially Assisted Pharmacological Treatment of Opioid Dependence WHO Gevenva, Switzerland 2009 Available online: https://www.who.int/publications/i/item/9789241547543 (accessed on 12 April 2022) 41. Enders C.K. Applied Missing Data Analysis The Guilford Press New York, NY, USA 2010 42. Öhrvik V. Lemming E.W. Nälsén C. Becker W. Ridefelt P. Lindroos A.K. Dietary intake and biomarker status of folate in Swedish adults Eur. J. Nutr. 2018 57 451 462 10.1007/s00394-016-1328-4 27787623 43. Rogers L.M. Cordero A.M. Pfeiffer C.M. Hausman D.B. Tsang B.L. De-Regil L.M. Rosenthal J. Razzaghi H. Wong E.C. Weakland A.P. Global folate status in women of reproductive age: A systematic review with emphasis on methodological issues Ann. N. Y. Acad. Sci. 2018 1431 35 57 10.1111/nyas.13963 30239016 44. Crider K.S. Bailey L.B. Berry R.J. Folic acid food fortification-its history, effect, concerns, and future directions Nutrients 2011 3 370 384 10.3390/nu3030370 22254102 45. Thuesen B.H. Husemoen L.L. Ovesen L. Jørgensen T. Fenger M. Linneberg A. Lifestyle and genetic determinants of folate and vitamin B12 levels in a general adult population Br. J. Nutr. 2010 103 1195 1204 10.1017/S0007114509992947 19968891 46. Monteagudo C. Scander H. Nilsen B. Yngve A. Folate intake in a Swedish adult population: Food sources and predictive factors Food Nutr. Res. 2017 61 1328960 10.1080/16546628.2017.1328960 28659736 47. Santolaria-Fernández F.J. Gómez-Sirvent J.L. González-Reimers C.E. Batista-López J.N. Jorge-Hernández J.A. Rodríguez-Moreno F. Martínez-Riera A. Hernández-García M.T. Nutritional assessment of drug addicts Drug Alcohol Depend. 1995 38 11 18 10.1016/0376-8716(94)01088-3 7648992 48. Nazrul Islam S.K. Jahangir Hossain K. Ahmed A. Ahsan M. Nutritional status of drug addicts undergoing detoxification: Prevalence of malnutrition and influence of illicit drugs and lifestyle Br. J. Nutr. 2002 88 507 513 10.1079/BJN2002702 12425731 49. Tang A.M. Forrester J.E. Spiegelman D. Flanigan T. Dobs A. Skinner S. Wanke C. Heavy injection drug use is associated with lower percent body fat in a multi-ethnic cohort of HIV-positive and HIV-negative drug users from three U.S. cities Am. J. Drug Alcohol Abus. 2010 36 78 86 10.3109/00952990903544851 20141402 50. Nolan L.J. Scagnelli L.M. Preference for sweet foods and higher body mass index in patients being treated in long-term methadone maintenance Subst. Use Misuse 2007 42 1555 1566 10.1080/10826080701517727 17918026 51. Medved D. Clausen T. Bukten A. Bjørnestad R. Muller A.E. Large and non-specific somatic disease burdens among ageing, long-term opioid maintenance treatment patients Subst. Abus. Treat. Prev. Policy 2020 15 87 10.1186/s13011-020-00311-4 33198799 52. Kozeniecki M. Ludke R. Kerner J. Patterson B. Micronutrients in liver disease: Roles, risk factors for deficiency, and recommendations for supplementation Nutr. Clin. Pract. 2020 35 50 62 10.1002/ncp.10451 31840874 53. Muro N. Bujanda L. Sarasqueta C. Gil I. Hijona E. Cosme A. Arenas J. Jr. Elosegui M.E. Sarasola M. Calpasoro J. Plasma levels of folate and vitamin B(12) in patients with chronic liver disease Gastroenterol. Hepatol. 2010 33 280 287 10.1016/j.gastrohep.2009.12.001 20206409

PMC009xxxxxx/PMC9099635.txt

==== Front J Clin Med J Clin Med jcm Journal of Clinical Medicine 2077-0383 MDPI 10.3390/jcm11092476 jcm-11-02476 Article Impact of Preterm Birth on Neurodevelopmental Disorders in South Korea: A Nationwide Population-Based Study https://orcid.org/0000-0002-7883-8714 Cha Jong Ho 1† Ahn Ja-Hye 1† Kim Yun Jin 2 Lee Bong Gun 3 https://orcid.org/0000-0002-2367-0934 Kim Johanna Inhyang 45 https://orcid.org/0000-0001-5956-9208 Park Hyun-Kyung 14 https://orcid.org/0000-0002-2403-3291 Kim Bung-Nyun 6*‡ https://orcid.org/0000-0002-9110-2963 Lee Hyun Ju 14*‡ Roccella Michele Academic Editor Valeriani Massimiliano Academic Editor 1 Department of Pediatrics, Hanyang University Hospital, Hanyang University College of Medicine, Seoul 04763, Korea; cjhchany@gmail.com (J.H.C.); mdscually@gmail.com (J.-H.A.); neopark@hanyang.ac.kr (H.-K.P.) 2 Biostatistical Consulting and Research Lab, Medical Research Collaborating Center, Hanyang University, Seoul 04763, Korea; yeun0148@hanyang.ac.kr 3 Department of Orthopedic Surgery, Hanyang University Hospital, Hanyang University College of Medicine, Seoul 04763, Korea; orthdr@naver.com 4 Clinical Research Institute of Developmental Medicine, Hanyang University Hospital, Seoul 04763, Korea; iambabyvox@hanmail.net 5 Department of Psychiatry, Hanyang University Hospital, Hanyang University College of Medicine, Seoul 04763, Korea 6 Division of Child and Adolescent Psychiatry, Department of Psychiatry, Seoul National University Hospital, Seoul 03080, Korea * Correspondence: kbn1@snu.ac.kr (B.-N.K.); blesslee77@hanmail.net (H.J.L.); Tel.: +82-2-2290-8399 (H.J.L.); Fax: +82-2-2297-2380 (H.J.L.) † These authors contributed equally to this work. ‡ These authors contributed equally to this work. 28 4 2022 5 2022 11 9 247614 2 2022 26 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Neurodevelopmental disorder (NDD) in preterm infants has become of great interest. We aimed to investigate the impact of preterm birth on the proportion of NDD using nationwide data provided by the Korean National Health Insurance Service. We included 4894 extremely preterm or extremely low-birth-weight (EP/ELBW; <28 weeks of gestation or birth weight < 1000 g) infants, 70,583 other preterm or low-birth-weight (OP/LBW; 28–36 weeks of gestation or birth weight < 2500 g) infants, and 264,057 full-term infants born between 2008 and 2015. We observed their neurodevelopment until 6 years of age or until the year 2019, whichever occurred first. Diagnoses of NDDs were based on the World Health Organization’s International Classification of Diseases 10th revision. An association between preterm birth and NDD was assessed using a multivariable logistic regression model. There was a stepwise increase in the risk of overall NDD with increasing degree of prematurity, from OP/LBW (adjusted odds ratio 4.46; 95% confidence interval 4.34–4.58), to EP/ELBW (16.15; 15.21–17.15). The EP/ELBW group was strongly associated with developmental delay (21.47; 20.05–22.99), cerebral palsy (88.11; 79.89–97.19), and autism spectrum disorder (11.64; 10.37–13.06). Preterm birth considerably increased the risk of NDD by the degree of prematurity. neurodevelopmental disorder preterm birth autism spectrum disorder developmental delay nationwide cohort Bio & Medical Technology Development Program of the National Research Foundation (NRF)Korean Government (MSIT)2020M3E5D9080787 2020R1F1A1048529 2019M3E5D1A01069345 This study was supported by the Bio & Medical Technology Development Program of the National Research Foundation (NRF), funded by the Korean Government (MSIT) (2020M3E5D9080787, 2020R1F1A1048529 and 2019M3E5D1A01069345). ==== Body pmc1. Introduction Recent advances in neonatal care have led to a marked increase in the survival rate of preterm infants. According to the Korean Statistical Information Service in 2020, preterm birth (gestational age (GA) < 37 weeks) and low birth weight (LBW; birth weight < 2500 g) accounted for 8.6% and 6.8% of total births, respectively, both of which were markedly higher than the figures for a decade ago, 5.8% and 4.9% [1]. As the number of survivors of preterm birth increases, their long-term neurodevelopmental issues have become of interest. Unfortunately, despite advancements in perinatal care, the neurodevelopmental outcomes of preterm infants have not been markedly improved [2,3]. In Australia, the neurodevelopmental outcomes of extremely preterm (EP; <28 weeks of GA) survivors were evaluated at eight years of age and found to be stationary over the past decade [4]. A recent French nationwide cohort study showed that rates of moderate-to-severe and mild neurodevelopmental impairments were 28% and 38.5%, respectively, in the EP population at preschool ages [5]. Neurodevelopmental disorder (NDD) is a group of lifelong conditions characterized by impairments in cognitive, communication, behavior, and motor skills [6]. Growing evidence suggests preterm birth has a high risk of a wide spectrum of NDD, including autism spectrum disorder (ASD), attention-deficit hyperactivity disorder (ADHD), intellectual disability (ID), and language disorder (LD) [7,8,9,10,11]. Even preterm infants born without visible brain injuries may result in later neurodevelopmental impairment [12,13]. Thus, every preterm population is at risk of developing NDD, and this risk should be thoroughly investigated. However, there is a limited literature regarding the risk of a wide spectrum of NDD in the preterm populations using nationwide data. As a definition of neurodevelopmental impairment varies across the nations or cohorts, definitions used in the study significantly influence the probability of NDD at diverse ages [14]. For instance, the incidence of neurodevelopmental impairment of EP survivors measured at 21 months of corrected age varied from 3.5% to 14.9% by the definition in a recent Canadian cohort study [15]. Moreover, the incidence of neurodevelopmental impairment is affected by the degree of prematurity. Prevalence of any neurodevelopmental impairment in EP survivors at two to five years of age is known to be substantial, as high as 11–37% in North America and 42% in the rest of the world [16]. For survivors with 28–31 weeks of GA, the incidence of neurodevelopmental disorder was 19% at five years of age [5]. In addition, the neurodevelopmental outcomes of the moderate-to-later preterm population are notable, reflecting the potential risk of the group [17]. Measured in two years of corrected age, the incidence of moderate-to-severe language delay was 13.7%, which was significantly higher than the term-born group in a previous Australian report [18]. Despite these methodological issues, the neurodevelopmental assessment of every preterm survivor is needed considering the high incidence of neurodevelopmental impairment. Herein, we investigated the overall prevalence of NDD in preschool-age children who were born prematurely, stratified by demographic and socioeconomic characteristics, using the nationwide database provided by Health Insurance Review and Assessment Service (HIRA) in South Korea. 2. Materials and Methods 2.1. Data Source and Case Definition We used the database obtained from the HIRA. The HIRA stores and reviews data on medical claims for the South Korean population, including diagnostic codes, medical visits, prescription records, and demographic information. In South Korea, each individual is registered with his/her health insurance identification number. More than 97% of the population is covered by national health insurance, and the remaining population is covered by the medical-aid program that provides healthcare services to low-income households. Thus, medical-aid beneficiaries can be defined as having the lowest socioeconomic status. Records with diagnostic codes were based on the World Health Organization’s International Classification of Diseases 10th revision (ICD-10). The study was approved by the Institutional Review Board of Hanyang University (IRB No. 2020-08-031). Informed consent was not required since public data from the HIRA were used. We analyzed health claims data of the pediatric population with ICD-10 birth codes recorded between 2008 and 2019 in the HIRA database. Our study population included newborns from 2008 to 2015 and observed their NDD until six years of age or until the year 2019, whichever occurred first. Since the database contain missing data in the ICD-10 birth codes for GA or birth weight, the study population was subdivided into three subgroups combining these measures; (1) EP or extremely low-birth-weight (ELBW) infants, with GA < 28 weeks or birth weight < 1000 g; (2) other preterm (OP) or LBW infants, with GA < 37 weeks or birth weight < 2500 g; and (3) full-term infants (FT), infants with GA ≥ 37 weeks. Exclusion criteria were as follows: (1) children diagnosed with congenital malformations of the nervous system (Q00, anencephaly and similar malformations; Q01, encephalocele; Q02, microcephaly; Q03, congenital hydrocephalus; Q04, other congenital malformation of the brain), (2) children diagnosed with chromosomal anomaly (Q09); and (3) children with missing information for GA or birth weight. The NDDs of interest included developmental delay (DD), cerebral palsy (CP), ASD, ADHD, LD, ID, and tic disorder (TD). Detailed information about ICD-10 codes for the definition of the study group and each sub-condition is summarized in Table 1. DD was diagnosed with children with a significantly delayed attainment of the expected physiological developmental stage. The diagnosis of NDD was made when children visited the outpatient clinic at least twice or those with more than one admission with a primary diagnosis. Each diagnosis was confirmed by clinical experts in pediatric rehabilitation, pediatric neurology, and pediatric psychiatry. 2.2. The Proportion of NDD The annual cumulative incidence of each sub-condition was calculated starting on the 1 January of each year from 2012 to 2017. To avoid bias in investigating temporal trends of NDDs, we tried to contain populations with more than three different birth years in each year of diagnosis. The proportion of diagnosed cases based on the population subgroup in each sub-condition was calculated. The age of diagnosis was computed in years, since our data uses the age of diagnosis only in years, not in months. 2.3. Statistical Analysis The prevalence of NDD was calculated by dividing the number of children diagnosed with NDD by the number of children who participated in the study. In addition, total cases were stratified by the degree of prematurity (categorized by individuals born EP/ELBW, OP/LBW, and FT), sex, and socioeconomic status (categorized by medium-to-highest and the lowest), and differences between each selected characteristic were evaluated. The age of diagnosis was found to be skewed, as assessed by the Anderson–Darling test, and was presented with interquartile ranges. A logistic regression model was used to estimate the adjusted odds ratio (aOR) and 95% confidence interval (C.I) to investigate significant factors for cases of NDD. The selected variables were the degree of prematurity, sex, and socioeconomic status. A univariable logistic regression analysis was performed to identify the association between the proportion of NDDs and preterm birth. A multivariable logistic regression analysis was performed after adjusting for sex and socioeconomic status. The probability of NDD based on age at the time of diagnosis were presented with Kaplan–Meier survival curves and compared among population subgroups using the log-rank test. Statistical analysis was performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). 3. Results 3.1. Demographics and the Proportion of NDD Figure 1 summarizes the flow of the study population selection process. First, 662,191 pediatric populations with ICD-10 birth codes in the HIRA database were enrolled. Then, we selected 351,665 newborns who were born between 2008 and 2015. After exclusion criteria, 339,534 newborns were included in the study group; 4894 infants born EP/ELBW, 70,583 infants born OP, and 264,057 infants born FT. The baseline demographic characteristics of children with NDD are summarized in Table 2. The prevalence of NDD in South Korea ranged from 0.6% to 4.1%, with that of DD and LD ranked 1 and 2, respectively, among the total population. In the prevalence of any NDD, males were more likely to be diagnosed than females. In socioeconomic status, except for TD, the lowest socioeconomic status group had more prevalent NDD cases than the medium-to-highest socioeconomic status group. Among sub-conditions, DD had the highest proportion of diagnosed cases, excluding CP, in both preterm populations, followed by ASD. CP, the most widely known NDD for preterm population, had the highest proportion of preterm infants than any other NDDs. Meanwhile, the median age of diagnosis differed among NDDs; indeed, DD, CP, and ASD were more prevalent before three years of age, while ADHD, LD, ID, and TD were more prevalent after three years of age. The annual cumulative incidence of NDD is summarized in Supplementary Table S1. Trends in the cumulative incidence also differed among the sub-conditions. In total, NDDs with decreasing trend over time was DD (0.76–0.58%), and CP (0.34–0.13%). On the other hand, the incidence of ADHD (0.03–0.32%), LD (0.28–0.47%), ID (0.06–0.14%), and TD (0.03–0.17%) increased and that of ASD (0.19–0.20%) was plateaued. The Kaplan–Meier survival curves for subgroups based on the age of diagnosis are presented in Figure 2. The NDD-free probability of each sub-condition was highest in the EP/ELBW group, followed by the OP/LBW and the FT group (p < 0.001). The NDD-free probability of DD and CP in both preterm groups was higher than that of the FT group at the beginning of the observation period; however, during the latter phase, the NDD-free probability of ADHD, LD, and ID in both preterm groups increased rapidly and was ultimately higher than that of the FT group. 3.2. Associations between Preterm Birth and NDD Table 3 summarizes the result of the logistic regression model. The aOR of overall NDD was 4.46 (95% C.I 4.34–4.58) in the OP/LBW group and 16.15 (95% C.I 15.21–17.15) in the EP/ELBW group, respectively. The aOR of the EP/ELBW group was highest in CP (88.11; 79.89–97.19), followed by DD (21.47; 20.05–22.99) and ASD (11.64; 10.37–13.06). The rank of aOR in the OP/LBW group was as same as that of EP/ELBW group; however, the aOR of the OP/LBW group was smaller. The aORs of CP, DD, and ASD were 18.09 (95% C.I 16.74–19.54), 6.19 (95% C.I 5.97–6.43), and 4.11(95% C.I 3.86–4.38), respectively. 4. Discussion This nationwide population-based study investigated the impact of preterm birth on NDD in South Korea. Our study revealed that preterm birth is an evident risk factor for NDD; the odds of an overall diagnosis were 16.15 times higher in the EP/ELBW group and 4.46 times higher in the OP/LBW group. In addition, the preterm population accounted for a considerable proportion of diagnosed cases. Lastly, preterm birth had the most significant impact on ASD and DD, following CP. To the best of our knowledge, this is the first study in Asian countries to address the association between preterm birth and a wide spectrum of NDD using nationwide birth cohort data. Although there is likely an inflated preterm population who were monitored more closely with earlier diagnosis of NDD than children born with FT, our findings provide a valuable understanding of the national epidemiology of preterm infants with NDD in South Korea. Nevertheless, our study has several limitations which are associated with the characteristics of the raw data. First, as we selected study populations who assigned the birth codes of ICD-10, selection bias can occur. It is not obligatory for clinicians to assign birth codes regarding both GA and birth weight. To obtain as many individuals as possible, we entered subjects who assigned either one of the information. Second, we limited the age of interest to 0–6 years. Newly diagnosed cases or individuals diagnosed with one or more morbidities later than six years of age could not be calculated. Additionally, the observed time period among individuals varied from three to six years. Therefore, a gap between our findings and actual incidences might exist. Third, we could not consider clinical (i.e., mechanical ventilation, surfactant use) and environmental information (i.e., maternal medication, familial structure) which are not registered in the form of ICD-10 codes. Additionally, our study design may have overlooked co-morbidity cases and sub-clinical neurodevelopmental problems. Lastly, we could not consider individuals with undernutrition, and growth restrictions. Risk factors associated with NDD are not yet completely understood and are expected to be multifactorial [19]. Regarding socioeconomic status, nationwide studies conducted in Taiwan [20] and in the USA [21] found that a lower socioeconomic status decreased the probability of NDD, which was contrary to our result. These contradictory findings may be implicated in the limited access to adequate medical services and the caregiver’s inability to recognize early signs of NDD. In South Korea, the government has implemented a ‘National Health Screening Program for infants and children’ since 2007, which consists of seven annual health check-ups until preschool age [22]. We suspect that this policy enabled early detection of NDD in the high-risk group, even in the population group with low socioeconomic status. However, given that a ‘medical blind spot’ still exists for health insurance services, the risk of NDD in the lowest socioeconomic status group could be underestimated. Previous studies have shown that preterm birth is a significant risk factor for NDD. Recent meta-analysis studies have shown that the preterm population has as much as a three-fold increase in the aOR of NDD compared to the control group [7,17,23]. Our study is an up-to-date nationwide study with the current status and trend of NDD. Previous nationwide studies measured neurodevelopmental outcomes with the diagnosis of CP for motor impairment, the result of the Bayley Scales of Infant and Toddler test for cognitive and language impairment in toddler age [14]. At preschool age or older, intelligent quotient, academic achievement, and the questionnaires by parents were widely used [4,21]. Compared to those studies, our findings would be more conservative. Since we only confirmed diagnoses made by clinicians, subclinical neurodevelopmental difficulties (i.e., academic performance, peer relationship) can be overlooked. A few nationwide studies have been conducted using diagnostic codes as an evaluation method [8,10,24,25,26]. Compared to those studies, our study covers an up-to-date preterm population reflecting timely detection of long-term issues following recent better survival rates and the increasing number of preterm. CP, which marked the highest aOR among NDDs, is a widely known sequela of preterm birth. Based on a nationwide study in South Korea [27], DD, ASD, and LD were the sub-conditions that most contributed to the recently increasing trend of NDD. We assume that preterm birth has considerable implications. In South Korea, DD (R62.0) is a diagnostic code for referral to tertiary hospitals when a delay in acquiring developmental skills is highly suspected for children < 3 years of age. It includes children in states who have not been diagnosed with the specific NDD and children with caregivers who avoid a certain diagnosis for reasons such as social myths and insurance issues. Thus, the decreasing probability of DD in the preterm population should not be interpreted as an improvement in the developmental outcome; rather, it suggests an emerging trend of early intervention with a precise evaluation of preterm infants with suspected DD. In addition, considering that DD has high comorbidity to ASD [28], it should be noted that a high risk of DD can be attributable to the risk of ASD. In 2011, the ASD diagnosis in the Diagnostic and Statistical Manual of Mental Disorders Fifth Edition was revised due to its ambiguity. After revision, the symptom definition of ASD narrowed and its incidence is expected to be lower in the toddler group [29,30]. In particular, the probability of ASD in preterm infants was higher compared to a study performed in previous decades [31]. Differences may have originated from methodological differences since we selected cases of NDD in which the diagnostic code concerning birth (preterm or FT) was used at baseline. We excluded NDD cases without information regarding GA or birth weight to quantify the prevalence of NDD based on the degree of prematurity. Moreover, there is a tendency that the diagnostic code for the FT group (Z38) is not entered when perinatal issues occur (e.g., asphyxia, transient tachypnea of newborn, jaundice). Therefore, our study design could overlook those affected populations and have a relatively large proportion of preterm infants. Although the current analysis showed a decreasing prevalence of ASD in both preterm groups, caution must be exercised when interpreting the results. However, the results show that the probability of ASD has increased over time and recently plateaued. For ASD and LD, the median age at the time of diagnosis was higher than that of DD, suggesting that early identification of these disorders is still challenging. Our group has recently reported, with the use of neuroimaging, that preterm infants have alterations in fronto-limbic circuitry maturation, especially the cingulum, which is related to the core symptoms of ASD [32,33]. In addition, the preterm population showed an increasing trend and significant impact on both ADHD and LD. This was in line with a previous Swedish nationwide study, which showed that the risk of ADHD was inversely related to GA [8]. As shown in the median age of diagnosis, both ADHD and LD are diagnosed when higher cortical functions develop to a specific level, which implies that such disorders would become more prominent as children reach school age. Our findings suggest that increasing awareness of ADHD and LD is needed, since surviving preterm infants are more likely to present more issues at school age [34]. Unlike other NDDs we studied, the pattern of TD was different; highest in OP/LBW infants, followed by FT and EP/ELBW infants. The mean onset of TD is between six and seven years and the annual average prevalence is known to be 0.2–0.3% [35,36]. Therefore, it is hard to say that the OP/LBW or FT infants have a significant risk of TD compared to EP infants, based on our data. Further study with sufficient follow-up duration is needed. 5. Conclusions Our study examined the prevalence and trend of NDD and the impact of preterm birth using nationwide data. The risk of NDD was considerably increased by the degree of prematurity, and this tendency was more prominent in ASD and DD. We emphasize the importance of NDD as preterm birth-related morbidity, and further intervention and management programs should be implemented. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/jcm11092476/s1, Supplementary Table S1: Trends in annual cumulative incidence of neurodevelopmental disorders in the pediatric population from 2012 to 2017 in South Korea. Click here for additional data file. Author Contributions Conceptualization, J.H.C., J.-H.A. and H.J.L.; methodology, J.H.C., Y.J.K. and B.G.L.; software, J.-H.A. and Y.J.K.; validation, J.I.K., H.-K.P., H.J.L. and B.-N.K.; formal analysis, J.H.C., Y.J.K. and H.J.L.; investigation, J.H.C. and J.-H.A.; resources, J.I.K., H.J.L. and B.-N.K.; data curation, J.H.C., J.-H.A., Y.J.K., B.G.L. and H.J.L.; writing—original draft preparation, J.H.C. and H.J.L.; writing—review and editing, J.-H.A., B.G.L., J.I.K., H.-K.P. and B.-N.K.; visualization, J.-H.A., Y.J.K.; supervision, B.-N.K. and H.J.L.; project administration, B.-N.K. and H.J.L.; funding acquisition, J.I.K., H.-K.P. and H.J.L. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of Hanyang University Hospital (HYUH 2020-08-031). Informed Consent Statement Patient consent was waived due to the historical cohort nature of the study using National Health Insurance data. Data Availability Statement The data presented in this study are available on request from the corresponding author. Conflicts of Interest The funders had no role in the design of the study; in the collection, analysis, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. Figure 1 Flow chart of the study population. Abbreviations: ICD-10, the World Health Organization’s International Classification of Diseases 10th revision; EP, extremely preterm; ELBW, extremely low birth weight; OP, other preterm; LBW, low birth weight; FT, full-term. Figure 2 Comparisons of the probability of neurodevelopmental disorders in a pediatric population stratified by the degree of prematurity. All subjects are plotted against follow-up years until six years of age. The solid blue line represents the EP/ELBW group, the red dotted line represents the OP/LBW group, and the green dotted line represents the FT group. The x-axis represents follow-up time (year) and the y-axis represents the NDD-free probability. The log-rank test was applied, and p < 0.001 was considered statistically significant. Abbreviations: NDD, neurodevelopmental disorder; EP, extremely preterm; ELBW, extremely low birth weight; OP, other preterm; LBW, low birth weight; FT, full-term; ASD, autism spectrum disorder; ADHD, attention-deficit hyperactivity disorder. jcm-11-02476-t001_Table 1 Table 1 ICD-10 diagnostic codes included to the definition of study groups and neurodevelopmental disorders. Definition ICD-10 Codes EP/ELBW group a P07.0 (P07.00, P07.01, P07.02) P07.2 (P07.20, P07.21, P07.22, P07.23, P07.24, P07.25) OP/LBW group b P07.1 (P07.10, P07.11, P07.12, P07.13, P07.14) P07.3 (P07.30, P07.31, P07.32) FT group Z38 (Z38.0 Z38.3 Z38.6) Developmental delay c R62.0 Cerebral palsy G8 (G80, G81, G82, G83) Autism spectrum disorder F84 (F84.0, F84.1, F84.2, F84.3, F84.4, F84.5, F84.8, F84.9) Attention-deficit hyperactivity disorder F90 (F90.0, F90.1, F90.8, F90.9) Language disorder F80 (F80.0, F80.1, F80.2, F80.3, F80.8, F80.9) Intellectual disability F7 (F70, F71, F72, F73, F78, F79) Tic disorder F95 (F95.0, F95.1, F95.2, F95.8, F95.9) a Contains preterm infants with P07.0 (extremely low birth weight; birth weight of less than 1000 g) and P07.2 (extremely preterm; <28 weeks of gestation). b Contains preterm infants with P07.1 (low birth weight; birth weight of less than 2500 g) and P07.3 (very preterm and moderate-to-later preterm; 28–36 weeks of gestation). c Diagnosed in children with a significant delay in acquiring developmental skill areas including gross motor, fine motor, verbal speech, language, and self-help. Abbreviations: ICD-10, the World Health Organization’s International Classification of Diseases 10th revision; EP, extremely preterm; ELBW, extremely low birth weight; OP, other preterm; LBW, low birth weight; FT, full-term. jcm-11-02476-t002_Table 2 Table 2 Baseline demographic findings of children diagnosed with neurodevelopmental disorders. DD CP ASD ADHD LD ID TD Population subgroup, N (%) Total, 339,534 (100) 13,872 (4.1) 5494 (1.6) 4310 (1.3) 4551 (1.3) 7189 (2.1) 2268 (0.7) 2184 (0.6) EP/ELBW group, 4894 (1.4) 1405 (28.7) 1023 (20.9) 370 (7.6) 160 (3.3) 284 (5.8) 161 (3.3) 22 (0.5) OP/LBW group, 70,583 (20.8) 7516 (10.7) 3676 (5.2) 2056 (2.9) 1538 (2.2) 2433 (3.5) 974 (1.4) 564 (0.8) FT group, 264,057 (77.8) 4951 (1.9) 795 (0.3) 1884 (0.7) 2853 (1.1) 4472 (1.7) 1133 (0.4) 1598 (0.6) Sex, N (%) Male, 175,328 (51.6) 8459 (4.8) 3149 (1.8) 2924 (1.7) 3640 (2.1) 5138 (2.9) 1583 (0.9) 1574 (0.9) Female, 164,206 (48.4) 5413 (3.3) 2345 (1.4) 1386 (0.8) 911 (0.6) 2051 (1.3) 685 (0.4) 610 (0.4) Socioeconomic status, N (%) Medium-to-highest, 336,364 (99.1) 13,691 (4.1) 5409 (1.6) 4232 (1.3) 4422 (1.3) 7020 (2.1) 2158 (0.6) 2167 (0.6) Lowest, 3170 (0.9) 181 (5.7) 85 (2.7) 78 (2.5) 129 (4.1) 169 (5.3) 110 (3.5) 17 (0.5) The proportion of the diagnosed cases, % EP/ELBW group 10.1% 18.6% 8.6% 3.5% 4.0% 7.1% 1.0% OP/LBW group 54.1% 66.9% 47.7% 33.8% 33.8% 43.0% 25.8% FT group 35.8% 14.5% 43.7% 62.7% 62.2% 49.9% 73.2% Median age of diagnosis 1.11 (0.41–2.92) 0.84 (0.43–1.75) 2.59 (0.64–3.77) 3.83 (3.41–4.60) 3.14 (2.47–3.96) 3.87 (3.25–4.82) 3.91 (3.28–4.71) Data are expressed as number (%) for categorical variables and median [Q1–Q3] for continuous variables. Abbreviations: DD, developmental delay; CP, cerebral palsy; ASD, autism spectrum disorder; ADHD, attention-deficit hyperactivity disorder; LD, language disorder; ID, intellectual disability; TD, tic disorder; EP, extremely preterm; ELBW, extremely low birth weight; OP, other preterm; LBW, low birth weight; FT, full-term. jcm-11-02476-t003_Table 3 Table 3 Multivariable logistic regression model presenting the association between the probability of neurodevelopmental disorders and preterm birth adjusted for sex and socioeconomic status. Univariable Multivariable Odds Ratio 95% C.I Odds Ratio 95% C.I DD EP/ELBW group 21.08 a 19.69–22.56 21.47 a 20.05–22.99 OP/LBW group 6.24 a 6.01–6.47 6.19 a 5.97–6.43 FT group reference reference CP EP/ELBW group 87.51 a 79.35–96.52 88.11 a 79.89–97.19 OP/LBW group 18.19 a 16.84–19.65 18.09 a 16.74–19.54 FT group reference reference ASD EP/ELBW group 11.39 a 10.15–12.78 11.64 a 10.37–13.06 OP/LBW group 4.18 a 3.92–4.45 4.11 a 3.86–4.38 FT group reference reference ADHD EP/ELBW group 3.09 a 2.63–3.64 3.20 a 2.72–3.77 OP/LBW group 2.04 b 1.91–2.17 1.98 b 1.86–2.11 FT group reference reference LD EP/ELBW group 3.58 a 3.16–4.05 3.67 a 3.24–4.15 OP/LBW group 2.07 b 1.97–2.18 2.03 b 1.93–2.13 FT group reference reference ID EP/ELBW group 7.90 a 6.68–9.33 8.07 a 6.81–9.54 OP/LBW group 3.25 b 2.98–3.54 3.16 b 2.90–3.45 FT group reference reference TD EP/ELBW group 0.74 0.48–1.13 0.76 0.42–1.15 OP/LBW group 1.32 a 1.20–1.46 1.30 a 1.18–1.43 FT group reference reference Overall EP/ELBW group 15.40 a 14.51–16.33 16.15 a 15.21–17.15 OP/LBW group 4.48 a 4.36–4.60 4.46 a 4.34–4.58 FT group reference reference Abbreviations: C.I, confidence interval; EP, extremely preterm; ELBW, extremely low birth weight; OP, other preterm; LBW, low birth weight; DD, developmental delay; CP, cerebral palsy; ASD, autism spectrum disorder; ADHD, attention-deficit hyperactivity disorder; LD, language disorder; ID, intellectual disability; TD, tic disorder. a p < 0.001, b 0.001 < p < 0.05. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. KOSIS Birth Statistics in 2020 Available online: https://kostat.go.kr/portal/korea/kor_nw/1/1/index.board?bmode=read&aSeq=391575 (accessed on 25 August 2021) 2. Potharst E.S. van Wassenaer A.G. Houtzager B.A. van Hus J.W. Last B.F. Kok J.H. High incidence of multi-domain disabilities in very preterm chilsdren at five years of age J. Pediatrics 2011 159 79 85 10.1016/j.jpeds.2010.12.055 21349538 3. Cheong J.L. Spittle A.J. Burnett A.C. Anderson P.J. Doyle L.W. Have outcomes following extremely preterm birth improved over time? Semin. Fetal Neonatal Med. 2020 25 101114 10.1016/j.siny.2020.101114 32451304 4. Cheong J.L. Anderson P.J. Burnett A.C. Roberts G. Davis N. Hickey L. Carse E. Doyle L.W. Group V.I.C.S. Changing neurodevelopment at 8 years in children born extremely preterm since the 1990s Pediatrics 2017 139 e20164086 10.1542/peds.2016-4086 28814550 5. Pierrat V. Marchand-Martin L. Marret S. Arnaud C. Benhammou V. Cambonie G. Debillon T. Dufourg M.-N. Gire C. Goffinet F. Neurodevelopmental outcomes at age 5 among children born preterm: EPIPAGE-2 cohort study BMJ 2021 373 n741 10.1136/bmj.n741 33910920 6. Mullin A. Gokhale A. Moreno-De-Luca A. Sanyal S. Waddington J. Faundez V. Neurodevelopmental disorders: Mechanisms and boundary definitions from genomes, interactomes and proteomes Transl. Psychiatry 2013 3 e329 10.1038/tp.2013.108 24301647 7. Laverty C. Surtees A. O’Sullivan R. Sutherland D. Jones C. Richards C. The prevalence and profile of autism in individuals born preterm: A systematic review and meta-analysis J. Neurodev. Disord. 2021 13 41 10.1186/s11689-021-09382-1 34548007 8. Lindström K. Lindblad F. Hjern A. Preterm birth and attention-deficit/hyperactivity disorder in schoolchildren Pediatrics 2011 127 858 865 10.1542/peds.2010-1279 21502231 9. Rabie N. Bird T. Magann E. Hall R. McKelvey S. ADHD and developmental speech/language disorders in late preterm, early term and term infants J. Perinatol. 2015 35 660 664 10.1038/jp.2015.28 25836321 10. Moster D. Lie R.T. Markestad T. Long-term medical and social consequences of preterm birth N. Engl. J. Med. 2008 359 262 273 10.1056/NEJMoa0706475 18635431 11. Limperopoulos C. Bassan H. Sullivan N.R. Soul J.S. Robertson R.L. Moore M. Ringer S.A. Volpe J.J. du Plessis A.J. Positive screening for autism in ex-preterm infants: Prevalence and risk factors Pediatrics 2008 121 758 765 10.1542/peds.2007-2158 18381541 12. Peterson B.S. Vohr B. Staib L.H. Cannistraci C.J. Dolberg A. Schneider K.C. Katz K.H. Westerveld M. Sparrow S. Anderson A.W. Regional brain volume abnormalities and long-term cognitive outcome in preterm infants JAMA 2000 284 1939 1947 10.1001/jama.284.15.1939 11035890 13. Nosarti C. Al-Asady M.H. Frangou S. Stewart A.L. Rifkin L. Murray R.M. Adolescents who were born very preterm have decreased brain volumes Brain 2002 125 1616 1623 10.1093/brain/awf157 12077010 14. Torchin H. Morgan A.S. Ancel P.-Y. International comparisons of neurodevelopmental outcomes in infants born very preterm Fetal Neonatal Med. 2020 25 101109 10.1016/j.siny.2020.101109 32354556 15. Haslam M.D. Lisonkova S. Creighton D. Church P. Yang J. Shah P.S. Joseph K. Synnes A. Harrison A. Ting J. Severe neurodevelopmental impairment in neonates born preterm: Impact of varying definitions in a Canadian cohort J. Pediatrics 2018 197 75 81.e74 10.1016/j.jpeds.2017.12.020 29398054 16. Sarda S.P. Sarri G. Siffel C. Global prevalence of long-term neurodevelopmental impairment following extremely preterm birth: A systematic literature review J. Int. Med. Res. 2021 49 03000605211028026 10.1177/03000605211028026 17. de Gamarra-Oca L.F. Ojeda N. Gómez-Gastiasoro A. Peña J. Ibarretxe-Bilbao N. García-Guerrero M.A. Loureiro B. Zubiaurre-Elorza L. Long-Term Neurodevelopmental Outcomes after Moderate and Late Preterm Birth: A Systematic Review J. Pediatrics 2021 237 168 176.e111 10.1016/j.jpeds.2021.06.004 18. Cheong J.L. Doyle L.W. Burnett A.C. Lee K.J. Walsh J.M. Potter C.R. Treyvaud K. Thompson D.K. Olsen J.E. Anderson P.J. Association between moderate and late preterm birth and neurodevelopment and social-emotional development at age 2 years JAMA Pediatrics 2017 171 e164805 10.1001/jamapediatrics.2016.4805 28152144 19. May T. Adesina I. McGillivray J. Rinehart N.J. Sex differences in neurodevelopmental disorders Curr. Opin. Neurol. 2019 32 622 626 10.1097/WCO.0000000000000714 31135460 20. Chen C.-Y. Liu C.-Y. Su W.-C. Huang S.-L. Lin K.-M. Factors associated with the diagnosis of neurodevelopmental disorders: A population-based longitudinal study Pediatrics 2007 119 e435 e443 10.1542/peds.2006-1477 17272605 21. Zablotsky B. Black L.I. Maenner M.J. Schieve L.A. Danielson M.L. Bitsko R.H. Blumberg S.J. Kogan M.D. Boyle C.A. Prevalence and trends of developmental disabilities among children in the United States: 2009–2017 Pediatrics 2019 144 e20190811 10.1542/peds.2019-0811 31558576 22. Moon J.S. Lee S.Y. Eun B.-L. Kim S.W. Kim Y.K. Shin S.M. Lee H.K. Chung H.J. One-year evaluation of the national health screening program for infants and children in Korea Korean J. Pediatrics 2010 53 307 313 10.3345/kjp.2010.53.3.307 23. Allotey J. Zamora J. Cheong-See F. Kalidindi M. Arroyo-Manzano D. Asztalos E. Van Der Post J. Mol B. Moore D. Birtles D. Cognitive, motor, behavioural and academic performances of children born preterm: A meta-analysis and systematic review involving 64 061 children BJOG Int. J. Obstet. Gynaecol. 2018 125 16 25 10.1111/1471-0528.14832 29024294 24. Leavey A. Zwaigenbaum L. Heavner K. Burstyn I. Gestational age at birth and risk of autism spectrum disorders in Alberta, Canada J. Pediatrics 2013 162 361 368 10.1016/j.jpeds.2012.07.040 22947654 25. Mathiasen R. Hansen B.M. Forman J.L. Kessing L.V. Greisen G. The risk of psychiatric disorders in individuals born prematurely in Denmark from 1974 to 1996 Acta Paediatr. 2011 100 691 699 10.1111/j.1651-2227.2011.02136.x 21244484 26. Hwang Y.-S. Weng S.-F. Cho C.-Y. Tsai W.-H. Higher prevalence of autism in Taiwanese children born prematurely: A nationwide population-based study Res. Dev. Disabil. 2013 34 2462 2468 10.1016/j.ridd.2013.05.019 23747937 27. Rah S.S. Hong S.-B. Yoon J.Y. Prevalence and incidence of developmental disorders in Korea: A nationwide population-based study J. Autism Dev. Disord. 2020 50 4504 4511 10.1007/s10803-020-04444-0 32347466 28. Koch S.V. Andersson M. Hvelplund C. Skovgaard A.M. Mental disorders in referred 0–3-year-old children: A population-based study of incidence, comorbidity and perinatal risk factors Eur. Child Adolesc. Psychiatry 2021 30 1251 1262 10.1007/s00787-020-01616-2 32815033 29. Worley J.A. Matson J.L. Comparing symptoms of autism spectrum disorders using the current DSM-IV-TR diagnostic criteria and the proposed DSM-V diagnostic criteria Res. Autism Spectr. Disord. 2012 6 965 970 10.1016/j.rasd.2011.12.012 30. Matson J.L. Kozlowski A.M. Hattier M.A. Horovitz M. Sipes M. DSM-IV vs DSM-5 diagnostic criteria for toddlers with autism Dev. Neurorehabilit. 2012 15 185 190 10.3109/17518423.2012.672341 31. Pinto-Martin J.A. Levy S.E. Feldman J.F. Lorenz J.M. Paneth N. Whitaker A.H. Prevalence of autism spectrum disorder in adolescents born weighing < 2000 grams Pediatrics 2011 128 883 891 22007018 32. Lee J.Y. Choi Y.-H. Cha J.H. Lee J.Y. Lee Y.-J. Kim B.-H. Kim I.-K. Lee J.-M. Lee H.J. Altered asymmetries of the structural networks comprising the fronto-limbic brain circuitry of preterm infants Sci. Rep. 2021 11 1318 10.1038/s41598-020-79446-0 33446672 33. Lee H.J. Kwon H. Kim J.I. Lee J.Y. Lee J.Y. Bang S. Lee J.-M. The cingulum in very preterm infants relates to language and social-emotional impairment at 2 years of term-equivalent age NeuroImage Clin. 2021 29 102528 10.1016/j.nicl.2020.102528 33338967 34. Hong M. Kwack Y.S. Joung Y.S. Lee S.I. Kim B. Sohn S.H. Chung U.S. Yang J. Bhang S.Y. Hwang J.W. Nationwide rate of attention-deficit hyperactivity disorder diagnosis and pharmacotherapy in K orea in 2008–2011 Asia-Pac. Psychiatry 2014 6 379 385 10.1111/appy.12154 25277653 35. Freeman R.D. Fast D.K. Burd L. Kerbeshian J. Robertson M.M. Sandor P. An international perspective on Tourette syndrome: Selected findings from 3500 individuals in 22 countries Dev. Med. Child Neurol. 2000 42 436 447 10.1017/S0012162200000839 10972415 36. Choi S. Lee H. Song D.-H. Cheon K.-A. Population-based epidemiology of pediatric patients with treated tic disorders from real-world evidence in Korea J. Child Adolesc. Psychopharmacol. 2019 29 764 772 10.1089/cap.2019.0050 31361509

PMC009xxxxxx/PMC9099636.txt

==== Front Animals (Basel) Animals (Basel) animals Animals : an Open Access Journal from MDPI 2076-2615 MDPI 10.3390/ani12091074 animals-12-01074 Article Canine Caregivers: Paradoxical Challenges and Rewards https://orcid.org/0000-0002-5816-9672 Kogan Lori R. 1* Wallace Jean E. 2 Hellyer Peter W. 1 https://orcid.org/0000-0003-1870-4244 Carr Eloise C. J. 3 Paterson Mandy Academic Editor 1 College of Veterinary Medicine and Biomedical Sciences, Colorado State University, Fort Collins, CO 80523, USA; peter.hellyer@colostate.edu 2 Department of Sociology, University of Calgary, 2500 University Drive NW, Calgary, AB T2N 1N4, Canada; jwallace@ucalgary.ca 3 Emeritus, Faculty of Nursing, University of Calgary, 2500 University Drive NW, Calgary, AB T2N 1N4, Canada; ecarr@ucalgary.ca * Correspondence: lori.kogan@colostate.edu 21 4 2022 5 2022 12 9 107425 3 2022 17 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Simple Summary People with aging or ill family members often fill the role of caretaker. Companion dogs are often viewed as family members and because they age more rapidly than people and have shorter lifespans, having a dog often includes caring for it during its senior years. Caring for an elderly dog can be physically and emotionally challenging, yet we know little about how caring for an aging dog impacts guardians’ lives. This study was designed to better understand dog guardians’ experiences and perceptions related to caring for their aging dog. We asked dog guardians to complete an online anonymous survey, resulting in a sample size of 284 participants. We found that the impact on guardians when caring for an aging dog appear to share many similarities with caregivers of human family members. Our results suggest that, for many guardians, caring for an aging dog is a complex dynamic with both positive and negative factors that offers an opportunity to deepen the human-animal bond and create positive, rewarding experiences and memories. Abstract Companion dogs are increasingly popular, 38.4% of households in the United States include at least one dog. There are numerous benefits to sharing one’s home with a dog, but because they age more rapidly than people and have shorter lifespans, acquiring a dog often includes caring for it during its senior years. Caring for an elderly dog can be physically and emotionally challenging, yet the impact on guardians’ lives when caring for an aging dog has received minimal scientific attention. This study was designed to better understand dog guardians’ experiences and perceptions related to caring for their aging dog. Utilizing an exploratory mixed methods design, this study asked dog guardians to complete an online anonymous survey. From a total of 284 participants, we found that the impact on guardians when caring for an aging dog appears to share many similarities with caregivers of human family members. Our quantitative and qualitative results suggest that, for many guardians, caring for an aging dog is a complex dynamic with both positive and negative factors that offers an opportunity to deepen the human-animal bond and create positive, rewarding experiences and memories. canine aging senior caregiver human animal bond ==== Body pmc1. Introduction An aging population, as well as an increased number of people living with chronic disease, has increased the number of family members acting as caregivers [1]. Similarly, with advances in veterinary medicine, caring for aging pets is also becoming a reality for many families. This paper focuses on caring for an aging dog, using the rich collection of research pertaining to human caretaking as an important theoretical foundation. 1.1. Caregivers and Family Members People with aging or ill family members often fill the role of caretaker; providing unpaid, ongoing assistance with activities of daily living [2]. Yet, these family members are often unprepared for this challenging new role, many times creating a wide range of unmet needs for themselves and those they care for [3]. In addition to the physical and emotional demands associated with caregiving, many caretakers find the people they provide care for, often times the very same people they have received emotional support from in the past, are no longer able to offer support [4]. Because of this, informal family caregiving is often described as a chronically stressful experience [5] with the resultant caregiver burden defined as the “multidimensional biopsychosocial reaction resulting from an imbalance of care demands relative to caregivers’ personal time, social roles, physical and emotional states, financial resources, and formal care resources given the other multiple roles they fulfill” [6]. Several studies have found that caregiving, a role that often spans many years, can affect caretakers’ physical and emotional health as well as their quality of life [7,8,9,10]. This includes increased levels of anxiety and depression, which can negatively impact both the caretaker and the recipient [7,11,12,13,14,15,16,17,18]. It is, therefore, not surprising that Schulz [19] suggests that caring for an elderly individual with disabilities is burdensome and stressful and that family caregivers perform this service at considerable cost to themselves. Yet, these negative consequences do not appear to tell the full story. Other research suggests that the effects and impact of caretaking are not always negative for the caregiver. It has been suggested that the subjective perceptions of the caregiver, unique for each person, play a critical role in their perceived burden, anxiety, and depression [13,20,21,22]. While it appears that when caregiving demands exceed psychological or social resources to cope, the result can prove detrimental to the caretaker’s emotional and physical health; yet, it has been suggested that this stress process model should also include the healthy caregiver hypothesis [23]. This model contends that healthier people are more likely to become caregivers and to remain in caregiving roles over time, and may actually experience health benefits from the prosocial behaviors that accompany this type of role [24]. Therefore, perhaps theoretical models on caregiver burden that focus exclusively on the negative impact of caregiving and suggest that caregivers, as a whole, are more stressed than non-caregivers, are missing critical pieces of the picture. Several studies, in fact, have found that many family caregivers report little, or no, strain associated with providing caregiving assistance. Schulz and Beach, for example, found that 44% of the spouse caregivers reported “no strain” in association with caregiving tasks [19], and similarly, Roth [25] found that 33% of caregivers reported “no strain” and only 17% reported “a lot of strain.” Furthermore, a recent survey by the National Opinion Research Center [26] found that 83% of caregivers viewed it as being a positive experience. For example, caregivers may appreciate the positive experiences that come from sharing the limited time remaining with their family member [27,28]. In fact, the positive experiences of caregiving could potentially buffer against some of the possible stress-related health consequences. One reason for these mixed results may have to do with caretakers’ level of life satisfaction prior to their new role. While some studies [1,29] have suggested that life satisfaction is influenced by the perceived burden of caregiving (meaning caregivers with a low degree of burden experience higher satisfaction), other studies [30,31] suggest that perceived burden of caregiving is instead influenced by low life satisfaction [32]. Furthermore, as noted by Lazarus [33], stress is subjective and can be defined as a “particular relationship between the person and the environment that is appraised by the person as taxing or exceeding his or her resources and endangering well-being”. Using this definition, people may not view caretaking as stressful if they are confident they have sufficient resources and feel the role is within their scope of knowledge [2]. Indeed, it would appear that many caregivers experience both rewards and strains simultaneously [34,35,36]. They may experience both emotional distress and psychological satisfaction and growth, effects that are not on opposite ends of the same continuum. Despite this growing amount of research pertaining to caregiving for human family members, little attention has been placed on caregiving for companion animals. 1.2. Caregivers and Companion Animals Companion animals are increasingly popular, with 70% of U.S. households, or about 90.5 million families, owning a pet, and 38.4% owning at least one dog [36]. Typically seen as more than “just a dog”, 85% of owners consider their dog to be a member of the family [37]. Numerous studies have demonstrated the benefits of pets [38,39,40]. Pet owners, compared with non-owners, are more physically fit [41,42,43,44], have lower levels of depression [45], higher social functioning [46] and enhanced social support [44,45,47,48,49]. Yet, because dogs age more rapidly than people and have a shorter lifespan, acquiring a dog often includes ultimately caring for a senior dog (the exact definition of senior varies, dependent on several factors including breed and size) [50]. A dog’s life can be divided into four stages, including pediatric, adult, senior (mature, middle age), and traditional geriatric (senior/super senior). The senior/middle-age years typically include the transition period between healthy adult years and the traditional geriatric stage in which serious age-related issues are more common [51,52]. As a result of recent advances in veterinary medicine, including senior diets, improved dental care, pain management plans, and new drugs, many animals are living longer [53]. A longer life span, however, means that more dogs are faced with chronic ailments like arthritis, cardiovascular problems, cognitive dysfunction, and sensory impairment [54]. Caring for an elderly dog can be physically and emotionally challenging, yet the impact on owners’ lives when caring for an aging dog has received minimal scientific attention [55]. The handful of studies that have been conducted on caring for an aging or sick dog suggest that many aspects mirror those witnessed in human caretaking. For example, some owners talk about necessary changes in their daily routines, including feeling an obligation to leave their dog alone as little as possible [56]. Others report feeling depressed, guilty, and sorry for their pet [57]. Yet other owners indicate that caring for their aging dog does not negatively impact their own quality of life [58,59], with some owners noting they appreciate the extra time they are able to enjoy with their pet [60]. One qualitative study that investigated the impact of caring for an aging dog noted that most owners referred to their dog as a child or family member and, similar to the commitment they would feel in caring for a human family members, felt a strong obligation to caring for their dog [55]. Many participants in this study felt that not providing care because it was inconvenient or challenging was simply not an option [55]. Furthermore, many owners found caring for their aging dog rewarding because of the ability to spend time together, and also because they were able to live up to their own perceptions of a good dog guardian [55]. These studies, in both human and companion animal literature, suggest that caring for an aging (senior) dog may offer rewards as well as challenges. This current study was designed to better understand dog guardians’ experiences and perceptions related to caring for their aging dog. 2. Materials and Methods 2.1. Survey Development This study used an exploratory mixed method design that first started with an exploratory online survey followed by a concurrent nested mixed methods design (where the quantitative and qualitative data are collected at the same time but one data set is dominant) [61]. In this study the quantitative survey data was the dominant data with the qualitative data adding further understanding and breadth [62]. A preliminary qualitative survey was given via social media and word of mouth to help inform the measures used in the current study. The qualitative survey yielded 25 responses to three open-ended questions. The questions asked respondents to describe what aging means in terms of living with an aging dog as well as ways in which their dog’s aging has had a positive or negative impact on them and/or their dog in terms of their relationship, activities and/or experiences. Responses to these questions were used to ensure the quantitative survey included the aspects most often identified as related to owning an aging dog. A thematic analysis was carried out for each question to identify the core themes that were then used to inform survey development. Thematic analysis is a qualitative research method that is used to describe and analyze consistent and competitive themes in human experience [63]. The transcripts were read to develop a general sense of the themes that arose from the data. All transcripts were reviewed and analyzed by all four authors to reduce bias and increase confirmability of the results by including multiple researcher perspectives. Any discrepancies in coding were discussed between the authors to ensure agreement and to refine the themes. Similar codes were initially clustered and then consolidated into larger themes to gain an understanding of the data. When asked what aging means to them regarding their aging dog, half of the respondents’ comments related to physical changes (e.g., reduced mobility). Yet, for many, aging was viewed in a positive light where “old is gold”. Examples include those who mentioned their dog is less anxious than when younger, or that they cherish the beautiful memories they are making with their dog. One frequently mentioned positive sentiment was a strengthening of the human animal bond and feeling that their dog is a member of the family. In addition, guardians also reported enjoying shared activities together and the companionship of their older dog. The most common comments about the negative effects of aging related to the general slowing down or decreased physical abilities of their dog. Other negative changes included increased vocalization, physical changes, and their own anxiety about their dog’s vulnerability and aging process. An anonymous, online Qualtrics (Qualtrics, Inc., Provo, UT, USA) survey was designed, reviewed, and tested by the co-investigators and distributed between August and September 2021. The themes from the qualitative survey were used to identify relevant scales in the literature (i.e., The Lexington Attachment to Pets Scale (LAPS) and Pre-Death Inventory of Complicated Grief-Caregiver Version (Pre-ICG)) and to develop scale items to capture both the negative and positive experiences of caregiving and perceived caretaker support. These scales are described in greater detail below. 2.2. Participants The resultant survey was piloted by a small group of individuals for ambiguity and potentially missing response options with applicable revisions made based on their feedback. The final survey (see Supplementary Materials) and study design were approved by the Colorado State University Institutional Review Board (IRB #2554, 30 July 2021). Survey respondents were recruited through Amazon’s Mechanical Turk (MTurk; Amazon Inc., Seattle, WA, USA) platform, an open online marketplace providing affordable access to potential survey respondents. Diversity of participants recruited through MTurk is higher than typical Internet samples or American college-based samples. The data from MTurk has been found to meet acceptable psychometric standards [64]. To control for in-attentive participants, bots, virtual private networks (VPNs), and multiple submissions [65], we included two attention questions, and utilized the resources that can be embedded into Qualtrics surveys. Adult (18 years or older) participants who were the current guardians of at least one aging dog and had owned the dog for at least 3 years were recruited for the study. In order to minimize the influence of geographic and cultural differences on respondent data, the survey was made available only to guardians living in the United States. 2.3. Procedures In addition to guardian demographics (e.g., age, gender, ethnicity, education level, employment status and workplace (at home, away from home, etc.) and living arrangement (live alone, with other adults and/or children), respondents were asked how long they had lived with their dog. Participants were next asked to complete The Lexington Attachment to Pets Scale (LAPS) [66]. The LAPS is a widely used instrument to measure attachment of people to their pets and contains 23 items, scored on a Likert scale from 0 (strongly disagree) to 3 (strongly agree) with a possible range between 0 and 69. Next, participants were given a series of questions about possible lifestyle changes due to their aging dog. They were then asked to indicate behavioral and physical changes in their dog due to aging, as well as a series of statements that might reflect their own emotional and behavioral responses regarding their dog’s aging. The next series of questions pertained to their perception of the support they have received in caring for their dog. Participants were also given an adaptation of the Pre-Death Inventory of Complicated Grief-Caregiver Version (Pre-ICG). The pre-death version of the Inventory of Complicated Grief (ICG) assesses grief over the expected loss of a loved one. The pre-death ICG has demonstrated high levels of internal consistency among caregivers (Cronbach’s alpha = 0.90) [67,68]. The version used in the current study consisted of four questions used by Tomarken (2008) and reported to have adequate reliability (0.76). The questions were answered using a 5-point Likert scale with 1 = Never to 5 = Always. The quantitative data were analyzed using SPSS (IBM, Armonk, NY, USA). Descriptive statistics were calculated to characterize guardians and household demographics. New scales were created to reflect the following aspects of living with an aging dog: negative caretaking aspects, positive aspects, worry and anxiety, and social support. We performed two multiple linear regression analyses: one to determine predictors of positive caretaking aspects of living with an aging dog and one to predict negative caretaking aspects. Results of exploratory univariate analysis of variances were used to guide the selection of predictors for both multiple regression models. Significance level (α) was p = 0.05 and all tests were two-tailed. Finally, participants were asked to share one story or example of something that stands out to them about living with an aging dog—either positive or negative. The rationale for including qualitative analysis was to help interpret and illustrate the results provided by the quantitative data. The stories and examples are used to help extend the quantitative results to help elucidate the paradoxical effect of caregiving—the potential to simultaneously be both a positive and negative experience for dog guardians. From the 285 completed surveys, we received 216 narratives about living with an aging dog that varied in length and details. First, the authors independently read through the quotes and selected those that they felt best captured the positive and negative experiences of caring for an aging dog. This reduced the 216 accounts to 53. Together, the authors reviewed the 53 comments and discussed their content and meaning. The authors coded the content of these selected statements to identify the specific nature of the caretakers’ experiences. Specific quotes and examples that offered the best representation of the positive and negative experiences were then selected. 3. Results A total of 284 participants completed the survey. The sample consisted of 140 (49.3%) female, 138 (48.6%) male (n = 129), and 6 (2.1%) nonbinary/other participants. The sample was 75.7% White, 8.5% African American/Black, 7.0% LatinX/Hispanic, 4.2% Asian, 2.1% Multi racial/multiethnic, 0.7% American Indian/Native Alaskan, 0.4% Middle eastern/north African (MENA), and 1.4% other or prefer to not say. The age of participants ranged from 18–29 years of age (133, 46.8%), 30–39 years (82, 28.9%), 40–49 years (43, 15.1%), and 50 and older (26, 9.2%). When asked about employment status, the majority were employed full time (203, 71.5%), followed by unemployed (30, 10.6%), employed part time (27, 9.5%), retired (5, 1.8%), furloughed (3, 1.1%), other (12, 4.2%), and prefer to not say (4, 1.4%). For those who reported working (n = 230), 97 (42.2%) reported working at home and away from home, 65 (28.3%) reported working mostly/all the time at home, and 63 (27.4%) reported working mostly/all the time away from home. Five (2.2%) people indicated they preferred to not say. It should be noted that the survey was completed during COVID-19, such that more people than usual may be not working or working from home. Most participants lived with other adults (116, 40.8%) or other adults and children under the age of 18 (78, 27.5%). Fewer participants reported living only with children (48, 16.9%) or alone (31, 10.9%). The majority of participants reported having a university degree (146, 51.4%) or some college (67, 23.6%), with fewer reporting having a higher degree (52, 18.3%), high school/GED (18, 6.3%) or prefer to not say (1, 0.4%). When asked how long they had lived with their current dog, 122 (43.0%) reported 3–5 years, 71 (25.0%) reported 5–7 years, and 91 (32.0%) reported more than 7 years. 3.1. The Lexington Attachment to Pets Scale (LAPS) Possible scores for the LAPS were between 0–69. In this study, the mean was 55.0 (SD 9.37), with a minimum of 20 and a maximum of 69. Cronbach’s alpha was 0.84. 3.2. Pre-Death Inventory of Complicated Grief-Caregiver Version (Pre-ICG) The mean score for the four questions of the Pre-Death Inventory of Complicated Grief-Caregiver Version (Pre-ICG) was 2.95 (SD 1.14), with a Cronbach’s alpha of 0.88. 3.3. Negative Aspects of Caretaking Scale A new scale was created to depict the negative aspects of caretaking for an aging dog. The items in the new scale included four items created for this survey and the four Pre-ICG questions (Table 1). The Cronbach’s alpha for this scale was 0.86. 3.4. Worry and Anxiety Scale Seven survey items were combined to create the Worry and Anxiety Scale, with a resultant Cronbach’s alpha of 0.69 (Table 2). 3.5. Positive Aspects of Caretaking Scale The scale for Positive Aspects of Caretaking was created by summing seven survey items, with a Cronbach’s alpha of 0.72 (Table 3). 3.6. Caretaker Support Scale The Caretaker Support Scale was created by summing six items, with a resultant Cronbach’s alpha of 0.72 (Table 4). 3.7. Quantitative Results: Negative Aspects of Caretaking Scale A Univariate Analysis of Variance test was performed to explore the relationship between negative aspects of caretaking (measured with the Negative Aspects of Caretaking Scale) and dog caretaker characteristics (gender, age, workplace) and the LAPS, Worry and Anxiety Scale, Positive Aspects of Caretaking Scale, and Caretaker Support Scale. The factors that were significantly associated with the Negative Aspects of Caretaking Scale included workplace, LAPS, Worry and Anxiety Scale, Positive Aspects of Caretaking Scale, and Caretaker Support Scale (Table 5). Based on these results, multiple linear regression was conducted using the significant factors to determine their relationship with negative aspects of caretaking for aging dogs. The multiple regression model (Table 6) predicting the Negative Aspects of Caretaking Scale using LAPS, Worry and Anxiety Scale, Positive Aspects of Caretaking Scale, and Caretaker Support Scale, and workplace was significant (F5 = 35.18, p < 0.001), with an R2 of 0.445. Significant predictors of the Negative Aspects of Caretaking Scale included LAPS (B = −0.185; p < 0.001), Worry and Anxiety Scale (B = 0.688; p < 0.001), Positive Aspects of Caretaking Scale (B = 0.743, p < 0.001), and Caretaker Support Scale (B= −0.426; p = 0.011). The largest predictors of negative aspects of caretaking were the Worry and Anxiety Scale and the Positive Aspects of Caretaking Scale. 3.8. Quantitative Results: Positive Aspects of Caretaking Scale A Univariate Analysis of Variance test was also performed to explore the relationship between positive aspects of caretaking (measured with the Positive Aspects of Caretaking Scale) and guardian characteristics (gender, age, workplace), and the LAPS, Worry and Anxiety Scale, Negative Aspects of Caretaking Scale, and Caretaker Support Scale. The factors that were significantly associated with the Positive Aspects of Caretaking Scale included LAPS, Negative Aspects of Caretaking Scale, and Caretaker Support Scale (Table 7). Multiple linear regression was conducted using the significant factors from the Univariate Analysis of Variance test to determine their impact on positive aspects of caretaking. The multiple regression model (Table 8) predicting the Positive Aspects of Caretaking Scale using LAPS, Negative Aspects of Caretaking Scale, and Caretaker Support Scale was significant (F3 = 65.11, p < 0.001), with an R2 of 0.411. Significant predictors of positive caretaking included LAPS (B = 0.203; p < 0.001), Caretaker Support Scale (B = 0.207; p = 0.016), and Negative Aspects of Caretaking Scale (B = 0.279, p < 0.001). The largest predictor of the Positive Aspects of Caretaking Scale was the Negative Aspects of Caretaking Scale. 3.9. Qualitative Results: Compasionate Care as a Double-Edged Sword One of our participants wrote: “Living with an aging dog makes me think about my companionship with animals much more, it also makes me think about how I am going to deal with loss and grief when they eventually pass away, and it makes me sad.” Caring for others, whether human or animal, can simultaneously be satisfying and fulfilling, as well as upsetting and overwhelming (Beach et al., 2000, Brønden et al., 2003, Christiansen et al., 2013, Harmell et al., 2011, Lawton et al., 1991). This premise is supported by our regression results, whereby the positive and negative aspects of caretaking are positively related to one another (Table 6 and Table 8). The qualitative stories and examples provided by our study participants are helpful in clarifying the complex nature of caring for an aging dog. First, we share some of the reported negative experiences of living with an aging dog including both physical and mental challenges. We then report on some of the positive changes (e.g., calmer, quieter, a need for less physical exercise, etc.) associated with an aging dog. Finally, we discuss how the bond between aging dogs and their caretakers creates a unique bittersweet moment in time of deepening affection and appreciation, coupled with anticipatory grief and sadness. As expected, participants described many different challenges of caring for an aging dog, as well as the anticipated painfulness of their dog’s death. As one participant commented: “As my dog ages, I’ve noticed more and more little health issues start to crop up and it makes me become painfully aware that the number of days I have left with my dog is dwindling. My dog has been with me through most of my major life milestones so far, and seeing his health deteriorate has taken a bigger mental toll on me than I’ve ever expected. My dog takes a long time to get out of bed and I worry every day that one day he will not wake up.” In addition, participants identified the physical changes in their aging companion such as slowing down because of arthritis, and stiffness or pain resulting in less energy and stamina for going up stairs, playing, and going for walks. For example: “I first really noticed that my dog was aging when I was going to take him on a walk and went to grab the leash. He usually runs up behind me, extremely excited to go for a walk. However, I looked back and realized he was having trouble standing up. It was a very sad moment for me. His legs began rapidly developing arthritis, and it became more difficult to get around.” Another noted: “Sometimes, when we go out for a walk, she gets tired a lot faster than she used to, and I have to pick her up and carry her home. It breaks my heart a little bit that she has some trouble moving on her own now, but that’s a part of life.” Other common physical challenges mentioned include managing more health concerns, changes in the dog’s appetite or diet, and the need to go outside more often or having more accidents in the house. Participants also described the challenges associated with their dog losing their hearing or sight, which was often linked to their dog becoming more anxious or easily confused or frightened. For example, “As he’s grown older, my dog has become increasingly deaf. If I’m not careful, I can walk up behind him and startle him. He jumps and seems genuinely frightened until he sees who it is. It’s a reminder that I need to adapt to take care of his needs.” Additionally, participants described emotional changes including greater separation anxiety or confusion. One person noted “Something that has really affected us as our senior dog ages is his episodes of canine cognitive dysfunction at night. He often wakes up extremely restless, anxious, and confused and not only does it affect my sleep, it is very upsetting to see and not be able to help much.” In addition to the negative aspects of living with an aging dog, participants also described some of the more positive changes. As one person noted: “Aging dogs are always like an elder family member so matured and calm, always understanding our emotions and feelings. We are so blessed to have such caring dogs around us.” Another commented: “She has become much more gentle with her interaction with people. She doesn’t jump up as much when greeting new visitors, and gives a soft paw when prompted. She has become much more affectionate.” Older dogs were often described as calmer, as well as more mature, relaxed, cuddly/affectionate, tolerant of strangers or other dogs, and attentive to their guardian’s emotions. One participant noted: “My dog is more mature, shows more affection and likes to relax.” Another stated: “As much as I get sad thinking about the day she is no longer with me, I love the new sense of calmness that has come with her age.” Participants also described how their older dog gets along better with other animals, dogs and cats: “He’s just perfect and sweet! He’s a lot more relaxed around the other animals, even when they play a little too hard. He’s an angel.” The participants’ comments also illustrate how the strength of the human-animal bond is vital to their willingness and patience in providing compassionate care to their dog, where their compassion reflects an awareness of the suffering of another and the desire to ease it. A strong human-animal bond, however, also makes observing the pain and suffering of their aging dog, and their inability to alleviate their suffering and the inevitability of their death, an upsetting experience. For example, one participant’s comments illustrate the conflicting feelings of joy and sadness that are experienced simultaneously: “I would say one thing that stands out is that you really understand how much of a gift to the dog in your life is, and that the sadness over their aging is easily trumped by the joy the dog brings you.” Many participants described how their bond with their dog strengthens and they feel more emotional attachment and affection as their dog ages: “Something that stands out to me the most are our times spent together. They’re more affectionate and seem to be more meaningful, like I know that moments like these are becoming limited, so each one becomes more and more special. My dog also isn’t the most cuddly, so for her to come lay by my side, those are the little moments that I love.” Another interesting theme from participants’ comments is how they explicitly refer to the fact that they do not resent their aging dog’s need for additional care. They often acknowledge the reciprocal relationship of the bond their share with their dog in where their younger dog was faithful, “there for them” and now it’s their turn to be there for their dog. For example, “It’s bittersweet and sad, yet I will gladly take care of my aging dog because he gave me the best years of his life and I want to repay him by being there for him when he needs me, like he was always there for me when I needed him.” Another suggests how the bond with their aging dog hasn’t diminished but strengthened as their dog requires more care: “I don’t consider myself an exceptionally patient person, but with him I do what I need to and feel no resentment. His old age has made me appreciate him so much more, and I rarely feel upset being a witness to it. I’m happy to have grown up with him and watch him grow old.” Participants also described significant changes and sacrifices that they have made to their lifestyle in order to provide extra care for their dog. Some examples include shorter or fewer walks, carrying their dog up the stairs or into the car, purchasing a wagon for the dog when it’s too tired to walk any further, getting up more often at night to let the dog outside, forgiving accidents in the house, letting the dog sleep on the bed, and cooking special meals that are easier to digest. Several participants mentioned that working from home allows them to offer better care for their dog: “I never really worried about my dog aging until he started having stomach issues and occasional incontinence. This one thing changed my lifestyle quite a bit—I started working from home and staying home as much as possible just in case. But I don’t resent him for it at all; I feel blessed that I have the kind of job where I can be here for him when he needs me.” Another wrote “I heated my pool so I could swim with my 14-year-old choc lab. I worked from home, so we swam daily.” These lifestyle changes and sacrifices reflect the genuine compassion and unwavering sense of duty that many animal guardians feel for caring for their dog as it ages. For dog guardians, caring for their aging dog, despite the inevitable outcome, can be both fulfilling and rewarding. 4. Discussion Sharing one’s home with a companion dog offers a multitude of both physical and psychological benefits [38,39,40,43,46,49], but due to a dog’s relatively short lifespan, also typically includes aging and ultimately, loss. The impact on guardians when caring for an aging dog appears to share many similarities with caregivers of human family members. These changes often include practical, pragmatic ones such as altering one’s lifestyle or daily routines to ensure they are able to care for their loved one [1,56,69]. As noted by Christiansen et al. [55], many dog guardians deal with these changes by accepting the fact that caring for their dog includes changes that can be time consuming, burdensome, and inconvenient. In addition to logistical changes, similarities between human and canine caregiving can be seen in the emotional impact of caretaking. Decades of research pertaining to human caregiving has found that many caregivers struggle with stress, anxiety and depression [3,18,70], often mitigated by numerous factors, including mutuality (the positive qualities of the relationship), perceived support and available resources [1,30,71]. Many pet guardians also report feeling depressed and burdened [55,57,72,73]. Yet, others report more positive feelings associated with a severely ill companion animal [54,60]. Our study found that the impact of caring for an aging dog is in fact a complex interwoven myriad of feelings that often include both positive and negative emotions and experiences. For example, we found that guardians who reported more negative thoughts and ruminations about their dog’s aging were more likely to feel worried and anxious. The stories shared by participants highlight this quantitative finding and illustrate how painful it can be for some guardians to witness their dog slowing down and facing increasing physical and mental health-related challenges. Additionally, many guardians struggle with feelings of anticipatory loss and grief, defined as the fear of losing a significant other [74]. This anticipatory loss, coupled with worry and anxiety, are similar to that reported by caretakers of human family members [13]. A unique factor for aging pets that can add to the stress associated with the aging process is the option for euthanasia. The decision to euthanatize, including an ongoing assessment of their dog’s quality of life, adds an additional element of stress and worry for many dog guardians. Euthanasia decisions involve complex issues, none of which are typically black or white. Trying to discern their dog’s quality of life, assessing the impact of medical interventions, and trying to assess what is in the dog’s best interest, can be overwhelming and exhausting. Yet, it is important to note that not all caregivers experience worry, anxiety and depression. Instead, it appears that these emotions are strongly correlated with subjective caregiver burden in caring for both humans and companion animals [13,20,21,22,75]. This perception of burden is related to numerous factors including whether the demands exceed the guardian’s psychological or social resources to cope, feelings of mutuality, and their own mental and physical health [19,25,71,72,76]. So, while for some, the role of family caregiver is overwhelmingly burdensome and has a negative impact on emotional and physical health [5], these responses are not universal and do not preclude more positive feelings. Another factor that impacts feelings associated with caregiving is the bond guardians share with their dog. Our results suggest that having a stronger emotional attachment to one’s aging dog reduces caretakers’ negative feelings (e.g., feelings of guilt, resentment, longing for when their dog was younger). We learned from our participants’ stories that many do not resent their aging dog’s additional care and they gave a variety of examples of the different sacrifices they willingly make to provide extra care. Both the quantitative and qualitative results also highlight the positive experiences and emotions associated with living with an aging dog (e.g., sense of purpose, caring and cherishing their dog). Not surprisingly, stronger emotional attachment and feeling support from others both predict a more positive experience for animal caregivers. Many participants described how they feel closer to their aging dog and find them to be more affectionate, cuddly, quieter, and mature as they age. Similar positive sentiments regarding the benefits of a caretaking role have been reported by human caretakers [26,27,28]. Furthermore, not only can caregivers experience both positive and negative emotional reactions to their caregiving role, our results suggest that the two are correlated. That is, the more animal caretakers feel a sense of purpose and cherish their time with their aging dog, the more guilt and resentment they appear to feel in caring for their dog and the more they long for the time when their dog was younger. In addition, the quantitative results also revealed that a stronger emotional attachment to their dog predicted increased negative and positive emotions related to living with an aging dog. We learned from the qualitative findings the impact of the human-animal bond in shaping caretakers’ bittersweet experiences of watching their long-time companion slowing down; feelings of sadness, while also cherishing the time they spend caring for them. There are several limitations to this study that should be noted. First, the results are based on a small sample of self-selected individuals answering a survey through Amazon Turk. Although our results contribute to our understanding of the complex nature of care giving to aging dogs, caution should be taken when generalizing to the entire population of dog guardians in the U.S. Additionally, the survey was conducted during the COVID-19 pandemic, and we do not know if the pandemic influenced the nature of the relationship between guardians and their aging dogs. Furthermore, the survey did not explore the degree of age-related changes in the participants’ dogs, only that the participants recognized changes in their dog’s behavior that they attributed to age. As such, we do not know the role that certain common disease states, such as osteoarthritis, cancer, cognitive dysfunction, and diabetes had in guardians’ perceptions of their dogs’ aging process. Finally, we did not collect any medical data on the dogs to determine overall health status and how that may have affected the relationship with their guardian. Future research exploring the effect of both the guardian’s and dog’s health and resources available to care for the dog on the relationship between guardians and their aging dogs would be of value. Additionally, expanding this line of inquiry to the guardians of other types of companion animals can help to determine if the relationship found in this study is unique to humans and dogs, or if it can be applied to other types of companion animals too. 5. Conclusions In most cases people will outlive their dogs, resulting in the need to care for an aging dog. The fact that a deep bond with an aging dog can increase both positive and negative feelings is vitally important in understanding the caregiving role. Our findings suggest that these feelings are not opposites on the same continuum, but instead, they correlate, whereby increased feelings of satisfaction, a sense of purpose and moments of contentment and happiness are often accompanied by increased worry and concern, often within the context of anticipatory grief. This knowledge can be used to help support companion dog caretakers. For example, helping guardians identify the positive aspects of their new role may help them focus on the benefits that can accompany caring for an aging dog. As noted by many of the participants, these benefits include positive feelings associated with being able to give back to, and provide for, their steadfast companion. The provision of support for caregiving guardians should also include addressing potential anticipatory grief and helping them identify strategies that can help in mitigating their anxiety. This might entail exploring ways to implement changes in their daily routines to enhance their dog’s quality of life (e.g., special treats, allowed on the furniture, etc.) or permit more time together. It may also involve the process of initiating thought and discussion about how they may want to create a lasting bond with their dog when he/she is no longer physically present. Creating videos or photo albums or other ways to memorialize their dog may be of value. Above all, it is important to recognize that the experience of caring for an aging dog is individualized and it should not be assumed that the caregiving role is filled only with hardship and pain. Instead, it would appear that for many guardians, caring for an aging dog is a complex dynamic with both positive and negative factors that offers an opportunity to deepen their bond and create positive, rewarding experiences and memories. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ani12091074/s1, S1: Canine Caregivers survey. Click here for additional data file. Author Contributions Conceptualization, L.R.K., J.E.W., E.C.J.C., P.W.H.; methodology, L.R.K., J.E.W., E.C.J.C., P.W.H.; formal analysis, L.R.K., J.E.W., E.C.J.C., P.W.H.; writing—original draft preparation, L.R.K., J.E.W., E.C.J.C., P.W.H.; writing—review and editing, L.R.K., J.E.W., E.C.J.C., P.W.H. All authors have read and agreed to the published version of the manuscript. Funding This research received no external funding. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Review Board (or Ethics Committee) of Colorado State University (IRB #2554, 30 July 2021). Informed Consent Statement Participant consent was waived due to the fact that the research involved no more than minimal risk to the subjects and could not practicably be carried out without the requested waiver or alteration. Data Availability Statement Data available upon request from corresponding author. Conflicts of Interest The authors declare no conflict of interest. animals-12-01074-t001_Table 1 Table 1 Survey items that constitute the Negative Aspects of Caretaking Scale. I would like to be able to walk/run further with my dog than he/she can now walk I feel guilty when I exercise or go for a walk and can no longer take my dog with me There are times I resent the changes I have had to make in my daily schedule to care for my dog I dread leaving my dog for any period because of his/her age * I feel myself longing and yearning for my dog as he/she was before aging * I feel that life is empty and meaningless without my dog being healthy * I am bitter over my dog’s aging * I think about my dog’s aging so much that it can be hard for me to concentrate on anything else or do the things I normally do * Pre-ICG question. animals-12-01074-t002_Table 2 Table 2 Survey items that constitute the Worry/Anxiety Scale. I worry how the loss of my aging dog will affect me and my family I worry that the number of remaining days with my dog are limited I worry a great deal about when my dog can no longer get around by him/herself I am worried other dogs will accidently hurt my aging dog I worry about my ability to afford veterinary care for my aging dog My dog gives my life purpose, and I am worried about what I will do without him/her I dread the day my dog is no longer with me animals-12-01074-t003_Table 3 Table 3 Survey items that constitute the Positive Aspects of Caretaking Scale. Caring for my aging dog gives me a sense of purpose I tend to bend my dog-related rules more as my dog ages (i.e., I let my dog sleep on the couch or bed, I give treats more often) I find I am more protective of my dog as he/she ages I cherish the time I spend with my aging dog The amount of time I spend with my dog His/her ability to understand your feelings and know what you are thinking How affectionate he/she is animals-12-01074-t004_Table 4 Table 4 Survey items that constitute the Caretaker Support Scale. I talk with friends or my family about my concerns related to my aging dog I have talked to my vet about my concerns related to my aging dog I feel my vet and I are a team when it comes to caring for my aging dog I wish I had someone to talk to about my aging dog * How much I socialize I feel my family and/or friends do not understand what is needed to care for an aging dog * * reverse coded. animals-12-01074-t005_Table 5 Table 5 Univariate Analysis of Variance test results assessing the association between the Negative Aspects of Caretaking Scale and LAPS, Worry and Anxiety Scale, Positive Aspects of Caretaking Scale, Caretaker Support Scale and workplace. ANOVA Model Sum of Squares df Mean Squares F Sig. Worry/anxiety 1428.88 1 1428.88 47.34 <0.001 Positive aspects 1566.03 1 1566.03 51.88 <0.001 Caregiver Support 160.96 1 160.96 5.33 =0.022 LAPS 351.89 1 351.89 11.66 =0.001 Gender 7.57 1 7.57 0.25 =0.617 Age 88.41 3 29.47 0.98 =0.405 Workplace 538.61 2 269.30 8.92 <0.001 Bold denotes significance. animals-12-01074-t006_Table 6 Table 6 Results of the multiple linear regression model predicting the Negative Aspects of Caretaking Scale. ANOVA Model Sum of Squares df Mean Squares F Sig. Regression Residual Total 5720.97 7123.19 12844.16 5 219 224 1144.19 32.53 35.18 <0.001 Coefficients Dependent Variable: the Negative Aspects of Caretaking Scale) Variable Coefficient (B) Std. Error t Sig. (Constant) 1.25 4.14 0.30 =0.763 LAPS −0.19 0.05 −3.96 <0.001 Worry/anxiety 0.69 0.10 6.68 <0.001 Positive aspects 0.74 0.10 7.62 <0.001 Caregiver Support −0.43 0.17 −2.58 =0.011 Workplace −0.94 0.52 −1.79 =0.074 Bold denotes significance. animals-12-01074-t007_Table 7 Table 7 Univariate Analysis of Variance test results assessing the association between Positive Aspects of Caretaking Scale and LAPS, Negative Aspects of Caretaking Scale, and Caretaking Support Scale. ANOVA Model Sum of Squares df Mean Squares F Sig. Worry/anxiety 58.75 1 24.30 1.96 =0.163 Caregiver Support 372.39 1 58.75 4.74 =0.031 LAPS 643.22 1 372.39 30.04 <0.001 Negative aspects 1.40 1 643.22 51.88 <0.001 Gender 28.04 1 1.40 0.11 =0.737 Age 4.25 3 9.35 0.75 =0.521 Workplace 58.75 2 2.13 0.17 =0.843 Bold denotes significance. animals-12-01074-t008_Table 8 Table 8 Results of the multiple linear regression model predicting the Positive Aspects of Caretaking Scale. ANOVA Model Sum of Squares df Mean Squares F Sig. Regression Residual Total 2553.51 3660.33 6213.84 3 280 283 851.17 13.07 65.11 <0.001 Coefficients (Dependent Variable: Positive Aspects of Caretaking Scale) Variable Coefficient (B) Std. Error t Sig. (Constant) 16.45 1.94 8.49 <0.001 LAPS 0.203 0.023 8.65 <0.001 Caregiver Support 0.207 0.086 2.41 =0.016 Negative Aspects 0.279 0.028 9.80 <0.001 Bold denotes significance. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Adelman R.D. Tmanova L.L. Delgado D. Dion S. Lachs M.S. Caregiver burden: A clinical review JAMA 2014 311 1052 1060 10.1001/jama.2014.304 24618967 2. Roth D.L. Fredman L. Haley W.E. Informal Caregiving and Its Impact on Health: A Reappraisal From Population-Based Studies Gerontologist 2015 55 309 319 10.1093/geront/gnu177 26035608 3. Applebaum A.J. Breitbart W. Care for the cancer caregiver: A systematic review Palliat. Support. Care 2013 11 231 252 10.1017/S1478951512000594 23046977 4. Francis L.E. Worthington J. Kypriotakis G. Rose J.H. Relationship quality and burden among caregivers for late-stage cancer patients Support. Care Cancer 2010 18 1429 1436 10.1007/s00520-009-0765-5 19898879 5. Schulz R. Sherwood P.R. Physical and Mental Health Effects of Family Caregiving Am. J. Nurs. 2008 108 23 27 10.1097/01.NAJ.0000336406.45248.4c 6. Given C.W. Given B. Azzouz F. Kozachik S. Stommel M. Predictors of pain and fatigue in the year following diagnosis among elderly cancer patients J. Pain Symptom Manag. 2001 21 456 466 10.1016/S0885-3924(01)00284-6 7. Geng H. Chuang D. Yang F. Yang Y. Liu W.M. Liu L.H. Tian H.M. Prevalence and determinants of depression in caregivers of cancer patients: A systematic review and meta-analysis Medicine 2018 97 e11863 10.1097/MD.0000000000011863 30278483 8. Loh A.Z. Tan J.S. Zhang M.W. Ho R.C. The Global Prevalence of Anxiety and Depressive Symptoms Among Caregivers of Stroke Survivors J. Am. Med. Dir. Assoc. 2017 18 111 116 10.1016/j.jamda.2016.08.014 27742585 9. Song J.I. Shin D.W. Choi J.Y. Kang J. Baik Y.J. Mo H. Park M.H. Choi S.E. Kwak J.H. Kim E.J. Quality of life and mental health in family caregivers of patients with terminal cancer Support. Care Cancer 2011 19 1519 1526 10.1007/s00520-010-0977-8 21479527 10. Yang X. Hao Y. George S.M. Wang L. Factors associated with health-related quality of life among Chinese caregivers of the older adults living in the community: A cross-sectional study Health Qual. Life Outcomes 2012 10 143 10.1186/1477-7525-10-143 23186128 11. Braun M. Mikulincer M. Rydall A. Walsh A. Rodin G. Hidden morbidity in cancer: Spouse caregivers J. Clin. Oncol. 2007 25 4829 4834 10.1200/JCO.2006.10.0909 17947732 12. Cliff A.M. MacDonagh R.P. Psychosocial morbidity in prostate cancer: II. A comparison of patients and partners BJU Int. 2000 86 834 839 10.1046/j.1464-410x.2000.00914.x 11069410 13. del-Pino-Casado R. Priego-Cubero E. López-Martínez C. Orgeta V. Subjective caregiver burden and anxiety in informal caregivers: A systematic review and meta-analysis PLoS ONE 2021 16 e0247143 10.1371/journal.pone.0247143 33647035 14. Gallagher J. Parle M. Cairns D. Appraisal and psychological distress six months after diagnosis of breast cancer Br. J. Health Psychol. 2002 7 365 376 10.1348/135910702760213733 12614506 15. Grunfeld E. Coyle D. Whelan T. Clinch J. Reyno L. Earle C.C. Willan A. Viola R. Coristine M. Janz T. Family caregiver burden: Results of a longitudinal study of breast cancer patients and their principal caregivers CMAJ 2004 170 1795 1801 10.1503/cmaj.1031205 15184333 16. Hauser J.M. Kramer B.J. Family caregivers in palliative care Clin. Geriatr. Med. 2004 20 671 688.vi 10.1016/j.cger.2004.07.003 15541619 17. McLean L.M. Jones J.M. Rydall A.C. Walsh A. Esplen M.J. Zimmermann C. Rodin G.M. A couples intervention for patients facing advanced cancer and their spouse caregivers: Outcomes of a pilot study Psychooncology 2008 17 1152 1156 10.1002/pon.1319 18213694 18. Rivera H.R. Depression symptoms in cancer caregivers Clin. J. Oncol. Nurs. 2009 13 195 202 10.1188/09.CJON.195-202 19349266 19. Schulz R. Beach S.R. Caregiving as a risk factor for mortality: The Caregiver Health Effects Study JAMA 1999 282 2215 2219 10.1001/jama.282.23.2215 10605972 20. García-Alberca J.M. Cruz B. Lara J.P. Garrido V. Lara A. Gris E. Anxiety and depression are associated with coping strategies in caregivers of Alzheimer’s disease patients: Results from the MÁLAGA-AD study Int. Psychogeriatr. 2012 24 1325 1334 10.1017/S1041610211002948 22340759 21. Hu P. Yang Q. Kong L. Hu L. Zeng L. Relationship between the anxiety/depression and care burden of the major caregiver of stroke patients Medicine 2018 97 e12638 10.1097/MD.0000000000012638 30290641 22. Pearlin L.I. Mullan J.T. Semple S.J. Skaff M.M. Caregiving and the Stress Process: An Overview of Concepts and Their Measures1 Gerontologist 1990 30 583 594 10.1093/geront/30.5.583 2276631 23. Fredman L. Cauley J.A. Hochberg M. Ensrud K. Doros G. Mortality associated with caregiving, general stress, and caregiving-related stress in elderly women: Results of caregiver-study of osteoporotic fractures J. Am. Geriatr. Soc. 2010 58 937 943 10.1111/j.1532-5415.2010.02808.x 20374395 24. Brown M. Brown S. Informal Caregiving: A Reappraisal of Effects on Caregivers Spec. Issues Policy Rev. 2014 8 74 102 10.1111/sipr.12002 25. Roth D.L. Perkins M. Wadley V.G. Temple E.M. Haley W.E. Family Caregiving and Emotional Strain: Associations with Quality of Life in a Large National Sample of Middle-Aged and Older Adults Qual. Life Res. 2009 18 679 688 10.1007/s11136-009-9482-2 19421895 26. Benz J. Tompson T. Willcoxon N. Long-Term Care in America: Expectations and Reality—The Long-Term Care Poll Associated Press-NORC Center for Public Affairs Research Chicago, IL, USA 2014 Available online: https://www.longtermcarepoll.org/project/long-term-care-in-america-expectations-and-reality/ (accessed on 19 December 2021) 27. Andershed B. Relatives in end-of-life care—Part 1: A systematic review of the literature the five last years, January 1999-February 2004 J. Clin. Nurs. 2006 15 1158 1169 10.1111/j.1365-2702.2006.01473.x 16911057 28. Kotkamp-Mothes N. Slawinsky D. Hindermann S. Strauss B. Coping and psychological well being in families of elderly cancer patients Crit. Rev. Oncol. Hematol. 2005 55 213 229 10.1016/j.critrevonc.2005.03.006 15886008 29. Yates M.E. Tennstedt S. Chang B.H. Contributors to and mediators of psychological well-being for informal caregivers J. Gerontol. B Psychol. Sci. Soc. Sci. 1999 54 12 22 10.1093/geronb/54B.1.P12 30. Fianco A. Sartori R.D.G. Negri L. Lorini S. Valle G. Fave A.D. The relationship between burden and well-being among caregivers of Italian people diagnosed with severe neuromotor and cognitive disorders Res. Dev. Disabil. 2015 39 43 54 10.1016/j.ridd.2015.01.006 25666898 31. Haley W.E. LaMonde L.A. Han B. Burton A.M. Schonwetter R. Predictors of depression and life satisfaction among spousal caregivers in hospice: Application of a stress process model J. Palliat. Med. 2003 6 215 224 10.1089/109662103764978461 12854938 32. Fauziana R. Sambasivam R. Vaingankar J.A. Abdin E. Ong H.L. Tan M.E. Chong S.A. Subramaniam M. Positive Caregiving Characteristics as a Mediator of Caregiving Burden and Satisfaction With Life in Caregivers of Older Adults J. Geriatr. Psychiatry Neurol. 2018 31 329 335 10.1177/0891988718802111 30260715 33. Lazarus R. Folkman S. Stress, Appraisal, and Coping Springer Publishing Company Berlin/Heidelberg, Germany 1984 34. Beach S.R. Schulz R. Yee J.L. Jackson S. Negative and positive health effects of caring for a disabled spouse: Longitudinal findings from the Caregiver Health Effects Study Psychol. Aging 2000 15 259 271 10.1037/0882-7974.15.2.259 10879581 35. Harmell A.L. Chattillion E.A. Roepke S.K. Mausbach B.T. A review of the psychobiology of dementia caregiving: A focus on resilience factors Curr. Psychiatry Rep. 2011 13 219 224 10.1007/s11920-011-0187-1 21312008 36. Lawton M.P. Moss M. Kleban M.H. Glicksman A. Rovine M. A two-factor model of caregiving appraisal and psychological well-being J. Gerontol. 1991 46 181 189 10.1093/geronj/46.4.P181 2071844 37. American Veterinary Medical Association U.S. Pet Ownership Statistics American Veterinary Medical Association Schaumburg, IL, USA 2018 Available online: https://www.avma.org/resources-tools/reports-statistics/us-pet-ownership-statistics (accessed on 20 December 2021) 38. Burns K. Pet Ownership Stable, Veterinary Care Variable American Veterinary Medical Association Schaumburg, IL, USA 2018 Available online: https://www.avma.org/javma-news/2019-01-15/pet-ownership-stable-veterinary-care-variable (accessed on 20 December 2021) 39. Chalmers D. Dell C. Dixon J. Dowling T. Hanrahan C. Recognizing Animals as an Important Part of Helping: A Survey Exploring Knowledge and Practice Among Canadian Social Workers Crit. Soc. Work 2020 21 2 29 10.22329/csw.v21i1.6224 40. Knight S. Edwards V. In the Company of Wolves: The Physical, Social, and Psychological Benefits of Dog Ownership J. Aging Health 2008 20 437 455 10.1177/0898264308315875 18448686 41. McConnell A.R. Paige Lloyd E. Humphrey B.T. We Are Family: Viewing Pets as Family Members Improves Wellbeing Anthrozoös 2019 32 459 470 10.1080/08927936.2019.1621516 42. Headey B. Health Benefits and Health Cost Savings Due to Pets: Preliminary Estimates from an Australian National Survey Soc. Indic. Res. 1999 47 233 243 10.1023/A:1006892908532 43. Headey B. Grabka M.M. Pets and Human Health in Germany and Australia: National Longitudinal Results Soc. Indic. Res. 2007 80 297 311 10.1007/s11205-005-5072-z 44. Maugeri A. Medina-Inojosa J.R. Kunzova S. Barchitta M. Agodi A. Vinciguerra M. Lopez-Jimenez F. Dog Ownership and Cardiovascular Health: Results From the Kardiovize 2030 Project Mayo Clin. Proc. Innov. Qual. Outcomes 2019 3 268 275 10.1016/j.mayocpiqo.2019.07.007 31485564 45. McConnell A.R. Brown C.M. Shoda T.M. Stayton L.E. Martin C.E. Friends with benefits: On the positive consequences of pet ownership J. Personal. Soc. Psychol. 2011 101 1239 1252 10.1037/a0024506 21728449 46. McConnell A.R. Lloyd E.P. Buchanan T.M. Animals as friends: Social psychological implications of human-pet relationships The Psychology of Friendship Oxford University Press New York, NY, USA 2017 157 174 47. Taniguchi Y. Seino S. Nishi M. Tomine Y. Tanaka I. Yokoyama Y. Amano H. Kitamura A. Shinkai S. Physical, social, and psychological characteristics of community-dwelling elderly Japanese dog and cat owners PLoS ONE 2018 13 e0206399 30427858 48. Allen K. Shykoff B.E. Izzo J.L. Pet ownership, but not ace inhibitor therapy, blunts home blood pressure responses to mental stress Hypertension 2001 38 815 820 10.1161/hyp.38.4.815 11641292 49. Brown C.M. Hengy S.M. McConnell A.R. Thinking about cats or dogs provides relief from social rejection Anthrozoös 2016 29 47 58 10.1080/20414005.2015.1067958 50. Wood L. Martin K. Christian H. Nathan A. Lauritsen C. Houghton S. Kawachi I. McCune S. The Pet Factor—Companion Animals as a Conduit for Getting to Know People, Friendship Formation and Social Support PLoS ONE 2015 10 e0122085 10.1371/journal.pone.0122085 25924013 51. Epstein M. Kuehn N.F. Landsberg G. Lascelles BD X. Marks S.L. Schaedler J.M. Tuzio H. AAHA Senior Care Guidelines for Dogs and Cats J. Am. Anim. Hosp. Assoc. 2005 41 81 91 15767650 52. Fortney W.D. Implementing a Successful Senior/Geriatric Health Care Program for Veterinarians, Veterinary Technicians, and Office Managers Vet. Clin. N. Am. Small Anim. Pract. 2012 42 823 834 10.1016/j.cvsm.2012.04.011 53. Landsberg G. Araujo J.A. Behavior Problems in Geriatric Pets Vet. Clin. Small Anim. Pract. 2005 35 675 698 10.1016/j.cvsm.2004.12.008 54. Fortney W.D. Geriatrics Vet. Clin. N. Am. Small Anim. Pract. 2012 42 xi xii 10.1016/j.cvsm.2012.05.002 55. Hart L. Dorairaj K. Camacho S. Nurturing older dogs: Attitudes and experiences of caregivers J. Am. Anim. Hosp. Assoc. 2001 37 307 310 10.5326/15473317-37-4-307 11450828 56. Christiansen S. Kristensen A. Sandøe P. Lassen J. Looking After Chronically III Dogs: Impacts on the Caregiver’s Life Anthrozoos 2003 26 519 533 10.2752/175303713X13795775536174 57. Berendt M. Gredal H. Ersbøll A.K. Alving J. Premature death, risk factors, and life patterns in dogs with epilepsy J. Vet. Intern. Med. 2007 21 754 759 10.1111/j.1939-1676.2007.tb03017.x 17708395 58. Slater M.R. Barton C.L. Rogers K.S. Peterson J.L. Harris C.K. Wallace K. Factors affecting treatment decisions and satisfaction of owners of cats with cancer J. Am. Vet. Med. Assoc. 1996 208 1248 1252 8635966 59. Lord L.K. Podell M. Owner perception of the care of long-term phenobarbital-treated epileptic dogs J. Small Anim. Pract. 1999 40 11 15 10.1111/j.1748-5827.1999.tb03246.x 10092036 60. Malik R. Gabor L.J. Foster S.F. McCorkell B.E. Canfield P.J. Therapy for Australian cats with lymphosarcoma Aust. Vet. J. 2001 79 808 817 10.1111/j.1751-0813.2001.tb10923.x 11837901 61. Brønden L.B. Rutteman G.R. Flagstad A. Teske E. Study of dog and cat owners’ perceptions of medical treatment for cancer Vet. Rec. 2003 152 77 80 10.1136/vr.152.3.77 12570310 62. Creswell J. Klassen A.C. Plano V. Smith K.C. Best Practices for Mixed Methods Research in the Health Sciences Office of Behavioral and Social Sciences Research (OBSSR) Bethesda, ML, USA 2011 9 63. Creswell J.W. Hirose M. Mixed methods and survey research in family medicine and community health Fam. Med. Community Health 2019 7 e000086 10.1136/fmch-2018-000086 32148709 64. Braun V. Clarke V. Reflecting on reflexive thematic analysis Qual. Res. Sport Exerc. Health 2019 11 589 597 10.1080/2159676X.2019.1628806 65. Buhrmester M. Kwang T. Gosling S.D. Amazon’s Mechanical Turk: A New Source of Inexpensive, Yet High-Quality, Data? Perspect. Psychol. Sci. 2011 6 3 5 10.1177/1745691610393980 26162106 66. Agley J. Xiao Y. Nolan R. Golzarri-Arroyo L. Quality control questions on Amazon’s Mechanical Turk (MTurk): A randomized trial of impact on the USAUDIT, PHQ-9, and GAD-7 Behav. Res. 2021 1 13 10.3758/s13428-021-01665-8 67. Johnson T.P. Garrity T.F. Stallones L. Psychometric Evaluation of the Lexington Attachment to Pets Scale (Laps) Anthrozoös 1992 5 160 175 10.2752/089279392787011395 68. Beery L.C. Prigerson H.G. Bierhals A.J. Santucci L.M. Newsom J.T. MacIejewski P.K. Rapp S.R. Fasiczka A. Reynolds C.F. III Traumatic Grief, Depression and Caregiving in Elderly Spouses of the Terminally Ill Omega 1997 35 261 279 10.2190/9QT0-XBD0-V8E9-U650 69. Van Doorn C. Kasl S.V. Beery L.C. Jacobs S.C. Prigerson H.G. The Influence of Marital Quality and Attachment Styles on Traumatic Grief and Depressive Symptoms J. Nerv. Ment. Dis. 1998 186 566 573 10.1097/00005053-199809000-00008 9741563 70. Bialon L.N. Coke S. A study on caregiver burden: Stressors, challenges, and possible solutions Am. J. Hosp. Palliat. Care 2012 29 210 218 10.1177/1049909111416494 21803785 71. Spitznagel M.B. Jacobson D.M. Cox M.D. Carlson M.D. Caregiver burden in owners of a sick companion animal: A cross-sectional observational study Vet. Rec. 2017 181 321 10.1136/vr.104295 28870976 72. Spitznagel M.B. Cox M.D. Jacobson D.M. Albers A.L. Carlson M.D. Assessment of caregiver burden and associations with psychosocial function, veterinary service use, and factors related to treatment plan adherence among owners of dogs and cats J. Am. Vet. Med. Assoc. 2019 254 124 132 10.2460/javma.254.1.124 30668290 73. Wittenberg-Lyles E. Demiris G. Parker Oliver D. Washington K.T. Burt S. Shaunfield S. Stress Variances Among Informal Hospice Caregivers Qual. Health Res. 2012 22 1114 1125 10.1177/1049732312448543 22673093 74. Spitznagel M.B. Jacobson D.M. Cox M.D. Carlson M. Predicting caregiver burden in general veterinary clients: Contribution of companion animal clinical signs and problem behaviors Vet. J. 2018 236 23 30 10.1016/j.tvjl.2018.04.007 29871745 75. Sallim A.B. Sayampanathan A.A. Cuttilan A. Ho R.C.-M. Prevalence of Mental Health Disorders Among Caregivers of Patients With Alzheimer Disease J. Am. Med. Dir. Assoc. 2015 16 1034 1041 10.1016/j.jamda.2015.09.007 26593303 76. Shyu Y.-I.L. Yang C.-T. Huang C.-C. Kuo H.-C. Chen S.-T. Hsu W.-C. Influences of Mutuality, Preparedness, and Balance on Caregivers of Patients With Dementia J. Nurs. Res. 2010 18 155 163 10.1097/JNR.0b013e3181ed5845 20808075

PMC009xxxxxx/PMC9099637.txt

==== Front Polymers (Basel) Polymers (Basel) polymers Polymers 2073-4360 MDPI 10.3390/polym14091852 polymers-14-01852 Article Extraction of Nanocellulose for Eco-Friendly Biocomposite Adsorbent for Wastewater Treatment https://orcid.org/0000-0002-8711-6019 Bassyouni Mohamed 12* Zoromba Mohamed Sh. 13 https://orcid.org/0000-0002-3075-4968 Abdel-Aziz Mohamed H. 14 https://orcid.org/0000-0002-5147-9221 Mosly Ibrahim 5 Bikiaris Dimitrios Academic Editor 1 Department of Chemical and Materials Engineering, King Abdulaziz University, Rabigh 21911, Saudi Arabia; mzoromba@kau.edu.sa (M.S.Z.); mhmossa@kau.edu.sa (M.H.A.-A.) 2 Department of Chemical Engineering, Faculty of Engineering, Port Said University, Port Said 42526, Egypt 3 Chemistry Department, Faculty of Science, Port Said University, Port Said 42521, Egypt 4 Chemical Engineering Department, Faculty of Engineering, Alexandria University, Alexandria 21544, Egypt 5 Department of Civil Engineering, King Abdulaziz University, Rabigh 21911, Saudi Arabia; ikmosly@kau.edu.sa * Correspondence: migb2000@gmail.com 30 4 2022 5 2022 14 9 185218 2 2022 26 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). In the present study, nanocellulose was extracted from palm leaves to synthesize nanocellulose/chitosan nanocomposites for the removal of dyes from textile industrial wastewater. Nanocellulose is of interest in water purification technologies because of its high surface area and versatile surface chemistry. Following bleach, alkali, and acid treatments on palm leaves, nanocellulose is obtained as a white powder. The produced nanocellulose was investigated. The adsorption capacity of chitosan, nanocellulose, and novel synthetic nanocellulose/chitosan microbeads (CCMB) for direct blue 78 dye (DB78) removal was studied. A series of batch experiments were conducted in terms of adsorbent concentration, mixing time, pH, dye initial concentration, and nanocellulose concentration in synthetic microbeads. The CCMB was characterized by using physicochemical analysis, namely Brunauer–Emmett–Teller (BET), scanning electron microscope (SEM), zeta potential analysis, and Fourier-transform infrared spectroscopy (FTIR). It was found that the surface area of synthetic CCMB is 10.4 m2/g, with a positive net surface charge. The adsorption tests showed that the dye removal efficiency increases with an increasing adsorbent concentration. The maximum removal efficiencies were 91.5% and 88.4%, using 14 and 9 g/L of CCMB-0.25:1. The initial dye concentrations were 50 and 100 mg/L under acidic conditions (pH = 3.5) and an optimal mixing time of 120 min. The equilibrium studies for CCMB-0.25:1 showed that the equilibrium data were best fitted to Langmuir isothermal model with R2 = 0.99. These results revealed that nanocellulose/chitosan microbeads are an effective eco-adsorbent for the removal of direct blue 78 dye and provide a new platform for dye removal. nanocellulose chitosan microbeads adsorption isothermal models direct dye removal Deanship of Scientific Research (DSR)King Abdulaziz University, JeddahG 327-829-38 This project was funded by the Deanship of Scientific Research (DSR), King Abdulaziz University, Jeddah, under grant No. (G 327-829-38). ==== Body pmc1. Introduction Dye contaminants in aquatic resources have become a significant problem as a result of the recent industrialization, urbanization, and growth of dyes-based operations such as the textile industry. Textile industries are also one of the world’s fastest-growing industrial sectors presently. They use a lot of water and produce a large amount of wastewater, which mostly consists of colors used in the dyeing process. The yearly water consumption by textile industries in the world is 40 × 109 m3 [1]. Approximately 10–15% of the applied dyes amount is gushed out as effluent. Such wastewaters severely affect land and water, resulting in pollution of the ecosystem [2,3]. The organic effluent disrupts the aquatic biosphere by obstructing light penetration, as well as posing major health risks to humans [4,5,6,7]. These dangers serve as ongoing reminders of the importance of finding effective color-removal technologies for wastewater. Many treatment technologies specialized in dye removal have been investigated, with several degree of success (i.e., chemical coagulation, membrane separation, catalytic, and chemical/physical adsorption) [8,9,10,11]. Among these processes, adsorption is considered one of the most commonly utilized and adaptable, allowing for the efficient and cost-effective removal of pollutants [12,13,14,15,16]. It was reported that biosorption process is one of the most promising technologies for wastewater treatment. Microorganisms such as bacteria, fungi, and algae have been investigated and used efficiently as bio-based sorbents for the removal of many contaminates from wastewater [17,18,19]. Various dyes adsorbents have been designed and examined, such as graphene-based composite, activated carbons, inorganic nanomaterials, microorganisms, and metal–organic frameworks. Because of their natural abundance, low cost, biocompatibility, and low environmental impact, biomass materials such as cellulose, chitin, chitosan, and lignin have been recognized as one of the most promising candidates in the past few years [20,21]. Cellulose is the most important plant component to substitute synthetic polymers, due to its being an inexpensive, nontoxic, and biodegradable polymer [22]. Cellulose is the main component of the plant cell wall [23,24]. Cellulose can be extracted from various sources, including wood, grasses, seed fibers, date palm seeds, algae, sisal fibers, fungi, argo-industrial waste, and bacteria [25,26,27]. On the other hand, cellulose is the most plentiful and renewable naturally occurring polymer in the world. It has been used in the manufacturing of energetic material for numerous supplies in the food and pharmaceutical industries, in paint, and in textiles [25]. In recent years, a wider application of cellulose has been suggested at the nanoscale level for developing various biocompatible products and a variety of cellulose derivatives [28,29,30]. A new area of nanocellulose applications is still under examination in some fields, including photonics, foams, surface modifications, nanocomposites, flexible, pharmaceutical industries, and optoelectronics. The most useful property of nanocellulose exploration is the green nature of the particles, as well as amazing chemical and physical properties. In addition, there are a variety of applications that can be taken from this vital material [31]. Cellulose is a carbohydrate polymer that has many monosaccharide units that are connected to each other through covalent bonds. Cellulose is considered to be a polymer with a linear backbone of anhydro-glucose monomer units connected through 1,4 β-linkages. Moreover, cellulose’s compact structure, mainly contributed to by the linkage of intra- and inter-hydrogen bonds as strong physical bonds, has a remarkable mechanical strength that protects the plant’s biological structure [32]. Date palm fiber biomass is a potential renewable resource which can contribute to energy sustainability. This may diminish the negative impacts of petroleum combustion in the environment. In the production of nanocellulose material from lignocellulosic biomass, various methods have been used, including acidic, basic, and ionic liquid treatment [33,34,35]. During the biomass treatment process, potassium hydroxide solution makes polymers (hemicellulose, lignin, and cellulose) swell, partially breaking the intra-hydrogen bonding in the biomass structure. The less ordered biomass structure leads to an increase in the number of existing hydroxyl groups and the availability of solvents for further hydrolysis reaction. Usually, sulfuric acid or hydrochloric acid as a strong acid were used for de-polymerization processing; acid makes hydrolytic cleave to the glycosidic linkages between the two adjacent anhydroglucose nits, dissolving the amorphous region of the cellulose by increasing the crystallinty of cellulose [36]. Because contamination of water by toxic dyes can affect human health and the ecology, it is necessary to remove dye wastewater. Natural adsorbents have attracted the attentions of many researchers throughout the world, due to their availability, ease of modification, and excellent adsorption capability. Heavy metals and dyes are removed from wastewater by using cellulose nanofibers (CNFs) [37]. This crystalline portion may be isolated from cellulose fibrils, which are typically 50–150 nm in length [38]. Nanocellulose (NC) may be versatilely modified with other materials to improve its adsorption capacity in the presence of hydroxymethyl functional groups. Some studies have found that using nanocrystalline cellulose (NC) to remove anionic and cationic dyes has satisfactory results [39]. More than 300 million tons of paper is produced yearly, and the demand is expected to increase by 2030 [40]. Paper uses 42 percent of the world’s wood, generating environmental issues. It was reported that a tris-azo dye (Direct Blue 71) was removed from aqueous solutions by using chitosan-based adsorbent gels. The chemical modification of chitosan was carried out by using crosslinking agents: sodium tripolyphosphate (TPP) and glutaraldehyde. The linear and nonlinear approaches for the Langmuir and Freundlich isotherms were compared. The adsorption studies revealed that the Langmuir model best described the experimental data from this study, with the maximal dye adsorption capacity of the adsorbent being 88.49 mg/g (linear form) and 92.22 mg/g (square form) (nonlinear form) [41,42]. Chitosan hydrogel beads were made by lowering the degree of crystallinity by generating a gel with the purpose of raising chitosan’s adsorption capability. Several strategies, including chemical crosslinking with crosslinking agents on their surface, have been developed to improve the commercial applicability of chitosan beads [43,44,45,46,47,48]. In recent years, the utilization of chitosan and cellulose for dye removal has attracted substantial interest as a potential eco-friendly adsorbent. Figure 1 shows the number of yearly published articles from 2013 to 2021. It was noted that the number of scientific publications on chitosan and cellulose in dye removal have significantly increased over the last 9 years. This study aimed to extract nanocellulose from palm leaves and develop a novel biocomposite adsorbent from the two largest natural resources (nanocellulose and chitosan), with an ionic liquid as the medium. The adsorption efficiency for direct blue 78 dye (anionic) removal using chitosan, nanocellulose, and nanocellulose/chitosan biocomposites was investigated. In this research, the nanocellulose particles were successfully extracted from palm leaves. A novel nanocellulose-based adsorbent was used for direct blue 78 dye removal with high efficiency (88–91%), optimal dose = (9–14) g/L, and high sedimentation rate, with an initial dye concentration of 50 and 100 mg/L respectively. Isothermal studies were also conducted for nanocellulose-based adsorbents, and the findings showed that the Langmuir isotherm best fit the adsorption results (R2 = 0.98). 2. Materials and Methods 2.1. Extraction of Nanocellulose from Palm Fiber Palm leaves were cleaned, shredded, and ground. Acetic acid, sodium hydroxide and sodium chlorite were purchased from Sigma-Aldrich (St. Louis, MI, USA). A solution of acetic acid, sodium chlorite, and distilled water was used for de-lignification. The suspension solution was refluxed at 80 °C. The solution was stirred for 24 h at 65 °C before rinsing with deionized water. The separation process was carried out by using centrifugation at 8000 r.p.m. The holocellulose was obtained at 70 °C. The remaining materials (hemicellulose and lignin) were removed by using diluted sodium hydroxide, using magnetic stirring for 7 h. The resultant suspended solution was dried for 24 h at 65 °C. Nanocellulose was extracted by using sulfuric solution (9 M) after stirring for 7 h at ambient temperature. The solution was rinsed with distilled water to avoid cellulose hydrolysis. Nanocellulose was obtained after centrifugation and freezing. Cellulose is a linear homopolymer made up of repeating units (called cellobiose) that are formed by connecting two anhydro-glucose rings via a -1,4 glycosidic linkage. It has s abundance of hydroxyl groups on the surface, according to its structure, as shown in Figure 2a. 2.2. Chitosan and Dyes Figure 2b shows chitosan, an amino-based polymer, synthesized in vast amounts by N-deacetylation of chitin. High-molecular-weight chitosan was used. The supplier reported that it is a white powder with a molecular-weight range from 140 to 220 kDa, degree of deacetylation (DAC) of 81.2%, viscosity of 36,000 cps, and density of 0.15 g/mL. The direct blue 78 (DB78) dye (Port Said, Egypt)was received by color print (Port Said, Egypt). Its relative molecular mass was 1059.95, max wavelength (λmax) was 604 nm, and solubility was up to 10 g/L at 25 °C. The direct blue 78 was selected for adsorption tests, as it is widely used in the textile industry. The chemical structure of direct blue 78 is shown in Figure 2c. Two synthetic dye solutions with different dye concentration (50 and 100 mg/L) were prepared for adsorption study. 2.3. Preparation of Nanocellulose/Chitosan Microbeads (CCMB) Chitosan was purchased from Sigma-Aldrich (St. Louis, MO, USA). Nanocellulose/chitosan microbeads (CCMBs) with different ratios of nanocellulose to chitosan were synthesized: CCMBz, where z refers to the ratio of nanocellulose to chitosan. The adsorption studies were conducted by using polymer biocomposite materials with different loading ratios: CCMB-0.1:1, CCMB-0.25:1, CCMB-0.5:1, and CCMB-1:1. In CCMB-1:1, the chitosan solution was prepared under magnetic stirring for 4 h by dissolving 2 g (2 wt.%) of chitosan powder into diluted acetic acid to form a 100 mL chitosan gel sample. A total of 500 mg (0.5 wt.%) of nanocellulose was added to the formed gel (50 mg of nanocellulose was added for every 10 mL of chitosan gel, under magnetic stirring, for 2.5 h, at a temperature of 50 °C). The final prepared gel was dropped into a 0.5 M NaOH solution (contact time 6 h), using a micropipette to form the beads. The formed beads were then washed with distilled water. Finally, the beads were oven-dried at 60 °C. The preparation process is illustrated in Figure 3. 2.4. Adsorption Studies This study was conducted by using a discontinuous batch adsorption system (lab scale) on a single-component synthetic wastewater. Nanocellulose samples used in this study were 0.1–2 g in 1000 mL of synthetic wastewater, with differing initial concentrations (50 and 100 mg/L), with mixing at 150 r.p.m. and contact times of 0–60 min, at a room temperature (22 ± 2 °C). Chitosan samples used in this study were 1–6 g in 1000 mL of synthetic wastewater, with differing initial concentrations (50 and 100 mg/L), with mixing at 150 r.p.m. for contact times of 0–60 min, at a room temperature (22 ± 2 °C). Nanocellulose/chitosan microbead samples used in this study were 1–15 g in 1000 mL of synthetic wastewater with differing initial concentrations (50 and 100 mg/L), with mixing at 150 r.p.m. for different contact times (0:150 min), at a room temperature (22 ± 2 °C). The spectrophotometrically analysis was applied to determine the removal efficiency by measuring dyes’ concentrations before and after the adsorption process at λmax = 600 nm for DB78. Nanocellulose and Chitosan Nanocomposites The nanocellulose suspension was diluted in water and ultrasonicated for 30 min in an ultrasonic bath, at 4% (w/v). Model USC-1400 is one-of-a-kind (40 kHz of ultrasound frequency). The Malvern 3000 Zetasizer NanoZS was used to make the measurements (Malvern Instruments, Malvern, WR14 1XZ. UK). This apparatus measures the diffusion of particles moving under Brownian motion and translates the data to size and size distribution, using dynamic light scattering. It also employs laser doppler micro-electrophoresis to provide an electric field to a dispersion of particles, which then move at a rate proportional to their zeta potential. The Smoluchowski algorithm was used to determine the particle size. The surface area of the CCMB-0.25:1 sample was measured in the presence of N2 adsorption at −195.65 °C, using surface area analyzers (Autosorb-l-C-8, Quantachrome, Boynton Beach, FL, USA). Prior to adsorption studies, the samples were degasified at 200 °C for 4 h. By applying the BET (Brunauer–Emmett–Teller) equation to the adsorption data, the BET surface area for the sample was determined. The colorimetric analysis was performed in this study, using a spectrophotometer (LAMOTTE smart spectrophotometer v3 2000-01-MN, Washington Ave. Chestertown, MD, USA). The pH values of nanocellulose, chitosan, and CCMB-0.25:1 solutions were determined by mixing 0.1 g from each sample with 100 mL of distilled water, at a mixing speed of 100 r.p.m., for a period of 1 h and temperature of 25 °C, using a digital pH meter (Omega CDS107, Taiwan). The surface morphology and porous microstructure of the CCMB-0.25:1 samples were investigated by SEM analysis, using a Quanta 250 FEG scanning electron microscope (Field Electron and Ion company, Hillsboro, OR, USA.). FTIR studies for the nanocellulose, chitosan, and CCMB-0.25:1 samples were observed by using a VERTEX 80v vacuum FTIR Spectrometer (Bruker corporation, Oberkochen, Germany). 3. Results and Discussion 3.1. Characterization of Nanocellulose The extracted nanocellulose was characterized by using zeta sizer to measure the particles size. Chemical structure was determined by using FTIR analysis. 3.1.1. Size of Nanocellulose To generate hydrodynamic diameter dimensions, light-scattering data were automatically evaluated and computed by using the built-in Zetasizer program. Figure 4 illustrates the particle size distribution acquired from DLS; it demonstrates that 95.5 percent of particles fall between nano-dimensions (up to 300 nm). 3.1.2. X-Ray Diffraction (XRD) Analysis Figure 5 shows the XRD pattern of the cellulose powder; the broad peaks indicate the amorphous nature of the cellulose powder. The XRD data (angle of diffracted beams, Miller indices (hkl), interplanar spacing, full width at half-maximum, and crystalline size (D) of cellulose powder) are listed in Table 1. Based on the diffraction peaks, a monoclinic 2 structure of cellulose was recorded with the following lattice parameters: a = 15.9634 Å, b = 7.85020 Å, c = 10.8664 Å, α = γ = 90, and β = 97.931°. The lattice parameters were calculated from the peak position, as given by the following relation [49]:(1) 1d2=1sin2θ(h2a2+k2sin2βb2+l2c2−2hkcosβac)+l2c2  As shown in Table 1, the estimated crystallite size (D) and miller index (hkl) are dependent on the absolute values of full width at half maximum (FWHM). The data in database code_amcsd 0,017,094 agree well with the interplanar distances’ d-spacing [50], according to the American Mineralogist Crystal Structure Database. The Debye–Scherrer method was applied to assessed XRD for cellulose powder, the range of 10 ≤ 2θ ≤ 90 with 1/dhkl=0.0566Å−1−0.7446Å−1, λ=1.540562 Å, I2/I1=0.5, polarization = 0.5, and function Pseudo-Voigt. From Scherer’s formula, we obtain the following:(2) D=0.9λFWHM.cosθ  where λ is the X-ray wavelength (1.541838 Å). As presented in Table 1, for cellulose powder, the XRD data from the XRD pattern were used to examine factors and features such as FWHM, the crystallite size (D), hkl indices, d-spacing (d), and peak intensity. The crystalline size Dav=164.71 nm was within the range of 83.04–183.56 nm for cellulose powder. While, for both the experimental and PXRD models, the intensity and location of specific peaks vary only slightly, the emphasis here is mostly on their overall resemblance. Only the important comparison characteristics between the measured and the experimental data should therefore be evaluated. It is also known that instrumentation and data-collection processes are only two of the many variables that can affect the experimental PXRD pattern. Employing X-ray powder diffraction to distinguish patterns of cellulose Iα and cellulose Iβ is exceedingly difficult, due to their overlap [51]. 3.1.3. Thermal Analysis of Nanocellulose The stability properties of cellulose are shown in Figure 6. At temperatures below 100 °C, moisture evaporation was observed in the cellulose samples. There was low weight loss within this stage because the amount of absorbed water or moisture in cellulose is low. Around 7% weight loss was recoded up to 100 °C. This process mostly relates to water moisture evaporation below 100 °C, as validated by the DrTGA investigation shown in Figure 7. It was found that cellulose breakdown began at 307 °C and lasted until 340 °C. These results are in a good agreement with the reported data [52]. The thermal stability of cellulose chains is enhanced by their highly ordered packing into systems (crystals) and by strong hydrogen bonding. The crystalline structure of cellulose is essential for its heat stability [53]. The observed narrow curve at DrTGA 307 °C might potentially be due to more surface area exposed to the heat and partial disruption when the temperature increased from 270 °C to 340 °C. Finally, the decomposition of cellulose was found from 330 to 500 °C. This stage can be attributed to cellulose oxidation. One medium peak can be found in the DrTGA curve at a temperature of 420 °C and refers to the decomposition of polymer chains of cellulose. A similar result was reported for nanocellulose decomposition using TGA at around 420 °C [54,55]. 3.1.4. FTIR Analysis The FTIR results of cellulose and palm fibers are shown in Figure 8. The results show that the cellulose has a broadband at the 3432 cm−1 region that can be attributed to O-H groups’ stretching vibration. As shown in the spectrum, the bands appearing in the regions (1322–1429), (2997–3766), (1561–1806), and (626–843) cm−1 can be attributed to the hydroxide group of water molecule and may be bending, stretching, rocking, and wagging vibrations. This indicates the presence of water molecules in the studied copolymer [56]. There were several absorption bands that were associated with the cellulose, and we it was also observed that 1160 and 1062 cm−1 were attributable to C-O bond stretching [57]. The cellulose with abundant surface hydroxyl groups was investigated by using FTIR, as shown in Figure 8. The appearance of a new peak at 1161cm−1 is associated with C=C stretching [58]. The absorption peak noticed at 1630 cm−1 refers to the O-H bond of absorbed water. The appearance of a new peak at 2354 cm−1 that was associated with ester groups was intensively observed on cellulose. Among the three kinds of OH groups, the OH group of the sixth position acts as a primary alcohol, where most of the modification predominantly occurs [59]. 3.2. Characterization of Nanocomposites Surface Morphology of Nanocomposites Because of the importance of surface morphology and its great influence on the adsorption process, SEM analysis was investigated for CCMB-0.25:1 in order to identify its surface morphology and nanocellulose particles’ distribution on the beads’ surface. Figure 7 shows the SEM image for the nanocellulose/chitosan microbead (CCMB) surface. The SEM images showed that all nanocellulose particles were incorporated effectively into chitosan network, and there is no agglomeration of large numbers of nanoparticles the on small surface area. This efficient distribution of nanocellulose particles on the chitosan microbeads’ outer surface resulted in the creation of a large number of adsorption active sites. Moreover, it can be noticed from the SEM analysis that CCMB has an average particle size of 2 µm, with a large number of micropores. The efficiency of the adsorption process is affected by the distribution of cellulose particles on the bead surface; with a uniform distribution of nanocellulose particles on beads’ surface, the adsorption behavior would be improved. The BET surface area for CCMB-0.25:1 was determined (SBET = 10.4 m2/g). In order to classify the main infrared (IR) bands of organics and determinate the adsorption mechanism (physisorption or chemosorption), pure and loaded samples of chitosan, nanocellulose, and CCMB-0.25:1 were investigated by FTIR analysis. As shown in Figure 9a, the peaks at 2919 and 2856 cm−1 can be attributed to C-H symmetric and asymmetric stretching, respectively. The cH2 bending and cH3 symmetrical deformations were confirmed by the presence of a peak at 1386 cm−1. The new peaks that appeared in the FTIR spectra for loaded chitosan when compared with the pure chitosan are attributed to the chemical bond formed between dyes molecules and -NH2 groups on the surface of chitosan particles after the adsorption process. These results were also reported by Reference [60]. Figure 9b shows the FTIR analysis of nanocellulose before and after dye adsorption. The peaks at 3446 and 2919 cm−1 can be attributed to an O-H stretching band caused by the hydrogen-bonded hydroxyl group variations of cellulose and aliphatic saturated symmetric C-H stretching variations in cellulose, respectively. The presence of bands at 1660 cm−1 can also be caused by O-H bending modes of adsorbed water. The C-O-C band was also confirmed by the pretense of a peak at 1064 cm−1. When comparing the two spectra, we noticed that there is no difference. New peaks have not yet appeared, but the peak wave number shifted to higher values. That means that chemical bonds were not formed, and the adsorption process was conducted due to the electrostatic interaction between anionic dyes molecules and H+ ions accumulated on the nanocellulose surface in acidic conditions [61]. Figure 9c shows the FTIR spectra for pure and loaded CCMB-0.25:1. It was observed that there is a significant difference. This difference is attributed to the chemical bonds formed between -SO3 groups on dye molecules and -NH3+ groups on the beads’ surface. The most relevant difference is the appearance of bands at 2915 and 2853 cm−1. These bands originated from dye molecules attached to the beads’ surface after adsorption process. Figure 9c also shows the -OH stretching vibration, which can be represented by the peak at 3328 cm−1. The stretching frequency of the –NH2 groups can be seen from the broad band at 16,330 cm−1. The peak at 1540 cm−1 could be due to the N−H stretching vibration [62]. 3.3. Adsorption Tests 3.3.1. Effect of Adsorbent Concentration on DB78 Dye Removal Efficiency The concentrations of chitosan, nanocellulose, and CCMB-0.25:1 were varied to investigate their effect on direct blue 78 dye removal efficiency. It was found that, by increasing the adsorbents concentration (increasing of adsorption active sites), the equilibrium loading would decrease, and the removal efficiency would increase until reaching the maximum efficiency and then approximately reach a constant value. The experiments were conducted by varying the concentration of chitosan powder from 1 to 6 g/L, nanocellulose from 0.25 to 2 g/L, and CCMB-0.25:1 from 1 to 15 g/L on DB78 dye solutions with initial concentrations of 50 and 100 mg/L, at a fixed temperature, pH, stirring speed, and mixing time. For chitosan, it was observed from batch adsorption tests that a removal percentage of 94% can be obtained by using a chitosan dose of 3 g/L for a solution with an initial concentration of 50 mg/L, as shown in Figure 10. The maximum equilibrium loading reached was 74.4 mg/g. For nanocellulose, it was observed that a removal percentage of 93.2% and equilibrium loading of 46.6 mg/g can be obtained by using a nanocellulose dose of 2 g/L for solutions with an initial concentration of 100 mg/L, as shown in Figure 11. The maximum equilibrium loading reached was 239 mg/g. For CCMB-0.25:1, it was observed that a removal percentage of 92.1% and equilibrium loading of 4.6 mg/g can be obtained by using a CCMB-0.25:1 dose of 10 g/L for solutions with an initial concentration of 50 mg/L, as shown in Figure 12. The maximum equilibrium loading reached was 13.9 mg/g. Table 2 presents the optimal adsorbent concentration for direct blue 78 dye removal. 3.3.2. Effect of Solution pH on DB78 Dye Removal Efficiency The effect of the initial pH of dye solution was experimentally investigated under a pH range from 1 to 10, and the results can be observed from Figure 13. For chitosan, the solution pH has a little effect on chitosan adsorption behavior. Chitosan reaches its maximum loading (maximum removal efficiency of 98.3%) under acidic conditions (pH = 3), in comparison with a 93% removal efficiency under alkaline conditions (pH = 9). This can be attributed to the presence of acidic conditions, where hydrogen ions (H+) could protonate the amine groups (–NH2) of chitosan. Chitosan–NH2 + H+ → chitosan–NH3+(3) The direct blue 78 dye was dissolved in aqueous solution, and the sulfonate groups were separated and converted into anionic dye ions. DSO3Na → DSO3− + Na+(4) The adsorption process then ensued due to the electrostatic interaction between these two ions [63]. Chitosan–NH3+ + DSO3− → chitosan–NH3+-O3SD(5) The pH value is one of the most important process variables when considering dye adsorption. The adsorption of a positive charged adsorbate is favored when the pH of the solution is greater than the point of zero charge (pHpzc) of the adsorbent. The point of zero charge indicates that the net charge on the whole particle surface (i.e., the surface of the absorbent) is zero. Whereas the adsorption of negatives charges, in turn, is favored at pH levels less than pHpzc. Therefore, the adsorption of the anionic dyes is expected to be favored in solutions with pH values less than the pHpzc of the adsorbent [64]. From this point, it can be concluded that the swelling of chitosan powder into chitosan beads in the presence of acidic conditions will protonate the amine groups (NH2) into NH3+. This process will improve the electrostatic interaction between chitosan particles and dye ions and enhance the chitosan ability for anionic dyes’ removal. The adsorption process was significantly improved in the acidic solution. Therefore, it is supposed that the adsorbent surface is positively charged, and this is favorable to the adsorption of anionic dyes. The enhanced electrostatic interactions were formed between the positively charged bioadsorbent’s surface and the negatively charged SO3 group of dyes in acidic solutions. As seen in Figure 13, the initial pH of the solution has a higher effect on the direct blue 78 dye solutions’ removal process. Nanocellulose reaches its maximum adsorption capacity for DB78 dye under acidic conditions, i.e., a pH range from 1 to 2, and then the adsorption capacity is decreased sharply during the increasing of the pH from 3 to 6, and then it decreases slowly down to a pH equal to 8. Figure 14 shows that nanocellulose particles and CCMB-0.25:1 have a negative net surface charge (Zeta potential = −63 mv) and positive net surface charge (Zeta potential = +66 mv), respectively. The negative Zeta potential value of −63 mV is attributed to the presence of highly electronegative sulfate groups on the surface of the cellulose nanoparticles. The maximum adsorption capacity for DB78 dye was found in acidic conditions within a pH range from 3 to 4, and then it decreased sharply as the pH increased from 5 to 8. 3.3.3. Effect of Mixing Time on DB78 Dye Removal Efficiency The effect of the mixing time on both the percentage of DB78 dye removal and adsorbent loading was investigated. It is experimentally observed that the percentage of dye removal increased with the increasing mixing time, even reaching the optimal removal efficiency (equilibrium concentration, Ce); at this time, the adsorbent is reaching its maximum loading capacity (equilibrium loading, qe). For 45 min of contact time and a 4 g/L adsorbent dose of chitosan, the equilibrium concentration decreased to 1.6 mg/L, the dye removal efficiency was 96.7%, and the optimum loading capacity was 12.1 mg/g for an initial dye concentration of 50 mg/L solution. For 30 min of contact time and 500 mg/L as the adsorbent dose of nanocellulose, the equilibrium concentration decreased to 7.2 mg/L, the dye removal efficiency was 85.7%, and the optimum loading capacity was 85.5 mg/g for an initial dye concentration of 50 mg/L solution, as shown in Figure 15. For 120 min of contact time and 10 g/L as the adsorbent dose of CCMB-0.25:1, the equilibrium concentration decreased to 3.9 mg/L, the dye removal efficiency was 92.1%, and the optimum loading capacity was 4.6 mg/g for an initial dye concentration of 50 mg/L solution, as shown in Figure 16. 3.3.4. Effect of Dye Initial Concentration Removal Efficiency The direct blue 78 (DB78) dye solutions had initial concentrations of 50 and 100 mg/L. Figure 17a shows that the removal efficiency at 80.3% and 77.2% from using chitosan dose 1g/L for the initial concentrations 50 and 100 mg/L, respectively. The removal efficiency was improved to 90.1% and 84.6% by using chitosan dose 2 g/L. At a higher concentration of neat chitosan (3 g/L), the removal efficiency of the direct blue 78 (DB78) dye reached 94.2% and 90.8% at the initial dye concentration 50 and 100 mg/L, respectively. Figure 17b shows that the removal efficiency was obtained at up to 85.4% and 69.7% for the dyes’ initial concentration of 50 and 100 mg/L, respectively, using 0.5 g/L of nanocellulose. A higher removal efficiency was observed when using 0.75 g/L nanocellulose. The removal efficiency of DB78 was found to be 91.7% and 79.4% when using 1 g/L of nanocellulose for initial concentrations of 50 and 100 mg/L, respectively. For CCMB-0.25:1 at a dose of 6 g/L, the removal efficiency was 86% and 63%, respectively, for the 50 and 100 mg/L initial concentrations of the dyes, as shown in Figure 17c. These results show that the adsorption process is highly dependent on the initial concentration of dyes. 3.3.5. Effect of Nanocellulose Concentration on Removal Efficiency To study the effect of nanocellulose loading using nanocellulose/chitosan microbeads (CCMBs) on dye-removal efficiency, a different ratios were applied: CCMB-0.1:1, CCMB-0.25:1, CCMB-0.5:1, and CCMB-1:1. The experimental results showed that the dye-removal efficiency increases with the increase nanocellulose loading, up to 0.5:1. At higher nanocellulose loads, the removal efficiency decreases, as shown in Figure 18. This can be attributed to blockage of internal porosities of chitosan by the incorporated higher nanocellulose loadings. 3.4. Adsorption Isotherm As shown in Figure 19, the adsorption isothermal curve indicates the quantity of adsorbate DB78 dye that can be adsorbed by the adsorbents (chitosan, nanocellulose, and CCMB-0.25:1), qeq, in comparison to the adsorbate concentration in the liquid state (Ceq). These are essential considerations in the design of adsorption systems. Moreover, the form of the equilibrium curve helps to describe other phenomena linked with the adsorbent–adsorbate interaction. The equilibrium curves are identified in four main classes, according to the primary slope, and the subgroups are described for each class based on the upper parts’ shapes and the slope changes: (a) S curves or vertical orientation isotherm, (b) L curves or normal or “Langmuir” isotherms, (c) H curves or high-affinity isotherms, and (d) C curves or constant partition isotherm [65]. The initial shape of the equilibrium curve (L shape) in Figure 19 follows the basic premise that, the higher the solute concentration, the greater the adsorption capacity, until the number of adsorption site clearance is limited, and competition occurs between the solute molecules for the available sites. This isotherm type indicates that the adsorption occurs due to relatively weak forces, such as “van der Waals forces”. There are several isothermal models (equations) available, and the two important isotherms are selected in this study, namely the Freundlich and Langmuir isotherms. The Freundlich isotherm believes that the adsorption happens on a heterogeneous surface, and the adsorbed mass increases exponentially with an increase in concentration [66]. This isotherm explains equilibrium on heterogeneous surfaces, and, hence, capacity is not presumed to be a monolayer. In liquid phase, this isotherm is given by Equation (6):Qe = Kf Ce 1/nf(6) where kF is the Freundlich fixed value (kF unit = mg/g, where c = 1/nF is the heterogeneity factor). This isotherm focuses on integrating the role of adsorbent–adsorbate surface interactions. Figure 20 indicates the application of equilibrium data according to the Freundlich isotherm. For chitosan as an adsorbent, the Freundlich constant kf values were 8.02 and 3.65, and the heterogeneity factor 1/nf values were 0.67 and 0.88, respectively, for solutions with the initial concentrations of 50 and 100 mg/L. For nanocellulose as an adsorbent, the Freundlich constant kf values were 30.6 and 13.49 mg/L, and the heterogeneity factor 1/nf values were 0.47 and 0.67 for the initial concentration of 50 and 100 mg/L, respectively. For CCMB-0.25:1 as an adsorbent, the Freundlich constant kf values were 2.18 and 2.12, and the heterogeneity factor 1/nf values were 0.51 and 0.43 with the initial concentrations of 50 and 100 mg/L. The Langmuir isotherm believes that sorption occurs within the adsorbent at different homogeneous sites, and it has been successfully applied to several processes of sorption. The isotherm’s physical simplicity is based on some assumptions: Adsorption cannot occur beyond monolayer coverage. Each site can hold only one adsorbate molecule. All sites are energetically equivalent, and the surface is uniform. The linear form of the Langmuir isotherm is given by Equation (7):(Ce/qe) = (1/Q0 b) + (Ce/Q0)(7) where Ce is the equilibrium concentration (mg/L), qe is the mass adsorbed at equilibrium (mg/g), Q0 is the adsorbent loading (mg/g), and b is the adsorption energy (Langmuir fixed value L/mg). The values of Q0 and b were determined from the slope and intercept of the linear plots Ce/qe versus Ce, resulting in a straight line of slope 1/Q0, corresponding to the total coverage of monolayer (mg/g), and the intercept is 1/Q0b [67,68]. Figure 21 indicates the application of equilibrium data according to the Langmuir isotherm. For chitosan, the adsorbent loading value (Q0) was 73.5 mg/g, and the Langmuir fixed value (b) was 0.107 L/mg for the initial concentration of 50 mg/L. For nanocellulose, the adsorbent loading value (Q0) was 175.4 mg/g, and the Langmuir fixed value (b) value was 0.16 and 1/mg. For CCMB-0.25:1, the adsorbent loading value (Q0) was 15.3 mg/g, and the Langmuir fixed value (b) was 1.03 L/mg. It was observed from the listed adsorption isothermal models in Table 3 that they follow the Freundlich isotherm for chitosan as an adsorbent. The Langmuir isotherm is for CCMB-0.25:1 as an adsorbent, and both the Freundlich and Langmuir are for nanocellulose as an adsorbent. 3.5. Adsorption Kinetics In order to understand the mechanism of adsorption process, the kinetic studies were conducted by extracting and analyzing the samples at time intervals of 10 min until the consecutive residue dye concentrations became closer. The kinetic data for the adsorption process of DB78 dye onto chitosan, nanocellulose, and CCMB-0.25:1 with an initial dye concentration of 50 mg/L were examined with the well-known kinetic models, namely pseudo first-order model (PFO) and pseudo second-order model (PSO). The plotting of these kinetic models is shown in Figure 22. Pseudo first-order equation: The pseudo first-order kinetic equation was used for the adsorption analysis. The linear form of this equation is as follows:ln (qe-qt) = ln qe–k1 t(8) where qe (mg/g) and qt (mg/g) are the amounts of adsorbed adsorbate at equilibrium and at time, t, respectively. K1 (min−1) is the rate constant of pseudo first-order model. Pseudo second-order equation: The adsorption kinetics can also be described by the pseudo second-order model. The linear form of the pseudo-second-order equation is expressed as follows:(t/qt) = (1/k2qe2) + (1/qe) t(9) where k2 (g/mg min) is the equilibrium rare constant of pseudo second-order adsorption; and qe (mg/g) and qt (mg/g) are the amounts of adsorbed adsorbate at equilibrium and at time, t, respectively [69,70]. Figure 22 shows the linear plots of PFO and PSO models of CCMB-0.25:1, nanocellulose, and chitosan. The kinetic parameters are listed in Table 4. On the basis of the low correlation coefficient for PSO and the high value for PFO, the adsorption abilities of CCMB-0.25:1 follow PFO rather than PSO; on the other hand, the adsorption behavior of nanocellulose and chitosan follows PSO rather than PFO. These results suggested that, for CCMB-0.25:1, PFO can best predict the kinetic process. The value of qe = 12.8 mg/g calculated by PFO was more similar to practical qe = 11.5 mg/g than PSO. For nanocellulose and chitosan, PSO can best predict the kinetic process. The values of mg/g calculated by PFO were more similar to practical qe = 11.5 mg/g than PFO, as shown in Table 4. For chitosan, the applicability of the PSO model indicates the interaction between dye molecules and amino groups. Hence, the adsorption system is chemical adsorption. It was reported that the adsorption process of DB78 dye onto chitosan is best fitted to pseudo second order with a chemical adsorption mechanism [71]. The chitosan in powder form showed a high adsorption capacity, as shown in Figure 23; this capacity can be attributed to its high surface area, but it needs longer sedimentation time (8 h). For the chitosan microbeads formed, the swelling of chitosan powder into microbeads in the presence of acidic conditions improved its adoption capacity, due to the protonation of amine groups (NH2) into NH3+. This modification process led to a significant decrease in sedimentation time. Surface area of chitosan in the form of beads is less than the surface area of chitosan in powder form. Therefore, the chitosan microbeads were loaded with cellulose nanoparticles in order to improve the surface area and increase its adsorption capacity. CCMB has shown a good adsorption capacity, in addition to remarkable short sedimentation time, with a low dose. Five minutes is sufficient to complete the sedimentation process. “Based on the experimental results, maximum removal efficiency 80% can be achieved using chitosan dose 1 g/L. While 65% removal was obtained using nano-cellulose dose 0.25 g/L. In the case of CCMB-0.25:1, 1 g/L of chitosan and 0.25 g/L of nano-cellulose can produce a microbead with a remarkable adsorption capacity of. Removal efficiency 95% with optimal dose 10 g/L was achieved using CCMB-0.25:1”, as shown in Figure 23. Figure 24 shows the adsorption process of DB78 dye, using adsorbents (chitosan, nanocellulose, and CCMB). It was observed that clear water can be obtained in the presence of chitosan powder after a long sedimentation time, i.e., 8 h. CCMB showed the lowest sedimentation time (5 min). Chitosan powder and cellulose nanoparticles showed longer sedimentation times than CCMB. 4. Conclusions The removal of direct blue 78 dye (anionic) from single-component synthetic wastewater by adsorption with nanocellulose, chitosan, and novel nanocellulose/chitosan microbeads (CCMB) was experimentally investigated. An altered microbead with different nanocellulose/chitosan ratios (0.1:1, 0.25:1, 0.5:1, and 1:1) was synthetized in order to study the effect of nanocellulose dose on removal efficiency. The removal efficiency increases with an increasing nanocellulose dosage in synthetic microbeads up to a nanocellulose/chitosan ratio 0.5:1, and then the efficiency decreases with an increasing nanocellulose dosage. The adsorption process is highly dependent on the initial solution pH; the CCMB-0.25:1 reaches its maximum loading under acidic conditions (pH 3.5). It was observed from experimental studies that removal efficiencies of 94%, 91.7%, and 92.1% can be obtained by using the adsorbents chitosan, nanocellulose, and CCMB-0.25:1, respectively, for dye solution with an initial concentration 50 mg/L. Equilibrium studies have shown that the initial shape of the equilibrium curve is an L-shape, meaning that the adsorption process resulted from electrostatic interaction between dyes molecules and adsorbent particles (physical forces). Adsorption studies were modeled by using the Langmuir and Freundlich isothermal models. Therefore, chitosan, nanocellulose, and CCMB could be highly efficient sorbents for removing anionic contaminants. The abovementioned excellent performances of chitosan, nanocellulose, and CCMB demonstrated them as promising dye adsorbents for wastewater treatment. Acknowledgments This project was funded by the Deanship of Scientific Research (DSR), King Abdulaziz University, Jeddah, under grant No. (G 327-829-38). The authors, therefore, acknowledge and thank DSR for technical and financial support. Author Contributions Conceptualization, M.B., M.S.Z., M.H.A.-A. and I.M.; methodology, M.B., M.S.Z. and M.H.A.-A.; software, M.B.; validation, M.B., M.S.Z., M.H.A.-A. and I.M.; formal analysis M.B., M.S.Z., M.H.A.-A. and I.M.; investigation, M.B., M.S.Z. and M.H.A.-A.; resources, M.B., M.S.Z. and M.H.A.-A.; data curation, M.B., M.S.Z. and M.H.A.-A.; writing—original draft preparation, M.B., M.S.Z. and M.H.A.-A.; writing—review and editing, M.B., M.S.Z., M.H.A.-A. and I.M.; visualization, M.B., M.S.Z., M.H.A.-A. and I.M.; supervision, M.B., M.S.Z., M.H.A.-A. and I.M.; funding acquisition, M.B. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Further data is available on request from the authors. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Published papers related to utilization of chitosan and cellulose for dye removal. Obtained from ScienceDirect. Search words, respectively, were chitosan adsorption for dye removal and cellulose adsorption for dye removal. Figure 2 (a) Chemical structure of nanocellulose. (b) Chemical structure of chitosan. (c) Chemical structure of direct blue 78 dye. Figure 3 The preparation process for nanocellulose/chitosan beads (CCMB-0.25:1). Figure 4 Particle size distribution of nanocellulose. Figure 5 XRD pattern of cellulose sample. Figure 6 Thermal analysis TG and DrTGA for cellulose at 20–500 °C. Figure 7 SEM image with different magnification for nanocellulose/chitosan microbead (CCMB-0.25:1). Figure 8 FTIR analysis of cellulose and palm leaves. Figure 9 FTIR spectra for (a) pure and loaded chitosan, (b) pure and loaded nanocellulose, and (c) BD78 dye and CCMB-0.25:1 before and after adsorption process. Figure 10 Effect of chitosan concentration on DB78 dye removal efficiency (temperature of 22 °C, pH of 8.5, mixing speed of 150 r.p.m., and contact time of 45 min). Figure 11 Effect of nanocellulose concentration on DB78 dye removal efficiency (temperature of 22 °C, pH of 2, mixing speed of 150 r.p.m., and contact time of 30 min). Figure 12 Effect of CCMB-0.25:1 concentration on DB78 dye removal efficiency (temperature of 22 °C, pH of 3.5, mixing speed of 150 r.p.m., and contact time of 120 min). Figure 13 Effect of pH value on DB78 dye removal efficiency (temperature of 22 °C, pH of 1–10, and mixing speed of 150 r.p.m.). Figure 14 Zeta potential analysis of nanocellulose and nanocellulose/chitosan microbead (CCMB-0.25:1). Figure 15 Effect of mixing time on DB78 dye removal efficiency. Adsorbents are chitosan and nanocellulose; temperature, 22 °C; pH, 8.5–2; and mixing speed, 150 r.p.m. Figure 16 Effect of mixing time on DB78 dye removal efficiency. Adsorbent, CCMB-0.25:1; temperature of 22 °C; pH of 3; and mixing speed of 150 r.p.m. Figure 17 Effect of dye initial concentration on removal efficiency using different adsorbents: (a) chitosan, (b) nanocellulose, and (c) CCMB-0.25:1. Figure 18 Effect of nanocellulose concentration in microbeads on removal efficiency. Figure 19 Adsorption isotherm for direct blue 78 dye removal using adsorbents (a) chitosan, (b) nanocellulose, and (c) CCMB-0.25:1. Figure 20 Freundlich adsorption isotherm for direct blue 78 dye removal using adsorbents (a) chitosan, (b) nanocellulose, and (c) CCMB-0.25:1. Figure 21 Langmuir adsorption isotherm for direct blue 78 dye removal using adsorbents (a) chitosan, (b) nanocellulose, and (c) CCMB-0.25:1. Figure 22 Adsorption kinetic studies: (a) pseudo first-order model and (b) pseudo second-order model. Figure 23 Effect of adsorbent type on dye removal efficiency. Figure 24 The adsorption process of DB78 dye by using adsorbents (chitosan, nanocellulose, and CCMB-0.25:1). polymers-14-01852-t001_Table 1 Table 1 The computation data resulted from the application Refine Version 3.0 Software Program (Kurt Barthelme’s and Bob Downs) for [Indol-4Ap]TF. Symmetry Compound 2θ d hkl Observed Calculated FWHM Dav 2θ 2d 2θ 2d Parameters 15.453 5.7296 111 15.331 5.882 0 1.3 × 10−5 5.7296 83.04 a = 15.96 Å; b = 7.85 Å; c = 10.87 Å 20.332 4.4253 051 20.332 4.425 0 4 × 10−6 1.7565 270.86 α = γ = 90, β = 97.931° 22.655 3.9201 020 22.635 3.974 0 0 3.9201 121.37 V = 1400 (22), rmse (a) = 1.05 × 10−3 34.575 2.5920 232¯ 34.575 2.613 0 −1 × 10−6 2.5920 183.56 Machine error = −0.48 Average 3.5995 164.71 (a) Root mean square error. polymers-14-01852-t002_Table 2 Table 2 Optimal adsorbent concentrations for direct blue 78 dye removal. Adsorbent Initial Concentration 50 mg/L Initial Concentration 100 mg/L Dose (g/L) Loading (mg/g) Removal % Dose (g/L) Loading (mg/g) Removal % Chitosan 3 15.7 94.2 5 18.77 93.85 Nanocellulose 1 48.3 96.6 2 46.6 935 CCMB-0.25:1 9 5.08 91.52 14 5.9 88.4 polymers-14-01852-t003_Table 3 Table 3 Comparison of adsorption isothermal models for adsorbents (chitosan, nanocellulose, and CCMB-0.25:1). Adsorbent DB78 Dye Initial Concentration (mg/L) Langmuir Isothermal Freundlich Isothermal Followed Q0 (mg/g) b (L/mg) R2 Kf (mg/g) 1/n R2 Chitosan 50 73.5 0.107 0.8514 8.02 0.67 0.9914 Freundlich 100 384.6 0.0074 0.1658 3.65 0.88 0.9406 Freundlich Nanocellulose 50 175.4 0.16 0.9867 30.6 0.47 0.9843 Langmuir 100 400 0.018 0.988 13.49 0.67 0.9959 Freundlich CCMB-0.25:1 50 15.31 1.0306 0.9867 2.18 0.51 0.9428 Langmuir 100 17.79 0.0404 0.9943 2.12 0.43 0.9841 Langmuir polymers-14-01852-t004_Table 4 Table 4 Kinetic models parameters. Kinetic Model Parameters Pseudo First-Order Model Pseudo Second-Order Model K1 (1/min) qe (mg/g) R2 K2 (g/mg min) qe (mg/g) R2 CCMB-0.25:1 0.0385 12.8 0.907 0.0003 13.2 0.689 Nanocellulose 0.1594 292 0.8807 0.0015 99 0.9947 Chitosan 0.0867 20.9 0.967 2.807 17.27 0.9824 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Ramya M. Karthika M. Selvakumar R. Raj B. Ravi K. A facile and efficient single step ball milling process for synthesis of partially amorphous Mg-Zn-Ca alloy powders for dye degradation J. Alloy. Compd. 2017 696 185 192 10.1016/j.jallcom.2016.11.221 2. Gómez V. Larrechi M. Callao M. Kinetic and adsorption study of acid dye removal using activated carbon Chemosphere 2007 69 1151 1158 10.1016/j.chemosphere.2007.03.076 17531288 3. Sandid A.M. Bassyouni M. Nehari D. Elhenawy Y. Experimental and simulation study of multichannel air gap membrane distillation process with two types of solar collectors Energy Convers. Manag. 2021 243 114431 10.1016/j.enconman.2021.114431 4. Zoromba M.S. Abdel-Aziz M.H. Ecofriendly method to synthesize poly (o-aminophenol) based on solid state polymerization and fabrication of nanostructured semiconductor thin film Polymers 2017 120 20 29 10.1016/j.polymer.2017.05.048 5. Elhenawy Y. Elminshawy N.A.S. Bassyouni M. Alhathal Alanezi A. Drioli E. Experimental and theoretical investigation of a new air gap membrane distillation module with a corrugated feed channel J. Membr. Sci. 2020 594 117461 10.1016/j.memsci.2019.117461 6. Nair A.K. Kumar B.V. Kalaiarasan G. Babu P.J. TiO2 nanosheet incorporated polysulfone ultrafiltration membranes for dye removal Desalination Water Treat. 2018 107 324 331 10.5004/dwt.2018.22157 7. Elhady S. Bassyouni M. Mansour R.A. Elzahar M.H. Abdel-Hamid S. Elhenawy Y. Saleh M.Y. Oily Wastewater Treatment Using Polyamide Thin Film Composite Membrane Technology Membranes 2020 10 84 10.3390/membranes10050084 8. El-Mehalmey W.A. Safwat Y. Bassyouni M. Alkordi M.H. Strong Interplay between Polymer Surface Charge and MOF Cage Chemistry in Mixed-Matrix Membrane for Water Treatment Applications ACS Appl. Mater. Interfaces 2020 12 27625 27631 10.1021/acsami.0c06399 32496035 9. Zoromba M.S. Ismail M.I. Bassyouni M. Abdel-Aziz M.H. Salah N. Alshahrie A. Memic A. Fabrication and characterization of poly (aniline-co-o-anthranilic acid)/magnetite nanocomposites and their application in wastewater treatment Colloids Surf. A Physicochem. Eng. Asp. 2017 520 121 130 10.1016/j.colsurfa.2017.01.075 10. Bassyouni M. Abdel-Aziz M.H. Zoromba M.S. Abdel-Hamid S.M.S. Drioli E. A review of polymeric nanocomposite membranes for water purification J. Ind. Eng. Chem. 2019 73 19 46 10.1016/j.jiec.2019.01.045 11. Elrasheedy A. Nady N. Bassyouni M. El-Shazly A. Metal Organic Framework Based Polymer Mixed Matrix Membranes: Review on Applications in Water Purification Membranes 2019 9 88 10.3390/membranes9070088 31330993 12. Abdel-Aziz M.H. El-Ashtoukhy E.-S. Zoromba M.S. Bassyouni M. Sedahmed G. Removal of nitrates from water by electrocoagulation using a cell with horizontally oriented Al serpentine tube anode J. Ind. Eng. Chem. 2020 82 105 112 10.1016/j.jiec.2019.10.001 13. Bassyouni M. Mansi A.E. Elgabry A. Ibrahim B.A. Kassem O.A. Alhebeshy R. Utilization of carbon nanotubes in removal of heavy metals from wastewater: A review of the CNTs’ potential and current challenges Appl. Phys. A 2019 126 38 10.1007/s00339-019-3211-7 14. Eteba A. Bassyouni M. Saleh M. Removal of Hazardous Organic Pollutants Using Fly Ash Environ. Ecol. Res. 2021 9 196 203 10.13189/eer.2021.090407 15. Qiao L. Wang S. Wang T. Yu S. Guo S. Du K. High-strength and low-swelling chitosan/cellulose microspheres as a high-efficiency adsorbent for dye removal Cellulose 2021 28 9323 9333 10.1007/s10570-021-04111-2 16. Chatterjee S. Chatterjee T. Woo S.H. Adsorption of Congo Red from Aqueous Solutions Using Chitosan Hydrogel Beads Formed by Various Anionic Surfactants Sep. Sci. Technol. 2011 46 986 996 10.1080/01496395.2010.538034 17. Staroń P. Chwastowski J. Raphia-Microorganism Composite Biosorbent for Lead Ion Removal from Aqueous Solutions Materials 2021 14 7482 10.3390/ma14237482 34885636 18. Staroń P. Płecka A. Chwastowski J. Lead Sorption by Chrysanthemum indicum: Equilibrium, Kinetic, and Desorption Studies Water Air Soil Pollut. 2021 232 1 14 10.1007/s11270-020-04956-6 19. Staroń P. Sorys P. Chwastowski J. Equilibrium and Kinetic Study of Ammonium Sorption by Raphia farinifera Water Air Soil Pollut. 2019 230 243 10.1007/s11270-019-4301-z 20. Ibrahim N.A. Abdellatif F.H. Hasanin M.S. Abdellatif M.M. Fabrication, characterization, and potential application of modified sawdust sorbents for efficient removal of heavy metal ions and anionic dye from aqueous solutions J. Clean. Prod. 2021 332 130021 10.1016/j.jclepro.2021.130021 21. Mateo S. Peinado S. Morillas-Gutiérrez F. La Rubia M.D. Moya A.J. Nanocellulose from Agricultural Wastes: Products and Applications—A Review Processes 2021 9 1594 10.3390/pr9091594 22. Hashem A.H. Hasanin M. Kamel S. Dacrory S. A new approach for antimicrobial and antiviral activities of biocompatible nanocomposite based on cellulose, amino acid and graphene oxide Colloids Surfaces B Biointerfaces 2021 209 112172 10.1016/j.colsurfb.2021.112172 34715596 23. Youssef A. Hasanin M. El-Aziz M.A. Turky G. Conducting chitosan/hydroxylethyl cellulose/polyaniline bionanocomposites hydrogel based on graphene oxide doped with Ag-NPs Int. J. Biol. Macromol. 2020 167 1435 1444 10.1016/j.ijbiomac.2020.11.097 33202266 24. Hammad A.A. El-Aziz M.A. Hasanin M. Kamel S. A novel electromagnetic biodegradable nanocomposite based on cellulose, polyaniline, and cobalt ferrite nanoparticles Carbohydr. Polym. 2019 216 54 62 10.1016/j.carbpol.2019.03.038 31047082 25. Dacrory S. Hashem A.H. Hasanin M. Synthesis of cellulose based amino acid functionalized nano-biocomplex: Characterization, antifungal activity, molecular docking and hemocompatibility Environ. Nanotechnol. Monit. Manag. 2021 15 100453 10.1016/j.enmm.2021.100453 26. Abitbol T. Rivkin A. Cao Y. Nevo Y. Abraham E. Ben-Shalom T. Lapidot S. Shoseyov O. Nanocellulose, a tiny fiber with huge applications Curr. Opin. Biotechnol. 2016 39 76 88 10.1016/j.copbio.2016.01.002 26930621 27. Morais J.P.S. Rosa M.F. Filho M.M.S. Nascimento L.D. Nascimento D.M. Cassales A.R. Extraction and characterization of nanocellulose structures from raw cotton linter Carbohydr. Polym. 2013 91 229 235 10.1016/j.carbpol.2012.08.010 23044127 28. Abdul Rahman N.H. Chieng B.W. Ibrahim N.A. Abdul Rahman N. Extraction and characterization of cellulose nanocrystals from tea leaf waste fibers Polymers 2017 9 588 10.3390/polym9110588 29. Iskandar M.A. Yahya E.B. Khalil H.P.S.A. Rahman A.A. Ismail M.A. Recent Progress in Modification Strategies of Nanocellulose-Based Aerogels for Oil Absorption Application Polymers 2022 14 849 10.3390/polym14050849 35267674 30. Beaumont M. Jahn E. Mautner A. Veigel S. Böhmdorfer S. Potthast A. Gindl-Altmutter W. Rosenau T. Facile Preparation of Mechanically Robust and Functional Silica/Cellulose Nanofiber Gels Reinforced with Soluble Polysaccharides Nanomaterials 2022 12 895 10.3390/nano12060895 35335708 31. Morcillo-Martín R. Espinosa E. Rabasco-Vílchez L. Sanchez L.M. de Haro J. Rodríguez A. Cellulose Nanofiber-Based Aerogels from Wheat Straw: Influence of Surface Load and Lignin Content on Their Properties and Dye Removal Capacity Biomolecules 2022 12 232 10.3390/biom12020232 35204733 32. Abdelraof M. El Saied H. Hasanin M.S. Green Immobilization of Glucanobacter xylinum onto Natural Polymers to Sustainable Bacterial Cellulose Production Waste Biomass-Valorization 2022 13 2053 2069 10.1007/s12649-021-01666-w 33. Ouellette R.J. Rawn J.D. Organic Chemistry: Structure, Mechanism, Synthesis Academic Press Cambridge, MA, USA 2018 34. Nia M.H. Tavakolian M. Kiasat A.R. van de Ven T.G. Hybrid Aerogel Nanocomposite of Dendritic Colloidal Silica and Hairy Nanocellulose: An Effective Dye Adsorbent Langmuir 2020 36 11963 11974 10.1021/acs.langmuir.0c02090 32937066 35. Mok C.F. Ching Y.C. Abu Osman N.A. Muhamad F. Hai N.D. Choo J.H. Hassan C.R. Adsorbents for removal of cationic dye: Nanocellulose reinforced biopolymer composites J. Polym. Res. 2020 27 1 15 10.1007/s10965-020-02347-3 36. Vivod V. Neral B. Mihelič A. Kokol V. Highly efficient film-like nanocellulose-based adsorbents for the removal of loose reactive dye during textile laundering Text. Res. J. 2018 89 975 988 10.1177/0040517518760752 37. Putro J.N. Santoso S.P. Soetaredjo F.E. Ismadji S. Ju Y.-H. Nanocrystalline cellulose from waste paper: Adsorbent for azo dyes removal Environ. Nanotechnol. Monit. Manag. 2019 12 100260 10.1016/j.enmm.2019.100260 38. Jamshaid A. Hamid A. Muhammad N. Naseer A. Ghauri M. Iqbal J. Shah N.S. Cellulose-based Materials for the Removal of Heavy Metals from Wastewater–An Overview ChemBioEng Rev. 2017 4 240 256 10.1002/cben.201700002 39. Marotta A. Luzzi E. Salzano de Luna M. Aprea P. Ambrogi V. Filippone G. Chitosan/Zeolite Composite Aerogels for a Fast and Effective Removal of Both Anionic and Cationic Dyes from Water Polymers 2021 13 1691 10.3390/polym13111691 34067280 40. Lopez-Barbosa N. Campaña A.L. Cruz J.C. Ornelas-Soto N. Osma J.F. Enhanced Catalytic Dye Decolorization by Microencapsulation of Laccase from P. Sanguineus CS43 in Natural and Synthetic Polymers Polymers 2020 12 1353 10.3390/polym12061353 41. Chiu C.-W. Wu M.-T. Lee J.C.-M. Cheng T.-Y. Isothermal Adsorption Properties for the Adsorption and Removal of Reactive Blue 221 Dye from Aqueous Solutions by Cross-Linked β-Chitosan Glycan as Acid-Resistant Adsorbent Polymers 2018 10 1328 10.3390/polym10121328 42. Sayyouh M.H. Al-Blehed M.S. Attia A.M. The effect of alkaline and polymer additives on phase behaviour of surfactant-oil-brine system at high salinity conditions Rev. De L’institut Français Du Pétrole 1993 48 359 369 10.2516/ogst:1993023 43. Hidayat E. Harada H. Mitoma Y. Yonemura S. A Halem H.I. Rapid Removal of Acid Red 88 by Zeolite/Chitosan Hydrogel in Aqueous Solution Polymers 2022 14 893 10.3390/polym14050893 35267716 44. Kyzas G.Z. Bikiaris D.N. Mitropoulos A. Chitosan adsorbents for dye removal: A review Polym. Int. 2017 66 1800 1811 10.1002/pi.5467 45. Sakkayawong N. Thiravetyan P. Nakbanpote W. Adsorption mechanism of synthetic reactive dye wastewater by chitosan J. Colloid Interface Sci. 2005 286 36 42 10.1016/j.jcis.2005.01.020 15848400 46. Pang Y.L. Tan J.H. Lim S. Chong W.C. A State-of-the-Art Review on Biowaste Derived Chitosan Biomaterials for Biosorption of Organic Dyes: Parameter Studies, Kinetics, Isotherms and Thermodynamics Polymers 2021 13 3009 10.3390/polym13173009 34503049 47. Alsamman M.T. Sánchez J. Chitosan- and Alginate-Based Hydrogels for the Adsorption of Anionic and Cationic Dyes from Water Polymers 2022 14 1498 10.3390/polym14081498 35458248 48. Sánchez-Duarte R.G. López-Cervantes J. Sánchez-Machado D.I. Correa-Murrieta M.A. Núñez-Gastélum J.A. Rodríguez-Núñez J.R. Chitosan-based adsorbents gels for the removal of tris-azo dye: Isotherms and kinetics studies Environ. Eng. Manag. J. 2016 15 2469 2478 49. Pal P. Pal A. Dye removal using waste beads: Efficient utilization of surface-modified chitosan beads generated after lead adsorption process J. Water Process Eng. 2019 31 100882 10.1016/j.jwpe.2019.100882 50. Abd-Elmageed A.A.I. Ibrahim S.M. Bourezgui A. Al-Hossainy A.F. Synthesis, DFT studies, fabrication, and optical characterization of the [ZnCMC] TF polymer (organic/inorganic) as an optoelectronic device New J. Chem. 2020 44 8621 8637 10.1039/D0NJ01719A 51. Arletti R. Quartieri S. Vezzalini G. Elastic behavior of zeolite boggsite in silicon oil and aqueous medium: A case of high-pressure-induced over-hydration Am. Miner. 2010 95 1247 1256 10.2138/am.2010.3482 52. Santmartí A. Lee K.-Y. Crystallinity and Thermal Stability of Nanocellulose Nanocellulose and Sustainability: Production, Properties, Applications, and Case Studies CRC Press Boca Raton, FL, USA 2018 67 86 10.1201/9781351262927-5 53. Jin L. Sun Q. Xu Q. Xu Y. Adsorptive removal of anionic dyes from aqueous solutions using microgel based on nanocellulose and polyvinylamine Bioresour. Technol. 2015 197 348 355 10.1016/j.biortech.2015.08.093 26344242 54. Yang H. Yan R. Chen H. Lee D.H. Zheng C. Characteristics of hemicellulose, cellulose and lignin pyrolysis Fuel 2007 86 1781 1788 10.1016/j.fuel.2006.12.013 55. Rebouillat S. Pla F. State of the Art Manufacturing and Engineering of Nanocellulose: A Review of Available Data and Industrial Applications J. Biomater. Nanobiotechnol. 2013 4 165 188 10.4236/jbnb.2013.42022 56. Movaghgharnezhad S. Mirabi A. Toosi M.R. Rad A.S. Synthesis of cellulose nanofibers functionalized by dithiooxamide for preconcentration and determination of trace amounts of Cd(II) ions in water samples Cellulose 2020 27 8885 8898 10.1007/s10570-020-03388-z 57. Evdokimova O.L. Evdokimova A.V. Kusova T.V. Agafonov A.V. Preparation of hybrid nanocomposites based on nanoscale cellulose and magnetic nanoparticles with photocatalytic properties Inorg. Mater. Appl. Res. 2020 11 371 376 10.1134/S2075113320020100 58. Tavakolian M. Wiebe H. Sadeghi M.A. Van De Ven T.G. Dye Removal Using Hairy Nanocellulose: Experimental and Theoretical Investigations ACS Appl. Mater. Interfaces 2019 12 5040 5049 10.1021/acsami.9b18679 59. Li M. Wang Z. Li B. Adsorption behaviour of congo red by cellulose/chitosan hydrogel beads regenerated from ionic liquid Desalination Water Treat. 2016 57 16970 16980 10.1080/19443994.2015.1082945 60. Wang Y. Zhao L. Peng H. Wu J. Liu Z. Guo X. Removal of Anionic Dyes from Aqueous Solutions by Cellulose-Based Adsorbents: Equilibrium, Kinetics, and Thermodynamics J. Chem. Eng. Data 2016 61 3266 3276 10.1021/acs.jced.6b00340 61. Khalil H.A. Yusra A.I. Bhat A.H. Jawaid M. Cell wall ultrastructure, anatomy, lignin distribution, and chemical composition of Malaysian cultivated kenaf fiber Ind. Crops Prod. 2010 31 113 121 10.1016/j.indcrop.2009.09.008 62. Fernandes Queiroz M. Melo K.R.T. Sabry D.A. Sassaki G.L. Rocha H.A.O. Does the Use of Chitosan Contribute to Oxalate Kidney Stone Formation? Mar. Drugs 2014 13 141 158 10.3390/md13010141 25551781 63. Chen Y.W. Lee H.V. Hamid S.B.A. Preparation of Nanostructured Cellulose via Cr(III)- and Mn(II)-Transition Metal Salt Catalyzed Acid Hydrolysis Approach BioResources 2016 11 7224 7241 10.15376/biores.11.3.7224-7241 64. Lamaming J. Hashim R. Leh C.P. Sulaiman O. Sugimoto T. Nasir M. Isolation and characterization of cellulose nanocrystals from parenchyma and vascular bundle of oil palm trunk (Elaeis guineensis) Carbohydr. Polym. 2015 134 534 540 10.1016/j.carbpol.2015.08.017 26428155 65. Morán J.I. Alvarez V.A. Cyras V.P. Vázquez A. Cyras, Analia Va’zquez, extraction of cellulose and preparation of nanocellulose from sisal fibers Cellulose 2008 15 149 159 10.1007/s10570-007-9145-9 66. Mazlita Y. Lee H.V. Hamid S.B.A. Hamid, Preparation of Cellulose Nanocrystals Bio-Polymer From Agro-Industrial Wastes: Separation and Characterization Polym. Polym. Compos. 2016 24 719 728 67. Nasir M. Hashim R. Sulaiman O. Nordin N.A. Lamaming J. Asim M. Laccase, an Emerging Tool to Fabricate Green Composites: A Review BioResources 2015 10 6262 6284 10.15376/biores.10.3.Nasir 68. Abraham E. Deepa B. Pothan L.A. Jacob M. Thomas S. Cvelbar U. Anandjiwala R. Extraction of nanocellulose fibrils from lignocellulosic fibres: A novel approach Carbohydr. Polym. 2011 86 1468 1475 10.1016/j.carbpol.2011.06.034 69. Hussain A. Farooq A. Bassyouni M.I. Sait H.H. El-Wafa M.A. Hasan S.W. Ani F.N. Pyrolysis of Saudi Arabian date palm waste: A viable option for converting waste into wealth Life Sci. J. 2014 11 667 671 10.7537/marslsj111214.126 70. Gutub S.A. Bassyouni M. Dissolved solids adsorption of freshwater using synthesized bio-foam composite Life Sci. J. 2013 10 464 471 71. Murcia-Salvador A. Pellicer J.A. Fortea M.I. Gómez-López V.M. Rodríguez-López M.I. Núñez-Delicado E. Gabaldón J.A. Adsorption of Direct Blue 78 Using Chitosan and Cyclodextrins as Adsorbents Polymers 2019 11 1003 10.3390/polym11061003

PMC009xxxxxx/PMC9099638.txt

==== Front Cancers (Basel) Cancers (Basel) cancers Cancers 2072-6694 MDPI 10.3390/cancers14092054 cancers-14-02054 Article A Six-Gene Prognostic and Predictive Radiotherapy-Based Signature for Early and Locally Advanced Stages in Non-Small-Cell Lung Cancer Peinado-Serrano Javier 123 Quintanal-Villalonga Álvaro 4 Muñoz-Galvan Sandra 12 https://orcid.org/0000-0002-0922-0371 Verdugo-Sivianes Eva M. 12 Mateos Juan C. 56 Ortiz-Gordillo María J. 3 https://orcid.org/0000-0003-4357-3979 Carnero Amancio 12* Alfieri Roberta Academic Editor 1 Instituto de Biomedicina de Sevilla, IBIS, Hospital Universitario Virgen del Rocío, Consejo Superior de Investigaciones Científicas, Universidad de Sevilla, Avda. Manuel Siurot s/n, 41013 Seville, Spain; jvrr18@gmail.com (J.P.-S.); smunoz-ibis@us.es (S.M.-G.); everdugo-ibis@us.es (E.M.V.-S.) 2 CIBERONC, Instituto de Salud Carlos III, 28029 Madrid, Spain 3 Department of Radiation Oncology, Hospital Universitario Virgen del Rocío, Avda. Manuel Siurot s/n, 41013 Seville, Spain; mjortizgordillo@yahoo.es 4 Molecular Pharmacology Program, Memorial Sloan Kettering Cancer Center, New York, NY 10065, USA; quintaa1@mskcc.org 5 Radiation Physics Department, Hospital Universitario Virgen del Rocío, Avda. Manuel Siurot s/n, 41013 Seville, Spain; jcmateos@us.es 6 Departamento de Fisiología Médica y Biofisica, Universidad de Sevilla, 41013 Seville, Spain * Correspondence: acarnero-ibis@us.es 19 4 2022 5 2022 14 9 205428 3 2022 13 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Simple Summary The search for prognostic and/or predictive gene signatures of the response to radiotherapy treatment can significantly aid clinical decision making. These signatures can condition the fractionation, the total dose to be administered, and/or the combination of systemic treatments and radiation. The ultimate goal is to achieve better clinical results, as well as to minimize the adverse effects associated with current cancer therapies. To this end, we analyzed the intrinsic radiosensitivity of 15 NSCLC lines and found the differences in gene expression levels between radiosensitive and radioresistant lines, resulting in a potentially applicable six-gene signature in NSCLC patients. The six-gene signature had the ability to predict overall survival and progression-free survival (PFS), which could translate into a prediction of the response to the cancer treatment received. Abstract Non-small-cell lung cancer (NSCLC) is the leading cause of cancer death worldwide, generating an enormous economic and social impact that has not stopped growing in recent years. Cancer treatment for this neoplasm usually includes surgery, chemotherapy, molecular targeted treatments, and ionizing radiation. The prognosis in terms of overall survival (OS) and the disparate therapeutic responses among patients can be explained, to a great extent, by the existence of widely heterogeneous molecular profiles. The main objective of this study was to identify prognostic and predictive gene signatures of response to cancer treatment involving radiotherapy, which could help in making therapeutic decisions in patients with NSCLC. To achieve this, we took as a reference the differential gene expression pattern among commercial cell lines, differentiated by their response profile to ionizing radiation (radiosensitive versus radioresistant lines), and extrapolated these results to a cohort of 107 patients with NSCLC who had received radiotherapy (among other therapies). We obtained a six-gene signature (APOBEC3B, GOLM1, FAM117A, KCNQ1OT1, PCDHB2, and USP43) with the ability to predict overall survival and progression-free survival (PFS), which could translate into a prediction of the response to the cancer treatment received. Patients who had an unfavorable prognostic signature had a median OS of 24.13 months versus 71.47 months for those with a favorable signature, and the median PFS was 12.65 months versus 47.11 months, respectively. We also carried out a univariate analysis of multiple clinical and pathological variables and a bivariate analysis by Cox regression without any factors that substantially modified the HR value of the proposed gene signature. NSCLC biomarkers radiation oncology prognosis predictive signature ==== Body pmc1. Introduction Lung cancer greatly influences the lives of patients and healthcare systems since it has the largest incidence and a nearly 95% mortality worldwide [1]. Based on its histological characteristics, it is divided into two large groups: small cell lung cancer (small-cell) and non-small-cell lung cancer (NSCLC). The latter accounts for approximately 85–90% of the total. In turn, this group is subdivided, according to histological and molecular characteristics, into the following: adenocarcinomas (the majority), squamous, large cells, neuroendocrine, and not otherwise specified. Currently, three clinical scenarios are considered regarding cancer treatment in newly diagnosed patients: resectable, locally advanced (not resectable), and metastatic. As shown in the main international therapeutic guidelines [2], radiotherapy treatment plays an important role in all settings. It is an alternative to surgical treatment in early stages (Stages I and II without nodal load) using stereotactic body radiotherapy (SBRT) and is complementary after surgical resections with affected edges and/or a positive nodal load (N2). In locally advanced stages (Stage III), where surgical resection is not possible, normofractionated radiotherapy treatment together with chemotherapy (concomitant or sequential) are the therapeutic standard. In patients with 1–5 thoracic and/or extrathoracic lesions, the option of local treatment with radical intention is considered an effective alternative. Last, most patients with metastatic spread in the central nervous system receive protocolized radiotherapy treatment if their health allows it. Focusing on unresectable locally advanced stages (Stages IIIA–C), there is currently no standard radiochemotherapy regimen, although the combination of a platinum-based regimen and chest radiotherapy has significantly improved the survival of these patients. These patients are treated with standard radiation doses of 60–66 Gy concomitantly, or sequentially, with combined chemotherapy [3,4,5]. Fractions can vary, but generally 1.8–2 Gy/fraction/day is used (normofraction). Despite the application of a combined cytotoxic treatment, we continue to observe local relapse rates of 30–50% in this group of patients [6,7]. This fact justifies, by itself, the need to continue delving into the biological keys that govern the poor clinical results obtained to date. Research focused on improving antineoplastic therapies in lung cancer has been based on the genomic and proteomic study of tumors with a specific known genetic basis, such as EGFR and KRAS mutations. This molecular classification influences the response to biological therapies based on monoclonal antibodies and tyrosine kinase inhibitors in patients with lung cancer [8,9,10]. On the other hand, a nonuniform response of patients to these therapies has been observed, which suggests a resistance model that could be mediated by other mutations in some relevant genes (insertions in EGFR [11] or KRAS [12], amplification of MET [13,14] or mutations in the HER-242 kinase domain). We also know that the tumor genetic profile has a relevant impact on the response to chemotherapy [15] and radiotherapy [16,17]. There are currently several studies that relate some of the mentioned mutations to the mechanisms of radioresistance or radiosensitivity in NSCLC [17,18]. However, today, there are still no biomarkers to consider as a condition for radiotherapy treatment to be administered, despite knowledge of the effects that some mutations could have on the response to ionizing radiation. Thus, the classification of NSCLC based exclusively on the clinicopathological characteristics was the only determinant of the therapy administered. The availability of biological material and advances in transcriptomic analysis techniques have made it possible to improve the subclassification of the group of neoplasms encompassed within NSCLC [19,20,21,22]. Based mainly on a transcriptomic analysis, many studies have proposed different gene signatures in adenocarcinoma [23,24,25,26,27,28,29,30,31], squamous cell carcinoma [32,33,34], or NSCLC in general [35,36,37,38,39,40,41,42,43,44,45,46]. Some of these studies have attempted to identify the prognostic and predictive biomarkers of the response to systemic treatments. Most focused on the identification of markers that help in clinical decision making, regarding the suitability of administering adjuvant systemic treatment in the early stages of NSCLC after surgery [47,48]. In contrast, very few published studies suggest predictive signatures of response to ionizing radiation in NSCLC. Thus, based on the research work carried out by the American National Cancer Institute (NCI)—in vitro analyses of the biological effects of different antineoplastic drugs in cell lines [49,50,51,52,53,54,55]—Scott et al. published, in 2017, a study proposing a model to adapt the radiotherapy prescription to the individual sensitivity of each patient’s tumor. The model, called GARD (genome-based model for adjusting radiotherapy dose), combines information derived from the radiosensitivity index (RSI) and the linear quadratic model (LQ model) [56]. On the other hand, some of the miRNAs identified to date have been considered diagnostic and prognostic biomarkers in numerous neoplastic entities, including NSCLC [57,58]. Sun et al. [59] proposed a total of 11 miRNAs, which, together with variables such as stage, age, radiation dose administered, systemic treatment, and Karnofsky general state scale, were used to determine the DFS of each patient, but it was not possible to demonstrate its predictive capacity on the local control of the disease in a statistically significant way [59]. Finally, there are other publications where the generation of prognostic signatures is considered through the joint consideration of multiple nonspecific biomarkers of lung cancer and multiple clinical variables [60]. Given the limited scientific evidence published, we consider that the search for prognostic and/or predictive gene signatures of the response to radiotherapy treatment can significantly help with clinical decision making. These signatures can condition the fractionation, the total dose to be administered, and/or the combination of systemic treatments and radiation. The ultimate goal is to achieve better clinical results, as well as to minimize the adverse effects associated with current cancer therapies. To this end, we carried out a basic translational study, which, by analyzing the intrinsic radiosensitivity of 15 NSCLC lines, allowed us to obtain the differences in gene expression levels between radiosensitive and radioresistant lines, resulting in potentially applicable gene signatures in NSCLC patients. 2. Materials and Methods 2.1. Cell Lines Fifteen NSCLC tumor cell lines from the American Type Culture Collection (ATCC) were used. Three corresponded to SCC (H226, H520, and Calu 1), two to bronchioloalveolar carcinomas (H1437 and H358), one to large cell carcinoma (H460), and nine were adenocarcinomas (A549, Calu-3, H1650, H1781, H1975, H2009, H2228, H3122, and HCC827). A RPMI culture medium with FBS and antibiotic was used, except for A549 (DMEM+sodium pyruvate, HEPES, and nonessential amino acids), Calu 1 (McCay + glutamine supplement), and Calu 3 (DMEM), according to the ATCC’s recommendations. Table S1 summarizes the histological information and molecular profile of some relevant NSCLC mutations of the cell lines used. 2.2. Clonogenicity Test for the Response to Ionizing Radiation by Determining the SF2 Parameter We seeded the cells in six-well plates and allowed them to grow until they reached 70–90% confluence. Then, we irradiated the cells in a linear photon accelerator of 6 Mv of energy at a rate of 400 monitor units (MU) per minute. After changing the medium and incubating for 36–48 h, we counted the cells and seeded them in 10-cm-diameter culture plates, between 300 and 2000 cells per plate, in triplicate for each dose of radiation used (0, 2, 4, 6, and 8 Gy). We incubated the cells for a period of 7–25 days in a standard atmosphere to allow the formation of colonies. We scanned the plates and counted the number of CFUs using the ImageJ® (NIH, Bethesda, MD, USA) computer program. The number of CFUs was normalized to the number of cells seeded in each case. Calculation of the radiobiological parameters was conducted as follows: (a) plating efficiency (PE): PE = number of colonies counted/number of cells seeded; and (b) surviving fraction at 2 Gy (SF2): SF2 = (number of colonies counted/number of cells seeded)/PE 9. The preparation of survival curves was based on the values of the surviving fraction in the control plates and those treated with 2, 4, 6, and 8 Gy. 2.3. RNA Extraction from Cells A commercial miRNA Extraction Kit (miRNeasy® Mini Kit (Qiagen, Barcelona, Spain)) was used following the manufacturer’s instructions. 2.4. cDNA Microarrays The RNA obtained was subjected to a reverse-transcription reaction in the presence of six nucleotide random primers (Invitrogen/Life Technologies, Carlsbad, CA, USA) using Cy3- or Cy5-labeled dCTPs (GE Healthcare, Chicago, IL, USA) and MultiScribeTM RT (Thermo Fisher Scientific, Waltham, MA, USA). RNA from normal lung epithelial cells was used as a reference. Equal amounts of cDNA were used to make the microarray. Prehybridization was performed in a humidified chamber at 42 °C for 16–20 h, and the hybridization was performed at 65 °C in a GeneTac Hybridization Station (Genomics Solutions, Oberhaching, Germany). Hybridized samples were analyzed with a confocal scanner (ArrayExpress, PerkinElmer, Waltham, MA, USA), and the data were quantified using QuantArray software (PerkinElmer, Waltham, MA, USA). The level of significance of the expression of each gene was determined using a Student’s t-test (10,000 permutations), using, as a correction, the false discovery rate (FDR) test to eliminate false positives. 2.5. Transcriptome Analysis After image acquisition and quantification, the mean signal intensity between replicates was determined for each sequence. At QT-02 (replicates 3–6 times represent the same gene), values with low signal intensity and low reproducibility between repeats were excluded (mean ± 2 standard deviations cutoff). The quantized signals were subjected to logarithmic and standardized transformation. We used a Student’s t-test of permutation (10,000 permutations) to determine the level of significance of the expression of each individual gene, and the false-positive rate was used as a correction for multiple analyses. Hierarchical cluster analysis and clustering reliability were evaluated using bootstrap techniques employing TMEV software (NIH, Bethesda, MD, USA). We identified altered mRNAs in cells that were resistant versus nonresistant to radiotherapy using targeted clusters, with subsequent enrichment of pathways. This process was carried out with the help of the bioinformatics service of the Institute of Biomedicine of Seville (IBIS). 2.6. Patient Cohort The publicly available Cancer Genome Atlas (TCGA) cohort focused on patients with NSCLC (version 2018) was used. Stage I to III patients were selected from this cohort based on TNM classification (tumor, lymph nodes, metastasis) at the time of diagnosis. The main requirement for the selection was to have received radiotherapy treatment. For the analysis of the impact of clinical/pathological variables on OS and PFS, univariate analysis by Cox regression was performed. Subsequently, we carried out a bivariate analysis with the variable “gene signature” as the main element and considered the rest of the variables as independent to identify any modifying relationship or confounding variables, which would condition the impact of the gene signature on the OS or SLP. We used the statistical package SPSS® version 20 (IBM, New York, NY, USA). To obtain the gene signatures, RNA sequencing data of tumor samples from TCGA patients were used, whose expression values were obtained as transcripts per million (TPM) and were subsequently transformed on a logarithmic scale. 2.7. Statistical Packages and Analysis Survival analysis was performed using R software version 3.6.3 (San Francisco, CA, USA, 29 February 2020) with the Survival package (version 3.2.7) for Cox regression, Survminer (version 0.4.6) for Kaplan–Meier curves, Glmnet (version 3.0.2) for cross-validation and obtaining signatures and SurvivalROC (version 1.0.3) for ROC curves. After identifying the differentially expressed genes between radiosensitive and radioresistant cell lines, we made use of RNA-seq data from the selected TCGA cohort. Next, a univariate Cox regression survival analysis was carried out. After selecting those genes whose impact on survival was statistically significant (p < 0.05), a Kaplan–Meier analysis of survival estimation was carried out, establishing the cutoff value of the median expression of each gene. We assessed differences in OS/PFS with a log-rank test. We selected those genes that were statistically significant in both analyses (p < 0.05) as potentially relevant elements for OS or PFS. 2.8. Obtaining Gene Signatures From the genes selected as individual elements with prognostic and/or predictive impact (HR with p < 0.05 and log-rank test with p < 0.05), we proceeded with a multivariate Cox regression analysis with L1 or Lasso type regularization. To do this, we separated the cohort of 107 TCGA patients into (1) a training cohort and (2) a trial cohort. We next carried out a cross-validation analysis of three iterations, randomly separating, for a maximum of 1000 times, 66.6% of the cohort for the training cohort and the remaining 33.3% for the test set. The prognostic and predictive signatures were constructed based on the linear combination of the regression coefficient obtained from the coefficients derived from the Cox regression analysis with L1 (β)-type regularization multiplied by the expression level of each gene (in units of transcripts per million). In this case, both the training and trial cohorts came from the same selected TCGA cohort. To calculate the survival prediction, we performed ROC curve (at 12, 24, and 60 months) and Kaplan–Meier analyses. Finally, we performed a Mann–Whitney U test to determine whether the differences in the expression of the genes of each signature in each risk group were statistically significant. Next, we separated the cohort based on histological type (adenocarcinoma or squamous cell) and tested the proposed signatures. The algorithm is shown in Figure 1. 3. Results 3.1. Survival Values at 2 Gy and Survival Curves of All Cell Lines We used the clonogenicity test to assess the response to ionizing radiation by determining the SF2 parameter (fraction surviving at 2 Gy) and counting the number of colony-forming units (CFU) at that dose. The mean value of the summation of all the SF2 Gy values of each of the cell lines was used to obtain a mean value (0.54), which we used as a cutoff point on which to establish a classification as sensitive lines (value of SF2 < 0.54) or resistant (SF2 ≥ 0.54) to radiation. This value, as well as the rest of the survival values at the different doses proposed (control, 4, 6, and 8 Gy), allowed us to generate the dose–response curves shown in Figure 2A. Table 1 shows the information on the SF values at 2 Gy of each cell line. The value of the H520 line was extracted from the literature (Table S2). 3.2. Identification of Genes with Differential Expression between Radiosensitive and Radioresistant Lines To identify differentially expressed genes between radiation-resistant and radiation-sensitive lines, we employed a cDNA-based microarray analysis. Our panel of cell lines was mainly composed of the histological subtype of adenocarcinoma, but we also included squamous and large cell lineages. To analyze the differential transcriptional expression in the subset of radioresistant versus radiosensitive cells, we used a supervised grouping. We considered as a cutoff point those with |log2FC| > 1 and p < 0.05. We obtained a total of 127 genes with differential expression, of which 76 were overexpressed and 51 were underexpressed (Table S3). A heatmap was generated to represent the results (Figure 2B). We established four subgroups of genes with differential expression between sensitive and resistant lines (Figure 2B). In gene groups 1, 2, and 4, there was a trend toward higher expression levels in the lines considered radioresistant versus sensitive. In contrast, group 3 showed a trend toward higher levels of expression in radiosensitive lines. 3.3. Individual Validation of Genes with Differential Expression as Prognostic Biomarkers in the TCGA Cohort of Radiation-Treated Patients We used the TCGA public cohort of patients with NSCLC (Pan Cancer Atlas) (2018 update) for the validation of our results. This cohort consisted of adenocarcinoma and squamous NSCLC. The main clinical criteria for the selection of patients were (1) having received radiotherapy treatment and (2) being classified as nonmetastatic at diagnosis. With these inclusion criteria, we selected a total of 107 patients. Table S4 shows the clinical data of the selected patient cohort. The mean age at diagnosis was 63 years (range: 39–86 years). Fifty-nine of the patients were male, and 48 were female. The initial diagnosis of the patients and the treatments were carried out between 1992 and 2013. Regarding the pathological classification by stages, based on the TNM assessment (from the 3rd to the 7th edition, depending on the year of the diagnosis), 24/107 patients were classified as stage I, 33/107 patients as stage II, and 50/107 patients as stage III. Among the latter group, 42/107 patients were classified as stage IIIa, and 7 patients were classified as stage IIIb. Regarding the therapeutic approach, all patients received radiotherapy treatment, without being able to specify whether it was administered as an adjuvant or in the presurgical context or as a radical modality alone or in combination with systemic treatment. At the time of the cohort data collection (TCGA 2018 update), 57/107 patients had died and 68/107 patients had experienced progression of their disease. The median PFS, cancer-specific survival (CSS), and OS were 22.9 months (range: 0.72–140), 41.6 months (range: 0.72–140), and 32.7 months (range: 0.72–140), respectively (Figure S1). OS, PFS, and CSS at 1, 2, and 5 years were 80%, 66%, and 33%; 67.5%, 50%, and 21%; and 82.5%, 72.5%, and 44%, respectively. To determine whether there are factors that affect OS and PFS, a series of clinical and pathological variables were considered to analyze their influence on these parameters. To facilitate the statistical analysis, the variables were dichotomized. These variables were sex (male versus female); age at diagnosis (less than or greater than 63 years); year of diagnosis (before or after 2008); histological subtype (adenocarcinoma versus squamous cell); T component (T1 and T2 versus T3 and T4); nodal load (n0 and n1 versus n2 and n3), and tumor stage (stage I and II versus stage III). Likewise, the gene signature was dichotomized into low- and high-risk patients based on whether the value of the individual signature was less than or greater than/equal to the median of the values of the established signature. The information regarding the estimation of OS and PFS as a function of the variables described above is shown in Table S5. A Kaplan–Meier analysis was used, and the curves were compared by means of the logarithmic rank test, considering a p-value of <0.05 as statistically significant. As shown, only patients with a high-risk versus a low-risk gene signature showed statistically significant differences in OS. Regarding PFS, the histological adenocarcinoma subtype and the proposed high-risk gene signature were statistically significantly associated with worse PFS. From the bioinformatics exploration of the results obtained considering all cell lines, a total of 127 genes were obtained, with differential expression between radiosensitive and radioresistant lines (absolute value of log2FC > 1 and adjusted p-value < 0.05) (Table S3). When analyzing the RNAseq data of the 127 genes reflected above for the selected TCGA cohort, we identified a lack of information for six of them (C4orf32, KIAA1324L, LINC00597, LOC100287896, SELENOP, and TMEM133). Thus, considering the expression levels of the 121 remaining genes, a survival analysis was carried out using univariate Cox regression, through which we obtained a total of 21 genes whose expression levels were significantly related to OS in the selected patient cohort (Table S6). Focusing on the previously exposed genes, we carried out an analysis of survival estimation using Kaplan–Meier curves, taking the median gene expression as the cutoff point to divide the samples into high and low expression levels. A total of 10 genes of the 21 identified above yielded a value of p < 0.05 by the log-rank test (Figure 3A). Thus, the genes FAM117A, KCNQ1OT1, KLHL24, SDR16C5, USP43, RHOBTB3, PXYLP1, APOBEC3B, PCDHB2, and GOLM1 achieved statistical significance in the univariate Cox regression analysis and in the OS estimation by Kaplan–Meier and SPL (PFS) by Kaplan–Meier. 3.4. Obtaining the Prognostic Signature We selected the 10 genes identified in the previous section as potential candidates to form part of the prognostic signature. We then performed a multiple Cox regression with a three-iteration cross-validation in which 2/3 of the data (71 patients) were used as the training cohort and 1/3 (36 patients) were used as the trial cohort. We also applied Lasso regularization (L1 type) to the regression to select the genes that influence survival in a more significant way. We obtained coefficients (betas in the Cox regression) for the genes APOBEC3B, GOLM1, FAM117A, KCNQ1OT1, SDR16C5, PCDHB2, RHOBTB3, and USP43. Next, we obtained a signature value for each of the patients, starting from the expression values (logTPM) of each gene of interest multiplied by their respective coefficients. Finally, we performed a univariate Cox regression analysis for all values of the firm (risk scores) as a result of the linear combination of the regression coefficient multiplied by the expression level of each gene in question:Risk Score = (−0.1353493) × Apobec3B expression + 0.0671513 × Golm1 expression + (−0.4002509) × Fam117A expression + 0.6455878 × KCNQ1OT1 expression + 0.0613337 × SDR16C5 expression + 0.0542689 × PCDHB2 expression + 0.0309122 × RHOBTB3 expression + 0.0834290 × USP43 expression. The expression of the gene in each group is shown by the box diagram in Figure 3B. All genes showed a statistically significant differential expression level depending on their relationship with each risk group. Using a univariate Cox regression analysis of the prognostic signature, patients associated with the adverse prognostic group showed a significantly poorer OS than those associated with the favorable group, with an HR of 3.9 (95% CI 2.39–6.37) and a p < 0.0001. The concordance index was 0.713 ± 0.0374. The survival prediction analysis based on the proposed gene signature using ROC curves at 12, 24, and 60 months yielded AUC values of 0.73, 0.72, and 0.79, respectively (Figure S2A). For the Kaplan–Meier analysis (Figure 3C), we considered the median of the risk scores as the cutoff point to separate the two prognostic groups. This value is −0.8. The patients whose gene signature showed expression levels lower than the cutoff point and who had better survival are identified in yellow (median OS of 71.47 months (95% CI 36.7–106.2)), and the patients in which the signature values were equal to or above said point and had a worse prognosis are shown in blue (median OS of 24.13 months (95% CI 11.98–36.27)). 3.5. Individual Validation of Genes with Differential Expression between Radiosensitive and Radioresistant Lines as a Predictive Biomarker of Response in the TCGA Cohort of Interest We considered PFS as a variable for evaluating the effect of cancer treatments in our cohort of interest. As in the OS section, we considered the 127 differentially expressed genes and identified 121 of them in the RNA-seq data from our TCGA cohort. We performed an impact analysis on PFS using the univariate Cox regression test. Table S7 shows the 29 genes whose impact on PFS was statistically significant. Focusing on the previously uncovered genes, we carried out a predictive analysis of SLP using Kaplan–Meier curves, taking the median of the expression of the gene as a cutoff point to divide the samples into high expression levels and low expression levels. Survival curves with p < 0.05 are shown in Figure 4A. Thus, the genes KCNQ1OT1, PCDHB2, GOLM1, USP43, JPH1, ABCC5, PXYLP1, ATP6AP1L, KLHL24, MLF1, APOBEC3B, SDR16C5, TUBB3, BASP1, PAIP2B, HECW2, FAM13B3A, FAM117A, and RHFAM137A individually showed a statistically significant impact on PFS by Cox regression analysis and maintained statistical significance in the estimation of PFS by Kaplan–Meier analysis. We selected these genes to develop the predictive response signature. We carried out a multivariate Cox regression analysis with L1-type regularization with cross-validation of three iterations. Only 7 out of 19 genes were kept in the model summarized in the following formula:Risk score = (−0.1462772) × APOBEC3B expression + 0.1824914 × GOLM1 expression + (−0.1364778) × FAM117A expression + 0.5059698 × KCNQ1OT1 expression + 0.0545114 × PAIP2B expression + 0.0624953 × PCDHB2 expression + 0.0317391 × USP43 expression. We calculated the risk score based on the seven genes of interest in each patient, and then, using the statistical package Survminer R, we identified the cutoff point for the risk score, classifying all patients in a high- or low-risk cohort based on the established cutoff point. The expression values of each gene in each of the two differentiated groups are shown in Figure 4B by means of a box diagram. All the genes of the signature except PAIP2B showed a statistically significant differential expression level depending on their relationship with each prognostic group (Mann–Whitney U test with p < 0.05). Next, a univariate Cox regression analysis of the predictive signature was carried out, obtaining an HR value of 5.04 (95% CI 3–8.47) and p < 0.001. The concordance index was 0.703. Finally, we carried out a prediction analysis of PFS based on the proposed gene signature. To do this, we performed an analysis using ROC curves at 12, 24, and 60 months (and subsequently a Kaplan–Meier analysis, considering the median Figure S2B) of the risk scores as the cutoff point, whose value was +0.72, to separate the two prognostic groups (Figure 4C). Patients whose gene signature showed expression levels lower or higher than this cutoff point obtained a median PFS of 51.55 (95% CI 20.1–83) and 15.45 months (95% CI 10.1–20.8), respectively. 3.6. Generation of a Common Gene Signature for Global Survival Prediction and Progression-Free Survival After obtaining the gene signatures with prognostic and predictive response values, we observed that six gene elements were shared between them (APOBEC3B, GOLM1, FAM117A, KCNQ1OT1, PCDHB2, and USP43), which led us to the generation of a single signature with prognostic ability for OS and PFS. In this case, using the OS data of the cohort, we obtained a risk score summarized in the following formula:Risk score = (−0.1783969) × APOBEC3B expression + 0.0877709 × GOLM1 expression + (−0.4162868) × FAM117A expression + 0.8842986 × KCNQ1OT1 expression + 0. 0859888 × PCDHB2 + 0.1117924 × USP43 expression. We then performed a signature univariate Cox regression analysis for OS and PFS of total NSCLC. As reflected in Table S8, the HRs for OS and PFS associated with the high-risk gene signature were 3.23 and 3.46, with concordance indices of 0.71 and 0.72, respectively. We carried out a prediction capacity analysis of the common signature using ROC curves, with AUCs at 12, 24, and 60 months for OS of 0.69, 0.70 and 0.81 and for PFS of 0.67, 0.78, and 0.79, respectively. For the Kaplan–Meier curves, we established a cutoff point based on the median of the firm’s values (risk score), whose value turned out to be −1.059, and obtained two curves compared by means of the log-rank test. The ROC curves are shown in Figure S3, and the Kaplan–Meier curves are shown in Figure 5A. Finally, we determined the expression levels of each gene based on the risk group considered in the signature (low or high risk). We compared these levels using the Mann–Whitney U test and presented them as box plots (Figure 5B). 3.7. Application of the Gene Signature for Prediction of OS and PFS Depending on the Histological Subtype Given that the proposed gene signature was generated without distinction by main histological subgroup (adenocarcinoma versus squamous cell), we proposed checking whether the signature was valid for each subtype individually. We proceeded methodologically in the same manner as in the previous sections for the application of the signature. Table S8 shows the values of the univariate Cox regression analysis of the common gene signature, distinguishing between adenocarcinoma and squamous cell carcinoma. When comparing the signature values with those obtained in the previous analysis without distinction by histological subtype, we observed similar HR values between them, without losing statistical power. The ROC curves with the AUC values at 12, 24, and 60 months are shown in Figure S4. We appreciate how the predictive capacity of the signature is maintained in both adenocarcinoma and squamous cell carcinoma (AUC values equal to or greater than 0.6 in all cases). Figure 6A shows the Kaplan–Meier curves, with statistically significant separation (log-rank test p < 0.05) between the patients assigned to the low- or high-risk group, separated according to the cutoff point used in the global cohort (−1.05). We also decided to carry out an additional Kaplan–Meier analysis, considering as the cutoff point the median value of the risk scores obtained independently in the adenocarcinoma or squamous carcinoma subsets. Thus, the cutoff points, when considering the median, for the adenocarcinoma cohort and squamous cancer cohort were −0.81 and −1.48, respectively. The KM curves are shown in Figure 6B. The differences in survival estimates for high- and low-risk patients are still statistically significant. Finally, we determined the expression levels of each gene based on the risk group considered in the signature (low or high risk). We compared these levels using the Mann–Whitney U test and presented them as box plots (Figure 6C). 3.8. Univariate and Multivariate Analysis of Clinical and Pathological Variables, Their Impact on Overall Survival and PFS and Their Influence on the Proposed Gene Signature To verify the potential impact of the variables considered on the OS of the patients, we first performed univariate Cox regression analysis, followed by a bivariate analysis to identify the influence of these variables on the impact of the gene signature on OS. In the univariate Cox regression analysis, only the gene signature proposed in our work showed a statistically significant impact on OS, with an HR for the high-risk group of 2.99 (95% CI for HR of 1.72–5.19) and p < 0.0001 (Table 2). The variable “tumor stage,” and, within it, stages I and II, was related to a decrease in the risk of death (HR = 0.60 with 95% CI for HR of 0.35–1.01), with a trend toward statistical significance (p = 0.051). We ruled out the need for a multivariate analysis after performing a bivariate analysis (File S1) where the variable “prognostic gene signature” was established as the main one. None of the variables considered had a statistically significant impact on survival or modified the impact of the prognostic gene signature in the selected cohort. 4. Discussion NSCLC, which encompasses the most frequent histological subtypes of lung cancer, is a neoplastic entity whose biological behavior is highly heterogeneous. Molecular heterogeneity could explain the different responses and behaviors before cancer treatments, as well as the impact of the neoplasm on the OS of patients. Considering this molecular heterogeneity, we believe that the identification of gene signatures with prognostic and predictive capacity of response to cancer treatment, in this case, focused on radiotherapy treatment, could be of great help in optimizing the therapeutic approach for patients. Currently, the tumor staging system by TNM classification continues to be the most powerful instrument for predicting patient survival and so is a focus for the oncology community in terms of therapeutic approaches in the case of NSCLC and most neoplasms [2,61,62]. Despite efforts to obtain clinical, pathological, and/or molecular information that could serve to predict responses to treatment and improve prognostic capacity, there are currently no validated biomarkers in NSCLC that facilitate the oncology community’s decisions regarding individualized treatment selection in the nonmetastatic setting. There are multiple proposals for gene signatures that attempt to predict survival or response to treatment (not radiotherapy], but these mainly focus on the early stages or metastatic stage at diagnosis. In contrast, there are few studies that propose such predictive and/or prognostic elements in locally advanced nonmetastatic stages, where treatment with ionizing radiation plays an important role. The radiation oncologist lacks molecular markers that serve to guide the radiotherapy treatment to be used, beyond the general recommendations derived, for example, from the pathological report, considering the status of the surgical margins or positive nodal load, among others [2,61]. One of the difficulties we face when we propose the identification of predictive and prognostic signatures in NSCLC is an inability to identify whether the clinical, therapeutic, histological, or molecular variables have the same weight when determining the sustained therapeutic response and global survival [63]. The main objective of our study was to develop and validate gene signatures with prognostic and predictive capacity for response to radiotherapy. We found a single signature that was valid as a prognostic element and as a predictor of therapeutic response with six gene elements (APOBEC3B, GOLM1, FAM117A, KCNQ1OT1, PCDHB2 and USP43, File S1). The HR values for OS and PFS associated with the high-risk gene signature were 3.23 and 3.46, with concordance indices of 0.71 and 0.72, respectively (p < 0.0001). In the case of the adenocarcinoma subgroup, the HR of the high-risk gene signature for OS and PFS was 3.52 and 3.76, respectively (p-value < 0.001). In the case of SCC, the values of the signature’s HR for OS and PFS were 3.87 and 3.78, respectively. These results indicate the value of the proposed gene signature regardless of the main histological subtype. There are several signatures reported to be diagnostic or prognostic for lung cancer in the literature [23,45,46,64,65,66] that were obtained by different methods, cohort types, or statistical strategies (File S1. In all cases, the proposed genes that make up the signatures do not coincide among them or with ours, probably due to notable differences in the methodology and materials used in these studies and in ours, differences in the patient cohorts, or the fact that most of the patients did not receive adjuvant treatment, unlike the cohort used in our study. In contrast to the works discussed above, articles published on the identification of predictive or prognostic gene markers focused on cohorts whose main treatment was ionizing radiation are scarcer and practically nonexistent in the specific case of NSCLC. Torres-Roca and collaborators [67] identified genetic elements common to all neoplasms, which could explain the differences in radiosensitivity observed both in vitro and in clinical practice. The generation of the so-called “radiosensitivity index” (RSI) forms a predictive signature of response to radiotherapy treatment composed of the AR, cJUN, STAT1, PKC, RELA, ABCc, SUMO1, CDK1, HDAC1, and IRF1 genes, which has been subsequently validated in cohorts of patients with breast cancer, head and neck cancer, esophageal cancer, rectal cancer, and glioblastoma multiforme [50,67,68]. The work by Scott and collaborators [56] proposes a model to adapt the radiotherapy prescription to the individual sensitivity of the tumor of each patient. The model, called GARD (genome-based model for adjusting radiotherapy dose), combines the information derived from the radiosensitivity index (RSI) and the linear quadratic model (LQ model), which proposes the existence of two parameters that impact the cytotoxic capacity of radiation, one of them being proportional to the dose of radiation administered (factor α) and the other being proportional to the square of the dose (factor β). This mathematical model has been used for decades to calculate the equivalent biological dose of different radiotherapy treatment schemes, taking into account the α/β index of each tumor, which has been used to propose altered radiotherapy fractionations that have allowed the attainment of biologically equal or superior results to the treatments based on daily normal fractionation (1.8/2 Gy/fraction) [69,70]. This work uses multiple cohorts of different neoplastic entities (breast, esophageal, head and neck, stomach, cervical, glioma, pancreatic, lung, and melanoma and nonmelanoma skin cancers) and establishes a numerical value for GARD (usually in the range of 1–200), with a higher GARD level being related to a greater therapeutic effect of radiotherapy treatment and vice versa. The authors emphasize that the GARD model is not useful for predicting survival. However, they insist that different GARD values adequately predict the differences in therapeutic response (to ionizing radiation) seen in daily clinical practice between the different tumor subtypes or different affected anatomical locations treated, which could be related to the impact in terms of locoregional control of the disease, as well as to patient survival. One of the potential criticisms of this extensive work is the presumption that the tumor response after ionizing radiation depends predominantly on the tumor biology itself, without considering other nontumor factors per se (tumor environment, comorbidities, homeostatic, immunological, etc.) or without considering the usual combination with other oncological treatments, which is usually the norm in the multidisciplinary approach of most of the neoplasms considered in the study. Based on the RSI and GARD models, more recent works continue to propose their application in routine clinical practice with the aim of predicting the response to ablative stereotaxic or normofractionated treatment at the lung or other locations and the adaptation of fractionation and DBE in those histological lines that usually show a worse response to ionizing radiation [71,72]. There are no strict genetic/molecular factors, such as the general condition of the patient, tumor size, nodal load, age, existence of comorbidities, previous treatments administered and TNM classification, or other elements identified in samples (markers of inflammation such as interleukins and C-reactive protein; indirect markers of hypoxia such as osteopontin, carbonic anhydrase IX and lactate dehydrogenase; or indirect markers of tumor burden such as carcinoembryonic antigen or cytokeratin 21–1 fragments) that have been used classically to predict the response to treatment and the vital prognosis, both in lung cancer and in other solid and hematological neoplasms [73,74,75]. These factors, generally validated in univariate studies and with significant biases, were considered in the work published by Dehing-Oberije et al., who proposed a combination of clinical factors (the WHO health status classification (WHO-PS), forced expiratory volume in the first second (FEV1), gross tumor volume (GTV) equivalent to the size of the main tumor component, nodal load, and sex) with biomarkers obtained from peripheral blood (CEA and IL-6), thus obtaining an improvement in the prognostic capacity at two years of patients affected and treated for NSCLC [60]. In our study, centered on a cohort of 107 TCGA patients, no variable considered had a statistically significant impact on OS in the univariate analysis by Cox regression. In the bivariate analysis, considering the gene signature as the main variable, none of the variables modified the HR value associated with the signature. In the case of the impact of the variables considered on PFS, only the variable “histological adenocarcinoma subtype” was statistically significant in the univariate analysis. However, in the bivariate analysis, the histological subtype lost statistical significance and did not substantially modify the HR value associated with the gene signature. In our work, we have proposed a basic approach with potential translational capacity, although there are several limitations: (1) The discovery set we used was a total of 15 commercial NSCLC lines, which may limit the statistical power during the bioinformatic analysis. (2) We assessed the response to ionizing radiation with the clonogenicity test, as this is the standard for the determination of survival in Rt. As reflected in the literature [51,76,77,78], there are some differences in the published SF2 values of the different cell lines, which implies that there may be possible small differences when establishing the classification of radiosensitive and radioresistant. (3) Most of the publicly available bioinformatics information databases on NSCLC contain samples of patients mainly in localized and/or metastatic stages, which generally have not received or do not reflect information on radiotherapy treatment, which has greatly limited the sample size used in our study (n = 107), as well as additional cohorts for the validation of our signatures. The generation of gene signatures of a prognostic and/or predictive nature of response to some treatment does not usually assess other biological factors not directly related to the biology of the tumor itself. These factors, whose genetic and epigenetic bases can condition the response to certain cancer treatments and even significantly condition the overall survival of patients, are undoubtedly the greatest biases when giving translational value to these signatures. In our study, we carried out a univariate analysis of multiple clinical and pathological variables, and a bivariate analysis by Cox regression without a factor will substantially modify the HR value of the proposed gene signature. The generation of new gene signatures with prognostic and/or predictive capacity in pathologies such as NSCLC can significantly benefit patients. In our study, we proposed a prognostic gene signature for OS with the ability to predict PFS. Knowledge about the respondent profile or the vital prognosis of the patient prior to the start of treatment can help us optimize the therapeutic approach and avoid dreaded and frequent iatrogenic events. However, our study still needs additional validation steps before clinical use. For example, validation of a reference cutoff against which to compare each patient, such as their blood, nontumor lung tissue, or a pool of tumor biopsies, is needed, remembering that our signature of high and low risk for each gene was generated based on the median of a cohort of tumor samples. Additionally, and very importantly, it would be necessary to calibrate the relative contribution of each of the genes to the predictive signature and organize a prioritization algorithm for these genes. All of this could be part of future work. We believe that our work is a good example of how it is possible to perform translational radio-oncology, and we trust that, together with the results already published and future work, we can contribute to achieving the final objective of the entire oncology community, which is to improve the quality of life and prognosis of cancer patients. 5. Conclusions The oncology community needs more tools and knowledge to improve the ability to predict response to cancer treatments and patient survival. This is transcendental in such incidental and fatal neoplasms as NSCLC. In this work, the characterization of the differential gene expression profile based on the radiophenotype in NSCLC cell lines allowed us to obtain a gene expression signature with prognostic capacity and therapeutic response prediction. This gene expression signature was further validated in a cohort of patients diagnosed in early and locally advanced stages and who had received radiotherapy among other treatments. We believe, pending its usefulness in the clinic, that this biological approach and a consideration of the treatment used is key to the potential identification of new biomarkers that may lead us to improve oncological results in this or other neoplasms. Acknowledgments We thank Juan Antonio Cordero for his help with statistics. Supplementary Materials The following are available online at https://www.mdpi.com/article/10.3390/cancers14092054/s1, Table S1: Histological information and the molecular profile of some relevant NSCLC mutations of the cell lines used in this study. Information on the histology and molecular profile of relevant genes in NSCLC. CBA: bronchioloalveolar carcinoma, CCG: large cell carcinoma. WT: wild type (native). MUT: mutated. Table S2: SF2 value obtained from the scientific literature, Table S3: List of differentially expressed genes between radiosensitive and radioresistant cell lines, Table S4: Summary of the characteristics of the 107 patients in the main study cohort, Table S5: Results of the univariate analysis (log rank test) for potential prognostic and predictive factors of response, Table S6: Genes with a significant impact (p < 0.05) on overall survival measured by univariate Cox regression. They are shown in descending order by p-value, Table S7: Genes with a significant impact (p < 0.05) on PFS measured by univariate Cox regression. Genes are sorted in descending order by p-value, Table S8: Results of the analysis of the common gene signature by univariate Cox regression for OS and PFS, Table S9: Bivariate Cox regression analysis. Assessment of the effect of independent categorical variables on the HR of the prognostic gene signature of OS, Table S10: Univariate Cox regression analysis. Impact of the independent categorical variables on the PFS, Table S11: Bivariate Cox regression analysis. Assessment of the effect of independent categorical variables on the HR of the prognostic gene signature of PFS, Figure S1: Kaplan–Meier curves showing progression-free survival (PFS), cancer-specific survival (SCE), and overall survival (OS) of the cohort of interest, Figure S2: (A) ROC curves as a function of time (T). AUC: area under the curve. OS; (B) ROC curves in function of time (T). AUC: area under the curve. PFS: progression free survival, Figure S3: ROC curves at 12, 24, and 60 months of the common gene signature for OS (a) and PFS (b), Figure S4: ROC curves at 12, 24, and 60 months of the common gene signature for OS (a and c) and PFS (b and d), according to the histologic group adenocarcinoma or squamous cell carcinoma (SCC). File S1: Supplementary information. Click here for additional data file. Author Contributions A.C. designed the experiments, supervised study, performed the analysis of results, and contributed to the writing and editing of this manuscript. J.P.-S., Á.Q.-V., J.C.M., S.M.-G. and E.M.V.-S. performed experiments and analysis of results and contributed to editing this manuscript. M.J.O.-G., contributed to the design and editing of this manuscript. All authors have read and agreed to the published version of the manuscript. Funding This work was supported by grants from the Ministerio de Ciencia, Innovación y Universidades (MCIU) Plan Estatal de I+D+I 2018, a la Agencia Estatal de Investigación (AEI) y al Fondo Europeo de Desarrollo Regional (MCIU/AEI/FEDER, UE): RTI2018-097455-B-I00; grant from AEI-MICIU/FEDER (RED2018-102723-T); from CIBER de Cáncer (CB16/12/00275), co-funded by FEDER from Regional Development European Funds (European Union); from Consejeria de Salud (PI-0397-2017) and Project P18-RT-2501 from 2018 competitive research projects call within the scope of PAIDI 2020 co-financed by the European Regional Development Fund (ERDF) from the Regional Ministry of Economic Transformation, Industry, Knowledge and Universities, Junta de Andalucía. Special thanks to the AECC (Spanish Association of Cancer Research) Founding Ref. GC16173720CARR and Fundacion Eugenio Rodriguez Pascual for supporting this work. E.S.-M. was funded by a fellowship from Junta de Andalucia. S.M.-G. was funded by Fundación AECC. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Dataset for transcriptional analysis of cell lines (code: GSE197109) was stored at GEO. Additional datasets used in this study are public; references and code can be consulted in the text. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Diagram of the steps followed in the present work. Figure 2 (A) Cell survival curves after radiation, measured by a clonogenicity assay. The ordinate axis represents the surviving fraction as a percentage, and the radiation dose administered (0–8 Gy) is represented on the abscissa axis. (A1) Radioresistant cell lines. (A2) radiosensitive cell lines. (B) Heatmap representing the result of the supervised clustering of the genes of interest. In the columns, the cell lines are presented (red represents the radiosensitive ones, and black represents the radioresistant ones), arranged according to the dendrogram located in the upper part of the figure. In the rows, the genes of interest are presented and grouped according to the dendrogram shown in the left part of the figure. See text for details. Figure 3 (A) Kaplan–Meier curves of the overall survival probability of the patients according to stratification by the expression levels of each individual gene. The information regarding the 10 genes of interest (see text) is ordered by level of statistical significance (p-value). The ordinate axis represents the probability of survival (0–1), and the abscissa axis represents the overall survival in months. Yellow represents patients whose expression level of the gene in question was lower than the median expression of the gene, and blue represents patients with expression levels equal to or greater than the median expression of the gene in question. (B). Box plot showing the expression levels (logTPM) (normalized) of each gene that makes up the signature based on the prognostic group (low and high risk). (C). Kaplan–Meier curves in which the overall survival probability of the patients is represented according to their grouping according to the established cutoff point, which corresponds to the median of the risk scores of each patient in the cohort, separating the latter into two groups (high or low risk, depending on whether their individual risk score is higher than/equal to or lower than the cutoff point, respectively). Likewise, the number of patients at risk in each group is represented according to the established time intervals. *, p < 0.05; **, p < 0.01. Figure 4 (A). Kaplan–Meier curves with the progression-free survival (PFS) probability of the 19 individual genes of interest (p-value < 0.05) ordered by level of statistical significance. The probability of PFS (0–1) is represented on the ordinate axis and the PFS in months on the abscissa axis. Yellow represents patients whose expression level of the gene in question was lower than the median, and blue represents patients with expression levels equal to or greater than the median. (B). Box diagram showing the expression levels (logTPM) (normalized) on the ordinate axis of each gene as a function of each prognostic group (low or high risk). (C). Kaplan–Meier curves of the probability of PFS of the patients according to their signature by the established cutoff point, which corresponds to the median of the risk scores of each patient in the cohort, separating the latter into two groups (high or low risk, depending on whether their individual risk score is higher than/equal to or lower than the cutoff point, respectively). Likewise, the number of patients at risk in each group is represented according to the established time intervals. *, p < 0.05; **, p < 0.01. Figure 5 (A). Common signature of six genes for OS and PFS. Kaplan–Meier curves in which the probability of OS (OS) and PFS (PFS) is presented as a function of the risk group of the proposed firm (low or high risk). (B). Box plot showing the expression levels (logTPM) on the ordinate axis of each gene as a function of each prognostic group (low or high risk). *, p < 0.05; **, p < 0.01. Figure 6 (A) Kaplan–Meier curves in which the probabilities of OS and PFS are represented as a function of the risk group associated with the gene signature for the subgroup of patients with adenocarcinoma, Adc, (a and b) and for the squamous, SCC, subtype (c and d). Each graph shows the number of patients at risk for different times. The value obtained from the analysis of the global cohort was used as a cutoff point to separate both groups, SCC and Adc. (B) Kaplan–Meier curves in which the probabilities of OS and PFS are represented as a function of the risk group associated with the gene signature for the subgroup of patients with adenocarcinoma, Adc, (a and b) and for the squamous, SCC, subtype (c and d). Each graph shows the number of patients at risk for different times. The median of the risk score values of each cohort, Adc or SCC, was independently used as a cutoff point to separate each respective group. (C) Box diagram showing the expression levels (logTPM) on the ordinate axis of each gene that makes up the gene signature according to the prognostic group (low or high risk). The results of the adenocarcinoma cohort analysis (a) and squamous cell carcinoma analysis (b) and the gene elements are ordered from left to right according to the p-value (Mann–Whitney U test). *, p < 0.05; **, p < 0.01. cancers-14-02054-t001_Table 1 Table 1 Values of the surviving fraction at 2 Gy (SF2) of all cell lines. Results of clonogenicity tests. Cell Line SF2 SD Radioresistant H1975 0.891 0.0104 A549 0.832 0.162 HC827 0.745 0.2 H358 0.676 - H2228 0.646 0.056 H3122 0.622 0.056 H460 0.617 0.129 H1650 0.570 0.156 Radiosensitive Calu1 0.454 0.079 H520 0.490 - H1437 0.448 - H226 0.430 0.09 Calu3 0.325 0.131 H2009 0.228 0.024 H1781 0.180 0.067 cancers-14-02054-t002_Table 2 Table 2 Univariate Cox regression analysis. The impact of independent categorical variables on OS was analyzed. Variables Beta ET Wald HR 95% CI for HR p-Value Inferior Superior Age at diagnosis −0.25 0.27 0.88 0.78 0.46 1.32 0.35 Year of diagnosis 0.46 0.31 2.28 1.59 0.87 2.91 0.13 Sex −0.21 0.27 0.65 0.81 0.48 1.36 0.42 Histological subtype 0.38 0.27 1.97 1.46 0.86 2.48 0.16 T component −0.38 0.29 1.71 0.68 0.39 1.21 0.19 Nodal load −0.24 0.27 0.80 0.78 0.46 1.34 0.37 Tumoral stage (AJCC) −0.52 0.27 3.73 0.60 0.35 1.01 0.05 Genetic signature OS 1.10 0.28 15.16 2.99 1.72 5.19 <0.0001 CI = confidence interval; HR = hazard ratio. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Ferlay J. Colombet M. Soerjomataram I. Mathers C. Parkin D.M. Piñeros M. Znaor A. Bray F. Estimating the global cancer incidence and mortality in 2018: GLOBOCAN sources and methods Int. J. Cancer 2019 144 1941 1953 10.1002/ijc.31937 30350310 2. National Comprehensive Cancer Network (NCCN) NCCN Clinical Practice Guidelines in Oncology Non-Small Cell Lung Cancer Version 1.2020 NCCN Plymouth Meeting, PA, USA 3. O’Rourke N. Roqué i Figuls M. Bernadó N.F. Macbeth F. Concurrent chemoradiotherapy in non-small cell lung cancer Cochrane Database Syst. Rev. 2010 6 CD002140 10.1002/14651858.CD002140.pub3 20556756 4. Aupérin A. Le Péchoux C. Pignon J.P. Koning C. Jeremic B. Clamon G. Einhorn L. Ball D. Trovo M.G. Groen H.J.M. Concomitant radio-chemotherapy based on platin compounds in patients with locally advanced non-small cell lung cancer (NSCLC): A meta-analysis of individual data from 1764 patients Ann. Oncol. 2006 17 473 483 10.1093/annonc/mdj117 16500915 5. Conibear J. on behalf of AstraZeneca UK Limited Rationale for concurrent chemoradiotherapy for patients with stage III non-small-cell lung cancer Br. J. Cancer 2020 123 10 17 10.1038/s41416-020-01070-6 33293671 6. Le Chevalier T. Arriagada R. Quoix E. Ruffle P. Martin M. Tarayre M. Marie-José L.-T. Douillard J.-Y. Laplanche A. Radiotherapy alone versus combined chemotherapy and radiotherapy in nonresectable non-small-cell lung cancer: First analysis of a randomized trial in 353 patients J. Natl. Cancer Inst. 1991 83 417 423 10.1093/jnci/83.6.417 1847977 7. Curran W.J. Jr. Paulus R. Langer C.J. Komaki R. Lee J.S. Hauser S. Movsas B. Wasserman T. Rosenthal S.A. Gore E. Sequential vs. concurrent chemoradiation for stage III non-small cell lung cancer: Randomized phase III trial RTOG 9410 J. Natl. Cancer Inst. 2011 103 1452 1460 10.1093/jnci/djr325 21903745 8. Pao W. Miller V. Zakowski M. Doherty J. Politi K. Sarkaria I. Singh B. Heelan R. Rusch V. Fulton L. EGF receptor gene mutations are common in lung cancers from “never smokers” and are associated with sensitivity of tumors to gefitinib and erlotinib Proc. Natl. Acad. Sci. USA 2004 101 13306 13311 10.1073/pnas.0405220101 15329413 9. Lynch T.J. Bell D.W. Sordella R. Gurubhagavatula S. Okimoto R.A. Brannigan B.W. Harris P.L. Haserlat S.M. Supko J.G. Haluska F.G. Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib N. Engl. J. Med. 2004 350 2129 2139 10.1056/NEJMoa040938 15118073 10. Paez J.G. Jänne P.A. Lee J.C. Tracy S. Greulich H. Gabriel S. Herman P. Kaye F.J. Lindeman N. Boggon T.J. EGFR mutations in lung cancer: Correlation with clinical response to gefitinib therapy Science 2004 304 1497 1500 10.1126/science.1099314 15118125 11. Yasuda H. Kobayashi S. Costa D. EGFR exon 20 insertion mutations in non-small-cell lung cancer: Preclinical data and clinical implications Lancet Oncol. 2012 13 e23 e31 10.1016/S1470-2045(11)70129-2 21764376 12. Pao W. Wang T.Y. Riely G.J. Miller V.A. Pan Q. Ladanyi M. Zakowski M.F. Heelan R.T. Kris M.G. Varmus H.E. KRAS mutations and primary resistance of lung adenocarcinomas to gefitinib or erlotinib PLoS Med. 2005 2 e17 10.1371/journal.pmed.0020017 15696205 13. Bean J. Brennan C. Shih J.-Y. Riely G. Viale A. Wang L. Chitale D. Motoi N. Szoke J. Broderick S. MET amplification occurs with or without T790M mutations in EGFR mutant lung tumors with acquired resistance to gefitinib or erlotinib Proc. Natl. Acad. Sci. USA 2007 104 20932 20937 10.1073/pnas.0710370104 18093943 14. Engelman J.A. Zejnullahu K. Mitsudomi T. Song Y. Hyland C. Park J.O. Lindeman N. Gale C.-M. Zhao X. Christensen J. MET amplification leads to gefitinib resistance in lung cancer by activating ERBB3 signaling Science 2007 316 1039 1043 10.1126/science.1141478 17463250 15. Eberhard D.A. Johnson B.E. Amler L.C. Goddard A.D. Heldens S.L. Herbst R.S. Ince W.L. Jänne P.A. Januario T. Johnson D.H. Mutations in the epidermal growth factor receptor and in KRAS are predictive and prognostic indicators in patients with non-small-cell lung cancer treated with chemotherapy alone and in combination with erlotinib J. Clin. Oncol. 2005 23 5900 5909 10.1200/JCO.2005.02.857 16043828 16. Das A.K. Sato M. Story M.D. Peyton M. Graves R. Redpath S. Girard L. Gazdar A.F. Shay J.W. Minna J.D. Non-small-cell lung cancers with kinase domain mutations in the epidermal growth factor receptor are sensitive to ionizing radiation Cancer Res. 2006 66 9601 9608 10.1158/0008-5472.CAN-06-2627 17018617 17. Raben D. Helfrich B. Bunn P.A. Jr. Targeted therapies for non-small-cell lung cancer: Biology, rationale, and preclinical results from a radiation oncology perspective Int. J. Radiat. Oncol. Biol. Physic. 2004 59 (Suppl. 2) S27 S38 10.1016/j.ijrobp.2004.01.054 18. Bergqvist M. Brattström D. Gullbo J. Hesselius P. Brodin O. Wagenius G. p53 status and its in vitro relationship to radiosensitivity and chemosensitivity in lung cancer Anticancer Res. 2003 23 1207 1212 12820372 19. Hayes D.N. Monti S. Parmigiani G. Gilks C.B. Naoki K. Bhattacharjee A. Socinski M.A. Perou C. Meyerson M. Gene expression profiling reveals reproducible human lung adenocarcinoma subtypes in multiple independent patient cohorts J. Clin. Oncol. 2006 24 5079 5090 10.1200/JCO.2005.05.1748 17075127 20. Travis W.D. Brambilla E. Riely G.J. New pathologic classification of lung cancer: Relevance for clinical practice and clinical trials J. Clin. Oncol. 2013 31 992 1001 10.1200/JCO.2012.46.9270 23401443 21. Choong N.W. Salgia R. Vokes E.E. Key signaling pathways and targets in lung cancer therapy Clin. Lung Cancer 2007 8 (Suppl. 2) S52 S60 10.3816/CLC.2007.s.002 17382025 22. Kris M.G. Johnson B.E. 3rd Kwiatkowski D.J. Iafrate A.J. Wistuba I.I. Aronson S.L. Engelman J.A. Shyr Y. Khuri F.R. Rudin C.M. Identification of driver mutations in tumor specimens from 1000 patients with lung adenocarcinoma: The NCI’s lung cancer mutation consortium (LCMC) J. Clin. Oncol. 2011 29 CRA7506 10.1200/jco.2011.29.15_suppl.cra7506 23. Beer D.G. Kardia S.L. Huang C.C. Giordano T.J. Levin A.M. Misek D.E. Lin L. Chen G. Gharib T.G. Thomas D.G. Gene-expression profiles predict survival of patients with lung adenocarcinoma Nat. Med. 2002 8 816 824 10.1038/nm733 12118244 24. Director’s Challenge Consortium for the Molecular Classification of Lung Adenocarcinoma Shedden K. Taylor J.M. Enkemann S.A. Tsao M.S. Yeatman T.J. Gerald W.L. Eschrich S. Jurisica I. Giordano T.J. Gene expression-based survival prediction in lung adenocarcinoma: A multi-site, blinded validation study Nat. Med. 2008 14 822 827 10.1038/nm.1790 18641660 25. Larsen J.E. Pavey S. Passmore L.H. Bowman R.V. Hayward N.K. Fong K. Gene expression signature predicts recurrence in lung adenocarcinoma Clin. Cancer Res. 2007 13 2946 2954 10.1158/1078-0432.CCR-06-2525 17504995 26. Li Y. Tang H. Sun Z. Bungum A.O. Edell E.S. Lingle W.L. Stoddard S.M. Zhang M. Jen J. Yang P. Network-based approach identified cell cycle genes as predictor of overall survival in lung adenocarcinoma patients Lung Cancer 2013 80 91 98 10.1016/j.lungcan.2012.12.022 23357462 27. Tomida S. Takeuchi T. Shimada Y. Arima C. Matsuo K. Mitsudomi T. Yatabe Y. Takahashi T. Relapse-related molecular signature in lung adenocarcinomas identifies patients with dismal prognosis J. Clin. Oncol. 2009 27 2793 2799 10.1200/JCO.2008.19.7053 19414676 28. Bianchi F. Nuciforo P. Vecchi M. Bernard L. Tizzoni L. Marchetti A. Buttitta F. Felicioni L. Nicassio F. Di Fiore P.P. Survival prediction of stage I lung adenocarcinomas by expression of 10 genes J. Clin. Investig. 2007 117 3436 3444 10.1172/JCI32007 17948124 29. Raz D.J. Ray M.R. Kim J.Y. Ricchetti R. Taron M. Skrzypski M. Segal M. Gandara D.R. Rosell R. Jablons D.M. A multigene assay is prognostic of survival in patients with early-stage lung adenocarcinoma Clin. Cancer Res. 2008 14 5565 5570 10.1158/1078-0432.CCR-08-0544 18765549 30. Park Y.-Y. Park E.S. Kim S.B. Kim S.C. Sohn B.H. Chu I.-S. Jeong W. Mills G.B. Byers L.A. Lee J.-S. Development and validation of a prognostic gene-expression signature for lung adenocarcinoma PLoS ONE 2012 7 e44225 10.1371/journal.pone.0044225 22970185 31. Akagi I. Okayama H. Schetter A.J. Robles A. Kohno T. Bowman E.D. Kazandjian D. Welsh J.A. Oue N. Saito M. Combination of protein coding and noncoding gene expression as a robust prognostic classifier in stage I lung adenocarcinoma Cancer Res. 2013 73 3821 3832 10.1158/0008-5472.CAN-13-0031 23639940 32. Raponi M. Zhang Y. Yu J. Chen G. Lee G. Taylor J.M. Macdonald J. Thomas D. Moskaluk C. Wang Y. Gene expression signatures for predicting prognosis of squamous cell and adenocarcinomas of the lung Cancer Res. 2006 66 7466 7472 10.1158/0008-5472.CAN-06-1191 16885343 33. Larsen J.E. Pavey S.J. Passmore L.H. Bowman R. Clarke B.E. Hayward N.K. Fong K.M. Expression profiling defines a recurrence signature in lung squamous cell carcinoma Carcinogenesis 2007 28 760 766 10.1093/carcin/bgl207 17082175 34. Skrzypski M. Jassem E. Taron M. Sanchez J.J. Mendez P. Rzyman W. Gulida G. Raz D. Jablons D. Provencio M. Three-gene expression signature predicts survival in early-stage squamous cell carcinoma of the lung Clin. Cancer Res. 2008 14 4794 4799 10.1158/1078-0432.CCR-08-0576 18676750 35. Landi M.T. Zhao Y. Rotunno M. Koshiol J. Liu H. Bergen A. Rubagotti M. Goldstein A.M. Linnoila I. Marincola F.M. MicroRNA expression differentiates histology and predicts survival of lung cancer Clin. Cancer Res. 2010 16 430 441 10.1158/1078-0432.CCR-09-1736 20068076 36. Hou J. Aerts J. den Hamer B. van Ijcken W. den Bakker M. Riegman P. Van Der Leest C. Van Der Spek P. Foekens J.A. Hoogsteden H.C. Gene expression-based classification of non-small cell lung carcinomas and survival prediction PLoS ONE 2010 5 e10312 10.1371/journal.pone.0010312 20421987 37. Mitra R. Lee J. Jo J. Milani M. McClintick J.N. Edenberg H. Kesler K.A. Rieger K.M. Badve S. Cummings O.W. Prediction of postoperative recurrence-free survival in non-small cell lung cancer by using an internationally validated gene expression model Clin. Cancer Res. 2011 17 2934 2946 10.1158/1078-0432.CCR-10-1803 21242119 38. Roepman P. Jassem J. Smit E.F. Muley T. Niklinski J. Van De Velde T. Witteveen A.T. Rzyman W. Floore A. Burgers S. An immune response enriched 72-gene prognostic profile for early-stage non-small-cell lung cancer Clin. Cancer Res. 2009 15 284 290 10.1158/1078-0432.CCR-08-1258 19118056 39. Xie Y. Xiao G. Coombes K. Behrens C. Solis L.M. Raso G. Girard L. Erickson H.S. Roth J. Heymach J.V. Robust gene expression signature from formalin-fixed paraffin-embedded samples predicts prognosis of non-small-cell lung cancer patients Clin. Cancer Res. 2011 17 5705 5714 10.1158/1078-0432.CCR-11-0196 21742808 40. Lee E.-S. Son D.-S. Kim S.-H. Lee J. Jo J. Han J. Kim H. Lee H.J. Choi H.Y. Jung Y. Prediction of recurrence-free survival in postoperative non-small cell lung cancer patients by using an integrated model of clinical information and gene expression Clin. Cancer Res. 2008 14 7397 7404 10.1158/1078-0432.CCR-07-4937 19010856 41. Boutros P.C. Lau S.K. Pintilie M. Liu N. Shepherd F.A. Der S.D. Tsao M.-S. Penn L.Z. Jurisica I. Prognostic gene signatures for non-small-cell lung cancer Proc. Natl. Acad. Sci. USA 2009 106 2824 2828 10.1073/pnas.0809444106 19196983 42. Chen D.-T. Hsu Y.-L. Fulp W.J. Coppola D. Haura E.B. Yeatman T.J. Cress W.D. Prognostic and predictive value of a malignancy-risk gene signature in early-stage non-small cell lung cancer J. Natl. Cancer Inst. 2011 103 1859 1870 10.1093/jnci/djr420 22157961 43. Guo N.L. Wan Y.-W. Tosun K. Lin H. Msiska Z. Flynn D.C. Remick S.C. Vallyathan V. Dowlati A. Shi X. Confirmation of gene expression-based prediction of survival in non-small cell lung cancer Clin. Cancer Res. 2008 14 8213 8220 10.1158/1078-0432.CCR-08-0095 19088038 44. Lu Y. Lemon W. Liu P.-Y. Yi Y. Morrison C. Yang P. Sun Z. Szőke J. Gerald W.L. Watson M. A gene expression signature predicts survival of patients with stage I non-small cell lung cancer PLoS Med. 2006 3 e467 10.1371/journal.pmed.0030467 17194181 45. Chen H.-Y. Yu S.-L. Chen C.-H. Chang G.-C. Chen C.-Y. Yuan A. Cheng C.-L. Wang C.-H. Terng H.-J. Kao S.-F. A five-gene signature and clinical outcome in non-small-cell lung cancer N. Engl. J. Med. 2007 356 11 20 10.1056/NEJMoa060096 17202451 46. Lau S.K. Boutros P.C. Pintilie M. Blackhall F.H. Zhu C.-Q. Strumpf D. Johnston M.R. Darling G. Keshavjee S. Waddell T.K. Three-gene prognostic classifier for early-stage non small-cell lung cancer J. Clin. Oncol. 2007 25 5562 5569 10.1200/JCO.2007.12.0352 18065728 47. Zhu C. Ding K. Strumpf D. Weir B.A. Meyerson M. Pennell N. Thomas R.K. Naoki K. Ladd-Acosta C. Liu N. Prognostic and predictive gene signature for adjuvant chemotherapy in resected non-small-cell lung cancer J. Clin. Oncol. 2010 28 4417 4424 10.1200/JCO.2009.26.4325 20823422 48. Tang H. Xiao G. Behrens C. Schiller J. Allen J. Chow C.-W. Suraokar M. Corvalan A. Mao J.-H. White M.A. A 12-gene set predicts survival benefits from adjuvant chemotherapy in non-small cell lung cancer patients Clin. Cancer Res. 2013 19 1577 1586 10.1158/1078-0432.CCR-12-2321 23357979 49. Shoemaker R.H. The NCI60 human tumour cell line anticancer drug screen Nat. Rev. Cancer 2006 6 813 823 10.1038/nrc1951 16990858 50. Torres-Roca J.F. Eschrich S. Zhao H. Bloom G. Sung J. McCarthy S. Cantor A.B. Scuto A. Li C. Zhang S. Prediction of radiation sensitivity using a gene expression classifier Cancer Res. 2005 65 7169 7176 10.1158/0008-5472.CAN-05-0656 16103067 51. Eschrich S. Zhang H. Zhao H. Boulware D. Lee J.-H. Bloom G. Torres-Roca J.F. Systems biology modeling of the radiation sensitivity network: A biomarker discovery platform Int. J. Radiat. Oncol. Biol. Phys. 2009 75 497 505 10.1016/j.ijrobp.2009.05.056 19735874 52. Strom T. Hoffe S.E. Fulp W. Frakes J. Coppola D. Springett G.M. Malafa M.P. Harris C.L. Eschrich S.A. Torres-Roca J.F. Radiosensitivity index predicts for survival with adjuvant radiation in resectable pancreatic cancer Radiother. Oncol. 2015 117 159 164 10.1016/j.radonc.2015.07.018 26235848 53. Amundson S.A. Do K.T. Vinikoor L.C. Lee R.A. Koch-Paiz C.A. Ahn J. Reimers M. Chen Y. Scudiero D.A. Weinstein J.N. Integrating global gene expression and radiation survival parameters across the 60 cell lines of the National Cancer Institute Anticancer Drug Screen Cancer Res. 2008 68 415 424 10.1158/0008-5472.CAN-07-2120 18199535 54. Ahmed K.A. Berglund A.E. Welsh E.A. Naghavi A.O. Kim Y. Yu M. Robinson T.J. Eschrich S.A. Johnstone P.A. Torres-Roca J.F. The radiosensitivity of brain metastases based upon primary histology utilizing a multigene index of tumor radiosensitivity Neuro-Oncology 2017 19 1145 1146 10.1093/neuonc/nox043 28379582 55. Jang B.-S. Han W. Kim I.A. Tumor mutation burden, immune checkpoint crosstalk and radiosensitivity in single-cell RNA sequencing data of breast cancer Radiother. Oncol. 2020 142 202 209 10.1016/j.radonc.2019.11.003 31767471 56. Scott J.G. Berglund A. Schell M.J. Mihaylov I. Fulp W.J. Yue B. Welsh E. Caudell J.J. Ahmed K. Strom T.S. A genome-based model for adjusting radiotherapy dose (GARD): A retrospective, cohort-based study Lancet Oncol. 2017 18 202 211 Erratum in Lancet Oncol. 2017, 18, e65 10.1016/S1470-2045(16)30648-9 27993569 57. Hammond S.M. An Overview of MicroRNAs Adv. Drug Deliv. Rev. 2015 87 3 14 10.1016/j.addr.2015.05.001 25979468 58. Boeri M. Verri C. Conte D. Roz L. Modena P. Facchinetti F. Calabrò E. Croce C.M. Pastorino U. Sozzi G. MicroRNA signatures in tissues and plasma predict development and prognosis of computed tomography detected lung cancer Proc. Natl. Acad. Sci. USA 2011 108 3713 3718 10.1073/pnas.1100048108 21300873 59. Sun Y. Hawkins P.G. Bi N. Dess R.T. Tewari M. Hearn J.W. Hayman J.A. Kalemkerian G.P. Lawrence T.S. Haken R.K.T. Serum MicroRNA Signature Predicts Response to High-Dose Radiation Therapy in Locally Advanced Non-Small Cell Lung Cancer Int. J. Radiat. Oncol. Biol. Phys. 2018 100 107 114 10.1016/j.ijrobp.2017.08.039 29051037 60. Dehing-Oberije C. Aerts H. Yu S. De Ruysscher D. Menheere P. Hilvo M. van der Weide H. Rao B. Lambin P. Development and validation of a prognostic model using blood biomarker information for prediction of survival of non-small-cell lung cancer patients treated with combined chemotherapy and radiation or radiotherapy alone (NCT00181519, NCT00573040, and NCT00572325) Int. J. Radiat. Oncol. Biol. Phys. 2011 81 360 368 20888135 61. Postmus P.E. Kerr K.M. Oudkerk M. Senan S. Waller D.A. Vansteenkiste J. Escriu C. Peters S. ESMO Guidelines Committee Early and locally advanced non-small-cell lung cancer (NSCLC): ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up Ann. Oncol. 2017 28 (Suppl. 4) iv1 iv21 10.1093/annonc/mdx222 28881918 62. Detterbeck F.C. Boffa D.J. Kim A.W. Tanoue L.T. The eighth edition lung cancer stage classification Chest 2017 151 193 203 10.1016/j.chest.2016.10.010 27780786 63. Subramanian J. Simon R. Gene expression-based prognostic signatures in lung cancer: Ready for clinical use? J. Natl. Cancer Inst. 2010 102 464 474 10.1093/jnci/djq025 20233996 64. Tomida S. Koshikawa K. Yatabe Y. Harano T. Ogura N. Mitsudomi T. Some M. Yanagisawa K. Takahashi T. Osada H. Gene expression-based, individualized outcome prediction for surgically treated lung cancer patients Oncogene 2004 23 5360 5370 10.1038/sj.onc.1207697 15064725 65. Zuo S. Wei M. Zhang H. Chen A. Wu J. Wei J. Dong J. A robust six-gene prognostic signature for prediction of both disease-free and overall survival in non-small cell lung cancer J. Transl. Med. 2019 17 152 10.1186/s12967-019-1899-y 31088477 66. He R. Zuo S. A Robust 8-Gene Prognostic Signature for Early-Stage Non-small Cell Lung Cancer Front. Oncol. 2019 9 693 10.3389/fonc.2019.00693 31417870 67. Torres-Roca J.F. A molecular assay of tumor radiosensitivity: A roadmap towards biology-based personalized radiation therapy Pers. Med. 2012 9 547 557 10.2217/pme.12.55 68. Ahmed K.A. Chinnaiyan P. Fulp W.J. Eschrich S. Torres-Roca J.F. Caudell J.J. The radiosensitivity index predicts for overall survival in glioblastoma Oncotarget 2015 6 34414 34422 10.18632/oncotarget.5437 26451615 69. Fowler J.F. 21 years of biologically effective dose Br. J. Radiol. 2010 83 554 568 10.1259/bjr/31372149 20603408 70. Miles E.F. Lee W.R. Hypofractionation for prostate cancer: A critical review Semin. Radiat. Oncol. 2008 18 41 47 10.1016/j.semradonc.2007.09.006 18082587 71. Ahmed K.A. Scott J.G. Arrington J.A. Naghavi A.O. Grass G.D. Perez B.A. Caudell J.J. Berglund A.E. Welsh E.A. Eschrich S.A. Radiosensitivity of Lung Metastases by Primary Histology and Implications for Stereotactic Body Radiation Therapy Using the Genomically Adjusted Radiation Dose J. Thorac. Oncol. 2018 13 1121 1127 10.1016/j.jtho.2018.04.027 29733909 72. Ahmed K.A. Grass G.D. Orman A.G. Liveringhouse C. Montejo M.E. Soliman H.H. Han H.S. Czerniecki B.J. Torres-Roca J.F. Diaz R. Personalizing Radiation Treatment Delivery in the Management of Breast Cancer Int. J. Breast Cancer 2018 2018 6729802 10.1155/2018/6729802 29984003 73. Le Q.-T. Chen E. Salim A. Cao H. Kong C.S. Whyte R. Donington J. Cannon W. Wakelee H. Tibshirani R. An evaluation of tumor oxygenation and gene expression in patients with early stage non-small cell lung cancers Clin. Cancer Res. 2006 12 1507 1514 10.1158/1078-0432.CCR-05-2049 16533775 74. Pujol J.-L. Molinier O. Ebert W. Daurès J.-P. Barlesi F. Buccheri G. Paesmans M. Quoix E. Moro-Sibilot D. Szturmowicz M. CYFRA 21-1 is a prognostic determinant in non-small-cell lung cancer: Results of a meta-analysis in 2063 patients Br. J. Cancer 2004 90 2097 2105 10.1038/sj.bjc.6601851 15150567 75. Swinson D.E. Jones J.L. Richardson D. Wykoff C. Turley H. Pastorek J. Taub N. Harris A. O’Byrne K.J. Carbonic anhydrase IX expression, a novel surrogate marker of tumor hypoxia, is associated with a poor prognosis in non-small-cell lung cancer J. Clin. Oncol. 2003 21 473 482 10.1200/JCO.2003.11.132 12560438 76. Tumati V. Kumar S. Yu L. Chen B. Choy H. Saha D. Effect of PF-02341066 and radiation on non-small cell lung cancer cells Oncol. Rep. 2013 29 1094 1100 10.3892/or.2012.2198 23254764 77. Jung J. Lee J.-S. Lee Y.-S. Lee K. Radiosensitivity of Cancer Cells Is Regulated by Translationally Controlled Tumor Protein Cancers 2019 11 386 10.3390/cancers11030386 78. Girdhani S. Rodriguez A. Dua R. Parry R. Correlation of radiosensitivity of lung carcinoma to CD137L endogenous expression levels J. Clin. Oncol. 2019 37 e14229 10.1200/JCO.2019.37.15_suppl.e14229

PMC009xxxxxx/PMC9099639.txt

==== Front Molecules Molecules molecules Molecules 1420-3049 MDPI 10.3390/molecules27092669 molecules-27-02669 Article Cannabis sativa CBD Extract Shows Promising Antibacterial Activity against Salmonella typhimurium and S. newington Gildea Logan 1 https://orcid.org/0000-0003-4561-8245 Ayariga Joseph Atia 2* https://orcid.org/0000-0002-3556-803X Ajayi Olufemi S. 3* Xu Junhuan 3 Villafane Robert 1 Samuel-Foo Michelle 3 Kim Ki Hyun Academic Editor Rateb Mostafa Academic Editor Hassan Hossam Academic Editor 1 The Microbiology Program, College of Science, Technology, Engineering, and Mathematics (C-STEM), Alabama State University, Montgomery, AL 36104, USA; lgildea5810@myasu.alasu.edu (L.G.); rvillafane@alasu.edu (R.V.) 2 The Biomedical Engineering Program, College of Science, Technology, Engineering, and Mathematics (C-STEM), Alabama State University, Montgomery, AL 36104, USA 3 Department of Biological Sciences, Alabama State University, Montgomery, AL 36104, USA; jxu@alasu.edu (J.X.); mfoo@alasu.edu (M.S.-F.) * Correspondence: ayarigajosephatia@yahoo.co.uk (J.A.A.); oajayi@alasu.edu (O.S.A.) 21 4 2022 5 2022 27 9 266926 3 2022 19 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Products derived from Cannabis sativa L. have gained increased interest and popularity. As these products become common amongst the public, the health and potential therapeutic values associated with hemp have become a premier focus of research. While the psychoactive and medicinal properties of Cannabis products have been extensively highlighted in the literature, the antibacterial properties of cannabidiol (CBD) have not been explored in depth. This research serves to examine the antibacterial potential of CBD against Salmonella newington and S. typhimurium. In this study, we observed bacterial response to CBD exposure through biological assays, bacterial kinetics, and fluorescence microscopy. Additionally, comparative studies between CBD and ampicillin were conducted against S. typhimurium and S. newington to determine comparative efficacy. Furthermore, we observed potential resistance development of our Salmonella spp. against CBD treatment. Salmonella novel antibacterial agents membrane disruption cannabidiol ==== Body pmc1. Introduction Cannabis sativa L., a member of the genus Cannabis and a species in the family Cannabaceae, originated from Central Asia and is one of the oldest psychoactive plants known to humans [1]. Since its discovery, C. sativa has been utilized recreationally, medicinally, and industrially for numerous applications. Early uses of C. sativa focused on its use as an industrial material to produce textiles, ropes, and paper products [2,3,4]. Throughout early history, the use of C. sativa as an industrial product spread from Asia throughout Europe and Africa [5]. As industrial use of C. sativa became common, its use as a therapeutic against rheumatic pain, intestinal constipation, disorders of the female reproductive system, malaria, and other common health problems can be traced back to ancient China where it was one of the oldest pharmaceuticals recorded [6]. Over time, the use of C. sativa as a therapeutic option was introduced across the western world and by the 19th century had gained attention from medical science. The first clinical conference on C. sativa in 1860 led to an increased interest in the research and development of medicinal products derived from C. sativa [2]. This interest in C. sativa drastically declined in 1942 when it was removed from the United States Pharmacopoeia and lost its status due to research suggesting a correlation between C. sativa and ‘insanity’ [7]. This decision eventually led to the United Nations designating Cannabis as an illegal psychotropic substance in 1971 at the Convention on Psychotropic Substances [8,9]. These designations and findings drastically reduced research progression on Cannabis and produced a negative stigma on the substance from the perspective of the public [10,11]. As we move towards the mid-21st century, the acceptance of Cannabis and the potential of its byproducts has increased [12]. Continued research demonstrated the efficiency of Cannabis in the treatment of medical conditions including epilepsy, multiple sclerosis, Tourette’s syndrome, and other neurological diseases [13,14,15]. Additionally, research has further characterized the active compounds—tetrahydrocannabinol (THC) and cannabidiol (CBD)—of Cannabis and their psychological and physiological effects on humans [16,17,18,19]. While clinical research has primarily focused on the efficacy of Cannabis against neurological disorders, a current gap in knowledge is the efficacy of Cannabis and its byproducts as antibacterial agents. The current literature does suggest that Cannabis and more specifically its active compound, CBD, shows antibacterial function [20,21]. However, further research and characterization of this antibacterial function is crucial to the development of novel therapeutics against clinically relevant bacteria [22,23]. The increased prevalence and occurrence of antimicrobial resistance is a major health concern worldwide, making the development of alternative therapies a necessity [24,25,26,27]. The steady decrease in antibiotic efficacy alongside the decreased development of new antibiotics present a major obstacle in the treatment of multidrug-resistant bacteria [28,29]. Antibiotic-resistant and multidrug-resistant bacterial infections account for roughly 2.8 million infections and 35,000 deaths annually in the United States alone [30]. Antibiotic-resistant bacterial infections, while a major threat to public health, also present a major economic burden to the United States, resulting in an annual cost of USD 7.7 billion [31]. As resistant bacterial species persist, the development of novel antibacterial treatment methods is essential to the preservation of public health. The purpose of this research is to determine if CBD extracted from C. sativa demonstrates antibacterial activity against a common Gram-negative bacterial pathogen, Salmonella. CBD, initially not recognized as an active compound of cannabinoids, has been shown to possess potential therapeutic benefits [32,33,34]. Cannabis is commonly associated with psychotic effects. We now know that these effects can be attributed to THC, whereas CBD “exhibits no effects indicative of any abuse or dependence-potential and to date, there is no evidence of public health related problems associated with the use of pure CBD” according to the World Health Organization [35]. The literature suggests that CBD’s antibacterial function is carried out through the disruption of the cell membrane of both Gram-positive and Gram-negative bacteria [36]. The antibacterial activity of CBD against several bacterial pathogens including Staphylococcus aureus, Streptococcus pneumoniae, Clostridiosis difficile, Neisseria spp., Moraxella catarrhalis, and Legionella pneumophila has been observed [20,21]. Study of CBD as an antibacterial agent requires further determination of the range of bacteria that CBD exhibits antibacterial activity against. Examining the potential of CBD against a clinically relevant pathogen such as Salmonella is essential to further expand our knowledge on CBD as a potential novel therapeutic. Salmonella species are one of the most common and prevalent foodborne pathogens worldwide, found in several food products including poultry, seafood, and other fresh or processed meats [26,27,37,38,39]. Salmonella accounts for 1.35 million infections, 26,500 hospitalizations, and 420 deaths annually in the United States alone [30]. Salmonella is a significant spoilage hazard. In addition, the increased prevalence of multidrug-resistant strains makes Salmonella a major threat to public health [25,30]. The CDC reported in 2019 that Salmonella typhimurium, one of the species examined in this study, accounted for 59% of ampicillin-resistant Salmonella infections in the United States [30]. The continued use of broad-spectrum antibiotics could drastically increase the prevalence of antibiotic and multidrug-resistant Salmonella strains, making the development of novel antibiotic alternatives a necessity [31,40]. In this study, we performed plate assays, fluorescence microscopy, and growth kinetic studies to determine the antibacterial activity of CBD extracted from C. sativa against S. typhimurium and S. newington. Additionally, we conducted comparative kinetic studies of the two Salmonella strains in the presence of CBD or a common broad-spectrum antibiotic, ampicillin. Finally, we examined resistance development of S. typhimurium and S. newington against CBD treatment over an extended time. We hypothesized that CBD would exhibit antibacterial activity against S. typhimurium and S. newington. The results of these studies suggest that CBD does exhibit antibacterial activity against these Salmonella species, further encouraging research and development of CBD as a potential antibacterial agent. 2. Results 2.1. Gas Chromatography For confirmation of pure CBD in our C. sativa extract, gas chromatography (GC) was conducted. Figure 1A represents the internal standard for CBD and THC, while Figure 1B represents our sample. In these readings, we observe a strong single peak at the 8.23 min retention time (RT) in our sample. This is consistent with the expected RT of CBD at 8.22 min as shown in the internal control sample. The absence of other peaks represents the absence of other compounds or contaminants. Overall, we conclude through the GC analysis that our sample is pure CBD. 2.2. Plate Assays To examine the potential antibacterial activity of our CBD against S. typhimurium and S. newington, the Kirby–Bauer and spot assays were conducted. These assays allow us to visualize and quantify the inhibitory effects of CBD against our Salmonella strains. Both assays confirmed inhibitory activity of CBD against S. typhimurium and S. newington. In the Kirby–Bauer assay, we observed zones of inhibition (ZOI) around the CBD-treated disks suggesting inhibition of bacterial growth due to exposure to CBD (Figure 2C). These results also suggest a dose-dependent inhibition due to CBD treatment with ZOIs decreasing as CBD concentration decreased (Figure 2A,B). The results of the spot assay further confirmed the dose-dependent nature of CBD’s inhibitory activity. As CBD concentration decreased, we saw an increase in the density of bacterial colonies in both S. typhimurium and S. newington (Figure 2D). These data also suggested that S. newington was more susceptible to CBD treatment than S. typhimurium due to the larger ZOIs (Figure 2B) and lower colony density (Figure 2D). 2.3. CBD Extract Reduces Bacterial Growth of Salmonella typhimurium and Salmonella newington To determine the effect of CBD on S. typhimurium and S. newington, cultures were treated with CBD at concentrations of 1.25, 0.125, 0.0125, or 0.00125 μg/mL. The optical density (OD600) was recorded using a spectrometer (Molecular Devices SpectraMax® ABS Plus) hourly for 6 h. Both S. typhimurium and S. newington cultures treated with CBD showed a significant reduction in OD600 within 6 h of treatment (Figure 3A,B). It was even observed that in S. newington, CBD at a concentration as low as 0.0125 μg/mL reduced bacterial growth over the 6 h period. These results suggest that S. newington might have a greater susceptibility to CBD than S. typhimurium, whose MIC was 0.125 μg/mL. It was observed that treatment of S. typhimurium with CBD at a concentration of 0.00125 μg/mL resulted in a higher OD600 than that of the control treatment. We infer that these results suggest a rapid development of resistance to low concentrations of CBD treatment resulting in an increase in bacterial fitness. The reduction of OD600 signifies that CBD extract does exhibit antibacterial characteristics. The antibacterial effect on S. typhimurium and S. newington was observed to be dose-dependent with the OD600 increasing in correlation with decreased concentrations of CBD. 2.4. Assessment of Bacterial Membrane Integrity via Fluorescent Staining of CBD-Treated Salmonella Cells To examine the effect of CBD treatment on Salmonella cells, S. typhimurium and S. newington samples were treated with CBD at concentrations of 1.25, 0.125, or 0.0125 μg/mL. These samples were visualized using immunofluorescent 4′,6-diamidino-2-phenylindole (DAPI) staining. This staining technique emits fluorescence when the stain binds to AT-rich DNA. This stain is membrane impermeable, and thus, fluorescence represents compromised membrane integrity. DAPI staining confirmed that CBD treatment resulted in degradation of membrane integrity after 5 min and 30 min of treatment (Figure 4A,B). Quantitative analysis using densitometry of these images showed increased DAPI fluorescence from the 5 min to the 30 min time points (Figure 4C). These images all indicate that CBD can successfully damage Salmonella cells outer lipopolysaccharide membranes. 2.5. Comparative Study of CBD and Antibiotic Treatment against Salmonella Salmonella infections are typically treated with broad-spectrum antibiotics such as ampicillin; however, the development of resistance to these treatments has become more prevalent, thus increasing the need for alternative treatments. To compare conventional antibiotics to CBD, we conducted comparative kinetic studies to observe antibacterial activity. In this study, Salmonella strains S. typhimurium and S. newington at an OD600 of 0.5 were treated with the minimum inhibitory concentration (MIC) of ampicillin (0.5 µg/mL) or CBD (0.125 µg/mL). Results suggested that CBD and ampicillin both successfully inhibited Salmonella growth in both strains (Figure 5). In comparison, both treatments resulted in a similar OD600 after 6 h of treatment suggesting that CBD was able to inhibit bacterial growth to an extent similar to ampicillin. 2.6. Developed Resistance of Salmonella to CBD Treatment A major concern with conventional antibiotic treatments is the development of resistance. To observe the development of resistance in Salmonella against CBD treatment, extended kinetic studies were conducted over a span of 48 h. Salmonella strains S. typhimurium and S. newington were treated with ampicillin (MIC 0.5 µg/mL) or CBD (1.25, 0.125, 0.0125, or 0.00125 μg/mL). Results confirmed a developed resistance to CBD in both Salmonella species (Figure 6). At concentrations of 1.25 and 0.125 μg/mL of CBD, resistance development was not observed until the 48 h time point. These results suggest that CBD was effective in killing Salmonella and decreasing the rate of resistance development in both Salmonella strains over the initial 24 h time period. At 0.0125, 0.00125, and 0.000125 μg/mL concentrations of CBD, there was no significant reduction in bacterial growth and resistance was developed after only 12 h of exposure in both Salmonella strains. These results suggest that both Salmonella strains were able to develop resistance quickly to low concentrations of CBD. 2.7. CBD Effectiveness against Salmonella Biofilm The literature has determined that most hospital-acquired infections are the result of bacterial biofilm formation [41]. Typically, a result of bacteria reaching a non-replicatory state, biofilms are a major health threat that are characterized by antibiotic resistance, reoccurring infection, and sepsis [42]. Understanding CBD activity against bacterial biofilm is an important aspect of determining CBD viability against Salmonella spp. To determine the efficacy of CBD treatment against biofilm formation, S. typhimurium biofilms were grown in vitro and treated with CBD at a concentration of 0.125 µg/mL. S. typhimurium biofilms were then visualized through scanning electron microscopy (SEM) (Figure 7). 3. Discussion The rapid increase of antibiotic- and multidrug-resistant bacteria over the early 21st century is a major threat to public health [30]. Knowing this, the study of potentially efficacious alternative therapeutics has been a topic of growing interest. C. sativa products, while typically associated with the treatment of neurological disorders, have shown promise as antibacterial agents against several notable pathogens [20,21]. Of the multiple metabolites of C. sativa, research suggests that CBD is the most promising [32,33,34]. This compound, unlike THC, holds no psychoactive properties but does possess antioxidant, anti-inflammatory, and antibacterial properties [22,43,44]. CBD’s antibacterial potential has shown some promise; however, research is relatively limited. Further investigation and confirmation of the specificity of this antibacterial activity is crucial as the field of medicine searches for new viable therapeutics for resistant bacterial infections. C. sativa products have exhibited a wide variety of applications in numerous fields. This study helps expand our knowledge of C. sativa-derived CBD as an antibacterial agent against two common Gram-negative pathogenic Salmonella strains. Salmonella spp. have significance in terms of resistance to antibiotics and standard treatment protocols making these infections a major threat to the preservation of public health [25,28,30]. Discovery and development of novel antibacterial agents such as CBD are a major step towards the future of therapeutics in a world where antibiotics are no longer efficacious and cost effective [20,21]. This study serves primarily to determine the antibacterial potential of CBD against S. typhimurium and S. newington through plate assays, fluorescence microscopy, and kinetic assays. Additionally, this study examines the comparative antibacterial activity of CBD and ampicillin as well as resistance development of S. typhimurium and S. newington against CBD treatments. In this study, we confirm that CBD does exhibit antibacterial activity against S. typhimurium and S. newington. This inference was derived from a combination of plate assays, fluorescence microscopy, and bacterial growth kinetic assays and thus allowed us to propose a potential mechanism of this antibacterial activity. Plate assays consisted of the Kirby–Bauer assay and the spot assay, both of which served as initial confirmation of bacterial inhibition of S. typhimurium and S. newington. Once this inhibitory activity was observed, a fluorescence microscopy assay for membrane integrity was utilized. This assay utilized DAPI stain, a membrane-impermeable DNA binding stain. Fluorescence of the DAPI stain confirmed that CBD treatment of S. typhimurium and S. newington resulted in a loss of membrane integrity and cell death, and thus proving our hypothesis that the CBD treatment leads to membrane disruption. Other researchers have reported this mechanism as the antibacterial mechanism of action of CBD [36]. Finally, kinetic assays were conducted to examine the effect of CBD treatment on lag-phase S. typhimurium and S. newington over a span of 6 h. Once again, it was observed that CBD was effective in growth inhibition of both species in a 6 h period with a MIC of 0.0125 μg/mL in S. typhimurium and 0.125 μg/mL in S. newington. These findings confirm that CBD does possess antibacterial activity through mechanisms similarly described in other Gram-negative bacterial species [20,21,36]. Within this study, we were able to exhibit the antibacterial effectiveness of CBD at a concentration of 0.125 μg/mL against a S. typhimurium biofilm (Figure 7). This study is relevant in terms of clinical application of CBD where most hospital-acquired infections are in the form of biofilms [41]. Biofilms typically associated with antibiotic resistance result in infections that are difficult to treat with an increased chance of sepsis and mortality in some cases [42]. Considering the threat that biofilms pose, it is essential that novel therapeutics are developed for the prevention and treatment of these harmful infections. Our results suggest that CBD features antibacterial activity that is not only effective against lag- and log-phase Salmonella spp. but also against Salmonella biofilms. An important facet of this study was the comparison between CBD and broad-spectrum antibiotic treatment. In this study, we examined the comparative kinetics between CBD treatment and ampicillin treatment of S. typhimurium and S. newington. We concluded from the results of these kinetic studies that both CBD and ampicillin at their MIC concentrations are successful at inhibition of S. typhimurium and S. newington growth. It is important to note that the MIC concentration of ampicillin (0.5 μg/mL) is significantly higher than the MIC of CBD (0.125 μg/mL) (Figure 5). To further investigate the comparative efficacy of a novel therapeutic agent (i.e., CBD in this study) and a standard treatment (such as ampicillin), it was important to examine the resistance development of the pathogen to the therapeutic agents. To examine the resistance of S. typhimurium and S. newington against CBD and ampicillin treatments, extended growth kinetics under different dosages of CBD were carried out over a 48 h period (see Figure 6). These results varied across the two strains of Salmonella that were used. In S. typhimurium, CBD concentrations of 1.25 and 0.125 μg/mL were able to inhibit bacterial growth over the span of 24 h after a single treatment. These results are similar to the growth kinetics we observed with ampicillin treatment after 24 h, whereas lower concentrations of CBD showed a rapid rise in growth following 12 h of treatment, suggesting a rapid development of resistance. After 48 h of exposure to CBD treatment, S. typhimurium had developed resistance, suggested by the rapid rise in OD600 at the 48 h time point, to all CBD treatments and the ampicillin treatment. This result is consistent with the literature on S. typhimurium, which accounts for 59% of all ampicillin-resistant Salmonella infections in the United States [30]. These results raise the question of what mechanism S. typhimurium develops to resist the antibacterial activity of CBD. The literature has suggested several mechanisms for S. typhimurium resistance to antibiotics including the multidrug efflux pump AcrAB, OXA-1 β -lactamase, and other beta-lactamase genes (bla). Ampicillin, a beta-lactam class antibiotic, kills bacteria through binding to penicillin-binding proteins in the cytoplasmic membrane [45,46,47]. This mechanism of resistance relies on the inner mechanisms of the bacterial cell, whereas we hypothesize that CBD’s antibacterial activity is a result of membrane disruption. In this work, we demonstrated that the antibacterial activity of CBD was generated through membrane integrity disruption. Therefore, we suggest that S. typhimurium resistance to CBD might be conferred through a resistance mechanism different from that of the ampicillin-resistance mechanism of this bacteria. This could be significant in the application of CBD as a therapeutic agent against ampicillin-resistant S. typhimurium or in a potential co-therapy designed with CBD and ampicillin. Co-therapies typically reduce the ability of bacteria to persist due to bacteria having to rapidly develop multiple resistance mechanisms [27,48,49,50,51]. Additionally, co-therapy strategies have been shown to ‘revert’ resistant bacteria back to a state that is susceptible to antibiotic treatment [52]. These factors further buttress the need to employ CBD in a potential CBD–antibiotic co-therapy once the resistance mechanism of CBD is fully understood. The extended kinetics of S. newington revealed that this strain was more susceptible to CBD treatments of 1.25, 0.125, and 0.0125 μg/mL 24 h after treatment. However, at the 48 h time point, there was a strong development of resistance in all CBD treatments but no resistance development to ampicillin. These results suggest that the mechanism of resistance is different between ampicillin and CBD (Figure 6). Further studies to characterize the resistance developed by S. typhimurium and S. newington against CBD are important to understand how to responsibly develop this potential therapeutic agent. The previous literature has outlined the potential of CBD against several, mostly Gram-positive, bacterial pathogens including Staphylococcus aureus, Streptococcus pneumoniae, and Clostridioides difficile [20,21]. In these studies, there was a reported MIC of 1–5 μg/mL against Staphylococci spp., Clostridioides spp., and Streptococci spp. [20]. More recent studies have reported a MIC of 0.5–2 μg/mL against several clinical isolates of methicillin-resistant Staphylococcus aureus (MRSA). This study focused on a relevant Gram-negative pathogen, Salmonella, whose tolerance of CBD has not previously been studied. Within our study, S. typhimurium and S. newington were shown to have a MIC of 0.125–1.25 μg/mL. These MIC results are consistent with the previous literature exploring the antibacterial capabilities of CBD [20,21]. Results from this study and the studies mentioned earlier [20,21] seem to suggest bacterial susceptibility to CBD within a MIC range of 0.1–5 μg/mL. As we continue to explore potential antibacterial agents, it is essential that we explore the effectiveness of these agents against a multitude of diverse bacterial pathogens. Considering this, our study confirms that CBD does have antibacterial activity against two Gram-negative Salmonella strains, filling a valuable gap in our knowledge of CBD as an antibacterial agent. While this study illuminates the potential of CBD as a therapeutic and fills a void in the current literature, future work is necessary for further development of this bioactive compound as a therapeutic agent. While our study was successful in determining the presence of CBD’s antibacterial activity against S. typhimurium and S. newington, there were limitations. One limitation was that the specific resistance mechanism of S. typhimurium and S. newington against CBD was not investigated in this study. This study simply determined the ability of our Salmonella strains to develop resistance against the CBD treatments administered and did not define the mechanism of resistance. The development of CBD as a novel therapeutic option will require further studies and characterization. Some relevant studies for the progression of CBD as a potential therapeutic include resistance mechanisms to CBD, cytotoxicity of CBD especially to co-therapy situations, immunological response to CBD treatment, CBD function in pathophysiological conditions, and in vivo models. These future studies will serve to further expand our knowledge of CBD as an antibacterial agent with potential therapeutic benefits. 4. Materials and Methods 4.1. CBD Extraction CBD extraction was carried out by Sustainable CBD LLC. C. sativa biomass was weighed, tagged, and recorded in a receiving trailer for processing. Following storage, the biomass was reduced to between 200 and 500 microns and underwent CO2 extraction in an Apeks Transformer for subcritical extraction (Gibraltar Industries Inc., Buffalo, NY, USA). Subcritical extraction was carried out at a target pressure of 1200 psi, chiller temperature of 20–25 °C, propylene glycol percentage of 10%, orifice size of 22, resultant separator pressure of 350–400 psi, resultant separator temperature of −6–4 °C, for an extraction time of approximately 2–3 h. Following subcritical extraction, samples were prepared for decarboxylation prior to supercritical extraction. Hemp biomass was placed in the oven for approximately 100 min at 265 °C to decarboxylate. Once decarboxylation was carried out, an Apeks Transformer was utilized for supercritical extraction (Gibraltar Industries Inc., Buffalo, NY, USA). Supercritical extraction was carried out at a target pressure of 1800 psi, chiller temperature of 37–42 °C, propylene glycol percentage of 10%, orifice size of 18, resultant separator pressure of 350–400 psi, resultant separator temperature of 0–10 °C, for an extraction time of approximately 1–2 h per pound of material. The resulting material then underwent winterization through addition of ethanol to the crude extract. This sample was frozen and then filtered through Buchner funnels and the remaining ethanol was evaporated using a Heidolph rotary evaporator (Heidolph Instruments GmbH & Co.KG, Kelheim, Germany). Distillation was carried out using the Lab Society 5 L short path distillation unit (Lab Society®, Boulder, CO, USA) for further refinement. The resulting product of this procedure was winterized cannabinoid [53,54,55]. 4.2. Gas Chromatography Analysis of CBD CBD extracts were analyzed by gas chromatograph (GC)/mass spectrometry (MS) using Agilent 6890N GC and Agilent 5975 MS (Agilent Technologies Inc., Santa Clara, CA, USA) with a Restek Rxi-5Sil MS with integra guard column (15 m, 0.250 mmID, 0.25 µm df) (Restek Corporation, Bellefonte, PA, USA). A solution of Restek Qualitative Retention Time Index Standard (Restek Corporation, Bellefonte, PA, USA) was used to create the retention time index. The temperature of the injection port was set at 280 °C and the helium gas flow was constant at 1.1 mL/min. The samples were injected in split mode (2:1) with a volume of 1 µL of sample. The GC oven temperature was programmed as follows: initial temperature of 70 °C for 4 min, ramp to 200 °C at 20 °C/min, ramp to 300 °C at 8 °C/min, ramp to 325 °C at 50 °C/min with a 5 min hold, thus requiring a total run time of 30.5 min. The MS transfer line was set to 250 °C, source to 230 °C, and quads to 150 °C. The raw data were processed and analyzed using Agilent Enhanced ChemStation software (Agilent Technologies Inc., Santa Clara, CA, USA). The NIST 2.0 library was used with AMDIS for compound identification. 4.3. Plate Assays for Antibacterial Screening of CBD To qualitatively determine the antibacterial potential of CBD against S. typhimurium LT2 strain MS1868 (a kind gift from Dr. Anthony R. Poteete, University of Massachusetts) and S. newington (also known as S. enterica serovar Anatum var 15+ strain UC1698, a kind gift from Dr. Sherwood R. Casjens, University of Utah), plate assays were utilized. Prior to the assays, CBD was serially diluted to the concentrations of 1.25, 0.125, 0.0125, or 0.00125 μg/mL. Bacterial strains S. typhimurium and S. newington were incubated until the late log-phase (OD > 1) in sterile Luria broth (BD Difco, Franklin Lakes, NJ, USA). To conduct the Kirby–Bauer assay, plates were inoculated with overnight cultures of either S. typhimurium or S. newington and top agar (BD Difco, Franklin Lakes, NJ, USA) to create a bacterial lawn. These plates were then divided into quarters representing the four CBD dilutions (1.25, 0.125, 0.0125, or 0.00125 μg/mL). Sterile paper discs were soaked in a designated dilution of CBD or treated with dH2O (control) and then applied to the agar plates in triplicate. The results were three discs per CBD dilution on the agar plate. Plates were then incubated at 37 °C for 24 h, and bacterial growth was observed to determine zones of inhibition (ZOI) around the CBD-treated discs [56,57,58]. ZOIs were quantitatively assessed using ImageJ (NIH Image, Bethesda, Maryland). This assay was completed in triplicate. For further confirmation of antibacterial activity, spot assays were conducted using lag-phase S. typhimurium and S. newington cultures (OD > 1). Sterile agar plates were inoculated with four 10 μL aliquots of S. typhimurium or S. newington. Each ‘dot’ was then inoculated with 10 μL of one of the four dilutions (1.25, 0.125, 0.0125, or 0.00125 μg/mL) of CBD. Each culture was also spotted and treated with 10 μL of dH2O as the control. The plate was then incubated for 24 h at 37 °C and observed for inhibition of bacterial growth [59]. This assay was completed in triplicate. 4.4. Fluorescence Microscopy of CBD-Treated Salmonella Cells To visualize the effects of CBD treatment on Salmonella cells, fluorescence microscopy was utilized. Salmonella cells were fixed and stained using DAPI (4′,6-diamidino-2-phenylindole) to determine cell membrane integrity. DAPI stain binds exclusively to dsDNA, which is only accessible as a result of compromised membrane integrity and expulsion of the cytoplasmic material including the nucleus out of the cell. Visualization of DAPI fluorescence under the fluorescence microscope (Biotek Cytation™ 3 Automated Fluorescence Microscope) (Agilent Technologies Inc., Santa Clara, CA, USA) confirmed cell membrane damage and hence cell death. Salmonella cells were treated with CBD dilutions (1.25, 0.125, or 0.0125 μg/mL) or left untreated (control). Following treatment, cells were fixed at either 5 min or 30 min and assessed using fluorescence microscopy [27]. 4.5. Bacterial Growth Kinetics To study bacterial growth kinetics, a 96-well plate (Fisherbrand™, Fisher Scientific, Fair Lawn, NJ, USA) was inoculated with 180μL of overnight bacterial cultures of either S. typhimurium or S. newington at an OD600 ≈ 0.5. The cultures were then incubated at 37 °C with rotary shaking at 121 rpm. Measurements of bacterial density (OD600) were taken every hour for 6 h and again at the 12, 24, and 48 h time points using a spectrometer (Molecular Devices SpectraMax® ABS Plus) (Molecular Devices LLC., San Jose, CA, USA). This experiment was completed in triplicate [27,60]. 4.6. Bacterial Growth Kinetics in the Prescence of CBD To study how CBD affects bacterial growth kinetics, three 96-well plates were inoculated with overnight bacterial cultures of either S. typhimurium or S. newington at an OD600 ≈ 0.5. These cultures were then treated with varying concentrations of CBD (1.25, 0.125, 0.0125, or 0.00125 μg/mL). The cultures were then incubated at 37 °C with rotary shaking at 121 rpm. Measurements of bacterial density (OD600) were taken hourly for 6 h and at the 12, 24, and 48 h time points using a spectrometer (Molecular Devices SpectraMax® ABS Plus). This experiment was completed in triplicate [27,60]. 4.7. Comparative Bacterial Growth Kinetics in the Prescence of Ampicillin or CBD To study how ampicillin affects bacterial growth kinetics, three 96-well plates were inoculated with overnight bacterial cultures of either S. typhimurium or S. newington at an OD600 ≈ 0.5. These cultures were then treated with either ampicillin (0.5 μg/mL) or CBD (0.125 μg/mL). The cultures were then incubated at 37 °C with rotary shaking at 121 rpm. Measurements of bacterial density (OD600) were taken hourly for 6 h and at 24 and 48 h time points using a spectrometer (Molecular Devices SpectraMax® ABS Plus). This experiment was completed in triplicate [27,60]. 4.8. SEM Imaging of CBD-Treated S. typhimurium Biofilm S. typhimurium was grown in 6-well plates (Fisherbrand™, Fisher Scientific, Fair Lawn, NJ, USA) supplemented with LB broth until a biofilm was developed. S. typhimurium biofilm samples were left untreated or treated with CBD at a concentration of 0.125 μg/mL, and the morphology and surface microstructure of the bacteria were observed using scanning electron microscopy. Untreated S. typhimurium served as control. Prior to microscopy, the specimens were dried in a vacuum, and sprayed with gold using an EMS Quorum (EMS 150R ES) ion-sputtering instrument and observed through an Analytical Scanning Electron Microscope (SEM) (JEOL JSM-6010LA, Tokyo, Japan) installed with IntouchScope software (JSM-IT200 InTouchScope™ SEM Series, JEOL Solutions, Tokyo, Japan). Sample preparation for SEM was carried out as described by Li et al., 2021 [41] with few modifications. In short, the specimens for SEM were fixed with 10% formaldehyde solution at room temperature for 10 min, washed with PBS solution thrice, and dehydrated serially in 50%, 70%, and 95% absolute ethanol solutions for 10 min each. Finally, the specimens were dried in a vacuum, and sprayed with gold using an EMS Quorum (EMS 150R ES) ion-sputtering instrument and observed through an Analytical Scanning Electron Microscope (SEM) (JEOL JSM-6010LA, Tokyo, Japan) installed with IntouchScope software (JSM-IT200 InTouchScope™ SEM Series, JEOL Solutions, Tokyo, Japan) [61]. 4.9. Statistical Analysis All experiments were performed on independent biological replicates. Statistical significance was determined for control and experimental groups using a paired sample t-test. Data points were excluded if contamination was identified. Statistical analyses were preformed using OriginPro Plus version 2021b (OriginLab Corporation, North Hampton, MA, USA). 5. Conclusions As this facultative anaerobe continues to cause serious public health problems, the need to investigate Salmonella has received much scientific scrutiny [61,62,63,64]. In this study, we demonstrated the antibacterial activity of CBD against two relevant pathogenic bacteria, S. typhimurium and S. newington. Despite the scarce knowledge of the molecular mechanisms of CBD’s mode of action, we propose that the antibacterial activity might be due to membrane integrity disruption, and this was verified through the utilization of plate assays, fluorescence microscopy, and kinetic studies. These experiments confirmed that CBD has antibacterial activity against our target bacteria. Additionally, our comparative studies showed that CBD has antibacterial activity similar to ampicillin with a MIC roughly one-fifth of ampicillin. We observed the resistance development of S. typhimurium and S. newington against CBD treatment. Resistance development was observed after 48 h. These results suggest that Salmonella resistance to CBD might be conferred through a different mechanism than antibiotic resistance. These results posed the question of CBD–antibiotic co-therapy as a potential novel application. Finally, we observed CBD to be effective against S. typhimurium biofilms. This study further progresses our current knowledge on the effectiveness of CBD as an antibacterial agent and demonstrates the effectiveness of CBD against Gram-negative bacteria, S. typhimurium and S. newington. Acknowledgments The authors acknowledge Benjamin Bramlett of Sustainable CBD LLC., Salem, AL for help extracting the CBD. We acknowledge Melissa Boersma of Auburn University for help with gas chromatography. In addition, we acknowledge Derrick Dean and Vida Dennis for their HBCU-RISE grant support. We also acknowledge Alabama State University, C-STEM for supplies and laboratory space. The authors acknowledge receiving funding from the United States Department of Education, Title III- HBGI-RES. Author Contributions Conceptualization, J.A.A.; methodology, L.G. and J.A.A.; software, L.G. and J.A.A.; validation, O.S.A. and M.S.-F.; formal analysis, L.G., J.A.A., O.S.A., J.X. and M.S.-F.; investigation, L.G., J.A.A., O.S.A., J.X. and M.S.-F.; resources, J.A.A., O.S.A., R.V. and M.S.-F.; data curation, L.G., J.A.A., O.S.A., J.X. and M.S.-F.; writing—original draft preparation, L.G.; writing—review and editing, L.G., J.A.A., J.X. and O.S.A.; visualization, L.G. and J.A.A.; supervision, J.A.A. and M.S.-F.; project administration, O.S.A. and M.S.-F.; funding acquisition, O.S.A. and M.S.-F. All authors have read and agreed to the published version of the manuscript. Funding This research was funded by United States Department of Education, Title III- HBGI-RES, and National Science Foundation; and HBCU-RISE; grant number: 1646729. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Data are contained within the article. Conflicts of Interest The authors declare no conflict of interest. Sample Availability Samples of the compounds are available from the authors. Figure 1 GC peaks confirm CBD as the primary compound present within our C. sativa extract. (A) represents CBD (in grey color) internal standard for GC with an 8.22000 expected RT (min). Peak detected at 10.09 min (A) represents the expected RT (min) of THC (in white color), which was not detected within our sample (B). (B) represents GC analysis of our CBD extract featuring an 8.2300 RT (min), confirming the sample is composed of CBD. Secondary readings around the base of the sample are consistent to those seen within our CBD internal standard (A). Figure 2 Plate assays confirm inhibitory activity of CBD. Quantitative analysis of Kirby–Bauer assay results of S. typhimurium (A) and S. newington (B). Images from Kirby–Bauer assay and CBD-produced ZOIs against S. typhimurium and S. newington (C). Images from spot assay featuring S. typhimurium and S. newington treated with decreasing concentrations (left to right) of CBD (D). (A) * p-value = 0.000961, ** p-value = 0.0117, *** p-value = 0.7179. (B) * p-value = 0.00000343, ** p-value = 0.000480, *** p-value = 0.00832. Figure 3 Treatment of lag-phase S. typhimurium (A) and S. newington (B) with CBD. The effect of several CBD dilutions on OD600 was recorded in triplicate. Figure 4 (A). Fluorescence microscopy of CBD-treated Salmonella cells stained with DAPI. Images were acquired at the time points of 5 min. DAPI staining was utilized to assess membrane integrity following treatment with CBD. (B). Fluorescence microscopy of CBD-treated Salmonella cells stained with DAPI. Images were acquired at the time points of 30 min. DAPI staining was utilized to assess membrane integrity following treatment with CBD. (C). Quantitative analysis of membrane integrity via densitometry analysis of fluorescence after DAPI staining. Figure 5 Comparative efficacy of CBD and antibiotic treatment of Salmonella typhimurium (A) and Salmonella newington (B). Salmonella cells were cultured overnight and then treated with either CBD or ampicillin. OD600 was measured hourly over a 6 h time period to observe bacterial growth following treatment. Figure 6 Development of resistance to CBD and ampicillin treatments by S. typhimurium (A) and S. newington (B) examined through extended kinetics. Bacterial cultures were treated with either ampicillin, CBD dilutions, or dH2O. OD600 was recorded over a period of 48 h. Figure 7 SEM images depicting the comparison between untreated (control) and treated (0.125 µg/mL CBD) S. typhimurium biofilms. Untreated S. typhimurium biofilm resulted in continued persistence (A). CBD treatment at 1.25 µg/mL resulted in a reduction of S. typhimurium characterized by fragmented cells suggesting S. typhimurium cell death (B). Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Long T. Wagner M. Demske D. Leipe C. Tarasov P. Cannabis in Eurasia: Origin of human use and Bronze Age trans-continental connections Veget. Hist. Archaeobot. 2017 26 245 258 10.1007/s00334-016-0579-6 2. Zuardi A.W. History of cannabis as a medicine: A review Rev. Bras. Psiquiatr. 2006 28 153 157 10.1590/S1516-44462006000200015 16810401 3. Crini G. Lichtfouse E. Chanet G. Crini N. Applications of hemp in textiles, paper industry, insulation and building materials, horticulture, animal nutrition, food and beverages, nutraceuticals, cosmetics and hygiene, medicine, agrochemistry, energy production and environment: A review Environ. Chem. Lett. 2020 18 1451 1476 10.1007/s10311-020-01029-2 4. Rehman M. Fahad S. Du G. Cheng X. Yang Y. Tang K. Liu L. Liu F.H. Deng C. Evaluation of hemp (Cannabis sativa L.) as an industrial crop: A review Environ. Sci. Pollut. Res. 2021 28 52832 52843 10.1007/s11356-021-16264-5 34476693 5. Aldrich M. History of therapeutic cannabis Cannabis in Medical Practice Mathre M.L. Mc Farland Jefferson, NC, USA 1997 35 55 6. Brand J.E. Zhao Z. Cannabis in Chinese medicine: Are some traditional indications references in ancient literature related to cannabinoids Front. Pharmacol. 2017 8 1663 9812 10.3389/fphar.2017.00108 7. Amar M.B. Cannabinoids in medicine: A review of their therapeutic potential J. Ethnopharmacol. 2006 105 1 25 10.1016/j.jep.2006.02.001 16540272 8. World Health Organization Annex 1: Extract from the Report of the 41th Expert Committee on Drug Dependence: Cannabis and Cannabis Related Substances, p. 4 2019 Available online: https://www.who.int/medicines/access/controlled-substances/Annex_1_41_ECDD_recommendations_cannabis_22Jan19pdf (accessed on 19 January 2022) 9. Clementi A. High time for a change: How the relationship between signatory countries and the United Nations Conventions governing narcotic drugs must adapt to foster a global shift in cannabis law Brook. J. Int’L Law 2021 46 603 10. Scadron T.E. Koelble T. Bongalong: Marketing Cannabis Products in a Stigmatized Environment GSB: 2018-5 Graduate School of Business, University of Cape Town Cape Town, South Africa 2018 11. Geiger-Oneto S. Sprague R. Cannabis regulatory confusion and its impact on consumer adoption Am. Bus. Law. J. 2020 57 735 772 10.1111/ablj.12171 12. Peters J. Chien J. Contemporary routes of cannabis consumption: A primer for clinicians J. Osteopath. Med. 2018 118 67 70 10.7556/jaoa.2018.020 13. Kalant H. Medicinal use of cannabis: History and current status Pain Res. Manag. 2001 6 12 10.1155/2001/469629 14. Solimini R. Rotolo M.C. Pichini S. Pacifici R. Neurological disorders in medical use of cannabis: An update CNS Neurol. Disord. Drug Targets 2017 16 527 533 10.2174/1871527316666170413105421 28412919 15. Fiani B. Sarhadi K.J. Soula M. Zafar A. Quadri S.A. Current application of cannabidiol (CBD) in the management and treatment of neurological disorders Neurol. Sci. 2020 41 3085 3098 10.1007/s10072-020-04514-2 32556748 16. Waskow I.E. Olsson J.E. Salzman C. Katz M.M. Psychological effects of tetrahydrocannabinol Arch. Gen. Psychiatry 1970 22 97 107 10.1001/archpsyc.1970.01740260001001 5410582 17. Karniol I.G. Shirakawa I. Takahashi R.N. Knobel E. Musty R.E. Effects of delta9-tetrahydrocannabinol and cannabinol in man Pharmacology 1975 13 502 512 10.1159/000136944 1221432 18. Gray R.A. Whalley B.J. The proposed mechanisms of action of CBD in epilepsy Epileptic Disord. 2020 22 10 15 10.1684/epd.2020.1135 32053110 19. Pichini S. Malaca S. Gottardi M. Pérez-Acevedo A.P. Papaseit E. Perez-Maña C. Farré M. Pacifici R. Tagliabracci A. Mannocchi G. F.P. UHPLC-MS/MS analysis of cannabidiol metabolites in serum and urine samples. Application to an individual treated with medical cannabis Talanta 2021 223 121772 10.1016/j.talanta.2020.121772 33298281 20. Appendino G. Gibbons S. Giana A. Pagani A. Grassi G. Stavri M. Smith E. Rahman M.M. Antibacterial cannabinoids from Cannabis sativa: A structure-activity study J. Nat. Prod. 2008 71 1427 1430 10.1021/np8002673 18681481 21. Blaskovich M.A.T. Kavanagh A.M. Elliott A.G. Zhang B. Ramu S. Amado M. Lowe G.J. Hinton A.O. Pham D.M.T. Zuegg J. The antimicrobial potential of cannabidiol Commun. Biol. 2021 4 7 10.1038/s42003-020-01530-y 33469147 22. van Klingeren B. ten Ham M. Antibacterial activity of Δ9-tetrahydrocannabinol and cannabidiol Antonie. Van. Leeuwenhoek 1976 42 9 12 10.1007/BF00399444 1085130 23. Iseppi R. Brighenti V. Licata M. Lambertini A. Sabia C. Messi P. Pellati F. Benvenuti S. Chemical characterization and evaluation of the antibacterial activity of essential oils from fibre-type Cannabis sativa L. (Hemp) Molecules 2019 24 2302 10.3390/molecules24122302 24. Livermore D.M. Has the era of untreatable infections arrived? J. Antimicrob. Chemother. 2009 64 i29 i36 10.1093/jac/dkp255 19675016 25. WHO Antimicrobial Resistance 2020 Available online: https://www.who.int/healthtopics/antimicrobial-resistance (accessed on 19 January 2022) 26. Ayariga J.A. Gildea L. Wu H. Villafane R. The Ɛ34 phage tailspike protein: An in vitro characterization, structure prediction, potential interaction with S. newington LPS and cytotoxicity assessment to animal cell line bioRxiv 2021 10.1101/2021.09.20.461090 27. Gildea L. Ayariga J.A. Robertson B.K. Villafane R. P22 phage shows promising antibacterial activity under pathophysiological conditions Arch. Microbiol. Immunol. 2022 6 81 100 10.26502/ami.93650078 28. Ardal C. Balasegaram M. Laxminarayan R. McAdams D. Outterson K. Rex J.H. Sumpradit N. Antibiotic development–economic, regulatory, and societal challenges Nat. Rev. Microbiol. 2020 18 267 274 10.1038/s41579-019-0293-3 31745330 29. Sharland M. Gandra S. Huttner B. Moja L. Pulcini C. Zeng M. Mendelson M. Cappello B. Cooke G. Magrini N. Encouraging AWaRe-ness and discouraging inappropriate antibiotic use-the new 2019 Essential medicines list becomes a global antibiotic stewardship tool Lancet Infect. Dis. 2019 19 1278 1280 10.1016/S1473-3099(19)30532-8 31782385 30. CDC Antibiotic Resistance Threats in the United States 2019 Available online: https://www.cdc.gov/DrugResistance/Biggest-Threats.html (accessed on 19 January 2022) 31. Slayton R.B. Toth D. Lee B.Y. Tanner W. Bartsch S.M. Khader K. Wong K. Brown K. McKinnell J.A. Ray W. Vital signs: Estimated effects of a coordinated approach for action to reduce antibiotic-resistant infections in health care facilities-United States MMWR Morb. Mortal. Wkly. Rep. 2015 64 826 831 10.15585/mmwr.mm6430a4 26247436 32. White C.M.A. Review of human studies assessing cannabidiol’s (CBD) therapeutic actions and potential J. Clin. Pharmacol. 2019 59 923 934 10.1002/jcph.1387 30730563 33. Watt G. Karl T. In vivo evidence for therapeutic properties of cannabidiol (CBD) for alzheimer’s disease Front. Pharmocol. 2017 8 20 10.3389/fphar.2017.00020 34. Solowij N. Broyd S.J. Beale C. Prick J.A. Greenwood L.M. van Hell H. Suo C. Galettis P. Pai N. Fu S. Therapeutic effects of prolonged cannabidiol treatment on psychological symptoms and cognitive function in regular cannabis users: A pragmatic open-label clinical trial Cannabis Cannabinoid Res. 2018 3 21 34 10.1089/can.2017.0043 29607408 35. World Health Organization Cannabidiol (CBD): Critical Review Report: Expert Committee on Drug Dependence, Fortieth Meeting World Health Organization Geneva, Switzerland 2018 36. Kosgodage U.S. Matewele P. Awamaria B. Kraev I. Warde P. Mastroianni G. Nunn A.V. Guy G.W. Cannabidiol is a novel modulator of bacterial membrane vesicles Front. Cell. Infect. Microbiol. 2019 9 2235 2988 10.3389/fcimb.2019.00324 37. Bryan F.L. Foodborne diseases in the United States associated with meat and poultry J. Food Prot. 1980 43 140 150 10.4315/0362-028X-43.2.140 30822910 38. Heinitz M.L. Ruble R.D. Wagner D.E. Tatini S.R. Incidence of Salmonella in fish and seafood J. Food. Prot. 2000 63 579 592 10.4315/0362-028X-63.5.579 10826714 39. Park S.Y. Woodward C.L. Kubena L.F. Nisbet D.J. Birkhold S.G. Ricke S.C. Environmental dissemination of foodborne Salmonella in preharvest poultry production: Reservoirs, critical factors, and research strategies Crit. Rev. Environ. Sci. Technol. 2008 38 73 111 10.1080/10643380701598227 40. White D.G. Zhao S. Sudler R. Ayers S. Friedman S. Chen S. McDermott P. McDermott S. Wagner D. Meng J. The isolation of antibiotic-resistant Salmonella from retail ground meats N. Engl. J. Med. 2001 345 1147 1154 10.1056/NEJMoa010315 11642230 41. Iseppi R. Mariani M. Condò C. Sabia C. Messi P. Essential oils: A natural weapon against antibiotic-resistant bacteria responsible for nosocomial infections Antibiotics 2021 10 417 10.3390/antibiotics10040417 33920237 42. Harrell J.E. Hahn M.M. D’Souza S.J. Vasicek E.M. Sandala J.L. Gunn J.S. McLachlan J.B. Salmonella biofilm formation, chronic infection, and immunity within the intestine and hepatobiliary tract Front. Cell. Infect. Microbiol. 2021 10 910 10.3389/fcimb.2020.624622 33604308 43. Atalay S. Jarocka-Karpowicz I. Skrzydlewska E. Antioxidative and anti-inflammatory properties of cannabidiol Antioxidants 2020 9 21 10.3390/antiox9010021 31881765 44. di Giacomo V. Recinella L. Chiavaroli A. Orlando G. Cataldi A. Rapino M. Di Valerio V. Politi M. Antolini M.D. Acquaviva A. Metabolomic profile and antioxidant/anti-inflammatory effects of industrial hemp water extract in fibroblasts, keratinocytes and isolated mouse skin specimens Antioxidants 2021 10 44 10.3390/antiox10010044 45. Nikaido H. Basina M. Nguyen V. Rosenberg E.Y. Multidrug efflux pump AcrAB of Salmonella typhimurium excretes only those β-Lactam antibiotics containing lipophilic side chains J. Bacteriol. 1998 180 17 10.1128/JB.180.17.4686-4692.1998 46. Ruiz J. Capitano L. Nuñez L. Castro D. Sierra J.M. Hatha M. Borrego J.J. Vila J. Mechanisms of resistance to ampicillin, chloramphenicol and quinolones in multiresistant Salmonella typhimurium strains isolated from fish J. Antimicrob. Chemother. 1999 43 699 702 10.1093/jac/43.5.699 10382892 47. Eguale T. Birungi J. Asrat D. Njahira M.N. Njuguna J. Gebreyes W.A. Gunn J.S. Djikeng A. Engidawork E. Genetic markers associated with resistance to beta-lactam and quinolone antimicrobials in non-typhoidal Salmonella isolates from humans and animals in central Ethiopia Antimicrob. Resist. Infect. Control 2017 6 13 10.1186/s13756-017-0171-6 28105330 48. Lukacik P. Barnard T.J. Keller P.W. Chaturvedi K.S. Seddiki N. Fairman J.W. Noinaj N. Kirby T.L. Henderson J.P. Steven A.C. Structural engineering of a phage lysin that targets gram negative pathogens Proc. Natl. Acad. Sci. USA 2012 109 9857 9862 10.1073/pnas.1203472109 22679291 49. Lin D.M. Koskella B. Lin H.C. Phage therapy: An alternative to antibiotics in the age of multi-drug resistance World J. Gastrointest. Pharmacol. Ther. 2017 8 162 173 10.4292/wjgpt.v8.i3.162 28828194 50. Ghosh C. Sarkar P. Issa R. Haldar J. Alternatives to conventional antibiotics in the era of antimicrobial resistance Trends Microbiol. 2019 27 323 338 10.1016/j.tim.2018.12.010 30683453 51. D’Accolti M. Soffritti I. Mazzacane S. Caselli E. Bacteriophages as a potential 360-degree pathogen control strategy Microorganisms 2021 9 261 10.3390/microorganisms9020261 33513949 52. Lorenzi V. Muselli A. Bernardini A.F. Berti L. Pagès J.M. Amaral L. Bolla J.M. Geraniol restores antibiotic activities against multidrug-resistant isolates from gram-negative species Antimicrob. Agents Chemother. 2009 53 2209 2211 10.1128/AAC.00919-08 19258278 53. Sustainable CBD LLC Sustainable CBD LLC Standard of Procedures (SOP) 1–6 Sustainable CBD LLC Salem, AL, USA 2020 54. Valizadehderakhshan M. Shahbazi A. Kazem-Rostami M. Todd M.S. Bhowmik A. Wang L. Extraction of cannabinoids from Cannabis sativa L. (Hemp)—Review Agriculture 2021 11 384 10.3390/agriculture11050384 55. Grijó D.R. Osorio I.A.V. Cardozo-Filho L. Supercritical extraction strategies using CO2 and ethanol to obtain cannabinoid compounds from Cannabis hybrid flowers J. CO2 Util. 2018 28 174 180 10.1016/j.jcou.2018.09.022 56. Kirby W.M.M. Yoshihara G.M. Sundsted K.S. Warren J.H. Clinical usefulness of a single disc method for antibiotic sensitivity testing Antibiot. Annu. 1957 892 897 Available online: https://pubmed.ncbi.nlm.nih.gov/13425478/ (accessed on 19 January 2022) 57. Clinical Laboratory Standards Institute Performance Standards for Antimicrobial Disk Susceptibility Tests 9th ed. CLSI Document M2-A9 Clinical Laboratory Standards Institute Malvern, PA, USA 2006 58. Jorgensen J.H. Turnidge J.D. Susceptibility test methods: Dilution and disk diffusion methods Man. Clin. Microbiol. 2007 9 1152 1172 59. Powell R. Neilan J. Gannon F. Plaque dot assay Nucleic Acids Res. 1986 14 1541 10.1093/nar/14.3.1541 2937017 60. Kalapala Y.C. Sharma P.R. Agarwal R. Antimycobacterial potential of mycobacteriophage under disease-mimicking conditions Front. Microbiol. 2020 11 583661 10.3389/fmicb.2020.583661 33381088 61. Ayariga J.A. Gildea L. Villafane R. Ɛ34 phage tailspike protein is resistant to trypsin and inhibits Salmonella biofilm formation Enliven Microb. Microb. Tech. 2022 9 002 62. Mittrücker H.W. Kaufmann S.H. Immune response to infection with Salmonella typhimurium in mice J. Leukoc. Biol. 2000 67 457 463 10.1002/jlb.67.4.457 10770276 63. Pui C.F. Wong W.C. Chai L.C. Tunung R. Jeyaletchumi P. Hidayah N. Ubong A. Farinazleen M.G. Cheah Y.K. Son R. Salmonella: A foodborne pathogen Int. Food Res. J. 2011 18 465 473 64. Ayariga J.A. Abugri D. Amrutha B. Villafane R. Capsaicin potently blocks Salmonella typhimurium invasion of Vero cells bioRxiv 2022 10.1101/2022.01.02.474733

PMC009xxxxxx/PMC9099640.txt

==== Front Polymers (Basel) Polymers (Basel) polymers Polymers 2073-4360 MDPI 10.3390/polym14091756 polymers-14-01756 Article Modeling and Experimental Verification of the Performance of Polymer Composite Reinforcing Bars of Different Types in Concrete of Different Density https://orcid.org/0000-0002-6173-9365 Beskopylny Alexey N. 1* https://orcid.org/0000-0002-0364-5504 Stel’makh Sergey A. 2 https://orcid.org/0000-0001-5376-247X Shcherban’ Evgenii M. 2 Mailyan Levon R. 3 Meskhi Besarion 4 Efremenko Innessa 5 Varavka Valery 6 Beskopylny Nikita 7 Dotsenko Natal’ya 8 Zheng Qingbin Academic Editor 1 Department of Transport Systems, Faculty of Roads and Transport Systems, Don State Technical University, Gagarin, 1, 344003 Rostov-on-Don, Russia 2 Department of Engineering Geology, Bases, and Foundations, Don State Technical University, 344003 Rostov-on-Don, Russia; sergej.stelmax@mail.ru (S.A.S.); au-geen@mail.ru (E.M.S.) 3 Department of Roads, Don State Technical University, 344003 Rostov-on-Don, Russia; lrm@aaanet.ru 4 Department of Life Safety and Environmental Protection, Faculty of Life Safety and Environmental Engineering, Don State Technical University, Gagarin, 1, 344003 Rostov-on-Don, Russia; reception@donstu.ru 5 Project Management Center, Don State Technical University; Gagarin, 1, 344000 Rostov-on-Don, Russia; i.efremenko@sci.donstu.ru 6 Research and Education Center “Materials”, Don State Technical University, Gagarin sq., 1, 344003 Rostov-on-Don, Russia; varavkavn@gmail.com 7 Department Hardware and Software Engineering, Don State Technical University, 344003 Rostov-on-Don, Russia; beskna@yandex.ru 8 Department of Technological Engineering and Expertise in the Construction Industry, Don State Technical University, 344003 Rostov-on-Don, Russia; natalya_1998_dotsenko@mail.ru * Correspondence: besk-an@yandex.ru; Tel.: +7-863-273-8454 26 4 2022 5 2022 14 9 175603 4 2022 20 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Currently, there is a scientific and practical deficit in new methods of integrated technological and design solutions based on improving the properties of concrete as the primary material that perceives compressive loads, and its joint work with various types of reinforcing rods. A new system using an integrated engineering approach to the design of building structures is proposed, which involves minimizing their cost and weight through numerical simulations and an experimental verification of the operation of reinforcing bars made of various materials in concrete of various densities. The control of the bearing capacity of reinforced building structures on the example of compressed elements is proposed to be carried out using the developed recipe-technological methods at the manufacturing stage. The economic and environmental efficiency of nano modification with the help of production waste and the use of lightweight dispersion-reinforced concrete to obtain such structures was revealed. The most effective concrete formulations showed strength gains ranging from 10% to 34%. Ultimately, this led to an increase in the bearing capacity of the elements up to 30%. The application of such an integrated lean approach will allow saving up to 20% of resources during construction. steel reinforcement polymer composite reinforcement fiber reinforcement nanomodified concrete lightweight dispersion-reinforced concrete This research received no external funding. ==== Body pmc1. Introduction Outdated technologies in the conditions of intensively growing volumes of buildings and structures being erected are a problem for modern construction, and they require improvement based on a scientific approach. For many decades, including the present, reinforced concrete has remained the main structural material for buildings and structures of various levels of responsibility. Many attempts by modern scientists and engineers to improve concrete technology have been successful; however, until now, the issues of transition to a qualitatively new level of structural composites remain open. Some of the most relevant areas in terms of the construction of buildings and structures made of reinforced concrete are the following:- searching for new solutions regarding concrete, from which reinforced concrete structures are built, namely: recipe-technological solutions that meet the latest science and technology criteria; - constructive solutions comprising geometric shapes and reinforcing cages, which are the basis of reinforced concrete and an integral part for the perception of loads, including ultra-high ones; - finally, engineering solutions related to the specific replacement and transition from traditional types of materials and structures to modified components, which are based on fundamentally new approaches in terms of technology, design, and calculations. In this regard, an analysis of the currently existing solutions in terms of concrete science and the reinforcement of concrete structures is required, as well as an analysis of the types of concrete and reinforcement used in various products, and structures made of heavy concrete, including the possibilities of their combinations, layout, and collaboration. An application of steel elements in concrete is considered in many works, among which [1,2,3,4,5,6,7,8,9,10,11,12] are highlighted. Concrete-filled rectangular steel tubes (CFRST) are currently popular in the construction of high-rise buildings and bridge structures. The study [2] analyzed the behavior of CFRSTs with different central angles under eccentric compression. A reasonable simplified approach to the calculation of the ultimate eccentric bearing capacity CFRST [2] is presented and confirmed. The influence of steel strength class, steel coefficient, types of clamps, and flexibility coefficient on the bearing capacity of such elements was studied in [3]. It was proven that the use of high strength steel and the improvement of the steel ratio can greatly improve the bearing capacity of SRC columns. A modified formula was proposed for calculating the bearing capacity of SRC columns, based on the limiting effect of stirrups and steel on concrete [3]. Steel fibers are used to improve the plasticity of high-strength lightweight [1] and ultra-high-performance fiber-reinforced concrete [9]. Steel fibers in the amount of 0–1.5% (by volume) were used in the concrete mix. High-strength fibrous lightweight concrete can perform better than ordinary concrete under extreme loads because it exhibited a much higher ductility [1]. A mixture with a large volume (6%), and a combination of two segments of steel fibers (3% each), a ratio of water and binder of 0.16%, and a ratio of superplasticizer and binder of 6.1%, showed the highest strength and deformability in the plastic region [9]. Reinforced concrete beams made of a fine-grained fiber composite with the addition of steel fibers in an amount of 1.2% by volume of the composite, using secondary filler, were studied for shear strength in terms of bending under the action of a transverse force, as well as cracking forces that cause the appearance of the first diagonal crack. The proposed new fibrous composite with fine aggregates can, in some cases, serve as an alternative to conventional concrete [10]. The behavior of Ultra High-Performance Concrete (UHPC) beams was studied using a finite element model that takes into account the response of UHPC and rebar to tensile and compressive stress-strain, the bond between concrete and reinforcing steel, and the effects of work hardening in bars and UHPC. The proposed model was claimed for application in the field of quantifying the contribution of stirrups and concrete to the shear strength of beams to study the possibility of removing transverse reinforcement in UHPC beams [13]. Hybrid fiber reinforced concrete, reinforced with two types of fibers (steel and polyvinyl alcohol), was studied to create a stress-strain model to predict the strength of steel fiber concrete in the range from 40 to 120 MPa. It was shown that an increase in the content of steel fiber to a certain level, at a constant amount of fiber from polyvinyl alcohol, leads to an increase in stress resistance values [14]. The effect of adding different types of fibers to concrete mixtures on the shear behavior of two-span fiber-reinforced concrete beams with or without shear reinforcement was studied by the authors in [15]. Steel or basalt fiber was used with dosages of 78.5 and 5 kg/m3. It is summarized that the shear capacity increased in concrete beams with steel or basalt fibers. In this case, the type of fracture changed from shear (brittle) to flexural shear (less brittle) [15]. Increasing the bearing capacity of reinforced concrete structures, reducing the consumption of materials, and ensuring quality, is possible using steel rods of increased strength and various types of fibers [16,17]. Research into High Performance Fiber Reinforced Concrete (HPFRC) is quite common at the present time. Much attention is paid to factors such as the dosage of fiber in the composition of concrete, as well as its surface and length, which have the greatest impact on the structure of fresh and hardened concrete, as well as its mechanical properties [18,19]. The works [20,21] contributed hugely to the development of the study of fibrous porous media. The proposed fractal models successfully described the dependences of microstructural and electrokinetic parameters of porous media [20], permeability, and the Kozeny–Carman constant [21] on porosity, equivalent particle diameter, ratio of the minimum pore radius to the maximum pore radius, molar concentration, zeta potential, fractal dimension tortuosity, and fiber diameter [20,21]. Fiber-reinforced concrete is widely used to increase the durability of reinforced structures. The redistribution behavior of bending and shear moments in fiberglass-reinforced and polymer-reinforced (GFRP) continuous concrete beams was studied in [22]. The paper [23] presents an estimate of the bending strength of concrete with the addition of 2.0 and 3.0 kg/m3 of synthetic fibers of various geometries and shapes. An increase in bending strength up to 13.5%, depending on the type of mixture, and an increase in plasticity were noted. It is summarized that the proposed fiber-reinforced concrete mixtures cannot replace traditional reinforcement as steel rods [23]. Modeling and calculation of hollow reinforced concrete elements were considered in [24,25]. The reinforcement of hollow concrete cylinders with nickel–titanium alloy (Ni–Ti) memory alloy (SMA) wires wound around them was considered using a model for analyzing thermal stresses in a concrete shell. Almost a year after prestressing with Ni–Ti SMA wire, it was confirmed that the residual stress in the wire is maintained and effective for a long time [24]. Simulation of the earthquake behavior of corroded reinforced concrete hollow poles by FB-FEM was used to study the effect at failure using collision calculations and stepwise dynamic analysis considering different corrosion rates [25,26]. Textile mortar reinforcement (TRM) of reinforced concrete (RC) columns through the shell, under combined axial and cyclic loads, showed that increasing the length of the shell improves the lateral deformation capacity, and linearly increases the length of the plastic hinge up to a constraint factor of 0.2. Mortars with higher flexural strength resulted in somewhat greater deformation capacity; however, the difference in the compressive strength of the solution did not affect the ultimate capacity for lateral deformation [27]. An increase in strength and deformation characteristics due to the use of polymer rods was considered in [6,7,28,29,30,31,32,33]. Replacing steel rods with polymer ones can significantly increase the ductility and strength of reinforced concrete columns. The results showed that carbon fiber rods increase the rigidity of the models, whereas steel rods increase the coefficient of energy absorption and ductility [28]. The effect of the joint work of polymer rods and glass fiber in the context of the relationship between the strength characteristics in bending and the density of concrete was studied in [29]. An increase in the experimental shear capacity of concrete beams by using a carbon fiber reinforced polymer with improved compressive strength was considered in [30]. The joint use of experimental and computational methods of high-strength concrete and mortar subjected to a compressive load was reflected in [34]. At a lower studied stress level, basalt coarse aggregate improved the fatigue characteristics of concrete. Signs of a negative effect were observed at a higher level of stress [34]. Analytical and numerical approaches to the calculation of elements of reinforced concrete structures with basalt fibers were considered based on [35]. It is possible to reduce environmental pollution by using sustainable materials to produce sustainable concrete. Such materials are reinforcing fibers (steel, polypropylene, carbon fiber), recycled materials (tire rubber, crushed glass, plastic, industrial waste), as well as organic and inorganic elements such as concrete aggregates and reinforcing elements. Some resistant materials added to cement can improve the compressive and bending strength of concrete elements [36,37,38]. Compressed and bendable concrete elements with bounding reinforcement grids can be calculated according to a common method [39], considering all the main factors affecting the mechanical properties of volumetrically compressed concrete [32,40]. Works that studied the modification of concretes with polymers in various dosages, with various sizes of large aggregates, at various sample testing temperatures [41], and calcium carbonate nanoparticles of a rationally selected dosage [42], were also aimed at improving the characteristics and structure of concretes. Summarizing the results of the review and its analysis, it should be noted that the scientific deficit and the practical gap based on it is associated with the applied development of new methods of complex technological and design solutions. They are based, firstly, on improving the properties of concrete as the main material that perceives compression loads and its joint work with various types of reinforcing rods. Focusing on the main principle of design improvement, maximum reduction of the weight and cost of the structure, and at the same time increasing its manufacturability, there is a need to check the compatibility of specific types of concrete of certain classes, and specific types of reinforcing bars made of certain materials with certain specified characteristics. The works studied during the literature review are relevant, but they need to develop the theoretical ideas and practical recommendations received by these authors. In the work of the considered authors, several scientific and applied problems are identified that should be solved. These problems stem from the poor adaptability of traditional building technologies for modern complicated construction conditions. In particular, these problems are concerned with the large weight of buildings, complex engineering geological conditions, and dense urban development. These problems can be effectively solved, and the data already obtained by the authors of the considered works should be developed in the direction of creating new lightweight types of structures with various types of reinforcement in various types of concrete. Such questions have not yet been investigated and are unresolved. Our study aims to address and answer these questions. After a theoretical review of experimental studies and analytical interpretation, the novelty of our work will be the developed integrated approach that will solve these problems. The objectives of the study are:- the formulation of a working hypothesis, development of an experiment plan, and based on a selection of the most significant factors that are based on the analysis, results that affect the bearing capacity of compressed reinforced concrete elements; - conducting large-scale numerical experiments aimed at identifying bottlenecks and points to strengthen solutions in terms of technology and design; - the determination of directions and vectors that can be verified by laboratory-physical experiments; - setting up a physical experiment in a laboratory to confirm a hypothesis in terms of technology; - carrying out a design check of the proposed technological solution and, on that basis, developing a design proposal with justification from the point of view of efficiency on operational and economic grounds. The scientific novelty of the research is:- from the point of view of theory, the study of compatibility and the development of existing ideas about such compatibility concerning various types of materials and their interaction from the point of view of the stress-strain state, and obtaining new knowledge about their joint work. Both stone materials (concretes), metals, and various kinds of composite materials are studied thusly, and the theory and empirical basis for the interactions of various materials in the body of a single composite under a load is developed; - various kinds of experimental data obtained with the help of special calculation programs and numerical experiments were investigated and applied, as well as physically verified and tested; - complex technological and design solutions are presented, combining simplified methods of technological design, applicable, first, for further scientific research in this direction, and second, for the engineering and construction industry in the construction of buildings and structures of a new type. 2. Materials and Methods One of the most important parts of the proposed integrated approach is the choice of research methodology. It should be understood that we are dealing with an even greater synergistic effect, which, in our opinion, arises from a simultaneous careful analysis of material science and the design components of the study. In this regard, our methodology is divided into the choice of initial components, recipes and technologies for laboratory physical experiments, as well as the appropriate material, software base, and verified calculation methods from the point of view of numerical experiments; therefore, in the section “Materials and Methods”, it is necessary to provide materials for conducting laboratory physical experiments, and the subsections detailing the methods should not only indicate the methods of testing and research in the conditions of the laboratory of the planned materials, but also the method of conducting numerical experiments using modern software. All this will allow us to obtain the necessary synergistic effect, which takes into account two components: materials science and design. 2.1. Materials Reinforced concrete columns with a cross section of 400 mm × 400 mm, and a length of 3000 mm, 6000 mm, and 9000 mm, made of class B30 concrete, were chosen as the basic object of study. The subject of the study was the characteristics of the columns depending on the changing technological and design factors. A previously developed composition with micro silica nano-modifier was used and adopted as heavy concrete for the experiments. It allows, using industrial waste, as an additive to heavy concrete, to achieve, first, a denser packing of particles, improving the quality of the structure of concrete, and allowing this concrete to acquire improved mechanical structural characteristics with an increase in the bearing capacity of elements obtained from such a concrete [43,44]. Samples of heavy concrete with nano-modified micro silica for compression testing are shown in Figure 1. Information about the initial components used in the manufacturing of concrete using a nanomodifier is presented below. When conducting research, we used Portland cement of the CEM 0 52.5N brand (Novoroscement, Novorossiysk, Russia), the physical and mechanical characteristics of which are presented in Table 1, the chemical composition—in Table 2, and the mineralogical composition—in Table 3. Granite crushed stone (Pavlovsknerud JSC, Pavlovsk, Russia) was used as a large dense aggregate, and slag pumice (Stroymir LLC, Lipetsk, Russia) was used as a porous aggregate. The physical and mechanical characteristics of the aggregate are presented in Table 4. Quartz sand (OOO Quartz Sands, Semenov, Russia) was used as a fine aggregate, the physical characteristics of which are presented in Table 5. In numerical experiments, steel reinforcing bars (Tyazhpromarmatura, Aleksin, Russia), and polymer composite reinforcement (PCR) (Yaroslavl Composites Plant, Yaroslavl, Russia), were used as reinforcing elements. Characteristics of steel reinforcement are presented in Table 6, and polymer composites—in Table 7. Micro silica grade MK-85 (OOO ZIPo, Lipetsk, Russia) was used as a nanomodifying additive in concrete. Table 8 shows the chemical composition of micro silica MK-85. The granulometric composition of the applied nanomodifier is shown in Figure 2. The micro silica particle size distribution plot shows that most of the particles (more than 80%) have a size of 2 to 40 µm, with the main peak at 20 µm. X-ray phase analysis of micro silica particles is shown in Figure 3. Micro silica (Figure 2) is represented by amorphous silica, minor impurities of iron, carbonaceous substances, and crystalline α-quartz. Lightweight fiber-reinforced concrete was also studied in accordance with [46,47,48,49,50]. Glass fiber (Armplast, Nizhny Novgorod, Russia) was used as dispersed reinforcing fibers, the physical and mechanical characteristics of which are presented in Table 9. Slag pumice (Stroymir LLC, Lipetsk, Russia) was used as a filler, the physical and mechanical characteristics of which are presented in Table 10. The mineralogical composition of slag pumice is characterized by light minerals (80%), which include calcium and magnesium carbonates (68%), quartz (12%), heavy minerals (18%), represented by onormanite in the form of short square prisms, and iron sulfate (2%). 2.2. Methods The calculation was performed using SP 52-101-2003 [51], SP 63.13330.2018 [52] and LIRA-SAPR 2016 R5 software (Lira Service LLC, Moscow, Russia). This version of the software corresponds to the Standard plus configuration with the superelement mode, it has the ability to perform a full dynamic analysis, and the ability to check the strength of sections. It allows you to select sections of reinforced concrete elements using all standards. The column was designed for central compression according to the formulas [51] p.6.2.17 for columns with a length of 3 m, 6 m, 9 m, concrete B30, B40, steel, and composite reinforcement. The accepted reinforcement is 4 longitudinal reinforcement bars with a diameter of 6 mm with a cross-sectional area of 1.13 × 10−4 m2. The plan of the numerical experiment is shown in Figure 4. According to [51], the limiting value of the longitudinal force (bearing capacity) was calculated:(1) Nult=φ(RbA+RscAs.tot) where Rb is the design resistance of concrete to axial compression for the limit states of the first group; A is the area of all concrete in cross section; Rsc is the resistance of reinforcement to compression; As.tot is the area of all longitudinal reinforcement in the section of the element; φ is the coefficient taken for a long-term load according to Table 6.2 from [51] depending on the flexibility of the element. The plan of the physical experiment is shown in Figure 5. For particle size analysis of micro silica, a Microsizer 201C laser particle analyzer (OOO VA Insatalt, St. Petersburg, Russia) was used in [43]. An X-ray phase analysis (XPA) of micro silica was carried out on an X-ray diffractometer HZG-4C (Freiberger Prazisionmechanik, Berlin, Germany). The concrete mixture was made in a laboratory concrete mixer BL-10 (ZZBO LLC, Zlatoust, Russia). First, the dry components were mixed for 60 s, then the mixture was mixed with water and mixed until a homogeneous consistency was obtained. The fiber-reinforced concrete mixture was prepared according to the algorithm described in [29]. In the manufacturing of nanomodified heavy concrete, in order to increase the homogeneity of the binder (a mixture of cement and micro silica), the mixture of powders was processed in an Activator-4M homogenizer (OOO Chemical Engineering Plant, Novosibirsk, Russia). Furthermore, the modified binder, together with other dry concrete components—sand and crushed stone—was mixed in a concrete mixer for 60 s. After that, the resulting mixture of dry components was mixed with water, and mixed until a homogeneous consistency was obtained. Then, cube samples were formed with a rib size of 100 mm, and the mixture was compacted in molds on a laboratory vibration platform SMZh-539-220A (OOO IMASH, Armavir, Russia). After that, concrete cube samples were placed in a normal hardening chamber and stored for 28 days at a relative air humidity of at least 95% and a temperature of (20 ± 2) °C. Twenty-four hours after being manufactured, the samples were removed from the molds. To obtain a lightweight fiber-reinforced concrete, some large and small dense aggregates were replaced by a porous aggregate—slag pumice—the fractions and dosages of which were taken in accordance with Tables 8 and 9 (composition S4) from [46]. The parameters of the composition of the concrete mix were applied similarly to Table 7 from [46]. The composition of heavy concrete nanomodified with micro silica was chosen according to the recommendations [43]. Compressive strength tests of specimens were carried out in accordance with GOST 10180 “Concretes. Methods for strength were determined using reference specimens” [53] on an IP-1000 hydraulic press (OOO NPK TEHMASH, Neftekamsk, Russia). All samples of the same series were tested at the age of 28 days for no more than 1 h. The samples were loaded continuously at a constant rate of load increase until failure. In this case, the loading time of the sample until its destruction was at least 30 s. The maximum force achieved during the test was taken as the breaking load. Sample cubes were installed on one of the selected faces on the lower support plate of the press which was centrally relative to its longitudinal axis, using the marks applied to the press plate. After placing the sample on the press support plate, the upper press plate was aligned with the upper support face of the sample so that their planes completely adjoined to one another. The sample was loaded to failure at a constant rate of load increase (0.6 ± 0.2) MPa/s. The compressive strength of concrete was calculated with an accuracy of 0.1 MPa using the formula:(2) R=αFA where F is the breaking load, N; A is the area of the working section of the sample, mm2; and α is a scale factor for converting the strength of concrete to the strength of concrete in samples of a basic size and shape (for cubes with an edge size of 100 mm, it is 0.95). The strength of concrete in a series of samples was determined as the arithmetic mean of the strength of the tested samples in a series of six samples—four samples with the highest strength. The average density of samples in a state of normal humidity was determined according to GOST 12730.1 “Concretes. Methods of determination of density” [54]. The volume of the samples was calculated from their geometric dimensions. The sample sizes were determined with a caliper with an error of no more than 1%. The mass of samples is determined by weighing an error of no more than 0.1%. The average concrete density of each sample in the series was calculated with an error of up to 1 kg/m3 using the formula:(3) ρ=mV where m is the mass of the sample, g; V is the sample volume, cm3. The average density of concrete was calculated as the arithmetic mean of the test results for all samples of the series. In total, 2 series of sample cubes of each type of concrete were made and tested (in one series—6 sample cubes). That is, a total of 48 sample cubes of four types of concrete were tested (12 of each type). The microstructure of fiber-reinforced samples was studied using a ZEISS CrossBeam 340 microscope equipped with an Oxford Instruments X-Max 80 X-ray microanalyzer (Carl Zeiss Microscopy GmbH Factory, Jena, Germany) [29,55]. The microstructure of micro silica-modified samples was studied using a VEGA II LMU microscope (Tescan, Brno, Czech Republic) [43,44]. 3. Results The results of a numerical experiment—calculation of the bearing capacity (limiting value of the longitudinal force N) of reinforced concrete columns with a cross section of 400 mm × 400 mm with a reinforcing bar diameter of 6 mm—are presented in Table 11 and Figure 6. The experiment varied the class of concrete, the length of the product, the type, and reinforcement class. From the data of the numerical experiment, the results of which are presented in Table 11 and Figure 6, it can be seen that: (1) the use of polymer composite reinforcement instead of steel, leads to an insignificant decrease (by 0.5–1.0%) in the bearing capacity of columns operating in compression; (2) the bearing capacity of elements in the form of columns working in compression does not depend on the type of polymer composite reinforcement used in the study; (3) the difference in bearing capacity between 3000 mm and 6000 mm columns is 9–10%, and between 6000 mm and 9000 mm columns it is 2% to 3%, regardless of other parameters considered; (4) increasing the concrete grade from B30 to B40 results in an increase in the bearing capacity of compressive columns by up to 30%. After analyzing the data obtained as a result of the numerical experiment in Figure 6, we will explain the resulting difference in the data obtained with different initial parameters. Moreover, polymer composite reinforcement leads to an insignificant decrease in the bearing capacity due to the high strength characteristics of most modern polymer composite reinforcements, that is, the high strength characteristics of polymer composite reinforcement, with all its unambiguous advantages, primarily that it is lightweight, also allows it to maintain the bearing capacity of the entire structure at a high level. As for the lack of dependence of indicators on the type of polymer composite reinforcement, this can be explained by the fact that the polymer composite rods used in the study are modern, have passed various stages of testing, and have high strength characteristics, regardless of their type. Some decreases in the difference of the bearing capacity between the columns depending on the length can be explained, among other things, by a change in the geometric parameters and the working design scheme when the structure is loaded. Undoubtedly, the most important parameter for compressible reinforced elements using heavy concrete is the concrete class, which is reflected as a result of a numerical experiments, namely, with an increase in the class, which we achieve using our developed methods. Moreover, the bearing capacity of the columns also increases, and significantly—up to 30 percent. All this allows us to talk about the effectiveness of the proposed integrated approach. The results of the numerical experiment are presented in Table 12 and Figure 7. Figure 8 shows photographs of the microstructure of cement stone reinforced with glass fibers. From the data of the physical experiment, the results of which are presented in Table 12 and Figure 7, it can be seen that: (1) the difference between lightweight and heavy concrete is up to 10% (concrete grade reduction from B30 to B25); (2) glass fiber reinforcement brings the compressive strength of lightweight concrete up to that of heavy concrete (upgrading concrete from B25 to B30); (3) nano-modification of heavy concrete with micro silica resulted in an increase in the compressive strength of concrete up to 35%, whereas the concrete grade increased from B30 to B40. Photographs of the microstructure of the fiber-reinforced sample are shown in Figure 8. The results of microscopic studies with high-resolution photographs and high magnification made it possible to determine the physical nature and mechanism of the destruction of fiber-reinforced composites, which are the basis for building elements. The microphotographs (Figure 8a–d) clearly show that the nature of crack development directly depends on the rational distribution of fibers in the body of the cement matrix. In areas that make it possible to evaluate the usefulness of a rational distribution of fibers (Figure 8a), there are no microcracks in the region of the fibers (Figure 8c,d); however, with improper homogenization and distribution of the reinforcing fiber, in the area of the resulting fiber bundles, defects appear at the same time, expressed in microcracks (Figure 8b), which, in principle, correlates with the analogue of this phenomenon at the macro level—excessive density of reinforcement of complex structures. Thus, microscopic studies confirm the thesis not only with regard to the initial characteristics and quality of the fibers used, but also the importance of their distribution and homogenization in the body of the matrix, since microcracking already occurs at this stage and can develop at the macro level. Figure 9 shows photographs of the microstructure of cement stone modified with micro silica. Figure 9a–d show that micro silica nanomodification [43,44] forms a denser and more uniform structure that looks ordered and connected. This is due to the formation of a structure mainly from low-basic calcium hydrosilicates with nuclei of crystalline phases and local accumulations of portlandite, as well as a decrease in recrystallization activity after a period of accelerated hydration and an increase in the degree of cement hydration. Photographs of the microstructure of samples nanomodified with micro silica confirm the increase in the compressive strength of concrete that was obtained by us in experimental studies, and accordingly, its bearing capacity. The description and analysis of photographs of the microstructure of samples nanomodified with micro silica were carried out, among other things, in accordance with [43,44]. 4. Discussion The compared criteria for determining the role and place of research in modern science included such parameters as the size of the section of products, the class of concrete, the length of the product, the types of reinforcement, the class of reinforcement, and the diameter of the rod, which ultimately influenced the limiting value of the longitudinal force for compressible elements. Thus, according to the indicated criteria, we evaluated the effectiveness of each of the considered methods when comparing different types of materials:- in relation to nanomodified concrete; - in relation to lightweight fiber-reinforced concrete; - in relation to lightweight concrete; - in relation to control heavy concrete. In terms of the result achieved, we note that the most effective compositions obtained showed strength gains from 10% to 34%. Ultimately, this led to an increase in bearing capacity of up to 30%. It is possible to explain the obtained quantitative increases in concrete indicators by the observed qualitative picture of these changes. Thus, we have obtained significant improvements, which, in an integrated approach, led to an improvement in the structure of concrete, a decrease in the weight of concrete and, thereby, loads on the base. At the same time, the bearing capacity due to a rationally selected recipe and technological factors continues to remain at a high level. Thus, we have achieved a significant improvement in the qualitative picture, which leads to a synergistic effect from the developed proposals. This ultimately leads to an improvement in the characteristics of materials, then to an increase in the bearing capacity, and ultimately, to greater economic efficiency of the created building structures. This is in agreement with the results of the authors who dealt with the previously mentioned issue [2,3,10,16,23,28]; however, unlike [3], where an increase in the bearing capacity of columns up to 35% was provided by high-strength steel, in our study we achieved a comparable effect, but by using much lighter and more ergonomic reinforcing rods in combination with one of the proposed recipe solutions. In addition, unlike [18,36], in our study, an exceptionally positive effect was observed without a decrease in the compressive strength of concrete and reinforced concrete. At the same time, in the case of lightweight fiber-reinforced concrete, we achieve greater versatility of the resulting concrete and elements by increasing the deformability, thus making the fracture pattern more viscous due to additional fiber reinforcement and causing the appearance of some damping effect, due to the combination of fiber and porous filler. In the case of concrete being nanomodified with micro silica and a reinforced element, we obtain a cheaper element with improved characteristics due to the use of industrial waste [37]. This leads to a reduction in waste and allows for more environmentally friendly and more economical construction, which, in general, will lead to a reduction in the cost of construction by up to 10–12%, according to preliminary estimates of industrial partners. Numerical experiments have established that one of the most important criteria in the calculation of elements is the class of concrete. At the same time, this characteristic has an effect for all types of reinforcement and all types of reinforcing bar material, for various section sizes, and bar lengths. Due to this, the compressible elements are significantly dependent on the class of concrete used. Therefore, we applied such a numerical physical approach when conducting experiments to produce a direct correlation between these quantities and to prove the effectiveness and possibility of obtaining reinforced building elements based on concrete using various types of reinforcing bars with different densities and weights. It has been confirmed that by increasing the class of concrete, it is possible to increase the bearing capacity of the element, or by reducing the mass of the element, it is possible to maintain the bearing capacity at the same level. All of this became possible thanks to the use of various recipe-technological methods at the stage of manufacturing a concreted reinforced structure, which works as a compressed element. Thus, due to inexpensive recipe methods, namely, nanomodification with the help of production waste, or the use of lightweight dispersed-reinforced concrete to obtain such structures, we achieve significant technological, design, and installation effects. The design effect reduces the importance of the rod reinforcement and supplements the improved element with dispersed fiber reinforcement. The technological effect is achieved using concrete modification, a more rational selection of its formulation, and by improving the characteristics of the concrete, an increased bearing capacity of elements arises. The construction and installation effect lies in the fact that such lightweight universal structures are more applicable and have a wider area of operation at construction sites of various levels of responsibility and different directions. The proposed integrated system approach has a number of advantages over existing approaches. In particular, our developed approach makes it possible to take into account both the characteristics of the materials used and the quality of the technological process of concrete production, that is, technological and design advantages, expressed in design and design efficiency, using software, and taking into account the characteristics of materials, especially in relation to lightweight structures of new types. Such structures are designed for the difficult conditions of modern construction and certainly require an integrated systematic approach in their creation, calculation, and design. 5. Conclusions A new integrated system engineering approach to the design of building structures is proposed, which involves minimizing their cost and weight through numerical simulations and experimental verificationss of the operation of reinforcing bars made of various materials in concretes of various densities. It has been confirmed that by increasing the class of concrete, it is possible to increase the bearing capacity of the element, or by reducing the mass of the element, it is possible to maintain the bearing capacity at the same level. The control of the bearing capacity of reinforced building structures, using the example of compressed elements, is proposed to be carried out by the developed recipe-technological methods at the manufacturing stage. The economic and environmental efficiency of the proposed engineering methods, namely, nanomodification with the help of production waste or the use of lightweight dispersed–reinforced concrete to obtain such structures, has been revealed. This allows significant effects to be achieved, such as:- design effect, which consists of reducing the importance of rod reinforcement and supplementing the improved element with dispersed fiber reinforcement; - technological effect, which is achieved through the use of concrete modification, a more rational selection of its formulation, and by improving the characteristics of a concrete by obtaining an increased bearing capacity of elements from it; - construction and installation effect, which consists of the fact that such lightweight universal structures are more applicable and have a wider area of operation at construction sites of various levels of responsibility and various directions. The most effective concrete formulations have demonstrated strength gains ranging from 10% to 34%. Ultimately, this led to an increase in the bearing capacity of the elements up to 30%. The use of such an integrated lean approach will save up to 20% of resources during construction. Future prospects, and ways to develop the study further, are planned in terms of bending elements with various types of concrete and reinforcement. Moreover, taking into account the improvement of the technology for calculating and designing lightweight building structures, it is of interest to develop the results obtained in the direction of concrete with an improved structure. Thus, in further research, we will apply the developed integrated system approach for concretes obtained using technologies other than standard vibration. Acknowledgments The authors would like to acknowledge the administration of Don State Technical University for their resources and financial support. Author Contributions Conceptualization, S.A.S., E.M.S. and A.N.B.; methodology, S.A.S., E.M.S., V.V. and N.D.; software, S.A.S., E.M.S., A.N.B. and N.B.; validation, I.E., S.A.S., E.M.S. and A.N.B.; formal analysis, N.D., S.A.S. and E.M.S.; investigation, I.E., L.R.M., S.A.S., E.M.S., A.N.B. and V.V.; resources, B.M.; data curation, S.A.S., E.M.S. and N.D.; writing—original draft preparation, S.A.S., E.M.S. and A.N.B.; writing—review and editing, S.A.S., E.M.S. and A.N.B.; visualization, S.A.S., E.M.S., V.V., A.N.B. and N.B.; supervision, L.R.M. and B.M.; project administration, L.R.M. and B.M.; funding acquisition, A.N.B. and B.M. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement The study did not report any data. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Samples of nano-modified micro silica heavy concrete for compression testing. Figure 2 Granulometric composition of micro silica. Figure 3 Micro silica diffraction pattern. Figure 4 Plan of the numerical experiment. Figure 5 Plan of the physical experiment. Figure 6 Dependence of the bearing capacity of reinforced columns of various lengths on the type and class of reinforcement when using concrete of class (a) B30; and (b) B40. Figure 7 Compressive strength of the studied types of concrete. Figure 8 Photographs of the microstructure of a fiber-reinforced sample with magnification: (a) 100×; (b) 300×; (c) 500×; and (d) 1000×. Figure 9 Photographs of the microstructure of micro silica-modified samples with magnification: (a) 1000×; (b) 2000×; (c) 3000×; (d) 7000×. polymers-14-01756-t001_Table 1 Table 1 Physical and mechanical characteristics of Portland cement. Density, kg/m3 Normal Cosistency, % Blaine Specific Surface Area, cm2/g Setting Time, min Compressive Strength at 28 Days, MPa Bending Strength at 28 Days, MPa Start End 3120 24.8 3586 140 260 55.3 6.95 polymers-14-01756-t002_Table 2 Table 2 Chemical composition of Portland cement. Cement Title Oxid Content, % L.O.I. Cl SiO2 Al2O3 CaO Fe2O3 MgO TiO2 P2O5 SO3 Na2O K2O Na2Oequiv. Additive-free Portland cement CEM 0 52.5N GOST 31108-2020 21.1 4.9 62.7 4.4 1.7 0.1 0.1 2.8 0.3 0.6 0.57 0.7 0.03 polymers-14-01756-t003_Table 3 Table 3 Mineralogical composition of Portland cement. Cement Title Mineral Content, % C3S C2S C3A C4AF CaOfr. Additive-free Portland cement CEM 0 52.5N GOST 31108-2020 75.5 8.1 4.5 11.4 0.5 polymers-14-01756-t004_Table 4 Table 4 Physical and mechanical characteristics of crushed granite. Fraction Bulk Density, kg/m3 True Density, kg/m3 Crushing, wt % Lamellar Grain Content and Needle-Shaped Forms, wt. % Void Index, % 5–20 1420 2640 10.8 7.5 43 polymers-14-01756-t005_Table 5 Table 5 Physical characteristics of dense fine aggregate. Fineness Modulus Content of Dust and Clay Particles, % True Density, kg/m3 Bulk Density, kg/m3 Clay Content in Lumps, % 1.72 1.4 2665 1422 0.1 polymers-14-01756-t006_Table 6 Table 6 Characteristics of the steel reinforcement. Characteristics Steel A400 Steel A600 Steel A800 Steel A1000 Yield strength, MPa 380 570 760 970 Tensile strength, MPa 580 860 1010 1220 Modulus of elasticity, GPa 200 Elongation, % 16 9 7 6 Density, t/m3 7.0 7.2 7.4 7.5 Note: Steel A400—steel reinforcement class A400; Steel A600—steel reinforcement class A600; Steel A800—steel reinforcement class A800; Steel A1000—steel reinforcement class A1000. polymers-14-01756-t007_Table 7 Table 7 Characteristics of the used polymer composite reinforcement. Indicator Title GCR BCR CaCR ACR CoCR Tensile strength, MPa 800 800 1400 1400 1000 Tensile modulus, GPa, not less than 50 50 130 70 100 Ultimate compressive strength, MPa, not less than 300 300 300 300 300 Ultimate strength at cross section, MPa, not less than 150 150 350 190 190 Note: GCR—glass composite reinforcement—a polymer composite containing a continuous reinforcing fiberglass filler; BCR—basalt composite reinforcement—a polymer composite containing a continuous reinforcing filler made of basalt fiber; CaCR—carbon composite reinforcement—a polymer composite containing a continuous carbon fiber reinforcing filler; ACR—aramid composite reinforcement—a polymer composite containing a continuous reinforcing filler of aramid fiber; CoCR—combined composite reinforcement—glass composite or basalt composite, or carbon composite, or aramid composite, additionally filled with a continuous reinforcing filler from another type or types of fiber [45]. polymers-14-01756-t008_Table 8 Table 8 Chemical composition of micro silica. Title Oxid Content, % SiO2 Al2O3 Fe2O3 CaO MgO R2O SO3 L.O.I. MS-85 81.9 1.5 2.8 1.2 0.3 0.9 3.8 7.2 polymers-14-01756-t009_Table 9 Table 9 Physical and mechanical characteristics of glass fiber. Density, g/cm3 Tensile Strength, GPa Elastic Modulus, GPa Fiber Length, mm Elongation, % 2.6 1.8 70 12 1.5 polymers-14-01756-t010_Table 10 Table 10 Physical and mechanical characteristics of slag pumice. Fraction, mm Bulk Density, kg/m3 True Density, kg/m3 Strength, MPa Void, % 5–10 608 1320 0.8 52 1.25–2.5 727 1408 - 54 polymers-14-01756-t011_Table 11 Table 11 Results of calculating the limiting value of the longitudinal force (bearing capacity). Product Section SIZE, mm × mm Concrete Class Product Length, mm Reinforcement Type Reinforcement Class Rod Diameter, mm Ultimate Value of Longitudinal Force N, kN 400 × 400 B30 3000 steel A400 6 2758.4 A600 2773.1 A800 2776.5 A1000 2776.5 GCR 800 × 50 2753.9 BCR 800 × 50 2753.9 CaCR 1400/130 2753.9 ACR 1400/70 2753.9 CoCR 1000/100 2753.9 B40 steel A400 3558.4 A600 3573.1 A800 3576.5 A1000 3576.5 GCR 800 × 50 3553.9 BCR 800 × 50 3553.9 CaCR 1400/130 3553.9 ACR 1400/70 3553.9 CoCR 1000/100 3553.9 B30 6000 steel A400 2502.2 A600 2515.6 A800 2518.7 A1000 2518.7 GCR 800 × 50 2498.2 BCR 800 × 50 2498.2 CaCR 1400/130 2498.2 ACR 1400/70 2498.2 CoCR 1000/100 2498.2 B40 steel A400 3228.0 A600 3241.3 A800 3244.4 A1000 3244.4 GCR 800 × 50 3223.9 BCR 800 × 50 3223.9 CaCR 1400/130 3223.9 ACR 1400/70 3223.9 CoCR 1000/100 3223.9 B30 9000 steel A400 2443.2 A600 2456.2 A800 2459.2 A1000 2459.2 GCR 800 × 50 2439.2 BCR 800 × 50 2439.2 CaCR 1400/130 2439.2 ACR 1400/70 2439.2 CoCR 1000/100 2439.2 B40 steel A400 3151.8 A600 3164.8 A800 3167.8 A1000 3167.8 GCR 800 × 50 3147.8 BCR 800 × 50 3147.8 CaCR 1400/130 3147.8 ACR 1400/70 3147.8 CoCR 1000/100 3147.8 polymers-14-01756-t012_Table 12 Table 12 Physical and mechanical parameters of the studied types of concrete. Type of Concrete Average Density, kg/m3 Compressive Strength, MPa Lightweight concrete 1880 34.3 ± 1.7 Heavy concrete (control) 2340 38.2 ± 1.9 Lightweight fiber concrete 1890 37.9 ± 1.9 Nano-modified heavy concrete 2430 51.2 ± 2.5 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Hosen M.A. Shammas M.I. Shill S.K. Al-Deen S. Jumaat M.Z. Hashim H. Ductility Enhancement of Sustainable Fibrous-Reinforced High-Strength Lightweight Concrete Polymers 2022 14 727 10.3390/polym14040727 35215640 2. Ren Z. Li Q. Wang G. Wei W. Abbas M.A.A.M. Eccentric Compressive Behavior of Round-Ended Rectangular Concrete-Filled Steel Tubes with Different Central Angles Materials 2022 15 456 10.3390/ma15020456 35057174 3. Wang J. Duan Y. Wang Y. Wang X. Liu Q. Analysis and Modification of Methods for Calculating Axial Load Capacity of High-Strength Steel-Reinforced Concrete Composite Columns Materials 2021 14 6860 10.3390/ma14226860 34832262 4. Hassan R.F. Al-Salim N.H. Mohammed N.S. Hussein H.H. Experimental study and theoretical prediction on torsional strength with different steel fiber reinforced concretes and Cross-Section areas Eng. Struct. 2022 251 113559 10.1016/j.engstruct.2021.113559 5. Chu S.H. Kwan A.K.H. A new bond model for reinforcing bars in steel fibre reinforced concrete Cem. Concr. Compos. 2019 104 103405 10.1016/j.cemconcomp.2019.103405 6. Zhou J.-K. Lin W.-K. Guo S.-X. Zeng J.-J. Bai Y.-L. Behavior of FRP-confined FRP spiral reinforced concrete square columns (FCFRCs) under axial compression J. Build. Eng. 2022 45 103452 10.1016/j.jobe.2021.103452 7. Sun Z. Chen Y. Xu S.-Y. Sun Y. Wu G. Shaking table test of concrete columns hybrid reinforced by steel/FRP bars J. Build. Eng. 2022 48 103938 10.1016/j.jobe.2021.103938 8. Chepurnenko A. Yazyev B. Meskhi B. Beskopylny A. Khashkhozhev K. Chepurnenko V. Simplified 2D Finite Element Model for Calculation of the Bearing Capacity of Eccentrically Compressed Concrete-Filled Steel Tubular Columns Appl. Sci. 2021 11 11645 10.3390/app112411645 9. Mina A.L. Trezos K.G. Petrou M.F. Optimizing the Mechanical Properties of Ultra-High-Performance Fibre-Reinforced Concrete to Increase Its Resistance to Projectile Impact Materials 2021 14 5098 10.3390/ma14175098 34501188 10. Lehmann M. Głodkowska W. Shear Capacity and Behaviour of Bending Reinforced Concrete Beams Made of Steel Fibre-Reinforced Waste Sand Concrete Materials 2021 14 2996 10.3390/ma14112996 34205948 11. Thermou G.E. Kappos A.J. Background to the Monolithicity Factors for the Assessment of Jacketed Reinforced Concrete Columns Buildings 2022 12 55 10.3390/buildings12010055 12. Wang J. Wang X. Duan Y. Su Y. Yi X. The Investigation on Mechanical Performances of High-Strength Steel Reinforced Concrete Composite Short Columns under Axial Load Materials 2022 15 329 10.3390/ma15010329 35009475 13. Solhmirzaei R. Kodur V. A Numerical Model for Tracing Structural Response of Ultra-High Performance Concrete Beams Modelling 2021 2 448 466 10.3390/modelling2040024 14. Abdulhameed A.A. Al-Zuhairi A.H. Al Zaidee S.R. Hanoon A.N. Al Zand A.W. Hason M.M. Abdulhameed H.A. The Behavior of Hybrid Fiber-Reinforced Concrete Elements: A New Stress-Strain Model Using an Evolutionary Approach Appl. Sci. 2022 12 2245 10.3390/app12042245 15. Krassowska J. Kosior-Kazberuk M. The Effect of Steel and Basalt Fibers on the Shear Behavior of Double-Span Fiber Reinforced Concrete Beams Materials 2021 14 6090 10.3390/ma14206090 34683681 16. Shilov A.V. Beskopylny A.N. Meskhi B. Mailyan D. Shilov D. Polushkin O.O. Ultimate Compressive Strains and Reserves of Bearing Capacity of Short RC Columns with Basalt Fiber Appl. Sci. 2021 11 7634 10.3390/app11167634 17. Yang L. Wang H. Gao S. Study on Axial Compression Behavior of Concrete Short Columns Confined by Flax/Glass Fiber Hybrid-Reinforced Epoxy Resin Composites Polymers 2022 14 517 10.3390/polym14030517 35160506 18. Grzesiak S. Pahn M. Schultz-Cornelius M. Harenberg S. Hahn C. Influence of Fiber Addition on the Properties of High-Performance Concrete Materials 2021 14 3736 10.3390/ma14133736 34279308 19. Liu X. Sun Q. Song W. Bao Y. Numerical modeling and parametric study of hybrid fiber-rebar reinforced concrete tunnel linings Eng. Struct. 2022 251 113565 10.1016/j.engstruct.2021.113565 20. Liang M. Fu C. Xiao B. Luo L. Wang Z. A fractal study for the effective electrolyte diffusion through charged porous media Int. J. Heat Mass Transf. 2019 137 365 371 10.1016/j.ijheatmasstransfer.2019.03.141 21. Xiao B. Wang W. Zhang X. Long G. Fan J. Chen H. Deng L. A novel fractal solution for permeability and Kozeny-Carman constant of fibrous porous media made up of solid particles and porous fibers Powder Technol. 2019 349 92 98 10.1016/j.powtec.2019.03.028 22. Ahmad H. Elnemr A. Ali N. Hussain Q. Chaiyasarn K. Joyklad P. Finite Element Analysis of Glass Fiber-Reinforced Polymer-(GFRP) Reinforced Continuous Concrete Beams Polymers 2021 13 4468 10.3390/polym13244468 34961019 23. Blazy J. Drobiec Ł. Wolka P. Flexural Tensile Strength of Concrete with Synthetic Fibers Materials 2021 14 4428 10.3390/ma14164428 34442952 24. Dębska A. Gwoździewicz P. Seruga A. Balandraud X. Destrebecq J.-F. The Longitudinal Compression Capacity of Hollow Concrete Cylinders Prestressed by Means of an SMA Wire Materials 2022 15 826 10.3390/ma15030826 35160772 25. Scattarreggia N. Qiao T. Malomo D. Earthquake Response Modeling of Corroded Reinforced Concrete Hollow-Section Piers via Simplified Fiber-Based FE Analysis Sustainability 2021 13 9342 10.3390/su13169342 26. Molaioni F. Di Carlo F. Rinaldi Z. Modelling Strategies for the Numerical Simulation of the Behaviour of Corroded RC Columns under Cyclic Loads Appl. Sci. 2021 11 9761 10.3390/app11209761 27. Parvin A. Alhusban M. Lateral Deformation Capacity and Plastic Hinge Length of RC Columns Confined with Textile Reinforced Mortar Jackets CivilEng 2021 2 670 691 10.3390/civileng2030037 28. Roudsari S.S. Ungureanu L.M. Soroushnia S. Abu-Lebdeh T. Petrescu F.I.T. Optimization of Fiber-Reinforced Polymer Bars for Reinforced Concrete Column Using Nonlinear Finite Element Algorithms Algorithms 2022 15 12 10.3390/a15010012 29. Stel’makh S.A. Shcherban’ E.M. Beskopylny A. Mailyan L.R. Meskhi B. Dotsenko N. Enchainment of the Coefficient of Structural Quality of Elements in Compression and Bending by Combined Reinforcement of Concrete with Polymer Composite Bars and Dispersed Fiber Polymers 2021 13 4347 10.3390/polym13244347 34960898 30. Abdel-Jaber M. Abdel-Jaber M. Katkhuda H. Shatarat N. El-Nimri R. Influence of Compressive Strength of Concrete on Shear Strengthening of Reinforced Concrete Beams with Near Surface Mounted Carbon Fiber-Reinforced Polymer Buildings 2021 11 563 10.3390/buildings11110563 31. Isleem H.F. Tayeh B.A. Alaloul W.S. Musarat M.A. Raza A. Artificial Neural Network (ANN) and Finite Element (FEM) Models for GFRP-Reinforced Concrete Columns under Axial Compression Materials 2021 14 7172 10.3390/ma14237172 34885326 32. Wang J. Xia J. Chang H. Han Y. Yu L. Jiang L. Experimental Study on Axial Compressive Behavior of Gangue Aggregate Concrete Filled FRP and Thin-Walled Steel Double Tubular Columns Coatings 2021 11 1404 10.3390/coatings11111404 33. Al-Rousan R. Nusier O. Abdalla K. Alhassan M. Lagaros N.D. NLFEA of Sulfate-Damaged Circular CFT Steel Columns Confined with CFRP Composites and Subjected to Axial and Cyclic Lateral Loads Buildings 2022 12 296 10.3390/buildings12030296 34. Oneschkow N. Timmermann T. Löhnert S. Compressive Fatigue Behaviour of High-Strength Concrete and Mortar: Experimental Investigations and Computational Modelling Materials 2022 15 319 10.3390/ma15010319 35009465 35. Zhelyazov T. Thorhallsson E. Analytical and Numerical Approaches for the Design of Concrete Structural Elements with Internal BFRP Reinforcement Materials 2022 15 1497 10.3390/ma15041497 35208037 36. Šantek Bajto J. Štirmer N. Cerković S. Carević I. Kostanić Jurić K. Pilot Scale Production of Precast Concrete Elements with Wood Biomass Ash Materials 2021 14 6578 10.3390/ma14216578 34772104 37. Zamora-Castro S.A. Salgado-Estrada R. Sandoval-Herazo L.C. Melendez-Armenta R.A. Manzano-Huerta E. Yelmi-Carrillo E. Herrera-May A.L. Sustainable Development of Concrete through Aggregates and Innovative Materials: A Review Appl. Sci. 2021 11 629 10.3390/app11020629 38. Aghdasi P. Ostertag C.P. Tensile fracture characteristics of Green Ultra-High Performance Fiber-Reinforced Concrete (G-UHP-FRC) with longitudinal steel reinforcement Cem. Concr. Compos. 2020 114 103749 10.1016/j.cemconcomp.2020.103749 39. Krishan A. Rimshin V. Troshkina E. Compressed and Bending Concrete Elements with Confinement Reinforcement Meshes IOP Conf. Ser.: Mater. Sci. Eng. 2020 753 022052 10.1088/1757-899X/753/2/022052 40. Zeng J. Long T. Compressive Behavior of FRP Grid-Reinforced UHPC Tubular Columns Polymers 2022 14 125 10.3390/polym14010125 41. Jafari K. Tabatabaeian M. Joshaghani A. Ozbakkaloglu T. Optimizing the mixture design of polymer concrete: An experimental investigation Constr. Build. Mater. 2018 167 185 196 10.1016/j.conbuildmat.2018.01.191 42. Safaei B. Davodian E. Fattahi A.M. Asmael M. Calcium carbonate nanoparticles effects on cement plast properties Microsyst. Technol. 2021 27 3059 3076 10.1007/s00542-020-05136-6 43. Shcherban’ E.M. Stel’makh S.A. Beskopylny A. Mailyan L.R. Meskhi B. Varavka V. Nanomodification of Lightweight Fiber Reinforced Concrete with Micro Silica and Its Influence on the Constructive Quality Coefficient Materials 2021 14 7347 10.3390/ma14237347 34885519 44. Beskopylny A.N. Stel’makh S.A. Shcherban’ E.M. Mailyan L.R. Meskhi B. Nano modifying additive micro silica influence on integral and differential characteristics of vibrocentrifuged concrete J. Build. Eng. 2022 51 104235 10.1016/j.jobe.2022.104235 45. GOST 31938-2012 Fibre-Reinforced Polymer Bar for Concrete Reinforcement. General Specifications Available online: https://docs.cntd.ru/document/1200101115 (accessed on 18 March 2022) 46. Mailyan L.R. Beskopylny A.N. Meskhi B. Stel’makh S.A. Shcherban E.M. Ananova O. Optimization of Composition and Technological Factors for the Lightweight Fiber-Reinforced Concrete Production on a Combined Aggregate with an Increased Coefficient of Structural Quality Appl. Sci. 2021 11 7284 10.3390/app11167284 47. Shcherban’ E.M. Stel’makh S.A. Beskopylny A. Mailyan L.R. Meskhi B. Shuyskiy A. Improvement of Strength and Strain Characteristics of Lightweight Fiber Concrete by Electromagnetic Activation in a Vortex Layer Apparatus Appl. Sci. 2022 12 104 10.3390/app12010104 48. Beskopylny A.N. Stel’makh S.A. Shcherban’ E.M. Mailyan L.R. Meskhi B. El’shaeva D. Varavka V. Developing Environmentally Sustainable and Cost-Effective Geopolymer Concrete with Improved Characteristics Sustainability 2021 13 13607 10.3390/su132413607 49. Beskopylny A.N. Shcherban’ E.M. Stel’makh S.A. Mailyan L.R. Meskhi B. El’shaeva D. The Influence of Composition and Recipe Dosage on the Strength Characteristics of New Geopolymer Concrete with the Use of Stone Flour Appl. Sci. 2022 12 613 10.3390/app12020613 50. Shcherban’ E.M. Stel’makh S.A. Beskopylny A. Mailyan L.R. Meskhi B. Increasing the Corrosion Resistance and Durability of Geopolymer Concrete Structures of Agricultural Buildings Operating in Specific Conditions of Aggressive Environments of Livestock Buildings Appl. Sci. 2022 12 1655 10.3390/app12031655 51. SP 52-101-2003 Concrete and Reinforced Concrete Structures without Prestressing Available online: https://docs.cntd.ru/document/1200037361 (accessed on 18 March 2022) 52. SP 63.13330.2018 Concrete and Reinforced Concrete Structures. General Provisions Available online: https://docs.cntd.ru/document/554403082 (accessed on 18 March 2022) 53. GOST 10180-2012 Concretes. Methods for Strength Determination Using Reference Specimens Available online: http://docs.cntd.ru/document/1200100908 (accessed on 18 March 2022) 54. GOST 12730.1 Concretes. Concretes. Methods of Determination of Density Available online: https://docs.cntd.ru/document/1200177299 (accessed on 18 March 2022) 55. Stel’makh S.A. Shcherban’ E.M. Beskopylny A. Mailyan L.R. Meskhi B. Varavka V. Quantitative and Qualitative Aspects of Composite Action of Concrete and Dispersion-Reinforcing Fiber Polymers 2022 14 682 10.3390/polym14040682 35215594

PMC009xxxxxx/PMC9099641.txt

==== Front Molecules Molecules molecules Molecules 1420-3049 MDPI 10.3390/molecules27093037 molecules-27-03037 Article Characterization of β-Glucan-Peanut Protein Isolate/Soy Protein Isolate Conjugates and Their Application on Low-Fat Sausage Zhang Manli Liu Hongzhi * Wang Qiang * Judeh Zaher Academic Editor Key Laboratory of Agro-Products Processing, Ministry of Agriculture and Rural Affairs, Institute of Food Science and Technology, Chinese Academy of Agricultural Science, Beijing 100081, China; manlizhang165@163.com * Correspondence: liuhongzhi@caas.cn or lhz0416@126.com (H.L.); wangqiang06@caas.cn (Q.W.) 09 5 2022 5 2022 27 9 303706 4 2022 06 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Polysaccharide–protein conjugates can improve the functional properties and expand the application field. The emulsifying, thermal properties of WSG-PPI conjugates and WSG-SPI conjugates were improved, compared to WSG, PPI and SPI. The Maillard reaction was confirmed by Fourier transform infrared spectroscopy (FT-IR). Circular dichroism (CD) exhibited that the structure of the conjugates was more expanded. Cryo-SEM and AFM demonstrated that the WSG, WSG-PPI and WSG-SPI conjugates had a morphology of a chain. When the conjugates were added as fat substitutes to low-fat sausage, the cooking yield, hardness and chewiness increased. The objective of this research was to study the emulsifying property, thermal property and structural changes of β-glucan-peanut protein isolate (WSG-PPI) conjugates and β-glucan-soy protein isolate (WSG-SPI) conjugates prepared through wet-heated Maillard reaction, and their effect on the texture of low-fat sausage. polysaccharide-protein conjugates emulsifying properties structure low-fat sausage National Key Research & Development Program2021YFD2100402 Hebei Oil Crop Innovation Team of Modern Agro-Industry Technology Research System of ChinaHBCT2019090203 National Key Research & Development Program (2021YFD2100402, Ministry of Science and Technology, PRC); Hebei Oil Crop Innovation Team of Modern Agro-Industry Technology Research System of China (HBCT2019090203, Department of Agriculture and Rural Affairs of Hebei Province). ==== Body pmc1. Introduction Saccharomyces cerecisiae β-glucan is a kind of polysaccharide, which is linked by a 1→3 glycosidic bond as the main chain and a 1→6 glycosidic bond as the branch chain [1,2]. It is present in the cell wall, which can contribute to the integrity of the cell wall, cell exchanges, and keep cells from pathogens and environmental stresses [3]. It has superior bioactivities, such as anti-tumor [4,5,6,7,8], antioxidant [9,10], hypoglycemic effect [11], immune regulation [12,13] and antibacterial effects [7]. However, β-glucan is hardly soluble in water and common solvents owing to its compact triple helix conformation [14]. It is difficult for β-glucan to react with protein in an aqueous solution. In this paper, a safe, efficient, and green enzymatic hydrolysis method was selected to prepare water-soluble β-glucan (WSG). WSG was the product obtained by the enzymatic hydrolysis of β-glucan. It was more soluble in water with a reduced molecular weight and had a highly ordered structure. Previous studies proved that WSG has an emulsifying property, antioxidant property [15,16], and that it stimulates immunity [17,18], which can be used in food, medicine, cosmetics and other fields [19]. The Maillard reaction is one of the major non-enzymatic browning reactions in food processes, which improves the functional properties of protein or polysaccharides [20,21,22]. For example, soy hull hemicelluloses-soy protein isolate conjugates demonstrated substantially better emulsification capacity and thermal stability than soy protein [23]. Myofibrillar protein-dextran conjugates exhibited emulsifying property and solubility [24]. Whey protein isolate-gum acacia possessed an improved emulsifying property, high solubility and stability heat-induced insolubility, and reduced surface hydrophobicity [25]. Oat β-glucan-dipeptide exhibited improved an apparent viscosity, emulsifying property, swelling power and fat binding capacity [26]. The previous studies were mainly about the functional properties and structures of protein during a Maillard reaction. Polysaccharides, as a substrate of the Maillard reaction, widely exist in plants. However, there are few studies on the changes of physicochemical properties of polysaccharides after a Maillard reaction. Sausage is one of the most popular meat products in the world, and has a high content of fat, with an unhealthy fatty acid profile. However, with the improvement of quality of life, people pay more and more attention to the relationship between diet and health. Consumers’ purchasing habits are changing, and low-fat food becomes more and more popular. Fat plays an important role in maintaining the sensory characteristics such as texture and juiciness of the meat product. The direct reduction of fat may lead to the poor quality of the sausage, as many ingredients had been studied as fat replacers to decline fat levels without compromising sausage quality. Pork skin-based emulsion gels elaborated with canola oil, bamboo fiber and inulin were promising alternatives to replace pork back fat in low-fat sausages [27]. Emulsion gels prepared with carrageenan, and emulsion stabilized by zein/carboxymethyl dextrin also served as a fat substitute in sausages, and could enhance the hardness and viscoelasticity of pork sausage [28]. The addition of 15% chicken skin, wheat fiber mixture or pineapple dietary fibers to sausage could significantly reduce the cooking yield [29,30]. The effect of pea protein isolate, pea low moisture extrudate and pea high moisture extrudate replacing 20% pork meat were compared; there are no significant differences between sausage with PPI and normal sausage, and the sausage with pea extrudate had a softer bite and significant color changes [31]. The reaction between WSG and amino acid/protein was inevitable in the process of food processing. The purpose of this paper was to study the functional properties and structural changes of WSG and peanut protein isolate (PPI) and soy protein isolate (SPI) during Maillard reaction, as well as the effect of WSG-PPI conjugates and WSG-SPI conjugates on the texture of low-fat sausage. 2. Results and Discussion 2.1. Functional Properties 2.1.1. Emulsifying Properties Emulsifying properties were related to the spatial conformation and physico-chemical properties of the substance itself. The conjugates bound the polyhydroxy carbohydrate to protein through the Maillard reaction, and the protein exposed a certain hydrophobic group to quickly and tightly absorb on the oil-water interface [32]. Otherwise, the formation of conjugates was affected by many factors, and then affected the emulsification. The emulsifying activity index (EAI) and emulsifying stability index (ESI) of different conjugates were compared (Figure 1). The EAI and ESI of all conjugates were higher than the EAI and ESI of the WSG. There were two reasons to explain the results. First, the access of hydrophobic groups allowed the conjugates to be more effectively absorbed to the oil–water interface by the reaction between protein with WSG [33]. Second, WSG could form a steric repulsion between the surface of emulsion particles and promote the formation of a stable membrane around the oil particles [34]. Moreover, the second structure of the molecules was extended by heat treatment [35]. Chen et al. [25] also reported that the emulsifying properties of WPI-GA conjugates were improved compared to the WPI and WPI-GA mixture. Sugar beet pectin-whey protein isolate/bovine serum albumin conjugates crosslinked by Genipin exhibited excellent emulsifying properties [36]. However, there were no significant differences between the emulsifying properties of WSG-PPI, WSG-arachin and WSG-conarachin conjugates. The same results were observed between WSG-SPI and WSG-7S globulin conjugates. Then, the WSG-PPI and WSG-SPI conjugates were selected for further study. 2.1.2. Contact Angle and Particle Size Distribution The amphiphilicity and particle size distribution were factors affecting emulsion stability. The wettability of the particles directly affected its adsorption on the oil–water interface [37]. The smaller the angle, the better the wettability. If the angle was <90°, the solid surface was hydrophilic; conversely, the solid surface was hydrophobic. The contact angle was used to measure the amphiphilicity and wettability of PPI, SPI, WSG, WSG-PPI and WSG-SPI conjugates (Figure 2). The contact angle of WSG, PPI, SPI, WSG-PPI and WSG-SPI conjugates was 50.1°, 101.9°, 117.3°, 94.1° and 85.9°. The contact angle of conjugates was higher than WSG, which was due to the increase in hydrophobic groups in the polysaccharide’s chain. However, the contact angle of conjugates was lower than SPI and PPI; it could be inferred that the hydrophilic groups in the protein molecules increased due to the combination of the polysaccharide chain and protein chain. The mean particle size reflected the aggregation behavior of WSG-PPI and WSG-SPI conjugates. The mean particle of WSG, WSG-PPI and WSG-SPI was 38.25 ± 0.17 μm, 50.07 ± 0.76 μm and 104.89 ± 3.79 μm, respectively. The mean particle size of WSG-PPI and WSG-SPI conjugates increased compared to the WSG and WSG-PPI and WSG-SPI mixture (Figure 3a). The particle size of the polysaccharide was related to its molecular weight, intrinsic viscosity, and physical stability [38]. WSG and PPI or SPI formed larger molecules through the Maillard reaction, which might explain the increase in particle size. Patel found that the particle size of the bio-polymer was larger [39]. 2.1.3. Thermal Properties The thermal property of conjugates could be studied by measuring the temperature and rate of pyrolysis. There were two thermal degradation process during the 30–500 °C heating treatment (Figure 3b). The initial weight loss was due to the evaporation of free and bound water during 30–150 °C. The second weight loss was attributed to the breaking of the intermolecular and intramolecular hydrogen bonds and electrostatic bonds, and the decomposition of the hydrophobic interaction during the temperature of 150–500 °C. WSG, PPI, SPI and WSG-PPI and WSG-SPI conjugates began to decompose drastically and substantially at 283.2 °C, 310.1 °C, 312.9 °C and 277.1 °C and 306.4 °C, respectively. However, the weight loss of WSG, PPI, SPI and WSG-PPI and WSG-SPI conjugates was 68.5%, 67.5%, 61.3% and 32.4% and 64.9%, respectively. It was inferred that thermal stability of WSG-PPI conjugates was improved compared to WSG and PPI, and the thermal stability of the WSG-SPI conjugates was lower than SPI, which was higher than WSG. The thermal stability of conjugates was related to the type of protein and the ratio of protein to polysaccharide. The thermal stability of caseinate-maltose conjugates was lower than caseinate [40]. The thermal stability of carboxymethyl cellulose(CMC)/SPI blend films improved due to the addition of the CMC content [41]. The improvement was also attributed to the Maillard reaction and the crosslinking between SPI and carboxymethyl cellulose. The thermal stability of soy hemicelluloses-SPI conjugates (1:9) was lower than conjugates with more SPI content [23]. Schizophyllan modified by α-amino also demonstrated better thermal stability compared to the unmodified schizophllan +poly(c) complex [42]. 2.1.4. Rheological Properties (1) Viscosity The viscosity of polysaccharide was affected by its molecular weight. The increase in molecular weight can make the molecular chain extend and the side chain increase, making it easier to become entangle between the molecules. The viscosity profiles of WSG and its conjugates were presented in Figure 3c. WSG and its conjugates demonstrated shear thinning behavior at 25 °C, which was also observed in oat β-glucan and its conjugates [43]. This behavior was due to the distribution of random coil polymers or their parallel alignment with the flow stream [44]. In addition, the viscosity of WSG-PPI and WSG-SPI conjugates apparently increased. The change of viscosity might be attributed to the change of structure, molecular weight and concentration of samples [26]. The molecular weight of WSG-PPI and WSG-SPI conjugates was higher than WSG′s molecular weight by the Maillard reaction; then, the viscosity of the WSG-PPI and WSG-SPI conjugates increased. Su et al. [41] also found that the reaction between soy protein isolate and carboxymethyl cellulose could cause an increase in viscosity. (2) Frequency sweep test. The storage modulus (G′) and the loss modulus (G″) represented the elastic property and the viscous property in the frequency sweep test, respectively. The G′ and G″ of WSG and its conjugates were shown in Figure 3c. The intersection between G′ and G″ of WSG indicated the transition from gelation to solation. The G′ of WSG-PPI conjugates and WSG-SPI conjugates was higher than the G″, which indicated a predominant solid-like behavior of the conjugates. The results indicated that the viscoelastic behaviors of WSG changed during the Maillard reaction. G′ and G″ of WSG and its conjugates increased after the Maillard reaction during the whole frequency range. The increase in G′ may be due to the strong molecular interaction stabilizing the closed-pack molecules system [45]. Sun et al. [43] found that the G′ of β-glucan conjugates was related to the reaction substrate, and the G′of β-glucan-amino acid conjugates was higher than β-glucan, but the β-glucan peptide conjugates was similar to β-glucan. 2.2. Structure 2.2.1. Fourier Transform Infrared Spectroscopy (FT-IR) FT-IR was a method to study the conformational changes of WSG and its conjugates in the dry powder state because the chemical fingerprints of polysaccharides and protein did not overlap [46]. The absorption band at the 3400 cm−1 corresponded to the stretching vibration of –OH in the constituent sugar residues [47]. The absorption band at 2920 cm−1 was due to the stretching vibration of –CH [48]. The absorption band at 1630 cm−1 was attributed to the stretching vibration of –CHO and –C=O [49]. The absorption band at 1030 cm−1 was associated with the stretching vibration of -C-O-C from the different units of (1→3) and (1→6) in WSG. However, the peak of WSG-PPI conjugates and WSG-SPI conjugates at 1630 cm−1 was lower than the peak of WSG, which might be explained by the fact that the Maillard reaction consumed the C=O and –CHO of WSG. The phenomenon also appeared in oat β-glucan-L-glutamic and β-glucan-collagen peptide I conjugates [26]. When the feruloyl groups in pectin molecule were hydrolyzed, a spectral change in the range of 1616–1634 cm−1 was observed, which was attributed to the esterified [50]. 2.2.2. Circular Dichroism (CD) CD was an effective method to study the structural changes of protein and polysaccharide. The conformational transition of samples was detected on the CD spectra (Figure 4b). All samples had a positive and negative cotton effect, which suggested that the samples were in a highly ordered structure [51]. However, the ellipticity and maximum/minimum peak position of the samples were different. The CD spectra of WSG had a positive cotton effect with a maximum absorption at 190.0 nm and a negative cotton effect with a minimum absorption at 210.2 nm among 190–250 nm. The maximum absorption of WSG-PPI mixture decreased and shifted to 192.0 nm. The maximum absorption of the WSG-SPI mixture and WSG-SPI conjugates did not change significantly. The minimum absorption peak of the WSG-PPI mixture and WSG-SPI mixture shifted at 208.2 nm and 206.6 nm, respectively. The maximum absorption peak of WSG-PPI conjugates shifted to 191.4 nm, which was lower than the WSG and its mixture. The minimum absorption peak of WSG-PPI conjugates and WSG-SPI conjugates appeared at 205.4 nm and 200.4 nm, remarkably indicating a conformational transformation among the WSG, WSG-PPI/SPI mixture and WSG-PPI/SPI conjugates. The difference between the WSGs and their conjugates’ spectra demonstrated a symmetrical peak at 205 nm, consisting with the nπ* absorbance maximum of esters [52]. β-lg-carboxymethyldextran conjugates were also found in the decrease in the peak in the presence of the polysaccharide, as well as a blue shift [53]. 2.2.3. Scanning Electron Microscope (SEM) The microstructure of WSG and its conjugates in solid state can be observed. WSG presented an irregular slice shape, the WSG-PPI conjugates and WSG-SPI conjugates exhibited aggregated clump shapes with small particles on the surface, and the surface was more uneven (Figure 5), which was due to WSG covalently bonded to PPI or SPI and improving the content of the hydrophilic radical on the surface of PPI and SPI [54]. Otherwise, the heating treatment expended the contact area between PPI/SPI and WSG, leading to the coarse surface of the WSG-PPI and WSG-SPI conjugates. The finding was consistent with Yu [54], who found that the lactose-high-temperature peanut protein isolate conjugates’ surface appeared with an uneven bulge and formed a sugar-surrounding protein structure. The high-temperature peanut protein isolate-sesbania gum conjugate also exhibited a rougher surface [55]. In general, the graft polysaccharide surface tended to be rough and non-uniform, which was favorable for the formation of a more hydrophilic protein surface structure and for dispersion in a water solution [56]. 2.2.4. Cryo-SEM and AFM Cryo-SEM could observe the frozen liquid samples on the pore scale. The sample was frozen at a low temperature and stored in the microscope, which made the sample retain moisture in a high vacuum and allowed the sample to preserve the original structure [57]. The microstructure of WSG, PPI, SPI and its conjugates in solution was observed by cryo-SEM (Figure 6). WSG demonstrated irregular chains, and WSG-PPI conjugates and WSG-SPI conjugates demonstrated an irregular chain with a spherical structure attached. PPI or SPI observed the spherical structure, so it could be inferred that the spherical structure of WSG-PPI and WSG-SPI conjugates chain was PPI or SPI. AFM was a technique that could characterize biopolymers on sub-nanometer scale, which had the advantages of visualizing the specimen by contouring the forces between the probe and the specimen surface [58]. As AFM figures demonstrated, the WSG exhibited a morphology of a chain and WSG-PPI conjugates and WSG-SPI conjugates exhibited a chain structure with some particles attached, while the PPI and SPI showed small particles. 2.3. Application 2.3.1. Cooking Yield of Sausage The cooking yield of low-fat sausages increased first and decreased as the additional weight of the WSG-PPI conjugates increased, comparing with the normal-fat (NF) sausage (Figure 7a). However, the cooking yield of the low-fat sausage-added WSG-SPI conjugates was slightly lower than the NF sausage. The WSG-PPI conjugate had good emulsifying properties, making the sausage compact and homogeneous. In addition, the WSG was a kind of macromolecular polysaccharide, which could form a network structure with protein. However, the cooking yield of WSG-SPI conjugates decreased, which indicated that the ability of WSG-SPI conjugates to strengthen the meat mixture was less than that of WSG-PPI conjugates. A similar result was found in sausages containing pineapple dietary fiber 60 [30]. Pig skin and wheat fiber mixture as a fat replacer for frankfurter-type sausages could improve their cooking yield and water-holding capability [59]. Low-fat sausage-added oatmeal or tofu increased water retention, and produced less cooking loss [60]. Polysaccharides had the capacity to improve fat and water retention in sausages [61,62]. The effects of CMC on the water-holding capacity of three batters with different viscosities were compared. CMC could improve the water retention of medium viscous batters and high viscous batters, but had no effect on low viscous batters [63]. 2.3.2. Water Loss of Sausages Low-fat sausages lose less water than the NF sausage (Figure 7b). The water-holding capacity of low-fat sausages was increased as a result of the addition of the WSG-PPI and WSG-SPI conjugates. The effect of WSG-PPI conjugates and WSG-SPI conjugates on the water-holding of sausages contained mainly two aspects: for one thing, the hydroxyl groups can absorb water through hydrogen bonds and hydrophobic interactions when the WSG-PPI conjugates and WSG-SPI conjugates were heated [64]; for another, the more compact gel structure could be formed between protein and WSG-PPI conjugates or WSG-SPI conjugates, which wrapped water inside the gel to prevent it from losing [65,66]. 2.3.3. Color of Sausages Color was an intuitive indicator of sausages, and it affected the choice of consumers. As shown in Figure 7c, the lightness (L*), redness (a*) and yellowness (b*) of the low-fat sausages were different to the NF sausage. WSG-PPI and WSG-SPI conjugates caused a decrease in lightness (L*) values. The redness (a*) values increased with the WSG-PPI and WSG-SPI conjugates’ addition. The yellowness (b*) values were higher than the NF sausages. The L* values of the meat product decreased as a decrease in the fat level, then the value of a* was the opposite [67,68]. Câmara et al. [69] reported that the values of L*, a* decreased and the value of b* increased, when chia (Salvia hispanica L.) mucilage was added to the meat model system. 2.3.4. Texture Profile Analysis (TPA) Fat content played a basic role in the texture, flavor, mouthfeel and bite [70]. Fat improved the water-binding capacity and stabilized the gel network of the protein-emulsified meat products [69]. The hardness of low-fat sausages was higher than NF sausages; the highest hardness was 6945.50 ± 460.21, 7045.78 ± 237.79, respectively, when the sausages with WSG-PPI or WSG-SPI replaced 20% fat (Figure 8). The increase in the hardness and chewiness was attributed to the addition of the WSG-PPI and WSSG-SPI conjugates, such as high-water retention capacity, high-fiber content, viscosity and viscoelastic properties, which increased the hardness [71,72]. The result agreed with Choe [29], who found that the addition of chicken skin and wheat fiber led to higher hardness in the sausage. The changes of cohesiveness were consistent with the result of Abbasi [73], who found that 0.5% gum tragacanth increased the cohesiveness of the sausage, while 1% gum tragacanth reduced the cohesiveness, which was related to the fat content and the water-holding capacity of the gum tragacanth. 3. Materials and Methods 3.1. Materials S. cerecisiae β-glucan was purchased from Angel Yeast Co., Ltd. (Hubei, China). Peanut protein powder was purchased from Qingdao Changshou Co., Ltd. (Shandong, China). Soy protein isolate was purchased from Solae Co., Ltd. (Zhengzhou, China). Pork loin was purchased from a local market. Unless otherwise specified, reagents including odium dodecane sulfonate solution (SDS), NaOH and sodium phosphate were analytical grade. 3.2. Extraction of PPI from Peanut Protein Powder Extraction of PPI was based on a previous study [74]. Briefly, peanut protein powder (50 g) was added to deionized water (100 mL). The mixture was adjusted to pH 9.0 and stirred at 150 rpm for 2 h. The supernatant was collected after centrifugation and kept for 1 h at pH 4.5 to precipitate the protein. The precipitation was collected after centrifugation and adjusted to pH 7.0. The suspension of PPI was freeze-dried. The content of PPI was 82.05 ± 0.36% (w/w) and measured by the Kjeldahl method. 3.3. Extraction of Arachin and Conarachin from PPI Arachin and conarachin from the PPI was prepared according to Feng et al. [75]. The PPI was mixed with phosphate buffer (0.3 mol/L, pH 7.5) at a ratio of 2:5 (w/v). The mixture was stirred for 1 h and centrifuged at 8000 rpm for 30 min. The supernatant was cooled to 4 °C for 4 h and centrifuged again. Arachin was the precipitation gathered from the second centrifugation. The supernatant was adjusted to pH 4.5, and centrifuged at 4500 rpm for 20 min. The precipitate obtained was conarachin. Then, the samples of arachin and conarachin were freeze-dried. 3.4. Extraction of 7S Globulin from Soy Protein Isolate (SPI) The 7S globulin was collected according to Nagano et al. [76] with light modification. SPI was added to deionized water at an ambient temperature and adjusted to pH 8.0. The mixture was stirred slowly and centrifuged to collect the supernatant. Sodium bisulfite was added to the supernatant and kept for 30 min. Then, the mixture was adjusted to pH 6.4 and kept at 4 °C overnight. The supernatant was collected after centrifugation and added NaCl with final concentration of 0.2 mol/L. The mixture was kept for 1 h at pH 5.0 and centrifuged to collect the supernatant. The pH of mixture was adjusted to 5.0 and centrifuged to collect the precipitation. The precipitation was mixed with little deionized water, then adjusted to pH 7.5 and freeze-dried. 3.5. Preparation of WSG from S. cerecisiae β-Glucan S. cerecisiae β-glucan (1.5 g) was mixed with deionic water (100 mL). The snail enzyme (0.06 g) was added to the mixture and kept at 45 °C for 80 min. Then, the mixture was put in boiling water to deactivate the enzyme. The supernatant was collected after centrifugation at 5000 rpm and freeze-dried. 3.6. Preparation of WSG-PPI/SPI/Arachin/Conarachin/7S Globulin Conjugates Conjugates were prepared based on a previous study [77] with light modification. WSG and PPI/arachin/conarachin with a proportion of 1:3 (w/w) were mixed with 0.02 M phosphate buffer and stirred for 2 h. The dispersion was adjusted to pH 9.0 and hydrated overnight. The dispersion was heated at 90 °C for 80 min and cooled down in an ice-water bath. The WSG and SPI/7S globulin were mixed with a 0.02 M phosphate buffer at a proportion of 1:2 (w/w) and stirred for 2 h. The mixture was adjusted to pH 9.0 and hydrated overnight. Then, the dispersion was heated at 90 °C for 3 h and cooled down. The samples were dialyzed at 4 °C for 48 h and freeze-dried. 3.7. Functional Properties 3.7.1. Emulsifying Properties EAI and ESI were determined by Chen et al. [78]. Briefly, the conjugates’ solution obtained a WSG concentration of 9 mg/mL and protein concentration of 3 mg/mL. The solution (7 mL) and oil (3 mL) were homogenized at 10,000 rpm for 2 min. A total of 50 μL of emulsions at 0 and 10 min were added to 5 mL 0.1% sodium dodecanesulfonate solution (SDS), respectively. The absorbance of diluted emulsion at 500 nm was recorded using a UVi-2050 spectrophotometer. The EAI and ESI were calculated by the following Equations:EAI (m2/g)=2×2.303×A0×DC×Φ×L×104 ESI (min)=A0×10A0−A10 where A0 and A10 were the absorbance of the diluted emulsion at 0 and 10 min, respectively. D, C, L, and Φ were the dilution factor (100), the concentration of protein, the width of light length, and proportion of oil phase, respectively. 3.7.2. Wettability Measurement The three-phase contact angle of WSG and its conjugates was measured according to Yang et al. [79]. The solid powder of WSG and its conjugates was compressed into pellets by a tablet press, then a droplet of water was dropped on the surface of the pellets through a high-precision injector, and the picture of contact interface was taken when the contact occurred for 10 s. The contact angle of the samples was analyzed by software. 3.7.3. Thermogravimetric Assay The thermal stability of samples was measured by a thermogravimetric analyzer (Q50, Tokyo, Japan). The weight of the samples was about 6 mg. The determination conditions were as follows: the heating program was 30–500 °C, the heating rate was 15 °C/min and the nitrogen flow rate was 50 mL/min. 3.7.4. Rheological Properties Rheological properties of WSG, WSG-PPI and WSG-SPI conjugates were studied using a DHR-2 rheometer (TA Instruments, New Castle, DE, USA). The concentration of all samples was 40 mg/mL. The flow and dynamic rheological tests were determined using a 40 mm parallel plate with a 1 mm gap. The temperature was 25 °C. 3.7.5. Particle Size The particle size distribution of samples was determined by a Matersize-3000 (Malvern Instruments Ltd., Worcestershire, UK). The refractive index of the continuous phase and dispersed phased was 1.330 and 1.375, respectively. 3.8. Structural Characterization 3.8.1. FT-IR WSG, WSG-PPI and WSG-SPI conjugates were mixed with KBr and pressed into a transparent sheet. The parameter was set to a resolution of 4 cm−1 and 64 scans in the range of 400–4000 cm−1. 3.8.2. CD The CD spectrum was obtained by a JASCO spectropolarimeter model J-1500 (JASCO, Tokyo, Japan). Each CD spectrum was performed at a scanning speed of 100 nm/min. The wavelength range was 190–250 nm with the bandwidth set at 0.1 nm. The deionized water was used as a solvent for all the samples. 3.8.3. SEM Samples were stuck to a carbon conductive tab and sprayed with gold. The morphological properties of the samples were observed by a scanning electron microscope at 12.5 kV voltages. 3.8.4. Cryo-Scanning Electron Microscope (Cryo-SEM) The samples were formulated as a 3 mg/mL solution. Then, a droplet of solution was loaded on the cryo-specimen holder and cryo-fixed in slush nitrogen (−210 °C). After freezing, the samples were fractured in the fracture chamber. Then, the cross-section was sublimed at −90 °C and deposited, and the samples were viewed in the cryo-SEM (FEI Helios NanoLab G3 UC, State of Oregon, USA). 3.8.5. AFM The method of AFM was referred to in a previous study [80]. The samples were prepared in 50 μg/mL solution and stirred at 80 °C for 4 h. A total of 10 μL of each solution was dropped on the fresh mica, and naturally aired. 3.9. Application of Soluble β-Glucan-PPI Conjugation 3.9.1. Preparation of Low-Fat Sausage The formulation of sausage was showed at Table 1. The lean pork was pickled with 2% salt for 24–48 h. Then, the lean pork was mixed with ice, WSG or WSG-PPI/SPI conjugates and chopped for 2 min at high speed. The fat was added to the mixture and chopped for 5 min (total time 7 min and final temperature <12 °C). Then, the batter was stuffed into collagen casings and vacuum-packed in polyethylene bags; the sausage was cooked at 80 °C for 45 min. 3.9.2. Cooking Yield of Sausage The surface of sample was dry and the exudate was removed after the sample was cooked. The quality of sausage was measured before and after cooking. Cooking yield (%)=final weight of sausage-weight of casinginitial weight of sausage-weight of casing×100 3.9.3. Color Measurement The sausage was cut into 3 mm slide. The model chromameter was used to determine the color of sausage among different treatments for each of the 6 replications. The L*, a* and b* values were determined. 3.9.4. Texture The texture profile analysis (TPA) of the samples was measured by a TA-xT2i Texture Analyzer (Stable Micro Systems, UK), according to a method by Pires et al. [81]. The samples were divided into cylinders with a height of 2 cm after stripping the casing. The cores were compressed into two consecutive cycles of 50% using a P-36R probe. The instrument settings were as follows: pre-test speed of 2 mm/s, test speed of 1 mm/s and the post-test speed of 2 mm/s. The parameters of hardness (g), springiness, cohesiveness and chewiness were studied. 3.10. Statistical Analysis All experiments were replicated 3 times, and the data were analyzed by a variance analysis (ANOVA) to determine the differences (p < 0.05) using SPSS 26. The graphs were made by Origin 8.5. 4. Conclusions In this study, WSG-PPI and WSG-SPI conjugates were formed in the specified conditions. A Maillard reaction occurred between amino and carbonyl groups in WSG and PPI or SPI, resulting in the consumption of some functional groups and the appearance of new groups in the conjugates. In the process of the graft, the formation of WSG-PPI and WSG-SPI conjugates was observed by SEM, AFM and Cryo-SEM. The functional properties and structure of WSG-PPI and WSG-SPI conjugates changed compared to WSG. WSG-PPI and WSG-SPI demonstrated a better emulsifying property and a higher thermal stability than WSG, which was because the polysaccharide was presumably the enhanced steric stabilization provided by the bulky hydrophilic moiety. In addition, the WSG-PPI and WSG-SPI conjugates replaced part of the fat in the low-fat sausages. The cooking yield and water-holding capability improved compared to the normal fat sausage, and the hardness, springiness, chewiness and cohesiveness of sausages replaced WSG-PPI instead of 20% fat, and were more in line with the requirements of commercially available products. The range of application for the conjugates was expended. Acknowledgments This work was financially supported by the National Natural Science Foundation of China (No. 31671817) and the Special Scientific Research Funds for Central Non-profit Institutes (No. S2016JC04). Author Contributions Investigation, resources, data, writing—original draft preparation, M.Z.; writing—review and editing, H.L. and Q.W. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Data are available from the corresponding author if requested. Conflicts of Interest The authors declare no conflict of interest. Sample Availability Samples of the compounds are not available from the author. Figure 1 Emulsifying properties of WSG and its conjugates. Figure 2 Contact angle of WSG and its conjugates: (a) WSG; (b) SPI; (c) PPI; (d) WSG-SPI conjugates; (e) WSG-PPI conjugates. Figure 3 Functional properties of WSG and its conjugates: (a) particle size of WSG-PPI conjugates; (b) TGA of WSG-PPI conjugates; (c) rheological properties of WSG-PPI conjugates; (d) particle size of WSG-SPI conjugates; (e) TGA of WSG-SPI conjugates; (f) rheological of WSG-SPI conjugates. Figure 4 FT-IR spectra and CD spectra of WSG and its conjugates. (a) FT-IR spectra of WSG-PPI conjugates; (b) FT-IR spectra of WSG-SPI conjugates; (c) CD spectra of WSG-PPI conjugate; (d) CD spectra of WSG-SPI conjugates. Figure 5 The SEM graphs of WSG and its conjugates: (a) WSG; (b) WSG-PPI conjugates; (c) WSG-SPI conjugates. Figure 6 Cryo-SEM and AFM graphs of WSG and its conjugates: (a) Cryo-SEM graph of WSG; (b) Cryo-SEM graph of WSG-PPI conjugates; (c) Cryo-SEM graph of WSG-SPI conjugates; (d) Cryo-SEM graph of PPI; (e) Cryo-SEM graph of SPI; (f) AFM graph of WSG; (g) AFM graph of WSG-PPI conjugates; (h) AFM graph of WSG-SPI conjugates; (i) AFM graph of PPI; (j) AFM graph of SPI. Figure 7 Cooking yield (a), water loss (b) and color (c) of low-fat sausages. Figure 8 The texture of low-fat sausages. molecules-27-03037-t001_Table 1 Table 1 Formulations of sausage. Treatment Lean Pork (%) Fat (%) Water/Ice (%) Salt (%) WSG-PPI/WSG-SPI (%) Normal 66.4 16.6 15 2 0 WSG-PPI 1 66.4 13.28 15 2 3.32 WSG-PPI 2 66.4 9.96 15 2 6.64 WSG-PPI 3 66.4 6.64 15 2 9.96 WSG-SPI 1 66.4 13.28 15 2 3.32 WSG-SPI 2 66.4 9.96 15 2 6.64 WSG-SPI 3 66.4 6.64 15 2 9.96 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Klis F.M. Mol P. Hellingwerf K. Brul S. Dynamics of cell wall structure in Saccharomyces cerevisiae FEMS Microbiol. Rev. 2002 26 239 256 10.1111/j.1574-6976.2002.tb00613.x 12165426 2. Aimanianda V. Bayry J. Bozza S. Kniemeyer O. Perruccio K. Elluru S.R. Clavaud C. Paris S. Brakhage A.A. Kaveri S.V. Surface hydrophobin prevents immune recognition of airborne fungal spores Nature 2009 460 1117 1121 10.1038/nature08264 19713928 3. Fang X.-H. Zou M.-Y. Chen F.-Q. Ni H. Nie S.-P. Yin J.-Y. An overview on interactions between natural product-derived β-glucan and small-molecule compounds Carbohydr. Polym. 2021 261 117850 10.1016/j.carbpol.2021.117850 33766346 4. Qi C. Cai Y. Gunn L. Ding C. Li B. Kloecker G. Qian K. Vasilakos J. Saijo S. Iwakura Y. Differential pathways regulating innate and adaptive antitumor immune responses by particulate and soluble yeast-derived β-glucans Immunobiology 2011 117 6825 6836 10.1182/blood-2011-02-339812 5. Mo L. Chen Y. Li W. Guo S. Wang X. An H. Zhan Y. Anti-tumor effects of (1→3)-β-d-glucan from Saccharomyces cerevisiae in S180 tumor-bearing mice Int. J. Biol. Macromol. 2017 95 385 392 10.1016/j.ijbiomac.2016.10.106 27838421 6. Salvador C. Li B. Hansen R. Cramer D.E. Kong M. Yan J. Yeast-Derived β-Glucan Augments the Therapeutic Efficacy Mediated by Anti–Vascular Endothelial Growth Factor Monoclonal Antibody in Human Carcinoma Xenograft Models Clin. Cancer Res. 2008 14 1239 1247 10.1158/1078-0432.CCR-07-1669 18281559 7. Andrade E.F. Lobato R.V. Araújo T.V. Zangerônimo M.G. Sousa R.V. Pereira L.J. Effect of beta-glucans in the control of blood glucose levels of diabetic patients: A systematic review Nutrcion Hosp. 2014 31 170 177 10.3305/nh.2015.31.1.7597 8. Wang N. Liu H. Liu G. Li M. He X. Yin C. Tu Q. Shen X. Bai W. Wang Q. Yeast β-D-glucan exerts antitumour activity in liver cancer through impairing autophagy and lysosomal function, promoting reactive oxygen species production and apoptosis Redox Biol. 2020 32 101495 10.1016/j.redox.2020.101495 32171725 9. Mei X. Tang Q. Huang G. Long R. Huang H. Preparation, structural analysis and antioxidant activities of phosphorylated (1 → 3)-β-d-glucan Food Chem. 2020 309 125791 10.1016/j.foodchem.2019.125791 31699554 10. Dietrich-Muszalska A. Olas B. Kontek B. Rabe-Jabłońska J. Beta-glucan from Saccharomyces cerevisiae reduces plasma lipid peroxidation induced by haloperidol Int. J. Biol. Macromol. 2011 49 113 116 10.1016/j.ijbiomac.2011.03.007 21421004 11. Cao Y. Sun Y. Zou S. Li M. Xu X. Orally administered baker’s yeast β-glucan promotes glucose and lipid homeostasis in the livers of obesity and diabetes model mice J. Agric. Food Chem. 2017 65 9665 9674 10.1021/acs.jafc.7b03782 29035040 12. Zhang M. Jin X. Yang Y.-F. β-Glucan from Saccharomyces cerevisiae induces SBD-1 production in ovine ruminal epithelial cells via the Dectin-1–Syk–NF-κB signaling pathway Cell. Signal. 2018 53 304 315 10.1016/j.cellsig.2018.10.018 30401641 13. Bose N. Chan A.S.H. Guerrero F. Maristany C.M. Qiu X. Walsh R.M. Ertelt K.E. Jonas A.B. Gorden K.B. Dudney C.M. Effects of β-glucan derivatives on the immunity ofwhite shrimp Litopenaeus vannamei and its resistance against white spot syndrome virus infection Aquaculture 2014 426–427 66 73 10.3389/fimmu.2013.00230 14. Liu H. Li Y. Gao J. Shi A. Liu L. Hu H. Putri N. Yu H. Fan W. Wang Q. Effects of microfluidization with ionic liquids on the solubilization and structure of β-d-glucan Int. J. Biol. Macromol. 2016 84 394 401 10.1016/j.ijbiomac.2015.12.014 26691385 15. Tang Q. Huang G. Zhao F. Zhou L. Huang S. Li H. The antioxidant activities of six (1 → 3)-β- d -glucan derivatives prepared from yeast cell wall Int. J. Biol. Macromol. 2017 98 216 221 10.1016/j.ijbiomac.2017.01.132 28161534 16. Zhang H. Zhang J. Fan Z. Zhou X. Geng L. Wang Z. Regenstein J.M. Xia Z. Chemical Synthesis of Sulfated Yeast (Saccharomyces cerevisiae) Glucans and Their In Vivo Antioxidant Activity Molecules 2017 22 1266 10.3390/molecules22081266 17. Long N.T. Anh N.T.N. Giang B.L. Son H.N. Luan L.Q. Radiation Degradation of β-Glucan with a Potential for Reduction of Lipids and Glucose in the Blood of Mice Carbohydr. Polym. 2019 11 955 10.3390/polym11060955 18. Shi F. Shi J. Li Y. Mechanochemical phosphorylation and solubilisation of β-D-Glucan from yeast Saccharomyces cerevisiae and its biological activities PLoS ONE 2014 9 e103494 10.1371/journal.pone.0103494 25075740 19. Liu Y. Wu Q. Wu X. Algharib S.A. Gong F. Hu J. Luo W. Zhou M. Pan Y. Yan Y. Structure, preparation, modification, and bioactivities of beta-glucan and mannan from yeast cell wall: A review Int. J. Biol. Macromol. 2021 173 445 456 10.1016/j.ijbiomac.2021.01.125 33497691 20. Cui Q. Zhang A. Li R. Wang X. Sun L. Jiang L. Ultrasonic treatment affects emulsifying properties and molecular flexi-bility of soybean protein isolate-glucose conjugates Food Biosci. 2020 38 100747 10.1016/j.fbio.2020.100747 21. Osano J.P. Hosseini-Parvar S.H. Matia-Merino L. Golding M. Emulsifying properties of a novel polysaccharide extracted from basil seed (Ocimum bacilicum L.): Effect of polysaccharide and protein content Food Hydrocoll. 2014 37 40 48 10.1016/j.foodhyd.2013.09.008 22. Sheng L. Tang G. Wang Q. Zou J. Ma M. Huang X. Molecular characteristics and foaming properties of ovalbu-min-pullulan conjugates through the Maillard reaction Food Hydrocoll. 2019 100 105384 10.1016/j.foodhyd.2019.105384 23. Wang L. Wu M. Liu H.-M. Emulsifying and physicochemical properties of soy hull hemicelluloses-soy protein isolate con-jugates Carbohydr. Polym. 2017 163 181 190 10.1016/j.carbpol.2017.01.069 28267495 24. Li Z. Zheng Y. Sun Q. Wang J. Zheng B. Guo Z. Structural characteristics and emulsifying properties of myofibrillar protein-dextran conjugates induced by ultrasound Maillard reaction Ultrason. Sonochemistry 2021 72 105458 10.1016/j.ultsonch.2020.105458 33453682 25. Chen W. Lv R. Wang W. Ma X. Muhammad A.I. Guo M. Ye X. Liu D. Time effect on structural and functional prop-erties of whey protein isolate-gum acacia conjugates prepared via Maillard reaction J. Sci. Food Agric. 2019 99 4801 4807 10.1002/jsfa.9735 30977143 26. Sun T. Qin Y. Xie J. Xu H. Gan J. Wu J. Bian X. Li X. Xiong Z. Xue B. Effect of Maillard reaction on rheological, physicochemical and functional properties of oat β-glucan Food Hydrocoll. 2018 89 90 94 10.1016/j.foodhyd.2018.10.029 27. Dos Santos M. Munekata P.E. Pateiro M. Magalhães G.C. Barretto A.C.S. Lorenzo J.M. Pollonio M.A.R. Pork skin-based emulsion gels as animal fat replacers in hot-dog style sausages LWT 2020 132 109845 10.1016/j.lwt.2020.109845 28. Li X.-L. Meng R. Xu B.-C. Zhang B. Cui B. Wu Z.-Z. Function emulsion gels prepared with carrageenan and ze-in/carboxymethyl dextrin stabilized emulsion as a new fat replacer in sausages Food Chem. 2022 389 133005 10.1016/j.foodchem.2022.133005 35490528 29. Choe J. Kim H.-Y. Quality characteristics of reduced fat emulsion-type chicken sausages using chicken skin and wheat fiber mixture as fat replacer Poult. Sci. 2019 98 2662 2669 10.3382/ps/pez016 30698807 30. Henning S.S. Tshalibe P. Hoffman L. Physico-chemical properties of reduced-fat beef species sausage with pork back fat replaced by pineapple dietary fibres and water LWT 2016 74 92 98 10.1016/j.lwt.2016.07.007 31. Broucke K. Van Poucke C. Duquenne B. De Witte B. Baune M.-C. Lammers V. Terjung N. Ebert S. Gibis M. Weiss J. Ability of (extruded) pea protein products to partially replace pork meat in emulsified cooked sausages Innov. Food Sci. Emerg. Technol. 2022 78 102992 10.1016/j.ifset.2022.102992 32. Zhang B. Guo X. Zhu K. Peng W. Zhou H. Improvement of emulsifying properties of oat protein isolate–dextran conju-gates by glycation Carbohydr. Polym. 2015 127 168 175 10.1016/j.carbpol.2015.03.072 25965470 33. Guo X. Xiong Y.L. Characteristics and functional properties of buckwheat protein–sugar Schiff base complexes LWT Food Sci. Technol. 2013 51 397 404 10.1016/j.lwt.2012.12.003 34. Qu W. Zhang X. Chen W. Wang Z. He R. Ma H. Effects of ultrasonic and graft treatments on grafting degree, structure, functionality, and digestibility of rapeseed protein isolate-dextran conjugates Ultrason. Sonochemistry 2018 42 250 259 10.1016/j.ultsonch.2017.11.021 35. Saatchi A. Kiani H. Labbafi M. A new functional protein polysaccharide conjugate based on protein concentrate from sesame processing by-products: Functional and physico-chemical properties Int. J. Biol. Macromol. 2019 122 659 666 10.1016/j.ijbiomac.2018.10.122 30342133 36. Lin J. Guo X. Ai C. Zhang T. Yu S. Genipin crosslinked sugar beet pectin-whey protein isolate/bovine serum albumin conjugates with enhanced emulsifying properties Food Hydrocoll. 2020 105 105802 10.1016/j.foodhyd.2020.105802 37. Shao P. Zhang H. Niu B. Jin W. Physical stabilities of taro starch nanoparticles stabilized Pickering emulsions and the potential application of encapsulated tea polyphenols Int. J. Biol. Macromol. 2018 118 Pt B 2032 2039 10.1016/j.ijbiomac.2018.07.076 30021133 38. Chen X. Qi Y. Zhu C. Wang Q. Effect of ultrasound on the properties and antioxidant activity of hawthorn pectin Int. J. Biol. Macromol. 2019 131 273 281 10.1016/j.ijbiomac.2019.03.077 30876895 39. Patel M.K. Tanna B. Gupta H. Mishra A. Jha B. Physicochemical, scavenging and anti-proliferative analyses of polysac-charides extracted from psyllium (Plantago ovata Forssk) husk and seeds Int. J. Biol. Macromol. 2019 133 190 201 10.1016/j.ijbiomac.2019.04.062 30981777 40. Xu W. Zhao X.-H. Structure and property changes of the maltose-glycated caseinate prepared through the Maillard reaction in an ionic liquid medium LWT 2018 99 483 489 10.1016/j.lwt.2018.10.021 41. Su J.-F. Huang Z. Yuan X.-Y. Wang X.-Y. Li M. Structure and properties of carboxymethyl cellulose/soy protein isolate blend edible films crosslinked by Maillard reactions Carbohydr. Polym. 2010 79 145 153 10.1016/j.carbpol.2009.07.035 42. Numata M. Matsumoto T. Umeda M. Koumoto K. Sakurai K. Shinkai S. Polysaccharide–polynucleotide complexes (15): Thermal stability of schizophyllan (SPG)/poly(C) triple strands is controllable by α-amino acid modification Bioorganic Chem. 2003 31 163 171 10.1016/S0045-2068(03)00020-8 43. Sun T. Xu H. Zhang H. Ding H.H. Cui S. Xie J. Xue B. Hua X. Maillard reaction of oat β-glucan and the rheological property of its amino acid/peptide conjugates Food Hydrocoll. 2018 76 30 34 10.1016/j.foodhyd.2017.07.025 44. Inglett G.E. Chen D. Liu S.X. Functional Properties of Teff and Oat Composites Food Nutr. Sci. 2015 6 1591 1602 10.4236/fns.2015.617164 45. Kim W.-W. Yoo B. Rheological behaviour of acorn starch dispersions: Effects of concentration and temperature Int. J. Food Sci. Technol. 2009 44 503 509 10.1111/j.1365-2621.2008.01760.x 46. Wang W.-Q. Bao Y.-H. Chen Y. Characteristics and antioxidant activity of water-soluble Maillard reaction products from interactions in a whey protein isolate and sugars system Food Chem. 2013 139 355 361 10.1016/j.foodchem.2013.01.072 23561117 47. Kanmani P. Kumar R.S. Yuvaraj N. Paari K. Pattukumar V. Arul V. Production and purification of a novel exopoly-saccharide from lactic acid bacterium Streptococcus phocae PI80 and its functional characteristics activity in vitro Bioresour. Technol. 2011 102 4827 4833 10.1016/j.biortech.2010.12.118 21300540 48. Tang J. Nie J. Li D. Zhu W. Zhang S. Ma F. Sun Q. Song J. Zheng Y. Chen P. Characterization and antioxidant activities of degraded polysaccharides from Poria cocos sclerotium Carbohydr. Polym. 2014 105 121 126 10.1016/j.carbpol.2014.01.049 24708960 49. Xu C. Lv J. Lo Y.M. Cui S.W. Hu X. Fan M. Effects of oat β-glucan on endurance exercise and its anti-fatigue properties in trained rats Carbohydr. Polym. 2013 92 1159 1165 10.1016/j.carbpol.2012.10.023 23399141 50. Chen H.-M. Fu X. Luo Z.-G. Effect of molecular structure on emulsifying properties of sugar beet pulp pectin Food Hydrocoll. 2015 54 99 106 10.1016/j.foodhyd.2015.09.021 51. Wang H. Chen J. Ren P. Zhang Y. Onyango S.O. Ultrasound irradiation alters the spatial structure and improves the antioxidant activity of the yellow tea polysaccharide Ultrason. Sonochemistry 2020 70 105355 10.1016/j.ultsonch.2020.105355 33007535 52. Jones C. An S.J. Yoon Y.K. Kothari S. Sahastrabuddhe S. Carbis R. Spectroscopic characterisation of a series of Salmonella Typhi Vi-diphtheria toxoid glycoconjugate antigens differing in polysaccharide-protein ratio J. Pharm. Biomed. Anal. 2020 181 113100 10.1016/j.jpba.2020.113100 31991319 53. Kato A. Industrial Applications of Maillard-Type Protein-Polysaccharide Conjugates Food Sci. Technol. Res. 2002 8 193 199 10.3136/fstr.8.193 54. Yu J.-J. Chen G.-Y. Zhang Y.-F. Zheng X.-C. Jiang P.-Y. Ji H. Li S.-H. Chen Y. Enhanced hydration properties and antioxidant activity of peanut protein by covalently binding with sesbania gum via cold plasma treatment Innov. Food Sci. Emerg. Technol. 2021 68 102632 10.1016/j.ifset.2021.102632 55. Yu J.-J. Ji H. Chen Y. Zhang Y.-F. Zheng X.-C. Li S.-H. Analysis of the glycosylation products of peanut protein and lactose by cold plasma treatment: Solubility and structural characteristics Int. J. Biol. Macromol. 2020 158 1194 1203 10.1016/j.ijbiomac.2020.04.255 32371128 56. Boostani S. Aminlari M. Moosavi-Nasab M. Niakosari M. Mesbahi G. Fabrication and characterisation of soy protein isolate-grafted dextran biopolymer: A novel ingredient in spray-dried soy beverage formulation Int. J. Biol. Macromol. 2017 102 297 307 10.1016/j.ijbiomac.2017.04.019 28392389 57. Echlin P. Low Temperature Microscopy and Analysis Plenum Press New York, NY, USA 1992 58. Morris V.J. Kirby A.R. Gunning A.P. Atomic Force Microscopy for Biologists World Scientific Singapore 2009 59. Choe J.-H. Kim H.-Y. Lee J.-M. Kim Y.-J. Kim C.-J. Quality of frankfurter-type sausages with added pig skin and wheat fiber mixture as fat replacers Meat Sci. 2013 93 849 854 10.1016/j.meatsci.2012.11.054 23313971 60. Yang H.-S. Choi S.-G. Jeon J.-T. Park G.-B. Joo S.-T. Textural and sensory properties of low fat pork sausages with added hydrated oatmeal and tofu as texture-modifying agents Meat Sci. 2007 75 283 289 10.1016/j.meatsci.2006.07.013 22063660 61. García-García E. Totosaus A. Low-fat sodium-reduced sausages: Effect of the interaction between locust bean gum, potato starch and κ-carrageenan by a mixture design approach Meat Sci. 2008 78 406 413 10.1016/j.meatsci.2007.07.003 22062459 62. Do Amaral D.S. Cardelle-Cobas A. do Nascimento B.M. Monteiro M.J. Madruga M.S. Pintado M.M.E. Development of a low fat fresh pork sausage based on chitosan with health claims: Impact on the quality, functionality and shelf-life Food Funct. 2015 6 2768 2778 10.1039/C5FO00303B 26158872 63. Gibis M. Schuh V. Allard K. Weiss J. Influence of molecular weight and degree of substitution of various carboxymethyl celluloses on unheated and heated emulsion-type sausage models Carbohydr. Polym. 2016 159 76 85 10.1016/j.carbpol.2016.12.012 28038756 64. Pietrasik Z. Janz J. Utilization of pea flour, starch-rich and fiber-rich fractions in low fat bologna Food Res. Int. 2010 43 602 608 10.1016/j.foodres.2009.07.017 65. Chen H. Huang Y. Rheological properties of HPMC enhanced surimi analyzed by small- and large-strain tests—II: Effect of water content and ingredients Food Hydrocoll. 2008 22 313 322 10.1016/j.foodhyd.2006.12.006 66. Zhuang X. Jiang X. Zhou H. Chen Y. Zhao Y. Yang H. Zhou G. Insight into the mechanism of physicochemical influence by three polysaccharides on myofibrillar protein gelation Carbohydr. Polym. 2020 229 115449 10.1016/j.carbpol.2019.115449 31826526 67. Lorenzo J.M. Franco D. Fat effect on physico-chemical, microbial and textural changes through the manufactured of dry-cured foal sausage Lipolysis, proteolysis and sensory properties Meat Sci. 2012 92 704 714 10.1016/j.meatsci.2012.06.026 22795774 68. Gómez M. Lorenzo J.M. Effect of fat level on physicochemical, volatile compounds and sensory characteristics of dry-ripened “chorizo” from Celta pig breed Meat Sci. 2013 95 658 666 10.1016/j.meatsci.2013.06.005 23811106 69. Câmara A.K.F.I. Okuro P.K. da Cunha R.L. Herrero A.M. Ruiz-Capillas C. Pollonio M.A.R. Chia (Salvia hispanica L.) mucilage as a new fat substitute in emulsified meat products: Technological, physicochemical, and rheological characterization LWT Food Sci. Technol. 2020 125 109193 10.1016/j.lwt.2020.109193 70. Colmenero F.J. Relevant factors in strategies for fat reduction in meat products Trends Food Sci. Technol. 2000 11 56 66 10.1016/S0924-2244(00)00042-X 71. Han M. Bertram H.C. Designing healthier comminuted meat products: Effect of dietary fibers on water distribution and texture of a fat-reduced meat model system Meat Sci. 2017 133 159 165 10.1016/j.meatsci.2017.07.001 28692849 72. Pintado T. Herrero A. Jimenez-Colmenero F. Ruiz-Capillas C. Strategies for incorporation of chia (Salvia hispanica L.) in frankfurters as a health-promoting ingredient Meat Sci. 2016 114 75 84 10.1016/j.meatsci.2015.12.009 26745305 73. Abbasi E. Sarteshnizi R.A. Gavlighi H.A. Nikoo M. Azizi M.H. Sadeghinejad N. Effect of partial replacement of fat with added water and tragacanth gum (Astragalus gossypinus and Astragalus compactus) on the physicochemical, texture, oxida-tive stability, and sensory property of reduced fat emulsion type sausage Meat Sci. 2018 147 135 143 10.1016/j.meatsci.2018.09.007 30243231 74. Jiao B. Shi A. Wang Q. Binks B.P. High-Internal-Phase Pickering Emulsions Stabilized Solely by Peanut-Protein-Isolate Microgel Particles with Multiple Potential Applications Angew. Chem. Int. Ed. 2018 57 9274 9278 10.1002/anie.201801350 29845713 75. Feng X.-L. Liu H.-Z. Shi A.-M. Liu L. Wang Q. Adhikari B. Effects of transglutaminase catalyzed crosslinking on phys-icochemical characteristics of arachin and conarachin-rich peanut protein fractions Food Res. Int. 2014 62 84 90 10.1016/j.foodres.2014.02.022 76. Nagano T. Hirotsuka M. Mori H. Kohyama K. Nishinari K. Dynamic viscoelastic study on the gelation of 7 S globulin from soybeans J. Agric. Food Chem. 1992 40 941 944 10.1021/jf00018a004 77. Wang Y. Yu J. Xu N. Wang G. Wang X. Influence of Protein Hydrolysis on the Freeze-thaw Stability of Emulsions Prepared with Soy Protein-Dextran Conjugates J. Oleo Sci. 2019 68 959 965 10.5650/jos.ess19076 31511465 78. Chen W. Ma X. Wang W. Lv R. Guo M. Ding T. Ye X. Miao S. Liu D. Preparation of modified whey protein isolate with gum acacia by ultrasound maillard reaction Food Hydrocoll. 2019 95 298 307 10.1016/j.foodhyd.2018.10.030 79. Yang T. Zheng J. Zheng B. Liu F. Wang S. Tang C.-H. High internal phase emulsions stabilized by starch nanocrystals Food Hydrocoll. 2018 82 230 238 10.1016/j.foodhyd.2018.04.006 80. Wang S. Zhao L. Li Q. Liu C. Han J. Zhu L. Zhu D. He Y. Liu H. Rheological properties and chain conformation of soy hull water-soluble polysaccharide fractions obtained by gradient alcohol precipitation Food Hydrocoll. 2019 91 34 39 10.1016/j.foodhyd.2018.12.054 81. Pires M. Munekata P.E.S. Baldin J.C. Rocha Y.J.P. Carvalho L.T. dos Santos I.R. Barros J.C. Trindade M. The effect of sodium reduction on the microstructure, texture and sensory acceptance of Bologna sausage Food Struct. 2017 14 1 7 10.1016/j.foostr.2017.05.002

PMC009xxxxxx/PMC9099642.txt

==== Front J Clin Med J Clin Med jcm Journal of Clinical Medicine 2077-0383 MDPI 10.3390/jcm11092587 jcm-11-02587 Article Association between Early Mobilization in the ICU and Psychiatric Symptoms after Surviving a Critical Illness: A Multi-Center Prospective Cohort Study Watanabe Shinichi 12 https://orcid.org/0000-0002-6867-1420 Liu Keibun 3* https://orcid.org/0000-0001-8481-0294 Nakamura Kensuke 4 https://orcid.org/0000-0002-9310-185X Kozu Ryo 5 Horibe Tatsuya 6 Ishii Kenzo 7 Yasumura Daisetsu 89 Takahashi You 10 Nanba Tomoya 11 Morita Yasunari 12 Kanaya Takahiro 1 Suzuki Shuichi 12 https://orcid.org/0000-0001-6673-5630 Lefor Alan Kawarai 13 https://orcid.org/0000-0003-2582-4324 Katsukawa Hajime 14 https://orcid.org/0000-0001-9504-0061 Kotani Toru 15 Ruetzler Kurt Academic Editor Spiel Alexander Oskar Academic Editor Andrès Emmanuel Academic Editor 1 Department of Rehabilitation, National Hospital Organization, Nagoya Medical Center, 4-1-1 Sannomaru, Naka-ku, Nagoya 460-0001, Japan; billabonghonor@yahoo.co.jp (S.W.); billabonghono@yahoo.co.jp (T.K.) 2 Department of Physical Therapy, Faculty of Rehabilitation, Gifu University of Health Science, 2-92 Higashiuzura, Gifu 500-8281, Japan 3 Critical Care Research Group, The Prince Charles Hospital, 627 Rode Rd, Chermside, QLD 4032, Australia 4 Department of Emergency and Critical Care Medicine, Hitachi General Hospital, 2-1-1 Jounann, Hitachi 317-0077, Japan; mamashockpapashock@yahoo.co.jp 5 Department of Cardiopulmonary Rehabilitation Science, Nagasaki University Graduate School of Biomedical Sciences, 1-14 Bunkyou-cho, Nagasaki 852-8521, Japan; ryokozu@nagasaki-u.ac.jp 6 Department of Rehabilitation, Tokyo Women’s Medical University, 8-1 Kawata-cho, Shinjuku-ku, Tokyo 162-8666, Japan; king.of.rehabilly@gmail.com 7 Department of Anesthesiology, Intensive Care Unit, Fukuyama City Hospital, 3-8-5 Zao-cho, Fukuyama 721-8511, Japan; keishii1101@gmail.com 8 Department of Rehabilitation, Naha City Hospital, 2-31-1 Furujima, Naha 902-8511, Japan; yasumuradai@yahoo.ne.jp 9 Department of Healthcare Administration, The University of Kyushu, 3-1-1, Maidashi, Higashi-ku, Fukuoka 812-8582, Japan 10 Department of Rehabilitation, Yuuai Medical Center, 50-5 Yone, Tomigusuku 901-0224, Japan; yo.takahashi7448@gmail.com 11 Department of Rehabilitation, Yao Tokushukai General Hospital, 1-17 Wakakusachou, Yao-shi, Osaka 581-0011, Japan; nanbatomoya@yahoo.co.jp 12 Department of Critical Care Medicine, National Hospital Organization, Nagoya Medical Center, 4-1-1 Sannomaru, Naka-ku, Nagoya 460-0001, Japan; moltlyme2@yahoo.co.jp (Y.M.); oltlyme2@yahoo.co.jp (S.S.) 13 Department of Surgery, Jichi Medical University, 3311-1 Yakushiji, Shimostuke-shi 329-0498, Japan; alefor@jichi.ac.jp 14 Japanese Society for Early Mobilization, 1-2-12 Kudankita, Tiyoda-ku, Tokyo 102-0073, Japan; winegood21@gmail.com 15 Department of Intensive Care Medicine, School of Medicine, Showa University, 1-5-8 Hatanodai, Shinagawa-ku, Tokyo 142-8666, Japan; trkotani@med.showa-u.ac.jp * Correspondence: keiliu0406@gmail.com; Tel.: +61-437-606-107 05 5 2022 5 2022 11 9 258711 3 2022 29 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). This is a prospective multicenter cohort study aiming to investigate the association between early mobilization (EM), defined as a rehabilitation level of sitting at the edge of the bed or higher within 72 h of ICU admission, and psychiatric outcome. Consecutive patients, admitted to the ICU for more than 48 h, were enrolled. The primary outcome was the incidence of psychiatric symptoms at 3 months after hospital discharge defined as the presence of any of three symptoms: depression, anxiety, or post-traumatic stress disorder (PTSD). Risk ratio (RR) and multiple logistic regression analysis were used. As a sensitivity analysis, two methods for inverse probability of treatment weighting statistics were performed. Of the 192 discharged patients, 99 (52%) were assessed. The patients who achieved EM had a lower incidence of psychiatric symptoms compared to those who did not (25% vs. 51%, p-value 0.008, odds ratio (OR) 0.27, adjusted p = 0.032). The RR for psychiatric symptoms in the EM group was 0.49 [95% Confidence Interval, 0.29–0.83]. Sensitivity analysis accounting for the influence of death, loss to follow-up (OR 0.28, adjusted p = 0.008), or potential confounders (OR 0.49, adjusted p = 0.046) consistently showed a lower incidence of psychiatric symptoms in the EM group. EM was consistently associated with fewer psychiatric symptoms. anxiety early mobilization depression ICU care mental health post-traumatic stress disorder Public Trust Foundation of Marumo ER Medicine & Research Institute of JapanSupported, in part, by grants from the Public Trust Foundation of Marumo ER Medicine & Research Institute of Japan. Morita received support for article research from the Public Trust Foundation of Marumo ER Medicine & Research Institute of Japan. A.K.L. reports personal fees from MERA and receives a salary from TXP Medical completely outside of the submitted work. K.N. reports personal fees from Abbott Laboratory, Nestle, TERUMO, GETINGE, Asahi Kasei Pharma, Ono Pharmaceutical, Japan Blood Products Organization, Nihon Pharmaceutical, Otsuka Pharmaceutical, Pfizer, Toray, and Baxter, and grants from Asahi Kasei Pharma outside of the submitted work. The remaining authors have disclosed that they do not have any potential conflicts of interest. ==== Body pmc1. Introduction Critical illness may result in severe psychiatric disorders, such as depression, anxiety, and post-traumatic stress disorder (PTSD) [1,2], which adversely affect the quality of life and prevent intensive care unit (ICU) survivors from returning to their original lives [3]. Psychiatric symptoms occur in 10–70% of ICU survivors [1,2,4,5] and could last for several months or years after hospital discharge [6,7]. Although research interest is growing as shown by the increasing body of literature, there is no strategy with definitive effects to prevent the development of these symptoms. Immobility induced by physical restriction is associated with the development of psychiatric disorders [8,9], while the potential benefits of regular physical exercise in non-ICU settings to decrease psychiatric symptoms have been described [10,11]. Active physical rehabilitation during the ICU stay, especially when initiated within the first 72 h of ICU stay, is recommended to prevent physical disabilities and improve clinical outcomes of ICU patients [12]. However, the effects of physical rehabilitation early in the ICU stay on psychiatric symptoms are unknown in the existing literature [13]. Therefore, we conducted a multicenter prospective cohort study to investigate the incidence rate of psychiatric symptoms at 3 months after hospital discharge and the association between active physical rehabilitation within 72 h of ICU admission and psychiatric symptoms. We focused on 3 months since a similar trend and incidence rate of psychiatric symptoms at 3 months and 1 year after hospital discharge have been observed [14,15]. 2. Materials and Methods 2.1. Study Design and Patient Selection This multicenter prospective cohort study was approved by the Ethics Committee of Nagoya Medical Center (2018093) and eight other participating hospitals (Hitachi General Hospital, Nagasaki University Hospital, Fukuyama City Hospital, Naha City Hospital, Yuuai Medical Center, Tokushukai General Hospital, Showa University Hospital, Tokyo Women’s Medical University Hospital) and registered in UMIN (ID: 000036503). We followed the STROBE guidelines [16], and all methods in this study were carried out in accordance with relevant guidelines and regulations. Informed consent was obtained from all patients. Consecutive patients, up to 25 patients in each participating hospital, who stayed in the ICU for more than 48 h between June and December in 2019, were eligible for enrollment. Patients less than 18 years of age, unable to walk independently before admission, with neurological complications, lacking communication skills due to pre-existing mental diseases, or in a terminal state were excluded (Table S1). Our study excludes patients with a history of psychiatric disorders (depression, anxiety, PTSD). Patients who died or did not complete the assessment at 3 months follow-up after hospital discharge, were also excluded. In the present study, mobilization was defined as physical rehabilitation at the level of sitting on the edge of the bed or higher [17], and patients were divided into two groups. The patients were allocated into the early mobilization (EM) group when they achieved rehabilitation at the level of mobilization within 72 h of ICU admission [17,18,19]. Patients were allocated into the non-EM group when they did not receive mobilization during the ICU stay or achieved mobilization more than 72 h after ICU admission. All patients equally aimed to receive mobilization daily under a protocol tailored to the circumstances of participating hospitals, though mobilization could not be achieved when the patient did not meet the criteria (Table S2). Patients were not randomized in this study, and whether patients could undergo mobilization depended on the EM protocol used in each participating hospital. 2.2. EM Protocol The early goal-directed protocol for rehabilitation [19,20,21] was developed more than 6 months before this study was initiated, and the details of the contents were arranged based on the situation of participating hospital. This protocol is used in multiple centers in routine practice and safety validation has already been reported [22]. The EM protocol details are shown in Table S2. The protocol provides patients with five rehabilitation levels (level 1, passive range of motion and respiration physical therapy; level 2, active range of motion; level 3, sitting exercise; level 4, standing exercise; and level 5, walking exercise) based on their medical condition. At each participating site, ICU physicians or physiotherapists made the final decision of the rehabilitation level based on the patient’s condition while referring to the protocol. All patients were supposed to receive at least one rehabilitation session per day for 20 min on weekdays. Regarding other ICU care, all participating hospitals followed the 2018 Clinical Practice Guidelines [23] and the clinical practice guideline for the management of ARDS [24]. After transfer out of the ICU, all patients underwent rehabilitation, such as muscle strengthening, balancing, walking, and stair exercises, for more than 20 min on weekdays by physical or occupational therapists according to the rehabilitation policy in the general ward of each hospital without any specific protocol. 2.3. Follow-Up Protocol Table S3 shows the details of the follow-up protocol. At 3 months after hospital discharge, a doctor, a nurse, or physiotherapist at each participating hospital made a phone call to the patient to confirm their survival. Once survival was confirmed, evaluation of the EuroQol-5 Dimensions-5 Levels (EQ-5D-5L) [25] was verbally performed during the same phone call. Then, questionnaires and response sheets for Hospital Anxiety and Depression Scale (HADS) and Impact of Events Scale-Revised (IES-R) were mailed to the patient’ home. The questionnaires were sent back to each participating hospital with the patient’s responses. Despite several phone calls, if the patient did not respond, they were excluded from analysis as lost to follow-up. 2.4. Data Collection Baseline characteristics were collected at the time of ICU admission and during the ICU stay by co-investigators at each hospital, including age, gender, body mass index, Charlson comorbidity index [26], Barthel Index before hospitalization [27], ICU admission diagnosis, Acute Physiology and Chronic Health Evaluation II score, Sequential Organ Failure Assessment score, and use of mechanical ventilation, continuous vasopressors, continuous analgesia, continuous sedation, steroids, neuromuscular blocking agents, dialysis, time to first out of bed mobilization from the time of ICU admission, highest ICU Mobility Scale (IMS) during ICU stay, and the number of daily rehabilitation sessions on the ward. Barthel index before hospitalization was scored at the time of ICU admission based on the information from the family or the patients if they were conscious. The average sedation level, described according to the Richmond Agitation-Sedation Scale (RASS), from days 1 to 3 was calculated based on data in the electronic medical record. The IMS provides a sensitive 11-point ordinal scale, ranging from nothing (lying/passive exercises in bed, score of 0) to independent ambulation (score of 10) [28]. 2.5. Study Outcomes The primary outcome was the incidence of psychiatric symptoms at 3 months after hospital discharge, which was defined as the presence of at least one of three symptoms: depression, anxiety, or PTSD. Depression and anxiety were assessed using the HADS that contains 14 items, seven for anxiety assessment and seven for depression, with a score of 0–3 for each item. Within a maximum score of 21 for each subset for depression or anxiety assessment, the presence of depression or anxiety was defined as a score of 8 or more [29]. PTSD was evaluated using the IES-R, a 22-item self-reported measure with scores ranging between 0 and 88 points (score 0–4 per item). The presence of PTSD was defined as a score of 25 or more [30]. Secondary outcomes were the scores on HADS subsets for depression or anxiety and the IES-R score for PTSD at 3 months after hospital discharge, as well as the change in each score between hospital discharge and at 3 months. Other variables included the EQ-5D-5L which is a standardized assessment for HRQoL [25] at 3 months follow-up and at hospital discharge, walking independence at discharge [21], duration of mechanical ventilation, length of ICU and hospital stays, Barthel Index at hospital discharge, incidence of delirium during ICU stay, incidence of ICU-acquired weakness (ICU-AW) at ICU discharge. Patients who could walk 45 m or more with or without braces were regarded as walking independent [19]. For the assessment of delirium, either Confusion Assessment Method for the Intensive Care Unit [31] or Intensive Care Delirium Screening Checklist [32] was used as delirium screening tool. ICU Acquired Weakness is defined that Medical Research Council-sum score evaluating by physical therapists is less than 48 at the time of ICU discharge [33]. 2.6. Statistical Analysis A sample size of 240 patients is needed with 80% power and a two-sided significance level of 0.05, under the assumption of a follow-up achievement rate of 80%, 60% of non-EM patients will develop psychiatric symptoms and EM patients will have a 20% reduction based on previous studies [14,15,34]. Data are presented as a median with interquartile range or as a number with percentage. The Mann–Whitney U test was used to analyze continuous variables and the χ2 test or Fisher’s exact test for nominal variables, as appropriate. Before using a non-parametric test, the distribution of each parameter was evaluated with the Shapiro–Wilk test. In addition to the comparison of baseline characteristics between the two groups, given the influence of death and loss to follow-up, we also compared them between patients discharged from the ICU and those who completed follow-up at 3 months. The same analysis was conducted among patients discharged from ICUs and those who completed follow-up both in the EM and non-EM groups. Multiple logistic regression analysis was performed to identify an association of the primary outcome with the following covariates: Age, male gender, Barthel index before hospitalization, ICU admission diagnosis, APACHE II score, and use of mechanical ventilation, continuous analgesia, continuous sedation, steroids, and neuromuscular blocking agents, and dialysis, which were considered as factors related to the primary outcome in previous reports (Table S4) [35,36,37,38,39,40,41,42,43,44,45,46,47]. To analyze secondary and other outcomes, multiple linear and logistic regression analyses were performed for log-transformed continuous and categorical variables, respectively, using the same covariates as used in analyzing the primary outcome. Each relative risk ratio (RR) for the incidence of depression, anxiety, or PTSD in the EM group against the non-EM group was described. Pearson’s correlation coefficient was used to assess the correlation between the number of days from ICU admission to first mobilization and the HADS subset scores for depression or anxiety or the IES-R score for PTSD. As a post hoc analysis, two methods for inverse probability of treatment weighting (IPTW) statistics were used. The first adjusted the influence of death and follow-up loss and the second adjusted as many confounding factors from baseline characteristics such as severity, baseline comorbidity, ICU admission diagnosis, consciousness level, which could affect the outcomes. The methodological details of the analysis are shown in Table S5 [48,49]. All analyses were performed using SPSS version 23.0 (IBM Corp., Armonk, NY, USA). Statistical tests were two-sided, and statistical significance was defined as p-values of <0.05. 3. Results 3.1. Baseline Characteristics Of the 1014 patients screened, 203 were enrolled (Figure 1). Of these, 192 patients were discharged from the ICU and were eligible for follow-up assessment. Excluding patients who died or who missed the 3-month follow-up, a total of 99 patients, including 60 in the EM group and 39 in the non-EM group, completed the 3-month follow-up. There were differences among the measured baseline characteristics between the two groups, for the Acute Physiology and Chronic Health Evaluation II score (17 vs. 21, p = 0.026) and the use of mechanical ventilation (53% vs. 74%, p = 0.034), steroids, neuromuscular blockade (0% vs. 13%, p = 0.008), and dialysis (12% vs. 28%, p = 0.037) (Table 1). The average RASS score within the first 3 days of the ICU stay was not significantly lower in the EM group. Among the rehabilitation items, the EM group had a shorter time from ICU admission to first rehabilitation (1.7 days vs. 5.3 days, p < 0.001), and a higher IMS score during their ICU stay (8 vs. 4, p < 0.001). Considering the bias due to death and loss to follow-up, no differences were observed in all baseline characteristics comparing patients discharged from the ICU and who completed the 3-month follow-up in each comparison (Table S6). 3.2. Primary and Secondary Outcomes The incidence of psychiatric symptoms at the 3-month follow-up point was significantly lower in the EM group than in the non-EM group [odds ratio (OR): 0.27, adjusted p = 0.032] even after adjustment for the baseline characteristics of the two groups, whereas the incidence at the time of hospital discharge was not different (OR: 1.03, adjusted p = 0.965) (Table 2). At 3 months follow-up, the EM group showed a significantly lower incidence of PTSD (OR: 0.06, adjusted p = 0.026) and had a significantly lower HADS subset score for anxiety (adjusted p = 0.004) and IES-R (adjusted p = 0.009) compared with the non-EM group. The EM group demonstrated no significant differences in the incidence of depression, anxiety, and PTSD, the HADS subset scores for depression and anxiety, and the IES-R score at the time of hospital discharge. When comparing the assessment scores at hospital discharge and at 3 months follow-up, only changes in the HADS subset scores for anxiety in the EM group were significantly higher (adjusted p = 0.032). The risk for developing psychiatric symptoms [RR: 0.49, confidence interval (CI): 0.29–0.83, p = 0.010], depression (RR: 0.52, CI: 0.27–0.99, p = 0.006), anxiety (RR: 0.27, CI: 0.10–0.71, p < 0.001), and PTSD (RR: 0.07, CI: 0.01–0.54, p < 0.001) at 3 months follow-up were lower in the EM group, while no significant difference in the risk ratios was observed at the time of hospital discharge (Table 3). The time from ICU admission to first mobilization correlated weak linearly with the HADS subset scores for depression (r = 0.244, p = 0.011) and anxiety (r = 0.350, p < 0.001), and the IES-R score (r = 0.289, p = 0.003) at 3 months follow-up (Figure S1). Multiple linear regression analysis also demonstrated that EM was significantly associated with the HADS subset score for anxiety [β coefficient: 0.311, adjusted p = 0.004] and the IES-R score (β coefficient: 0.278, adjusted p = 0.009) at 3 months follow-up. Changes in the HADS subset scores for anxiety were significantly associated with achieving EM (β coefficient: −0.242, adjusted p = 0.032) (Table S7). 3.3. Other Variables There was no difference between the two groups in the EQ-5D-5L at 3 months follow-up and at hospital discharge. The EM group had a shorter length of ICU and hospital stays, a higher Barthel Index at hospital discharge, and a lower incidence of delirium and ICU-AW. (Table 4). 3.4. Post Hoc Sensitivity Analysis The model created by IPTW considering the influence of death and follow-up loss showed that EM correlated significantly with a lower incidence of psychiatric symptoms (OR: 0.28, CI: 0.11–0.73, adjusted p = 0.008), lower incidence of anxiety and PTSD, and a lower HADS subset score for anxiety and IES-R score at 3 months follow-up (Table S8). In the IPTW analysis using the propensity score with adjustment of 19 baseline variables, EM correlated significantly with a lower incidence of psychiatric symptoms at 3 months (OR: 0.49, CI: 0.14–0.93, adjusted p = 0.046) and a lower incidence of PTSD and a lower HADS subset score for anxiety and IES-R score (Table S9). 4. Discussion The incidence of psychiatric symptoms was significantly lower in the EM group and EM in the ICU was significantly associated with a lower incidence of psychiatric symptoms at 3 months follow-up after hospital discharge even after adjustment for significant differences in the baseline characteristics including severity and treatments received. The results were consistent even if the influence of patient death, loss to follow-up or potential confounding factors were included in the analysis. Furthermore, the interval from ICU admission to first mobilization correlated significantly with the score of HADS subsets for Depression and Anxiety and the score of IES-R for PTSD. Mobilization, rehabilitation at the level of sitting on the edge of the bed or higher was significantly associated with improved psychiatric symptoms. A previous systematic review using mobilization in the ICU as an intervention did not show a significant improvement in psychiatric symptoms after ICU discharge [50]. Of note, most studies in that systematic review did not define the time point for mobilization, despite the potential benefit of mobilization within 72 h of ICU admission (Early Mobilization) which is described in current literature [13,51]. For instance, providing EM improves independent physical function at the time of discharge, shortens delirium duration, and increases ventilator-free days [52], whereas mobilization initiated more than 72 h after ICU admission has no beneficial effect [13,53]. The present study supports the idea that the effect of EM on outcomes related to psychiatric symptoms after ICU discharge as well as other clinical outcomes might be maximized if mobilization is initiated within 72 h of ICU admission. The significant correlation between time from ICU admission to achieve EM and scores for diagnosing psychiatric symptoms support this idea. The interval until initiation of mobilization might be an essential aspect of physical rehabilitation in the ICU to maximize the effect on patient outcomes. The EM group had a lower incidence of psychiatric symptoms at 3 months follow-up, but there was no significant difference between the groups at discharge. It has been reported that there is a time lag for the onset of anxiety and depression [54,55]. Therefore, the time of onset of each psychiatric symptom, such as depression, anxiety, and PTSD, could be different [56]. The time for the effect of EM on outcomes after hospital discharge to appear could vary among the psychiatric symptoms. Comparing the difference at 3 months and at hospital discharge, anxiety was the most improved among the psychiatric symptoms assessed. As previous studies have shown, an exercise period of at least 8 weeks is required to reduce depression, whereas anxiety disorders are improved with exercise in a relatively short period [13,57]. Considering these, the follow-up period might be adjusted depending on outcomes and interventions of the study related to psychiatric symptoms after ICU admission. Several studies suggest an interaction between physical and psychiatric factors and these outcomes could be linked [58,59]. Physical impairment could result in poor psychiatric function and vice versa. For example, ICU-AW could lead to the development of delirium which is a potential risk factor for the development of psychiatric disorders [60]. There is evidence to support the idea that EM could reduce not only physical impairment, such as ICU-AW, but also delirium similar to the results of the present study [61] and might result in a synergistic and beneficial effect on psychiatric function. It has been suggested that exercise or mobilization improves mental health [62] and self-efficacy [63], increase endorphin and monoamine levels [63], decrease cortisol levels [64], and increase brain-derived neurotrophic factor [65], resulting in better psychiatric outcomes as a potential mechanism for mobilization improving psychiatric symptoms. In addition, there were several differences in outcomes at the time of hospital discharge, such as physical function, that could be not only intervening factors but also mediators of psychiatric symptoms after a critical illness. Further studies are warranted to validate the relationship between physical and psychiatric function, and the background mechanisms to capture the interaction between interventions such as mobilization, physical impairment at hospital discharge, and psychiatric symptoms at follow-up. The comparability between the two groups is a primary limitation of this study. The calculated sample size could not be achieved, there was a relatively large influence of death and loss to follow-up and the study was not randomized. These factors may limit generalizability to other ICUs. Second, there are confounding factors which could not be adjusted for, such as educational history and clinical frailty. However, multivariate analysis with important clinical factors, IPTW, and propensity score matching adjusted by potential confounders showed consistent results even though there are several differences in important baseline characteristics of the two groups, for example the use of mechanical ventilation, steroids, neuromuscular blocking agents and dialysis. Third, outcomes were limited to short-term follow-up. Fourth, whether the patient could receive rehabilitation at the level of sitting on the edge of the bed or higher depended on the rehabilitation policy used in each participating hospital. Therefore, whether EM could not be provided due to poor general condition or other factors was not identified. Fifth, the psychiatric symptoms of the primary outcome of this study are diagnosed by scoring and are not based on criteria by psychiatrists or psychiatric liaison teams. Finally, patients who died in the ICU were excluded from the analysis in this study because we targeted ICU survivors. This could result in selection bias and these findings should be interpreted with caution. To further validate these results and investigate causality, a multicenter randomized controlled trial with more patients is needed. 5. Conclusions EM in the ICU is significantly associated with lower rates of psychiatric symptoms, including depression, anxiety, and PTSD, at 3 months follow-up after hospital discharge. The interval from ICU admission to mobilization might be an important parameter to maximize the beneficial effects on patient outcomes. Acknowledgments The authors would like to thank the study coordinators Tetsuo Ikai, Gen Kudo, Masako Shimada, Syouhei Yokota, Naoko Shima, Maiko Mori, Yayoi Honjo, Ohno Mika, Takahiro Kanaya, Yasuko Muranaka, Hiromasa Harada, Masahiro Tamashiro, Shohei Miyazato, Shogo Sakihama, Ryo Nagato, Shuhei Ikeguchi, Yoshiyuki Teranobu, Tsubasa Watanabe, Yuuiichi Miyagi, Hiroyuki Touyama, Moromizato, Ayako Kawasaki, Noriyo Suzuki, Sayaka Hosoi, Takahiro Fujita, Syohei Hachisu, Hidehiko Nakano, and Hiromu Naraba. The authors would also like to thank the entire ICU staff at all the participating hospitals. The authors thank Akiko Kada, a biostatistician at Nagoya Medical Center, for providing assistance in reviewing the manuscript. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/jcm11092587/s1, Table S1. Exclusion criteria; Table S2. Early Mobilization Protocol; Table S3. Details of follow-up; Table S4. The rationale for collected data; Table S5. Details of the post hoc sensitivity analysis; Table S6. Comparison of baseline characteristics of patients considering bias related to loss to follow-up and death; Figure S1. Correlation between the time from ICU admission to first mobilization and each assessment score for psychiatric disorders at 3 months follow-up. HADS = Hospital anxiety and depression scale, IES-R = Impact of event scale-revised. Mobilization is defined as physical rehabilitation at the level of sitting on the edge of the bed or higher; Table S7. Association between early mobilization and psychiatric disorder assessment score; Table S8. Sensitivity analysis using inverse probability of treatment weighted statistics considering bias from death and loss to follow-up; Table S9. Sensitivity analysis using a propensity score with 19 confounding factors. Click here for additional data file. Author Contributions S.W. and K.L. conducted the study design. S.W., A.K.L., K.N., R.K., T.H., K.I., D.Y., Y.T., T.N., Y.M., T.K. (Takahiro Kanaya), S.S., H.K. and T.K (Toru Kotani). participated in creating the protocol and introducing the protocol in our ICU. SW and KL helped in the data collection and the statistical analysis. S.W., K.L., K.N., A.K.L. and T.K. (Toru Kotani). helped in the development of this manuscript, and A.K.L. also checked the English grammar. A.K.L. advised on the statistical methods. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement This research was approved by the ethics committee of each participating hospital and the ethics committee of the Nagoya Medical Center Hospital (institutional review board approval number 2018093, date of approval 10 April 2018). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The dataset generated and analyzed during the current study is not publicly available but is available from the corresponding author on reasonable request. Conflicts of Interest Some authors report potential conflicts of interest outside of the current study: YM received support for article research from the Public Trust Foundation of Marumo ER Medicine & Research Institute of Japan. H.K. receives a salary from the Japanese Society for Early Mobilization (nonprofit society) as a chair (full time). A.K.L. received personal fees from MERA and receives a salary from TXP Medical. T.K. (Toru Kotani). is on the medical advisory board for Toray Medical, Healios, and C-Ventec; and received lecture fees from Paramount bed, Ono Pharmaceutical, and Pfizer. The remaining authors have disclosed that they do not have any potential conflicts of interest. Figure 1 Study flow chart. ICU = intensive care unit. a Neurological complications include cerebral infarction, cerebral hemorrhage, acute subdual hematoma, acute epidural hematoma, traumatic subarachnoid hemorrhage, and encephalitis. b Diseases include depression, anxiety, schizophrenia, dementia, cerebral infarction, cerebral hemorrhage, and alcoholism. c Four out of nine hospitals could not enroll 25 patients and enrolled 14, 21, 22, and 21 patients. jcm-11-02587-t001_Table 1 Table 1 Baseline characteristics. Variables Early Mobilization Group (n = 60) Non-Early Mobilization Group (n = 39) p Value Age (years), median (IQR) 70 (61–75) 73 (57–79) 0.416 Gender (male), n (%) 39 (65) 23 (59) 0.545 BMI (kg/m2), median (IQR) 23 (21–25) 24 (21–27) 0.439 Charlson Comorbidity Index, median (IQR) 2 (1–2) 1 (0–2) 0.591 Barthel index before hospitalization, median (IQR) a 100 (100–100) 100 (100–100) 0.638 ICU admission diagnosis, n (%) 0.652      Acute respiratory failure (including pneumonia) 6 (10) 6 (15)      Cardiovascular disease 30 (50) 18 (46)      Gastric or colonic surgery 10 (17) 4 (10)      Sepsis, non-pulmonary 9 (15) 5 (13)      Other diagnoses 5 (8) 6 (15) APACHE II score, median (IQR) 17 (12–22) 21 (16–26) 0.026 SOFA at ICU admission, median (IQR) 7 (3–8) 7 (4–11) 0.121 The use of mechanical ventilation during ICU stay, n (%) 32 (53) 29 (74) 0.034 The use of continuous vasopressor during ICU stay, n (%) 34 (57) 25 (64) 0.461 The use of continuous analgesia during ICU stay, n (%) 37 (62) 26 (67) 0.613 The use of continuous sedation during ICU stay, n (%) 45 (75) 28 (71) 0.723 The use of steroid during ICU stay, n (%) 7 (12) 14 (35) 0.006 The use of neuromuscular blocking agent during ICU stay, n (%) 0 (0) 5 (13) 0.008 The use of dialysis during ICU stay, n (%) 7 (12) 11 (28) 0.037 Average RASS score during the day shift from ICU day 1 to ICU day 3, median (IQR) b 0 (0–0) 0 (−2–0) 0.070 Time to first out of bed mobilization after ICU admission (days) 1.7 (0.9–2.0) 5.3 (4.0–8.0) <0.001 Highest ICU mobility scale score during ICU stay 8 (6–10) 4 (3–7) <0.001 Number of daily rehabilitations per person on the ward (minute/time) 31 (22–43) 34 (27–40) 0.331 Data are presented as median (interquartile range) or number (%). IQR = interquartile range; BMI = Body mass index; ICU = Intensive Care Unit; APACHE = Acute Physiology and Chronic Health Evaluation; SOFA = Sequential Organ Failure Assessment; RASS = Richmond agitation sedation scale. a Barthel index before hospitalization was scored at the time of ICU admission based on the information from the family or the patients if they were conscious. b In all participating hospitals, RASS score, as a sedation scale, was monitored every 2 h during the day shift by nurses and recorded in the medical record. The best RASS score, which means the recorded number closest to zero during the day, of each day from days 1 to 3 was used to calculate the average value of RASS. jcm-11-02587-t002_Table 2 Table 2 Outcomes: psychiatric disorders in ICU survivors. Outcomes Early Mobilization Group (n = 60) Non-Early Mobilization Group (n = 39) p Value Adjusted b p Value Primary Outcome Follow-up at 3 months after discharge Patients with psychiatric symptoms, n (%) a 15 (25) 20 (51) 0.008 0.032 At the time of hospital discharge Patients with psychiatric symptoms, n (%) a 20 (33) 17 (46) 0.214 0.965 Secondary Outcomes Follow-up at 3 months after discharge   Patients who scored HADS subset for depression ≥8, n (%) 12 (20) 15 (39) 0.044 0.107   HADS subset score for depression, median (IQR) 4 (2–7) 6 (2–9) 0.142 0.223   Patients who scored HADS subset for anxiety ≥8, n (%) 5 (8) 12 (30) 0.004 0.104   HADS subset score for anxiety, median (IQR) 3 (1–5) 6 (4–8) <0.001 0.004   Patients who scored IES-R ≥25, n (%) 1 (2) 9 (23) <0.001 0.026   IES-R score, median (IQR) 4 (1–9) 9 (4–20) <0.001 0.009 At the time of hospital discharge Patients who scored HADS subset for depression ≥8, n (%) 16 (27) 15 (40) 0.155 0.917 HADS subset score for depression, median (IQR) 4 (2–8) 6 (4–9) 0.086 0.471 Patients who scored HADS subset for anxiety ≥8, n (%) 11 (18) 7 (19) 0.943 0.772 HADS subset score for anxiety, median (IQR) 3 (1–6) 4 (2–7) 0.392 0.448 Patients who scored IES-R ≥25, n (%) 2 (3) 5 (14) 0.102 0.266 IES-R score, median (IQR) 6 (2–12) 10 (4–17) 0.052 0.163 Changes between follow-up at 3 months and hospital discharge HADS subset score for depression, median (IQR) 1 (−2–3) 1 (−1–3) 0.928 0.418 HADS subset score for anxiety, median (IQR) 1 (−2–4) −1 (−4–1) 0.006 0.032 IES-R score, median (IQR) 1 (−2–7) −1 (−7–5) 0.053 0.131 Data are presented as number (%) or median (interquartile range). ICU = Intensive Care Unit, HADS = Hospital anxiety and depression scale, IQR = interquartile range, IES-R = Impact of event scale-revised, EQ-5D-5L = EuroQol-5 Dimensions-5 Levels. a Psychiatric symptoms were defined as the presence of at least one of three symptoms; depression, anxiety, and PTSD. b Multiple linear for continuous variable or multiple logistic regression analysis for nominal variables were performed to identify an association of the primary outcome with the following covariates. The covariates in the multi-variates analysis included age, male gender, Barthel index before hospitalization, ICU admission diagnosis (acute respiratory failure, cardiovascular disease, gastric or colonic surgery, sepsis, other), acute physiology and chronic health evaluation II score, use of mechanical ventilation, use of continuous analgesia, use of continuous sedation, use of steroids, use of neuromuscular blocking agents, and use of dialysis. jcm-11-02587-t003_Table 3 Table 3 Association between early mobilization and the presence of psychiatric symptoms. Variables Risk Ratio (95% CI) Unadjusted Odds Ratio (95% CI) Adjusted b Odds Ratio (95% CI) Follow-up at 3 months after discharge Presence of psychiatric Symptoms a 0.49 (0.29–0.83) * 0.32 (0.13–0.74) ** 0.27 (0.08–0.89) * HADS depression score ≥8 0.52 (0.27–0.99) ** 0.40 (0.16–0.98) * 0.37 (0.11–1.24) HADS anxiety score ≥8 0.27 (0.10–0.71) *** 0.20 (0.06–0.61) ** 0.23 (0.06–1.31) IES-R score ≥25 0.07 (0.01–0.54) *** 0.06 (0.01–0.32) *** 0.06 (0.01–0.70) * At hospital discharge Presence of psychiatric Symptoms a 0.73 (0.44–1.20) 0.59 (0.25–1.36) 1.03 (0.33–3.14) HADS depression score ≥8 0.66 (0.37–1.16) 0.53 (0.22–1.27) 1.07 (0.29–3.91) HADS anxiety score ≥8 0.97 (0.41–2.27) 0.96 (0.34–2.87) 1.23 (0.30–5.05) IES-R score ≥25 0.25 (0.05–1.20) 0.22 (0.03–1.09) 0.01 (0.01–52.4) The data are presented as risk ratio or odds ratio with 95% confidence interval. * <0.05, ** <0.01, *** <0.001. HADS = Hospital anxiety and depression scale, IES-R = Impact of event scale-revised, CI = Confidence interval. a Psychiatric symptoms are defined as the presence of at least one of three symptoms; depression, anxiety, and PTSD. b Multiple logistic regression analysis was performed to identify an association of the primary outcome with the following covariates. The covariates in the multi-variates analysis included age, male gender, Barthel index before hospitalization, ICU admission diagnosis (acute respiratory failure, cardiovascular disease, gastric or colonic surgery, sepsis, other), acute physiology and chronic health evaluation II score, use of mechanical ventilation, use of continuous analgesia, use of continuous sedation, use of steroids, use of neuromuscular blocking agents, and use of dialysis. jcm-11-02587-t004_Table 4 Table 4 Comparison of clinical outcomes between early mobilization group and non-early mobilization group. Outcomes Early Mobilization Group (n = 60) Non-Early Mobilization Group (n = 39) p Value Adjusted a p Value a Health-related quality of life The EQ-5D-5L index at the time of 3 month after hospital discharge 0.89 (0.70–0.94) 0.82 (0.70–0.94) 0.235 0.952 The EQ-5D-5L index at the time of hospital discharge 0.81 (0.71–0.89) 0.70 (0.44–0.94) 0.384 0.926 Clinical outcomes and physical assessment The number of patients who can walk independently at the time of hospital discharge 58 (97) 33 (85) 0.032 0.136 Duration of mechanical ventilation (days) 1.4 (1–2) 4.4 (2–8.7) <0.001 0.326 ICU length of stay (days) 4 (3–5) 7 (5–13) <0.001 <0.001 Hospital length of stay (days) 22 (19–29) 32 (22–51) 0.004 0.009 Barthel index at hospital discharge 100 (90–100) 90 (70–100) 0.022 0.020 The number of patients who diagnosed as delirium during ICU stay 11 (18) 16 (41) 0.013 0.001 The number of patients who diagnosed as ICU acquired weakness at the time of ICU discharge 2 (3) 8 (21) 0.006 0.004 Data are presented as median (interquartile range) or number (%). IQR = interquartile range; ICU = Intensive Care Unit; IMS = ICU mobility scale. a Multiple linear for continuous variable or multiple logistic regression analysis for nominal variable were performed to identify an association of the primary outcome with the following covariates. The covariates in the multi-variates analysis included age, male gender, Barthel index before hospitalization, ICU admission diagnosis (acute respiratory failure, cardiovascular disease, gastric or colonic surgery, sepsis, other), acute physiology and chronic health evaluation II score, use of mechanical ventilation, use of continuous analgesia, use of continuous sedation, use of steroids, use of neuromuscular blocking agents and use of dialysis. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Kawakami D. Fujitani S. Morimoto T. Dote H. Takita M. Takaba A. Hino M. Nakamura M. Irie H. Adachi T. Prevalence of post-intensive care syndrome among Japanese intensive care unit patients: A prospective, multicenter, observational J-PICS study Crit. Care 2021 25 69 10.1186/s13054-021-03501-z 33593406 2. Jackson J.C. Pandharipande P.P. Girard T.D. Brummel N.E. Thompson J.L. Hughes C.G. Pun B.T. Vasilevskis E.E. Morandi A. Shintani A.K. Depression, post-traumatic stress disorder, and functional disability in survivors of critical illness in the BRAIN-ICU study: A longitudinal cohort study Lancet Respir. Med. 2014 2 369 379 10.1016/S2213-2600(14)70051-7 24815803 3. Needham D.M. Davidson J. Cohen H. Hopkins R.O. Weinert C. Wunsch H. Zawistowski C. Bemis-Dougherty A. Berney S.C. Bienvenu O.J. Improving long-term outcomes after discharge from intensive care unit: Report from a stakeholders’ conference Crit. Care Med. 2012 40 502 509 10.1097/CCM.0b013e318232da75 21946660 4. Harvey M.A. The truth about consequences--post-intensive care syndrome in intensive care unit survivors and their families Crit. Care Med. 2012 40 2506 2567 10.1097/CCM.0b013e318258e943 22809925 5. Myers E.A. Smith D.A. Allen S.R. Kaplan L.J. Post-ICU syndrome: Rescuing the undiagnosed JAAPA 2016 29 34 37 10.1097/01.JAA.0000481401.21841.32 27023654 6. Parker A.M. Sricharoenchai T. Raparla S. Schneck K.W. Bienvenu O.J. Needham D.M. Posttraumatic Stress Disorder in Critical Illness Survivors: A Metaanalysis Crit. Care Med. 2015 43 1121 1129 10.1097/CCM.0000000000000882 25654178 7. Huang M. Parker A.M. Bienvenu O.J. Dinglas V.D. Colantuoni E. Hopkins R.O. Needham D.M. National Institutes of Health, National Heart, Lung, and Blood Institute Acute Respiratory Distress Syndrome Network Psychiatric Symptoms in Acute Respiratory Distress Syndrome Survivors: A One-Year National Multi-Center Study Crit. Care Med. 2016 44 954 965 10.1097/CCM.0000000000001621 26807686 8. Styf J.R. Hutchinson K. Carlsson S.G. Hargens A.R. Depression, mood state, and back pain during microgravity simulated by bed rest Psychosom. Med. 2001 63 862 864 10.1097/00006842-200111000-00002 11719622 9. Ishizaki Y. Ishizaki T. Fukuoka H. Kim C.S. Fujita M. Maegawa Y. Fujioka H. Katsura T. Suzuki Y. Gunji A. Changes in mood status and neurotic levels during a 20-day bed rest Acta Astronaut. 2002 50 453 459 10.1016/S0094-5765(01)00189-8 11924678 10. Bridle C. Spanjers K. Patel S. Atherton N.M. Lamb S.E. Effect of exercise on depression severity in older people: Systematic review and meta-analysis of randomised controlled trials Br. J. Psychiatry 2012 201 180 185 10.1192/bjp.bp.111.095174 22945926 11. Cooney G.M. Kerry D. Carolyn A.G. Debbie A.L. Jane R. Fiona R.W. Marion M. Gillian E.M. Exercise for depression Cochrane Database Syst. Rev. 2013 12 CD004366 10.1002/14651858.CD004366.pub6 12. Nishida O. Ogura H. Egi M. Fujishima S. Hayashi Y. Iba T. Imaizumi H. Inoue S. Kakihana Y. Kotani J. The Japanese Clinical Practice Guidelines for Management of Sepsis and Septic Shock 2016 Acute Med. Surg. 2018 5 3 89 10.1002/ams2.322 29445505 13. Fuest K. Schaller S.J. Recent evidence on early mobilization in critical-ill patients Curr. Opin. Anaesthesiol. 2018 31 144 150 10.1097/ACO.0000000000000568 29351145 14. Hatch R. Young D. Barber V. Griffiths J. Harrison D.A. Watkinson P. Anxiety, Depression and Post Traumatic Stress Disorder after critical illness: A UK-wide prospective cohort study Crit. Care 2018 22 310 10.1186/s13054-018-2223-6 30466485 15. Marra A. Pandharipande P.P. Girard T.D. Patel M.B. Hughes C.G. Jackson J.C. Thompson J.L. Chandrasekhar R. Ely E.W. Brummel N.E. Co-Occurrence of Post-Intensive Care Syndrome Problems Among 406 Survivors of Critical Illness Crit. Care Med. 2018 46 1393 1401 10.1097/CCM.0000000000003218 29787415 16. Vandenbroucke J.P. von Elm E. Altman D.G. Gøtzsche P.C. Mulrow C.D. Pocock S.J. Poole C. Schlesselman J.J. Egger M. STROBE Initiative Strengthening the reporting of observational studies in epidemiology (STROBE): Explanation and elaboration Epidemiology 2007 18 805 835 10.1097/EDE.0b013e3181577511 18049195 17. Harrold M.E. Salisbury L.G. Webb S.A. Australia and Scotland ICU Physiotherapy Collaboration. Early mobilisation in intensive care units in Australia and Scotland: A prospective, observational cohort study examining mobilisation practises and barriers Crit. Care 2015 19 336 10.1186/s13054-015-1033-3 26370550 18. Kuhn K.F. Schaller S.J. Comment on Early versus delayed mobilization for in-hospital mortality and health-related quality of life among critically ill patients: A systematic review and meta-analysis J. Intensive Care 2020 8 21 10.1186/s40560-020-0436-7 32190331 19. Watanabe S. Liu K. Morita Y. Kanaya T. Naito Y. Arakawa R. Suzuki S. Katsukawa H. Lefor A.K. Hasegawa Y. Changes in barriers to implementing early mobilization in the intensive care unit: A single center retrospective cohort study Nagoya J. Med. Sci. 2021 83 443 464 10.18999/nagjms.83.3.443 34552282 20. Liu K. Ogura T. Takahashi K. Nakamura M. Ohtake H. Fujiduka K. Abe E. Oosaki H. Miyazaki D. Suzuki H. A Progressive Early Mobilization Program Is Significantly Associated With Clinical and Economic Improvement: A Single-Center Quality Comparison Study Crit. Care Med. 2019 47 e744 e752 10.1097/CCM.0000000000003850 31162197 21. Watanabe S. Kotani T. Taito S. Ota K. Ishii K. Ono M. Katsukawa H. Kozu R. Morita Y. Arakawa R. Determinants of gait independence after mechanical ventilation in the intensive care unit: A Japanese multicenter retrospective exploratory cohort study J. Intensive Care 2019 7 53 10.1186/s40560-019-0404-2 31798888 22. Katsukawa H. Ota K. Liu K. Morita Y. Watanabe S. Sato K. Ishii K. Yasumura D. Takahashi Y. Tani T. Risk Factors of Patient-Related Safety Events during Active Mobilization for Intubated Patients in Intensive Care Units-A Multi-Center Retrospective Observational Study J. Clin. Med. 2021 10 2607 10.3390/jcm10122607 34199207 23. Devlin J.W. Skrobik Y. Gélinas C. Needham D.M. Slooter A. Pandharipande P.P. Watson P.L. Weinhouse G.L. Nunnally M.E. Rochwerg B. Clinical Practice Guidelines for the Prevention and Management of Pain, Agitation/Sedation, Delirium, Immobility, and Sleep Disruption in Adult Patients in the ICU Crit. Care Med. 2018 46 e825 e873 10.1097/CCM.0000000000003299 30113379 24. Hashimoto S. Sanui M. Egi M. Ohshimo S. Shiotsuka J. Seo R. Tanaka R. Tanaka Y. Norisue Y. Hayashi Y. ARDS clinical practice guideline committee from the Japanese Society of Respiratory Care Medicine and the Japanese Society of Intensive Care Medicine. The clinical practice guideline for the management of ARDS in Japan J. Intensive Care 2017 5 50 10.1186/s40560-017-0222-3 28770093 25. Shiroiwa T. Ikeda S. Noto S. Igarashi A. Fukuda T. Saito S. Shimozuma K. Shiroiwa T. Comparison of Value Set Based on DCE and/or TTO Data: Scoring for EQ-5D-5L Health States in Japan Value Health 2016 19 648 654 10.1016/j.jval.2016.03.1834 27565282 26. Charlson M.E. Pompei P. Ales K.L. Mackenzie C.R. A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation J. Chronic. Dis. 1987 40 373 383 10.1016/0021-9681(87)90171-8 3558716 27. Katz P.P. Measures of Adult General Functional Status: The Barthel Index, Katz Index of Activities of Daily Living, Health Assessment Questionnaire (HAQ), MACTAR Patient Preference Disability Questionnaire, and Modified Health Assessment Questionnaire (MHAQ) Arthritis Care Res. 2003 49 15 27 10.1002/art.11415 28. Hodgson C. Needham D. Haines K. Bailey M. Ward A. Harrold M. Young P. Zanni J. Buhr H. Higgins A. Feasibility and inter-rater reliability of the ICU Mobility Scale Heart Lung. 2014 43 19 24 10.1016/j.hrtlng.2013.11.003 24373338 29. Zigmond A.S. Snaith R.P. The hospital anxiety and depression scale Acta Psychiatr. Scand. 1983 67 361 370 10.1111/j.1600-0447.1983.tb09716.x 6880820 30. Asukai N. Kato H. Kawamura N. Kim Y. Yamamoto K. Kishimoto J. Miyake Y. Nishizono-Maher A. Reliability and validity of the Japanese-language version of the Impact of Event Scale-Revised (IES-R-J): Four studies on different traumatic events J. Nerv. Ment. Dis. 2002 190 175 182 10.1097/00005053-200203000-00006 11923652 31. Ely E.W. Margolin R. Francis J. May L. Truman B. Dittus R. Speroff T. Gautam S. Bernard G.R. Inouye S.K. Evaluation of delirium in critically ill patients: Validation of the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) Crit. Care Med. 2001 29 1370 1379 10.1097/00003246-200107000-00012 11445689 32. Bergeron N. Dubois M.J. Dumont M. Dial S. Skrobik Y. Intensive Care Delirium Screening Checklist: Evaluation of a new screening tool Intensive Care Med. 2001 27 859 864 10.1007/s001340100909 11430542 33. Patel B.K. Pohlman A.S. Hall J.B. Kress J.P. Impact of early mobilization on glycemic control and ICU-acquired weakness in critically ill patients who are mechanically ventilated Chest. 2014 146 583 589 10.1378/chest.13-2046 25180722 34. Choo C.C. Chew P. Lai S.M. Soo S.C. Ho C.S. Ho R.C. Wong R.C. Effect of Cardiac Rehabilitation on Quality of Life, Depression and Anxiety in Asian Patients Int. J. Environ. Res. Public Health 2018 15 1095 10.3390/ijerph15061095 35. Mikkelsen M.E. Christie J.D. Lanken P.N. Biester R.C. Thompson B.T. Bellamy S.L. Localio A.R. Demissie E. Hopkins R.O. Angus D.C. The adult respiratory distress syndrome cognitive outcomes study. Long-term neuropsychological function in survivors of acute lung injury Am. J. Respir Crit Care Med. 2012 185 1307 1315 10.1164/rccm.201111-2025OC 22492988 36. Shima N. Miyamoto K. Shibata M. Nakashima T. Kaneko M. Shibata N. Shima Y. Kato S. W-PICS investigators Activities of daily living status and psychiatric symptoms after discharge from an intensive care unit: A single-center 12-month longitudinal prospective study Acute Med. Surg. 2020 7 e557 10.1002/ams2.557 32995017 37. LaBuzetta J.N. Rosand J. Vranceanu A.M. Review: Post-Intensive Care Syndrome: Unique Challenges in the Neurointensive Care Unit Neurocrit. Care 2019 311 534 545 10.1007/s12028-019-00826-0 38. Inoue S. Hatakeyama J. Kondo Y. Hifumi T. Sakuramoto H. Kawasaki T. Taito S. Nakamura K. Unoki T. Kawai Y. Post-intensive care syndrome: Its pathophysiology, prevention, and future directions Acute Med. Surg. 2019 6 233 246 10.1002/ams2.415 31304024 39. Rabiee A. Nikayin S. Hashem M.D. Huang M. Dinglas V.D. Bienvenu O.J. Turnbull A.E. Needham D.M. Depressive Symptoms After Critical Illness: A Systematic Review and Meta-Analysis Crit. Care Med. 2016 44 1744 1753 10.1097/CCM.0000000000001811 27153046 40. Easton K. Coventry P. Lovell K. Carter L.A. Deaton C. Prevalence and Measurement of Anxiety in Samples of Patients with Heart Failure: Meta-analysis J. Cardiovasc. Nurs. 2016 31 367 379 10.1097/JCN.0000000000000265 25930162 41. Kondo I. Arai Y. Hamada A. Yamada K. Shioji S. Sakamoto E. Katagiri D. Tada M. Hinoshita F. A Case of Frequently Relapsing Minimal-Change Nephrotic Syndrome with Steroid-Induced Psychiatric Syndrome Treated by Low-Dose, Short-Term Steroid Therapy in Combination with Cyclosporine Case Rep. Nephrol. Dial. 2020 10 1 8 10.1159/000505313 32002410 42. Hedayati S.S. Finkelstein F.O. Epidemiology, diagnosis, and management of depression in patients with CKD Am. J. Kidney Dis. Off. J. Natl. Kidney Found. 2009 54 741 752 10.1053/j.ajkd.2009.05.003 19592143 43. Narimatsu E. Nakayama Y. Sumita S. Iwasaki H. Fujimura N. Satoh K. Namiki A. Sepsis attenuates the intensity of the neuromuscular blocking effect of d-tubocurarine and the antagonistic actions of neostigmine and edrophonium accompanying depression of muscle contractility of the diaphragm Acta Anaesthesiol. Scand. 1999 43 196 201 10.1034/j.1399-6576.1999.430213.x 10027028 44. Nikayin S. Rabiee A. Hashem M.D. Huang M. Bienvenu O.J. Turnbull A.E. Needham D.M. Anxiety symptoms in survivors of critical illness: A systematic review and meta-analysis Gen. Hosp. Psychiatry 2016 43 23 29 10.1016/j.genhosppsych.2016.08.005 27796253 45. Caiuby A.V. Andreoli P.B. Andreoli S.B. Post-traumatic stress disorder in intensive care unit patients. Transtorno de estresse pós-traumático em pacientes de unidade de terapia intensiva Revista Brasileira de Terapia Intensiva 2010 22 77 84 10.1590/S0103-507X2010000100013 25303703 46. Davydow D.S. Gifford J.M. Desai S.V. Needham D.M. Bienvenu O.J. Posttraumatic stress disorder in general intensive care unit survivors: A systematic review Gen. Hosp. Psychiatry 2008 30 421 434 10.1016/j.genhosppsych.2008.05.006 18774425 47. Tang F. Tan J. Guo X. Huang J. Yi J. Wang L. Risk factors for post-traumatic stress disorder in acute trauma patients: A protocol for systematic review and meta-analysis Medicine 2021 100 e25616 10.1097/MD.0000000000025616 33907115 48. Fukunaga S. Nagami Y. Shiba M. Ominami M. Tanigawa T. Yamagami H. Tanaka H. Muguruma K. Watanabe T. Tominaga K. Long-term prognosis of expanded-indication differentiated-type early gastric cancer treated with endoscopic submucosal dissection or surgery using propensity score analysis Gastrointest. Endosc. 2017 85 143 152 10.1016/j.gie.2016.06.049 27365265 49. Nagami Y. Tominaga K. Machida H. Nakatani M. Kameda N. Sugimori S. Okazaki H. Tanigawa T. Yamagami H. Kubo N. Usefulness of non-magnifying narrow-band imaging in screening of early esophageal squamous cell carcinoma: A prospective comparative study using propensity score matching Am. J. Gastroenterol. 2014 109 845 854 10.1038/ajg.2014.94 24751580 50. Fuke R. Hifumi T. Kondo Y. Hatakeyama J. Takei T. Yamakawa K. Inoue S. Nishida O. Early rehabilitation to prevent postintensive care syndrome in patients with critical illness: A systematic review and meta-analysis BMJ Open 2018 8 e019998 10.1136/bmjopen-2017-019998 51. Bein T. Bischoff M. Brückner U. Gebhardt K. Henzler D. Hermes C. Lewandowski K. Max M. Nothacker M. Staudinger T. S2e guideline: Positioning and early mobilisation in prophylaxis or therapy of pulmonary disorders: Revision 2015: S2e guideline of the German Society of Anaesthesiology and Intensive Care Medicine (DGAI) Anaesthesist 2015 64 1 26 10.1007/s00101-015-0071-1 26335630 52. Schweickert W.D. Pohlman M.C. Pohlman A.S. Nigos C. Pawlik A.J. Esbrook C.L. Spears L. Miller M. Franczyk M. Deprizio D. Early physical and occupational therapy in mechanically ventilated, critically ill patients: A randomised controlled trial Lancet 2009 373 1874 1882 10.1016/S0140-6736(09)60658-9 19446324 53. Wright S.E. Thomas K. Watson G. Baker C. Bryant A. Chadwick T.J. Shen J. Wood R. Wilkinson J. Mansfield L. Intensive versus standard physical rehabilitation therapy in the critically ill (EPICC): A multicentre, parallel-group, randomised controlled trial Thorax 2018 73 213 221 10.1136/thoraxjnl-2016-209858 28780504 54. Bienvenu O.J. Colantuoni E. Mendez-Tellez P.A. Shanholtz C. Dennison-Himmelfarb C.R. Pronovost P.J. Needham D.M. Cooccurrence of and remission from general anxiety, depression, and posttraumatic stress disorder symptoms after acute lung injury: A 2-year longitudinal study Crit. Care Med. 2015 43 642 653 10.1097/CCM.0000000000000752 25513784 55. Brück E. Larsson J.W. Lasselin J. Bottai M. Hirvikoski T. Sundman E. Eberhardson M. Sackey P. Olofsson P.S. Lack of clinically relevant correlation between subjective and objective cognitive function in ICU survivors: A prospective 12-month follow-up study Crit. Care 2019 23 253 10.1186/s13054-019-2527-1 31300016 56. Taylor A.H. Physical Activity, Stress and Anxiety: A Review Physical Activity and Psychological Well-Being Biddle S.J.H. Fox K. Boutcher S. Routledge London, UK 2000 10 45 57. Penninx B.W. Rejeski W.J. Pandya J. Miller M.E. Di Bari M. Applegate W.B. Pahor M. Exercise and depressive symptoms: A comparison of aerobic and resistance exercise effects on emotional and physical function in older persons with high and low depressive symptomatology J. Gerontol. B Psychol. Sci. Soc. Sci. 2002 57 124 132 10.1093/geronb/57.2.P124 58. Zhao J.L. Jiang W.T. Wang X. Cai Z.D. Liu Z.H. Liu G.R. Exercise, brain plasticity, and depression CNS Neurosci. Ther. 2020 26 885 895 10.1111/cns.13385 32491278 59. Cahn B.R. Goodman M.S. Peterson C.T. Maturi R. Mills P.J. Yoga, Meditation and Mind-Body Health: Increased BDNF, Cortisol Awakening Response, and Altered Inflammatory Marker Expression after a 3-Month Yoga and Meditation Retreat Front. Hum. Neurosci. 2017 26 315 10.3389/fnhum.2017.00315 60. Vasilevskis E.E. Ely E.W. Speroff T. Pun B.T. Boehm L. Dittus R.S. Reducing iatrogenic risks: ICU-acquired delirium and weakness--crossing the quality chasm Chest 2010 138 1224 1233 10.1378/chest.10-0466 21051398 61. Girard T.D. Shintani A.K. Jackson J.C. Gordon S.M. Pun B.T. Henderson M.S. Dittus R.S. Bernard G.R. Ely E.W. Risk factors for post-traumatic stress disorder symptoms following critical illness requiring mechanical ventilation: A prospective cohort study Crit. Care 2007 11 R28 10.1186/cc5708 17316452 62. Geense W.W. van den Boogaard M. van der Hoeven J.G. Vermeulen H. Hannink G. Zegers M. Nonpharmacologic Interventions to Prevent or Mitigate Adverse Long-Term Outcomes Among ICU Survivors: A Systematic Review and Meta-Analysis Crit. Care Med. 2019 47 1607 1618 10.1097/CCM.0000000000003974 31464769 63. Craft L.L. Exercise and clinical depression: Examining two psychological mechanisms Psychol. Sport Exerc. 2005 6 151 171 10.1016/j.psychsport.2003.11.003 64. Chen M.J. The neurobiology of depression and physical exercise Handbook of Physical Activity and Mental Health Routledge London, UK 2013 169 184 65. Cotman C.W. Berchtold N.C. Exercise: A behavioural intervention to enhance brain health and plasticity Trends Neurosci. 2002 25 295 301 10.1016/S0166-2236(02)02143-4 12086747

PMC009xxxxxx/PMC9099643.txt

==== Front Materials (Basel) Materials (Basel) materials Materials 1996-1944 MDPI 10.3390/ma15093375 materials-15-03375 Article The Use of Barley Malt as a Binder in Molding Sand Technology https://orcid.org/0000-0001-8594-7583 Samociuk Bartłomiej 1 Medyński Daniel 2* https://orcid.org/0000-0001-8980-1303 Nowak Daniel 1 Kawa-Rygielska Joanna 3 https://orcid.org/0000-0002-9817-6324 Świechowski Kacper 4 https://orcid.org/0000-0002-4614-7177 Gasiński Alan 3 Janus Andrzej 2 Xu Chao Academic Editor 1 Department of Light Element Engineering, Foundry and Automation, Wroclaw University of Science and Technology, Wybrzeże Wyspiańskiego 27, 50-370 Wroclaw, Poland; bartlomiej.samociuk@pwr.edu.pl (B.S.); daniel.nowak@pwr.edu.pl (D.N.) 2 Faculty of Technical and Economic Sciences, Witelon Collegium State University, Sejmowa 5A, 59-220 Legnica, Poland; andrzej.janus@pwr.edu.pl 3 Department of Fermentation and Cereals Technology, Wroclaw University of Environmental and Life Sciences, pl. Grunwaldzki 24A, 50-363 Wroclaw, Poland; joanna.kawa-rygielska@upwr.edu.pl (J.K.-R.); alan.gasinski@upwr.edu.pl (A.G.) 4 Department of Applied Bioeconomy, Wroclaw University of Environmental and Life Sciences, Chełmońskiego 37A, 51-630 Wrocław, Poland; kacper.swiechowski@upwr.edu.pl * Correspondence: daniel.medynski@collegiumwitelona.pl 08 5 2022 5 2022 15 9 337521 3 2022 06 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). The aim of this study was to attempt to use barley malt as a natural, organic binder in the technology of molding sand. TGA analysis of the binder was performed, during which temperatures of thermal decomposition of its components were determined. The results of TG/DTG analysis show that a loss of ~75% of mass of the MB binder is organic matter. Over 50% of this is starch. The results indicate the possibility of using a binder made of barley malt as a binding material for quartz sand grains. This fact was confirmed by tests carried out with use of SEM. During the observations, it was found that barley malt forms smooth bridges connecting individual grains of quartz sand. The typical properties of molding sands with barley malt were also determined, compared to sands containing commonly used binders. At the same time, the influence of the content of this binder on flowability, permeability, strength properties, and wear resistance was assessed. It has been found that increasing the binder content in molding mass results in an increase in strength and wear resistance, as opposed to flowability and permeability. Test castings were also made. It was found that the addition of a binder made of barley malt has a positive effect on the surface quality of castings. This was confirmed by roughness measurements of the test castings. At the same time, a tendency to excessive gas evolution during pouring was shown, with higher contents of this binder. Moreover, greater amounts of barley malt in the molding sand (MB 5%) as compared to the lower content (MB 2%) increased the thickness of the burnt layer of the sand by 25%. This is due to the exothermic reaction when more binder is burnt. It is extremely important from the point of view of the regeneration of molding sand. casting alloys molding sands casting binders quality of castings Witelon Collegium State UniversityDB/GW/2022 The work was financed by internal grant of Witelon Collegium State University DB/GW/2022. ==== Body pmc1. Introduction The foundry industry has been looking for new solutions aimed at improving the quality of finished products for a long time. The introduction of new technologies and materials was usually associated with an increased risk of factors harmful to people and the environment. Employees employed directly in the foundry process have for many years belonged to the group with an increased occupational risk. Molding sands used during the pouring of molds with liquid metal are a source of emission of harmful, toxic, and carcinogenic substances (benzene, PAH). This is mainly caused by use of synthetic binders, such as, e.g., phenol-formaldehyde, urea-formaldehyde, furan, and furfuryl resins [1,2]. However, for several decades, the guidelines of international bodies have forced manufacturers to seek solutions that are friendly to humans and the environment. The new guidelines for the foundry industry limit the use of harmful, toxic materials that causethe emission of hazardous gases, noise, and vibrations, as well as those that cannot be reused after the recycling process [3,4,5]. As a result, these activities led to the emergence of research on the use of new or previously unused molding materials. In the last decade, research has appeared on the possibility of replacing synthetic binders with materials of plant origin [6,7,8]. Literature reports indicate the possibility of using materials such as wood rosin, molasses, oils, dextrin, starch, cellulose, and natural latex [9,10,11,12]. The presented materials may be an intrinsic binder shaping the properties of molding sand or a special additive influencing selected properties of the sand. This is carried out in such a way as to meet the requirements of the foundry and to meet the new environmental protection and health and safety regulations [3]. When comparing masses based on barley malt binders with other masses made with organic binders of vegetable origin, it can be seen that the mass with malt has better or comparable permeability. Depending on the binder content, its permeability varies between 317–240 × 10−8 m2/Pa∙s. On the other hand, the mass with a cassava binder (6%) obtained the result of 126 × 10−8 m2/Pa∙s, and the mass with rice starch 122–341 × 10−8 m2/Pa∙s and corn starch 156–350 × 10−8 m2/Pa∙s [6,13]. On the other hand, the compressive strength is definitely better, as it is above 5 MPa, compared to the mass with addition of: cassava—0.51 MPa; rice—0.065 MPa; corn—0.050 MPa [6,13]. Other parameters were not compared due to the lack of data. Therefore, authors of the work undertook research attempts aimed at assessing the possibility of using barley malt as a binder for molding and core sand, in relation to typical sand. The fact that malt is an organic material, derived from renewable sources and with a low environmental impact, is why it fits in the current trend of sustainable development. 2. Materials and Methods 2.1. Determination of Starch Content in Barley Malt For the research, malt in form of dried and finely ground malted barley flour was used, similarly to the use of other organic materials used as a binder in the technology of molding sand [14,15]. From a chemical point of view, malt consists of carbohydrates (especially starch, sugars, and dextrins), protein compounds (amylolytic, proteolytic, and cytolytic enzymes) and small amounts of fatty compounds [16]. The determination of starch in barley malt as a binder for molding sand was performed with the polarimetric method using an AA-55 polarimeter (Optical Activity Ltd., Huntingdon, UK). This method consists in dissolving the malt in diluted hydrochloric acid and, after clarifying the solution, measuring the twist of the plane of polarized light. Starch content was calculated from the Biot formula, assuming that the specific rotation of starch dissolved in HCl is 183.7°. Marking was made in accordance with PN-EN-ISO 10520: 2002—Native starch—Determination of starch content—Ewers polarimetric method [17,18]. 2.2. Determination of Protein Content in Barley Malt The total amount of protein in barley malt was determined using the Kjeldahl method using the Kjeltec 8100 analyzer (Foss, Copenhagen, Denmark) [19,20]. This method consists in the mineralization of the sample in a medium of concentrated sulfuric acid (VI) in the presence of catalysts. Protein nitrogen was converted under these conditions to the ammonium ion which, after alkalinization, was distilled in the form of ammonia. The ammonia was determined by acid-base titration with 0.01 MHCl. The standard PN-EN-ISO 20483: 2014-02 was used for marking—Cereals and pulses—Determination of the nitrogen content and calculation of the crude protein content—Kjeldahl method. 2.3. Thermogravimetric Analysis of Barley Malt Thermogravimetric analysis (TGA or TG) is a method on the basis of which it is possible to assess the temperature at which the test sample decomposes, the loss of binding properties, and burning. The authors proved that thermal analysis can be used in research on casting materials—in this case, the binder of barley malt [21,22,23,24,25,26]. Tests were carried out using the RST 40 × 200/100 tube furnace (Czylok, Jastrzębie-Zdrój, Poland) connected to the PS 750.3Y analytical balance (Radwag, Radom, Poland). Samples weighing approx. 2 g were heated to 850 °C at three different rates: 12.5 °C/min, 25 °C/min, 50 °C/min. Changes in samples masses were recorded at one-second intervals with an accuracy of 0.001 g. The obtained TGA curves were smoothed (with a scatter parameter of 0.1) by the local estimated scatter plot smoothing (LOESS) method using the OriginPro 2019b software (OriginLab, Northampton, MA, USA). Then, on the basis of the smoothed TG plot, the DTG derivative was calculated dTG/dT, where dTG (%) is a change in sample mass and dT (°C) a change in its temperature. In this way, the change in mass of the sample was determined, with the change in its temperature by 1 °C. 2.4. Preparation of Molding Sand Molding sands were prepared according to the instructions in the literature [3]. Preparation of the masses began with mixing dry ingredients together. Dry mixing time was one minute. For this purpose, the LM-1 laboratory roller mixer (Multiserw-Morek, Marcyporęba, Poland) was used. Then distilled water was added and all the ingredients were mixed for another 3 min. For all prepared mixes, the ambient conditions were the same, i.e., a temperature of 21 °C and an air humidity of about 40%. After mixing, the molding sand was stored for 60 min in a tightly closed container. The total mass of the dry ingredients was 5 kg. Five molding sands based on quartz sand from Grudzeń Las mine, with a main fraction of 1K class 0.20/0.16/0.10, were tested, in accordance with the requirements of PN-85/H-11001. The composition of the prepared molding sands is presented in Table 1. Two masses contained the addition of barley malt binder in the amount of 2% (MB 2%) and 5% (MB 5%) and water (2% and 5%, respectively). The third mass was made with 5% sodium water glass (WG 5%). The fourth mass (B 8%) contained Specjal bentonite in an amount of 8% and 0.8 parts by weight of water. On the other hand, the fifth thermosetting mass (RCS) was factory-prepared by Zębiec from quartz sand with the same main fraction as the other four masses. The binder of this mass was novolak resin. The commercial designation of the sand was ZGM D0128. The results of all the tests performed were the average value obtained from at least three measurements, taking into account the permissible measurement error. 2.5. Preparation of Samples for Testing the Properties of Molding Sand The preparation of samples for determining the properties of molding sands consisted in making three standard types of measuring samples: cylindrical, elongated, and octal (dog-bone) [3]. The samples were compacted on an LM-1 standard (Multiserw-Morek, Marcyporęba, Poland) hand compactor. It consisted in hitting the mass in a specific modeling block with a rammer three times. The compaction work was about 9.8 J. Then samples were dried in a laboratory dryer SLW 115 (Pol-Eko-Aparatura, Wodzislaw Śląski, Poland) with forced air circulation at the temperature of 150 °C for 60 min. 2.6. Methods of Determining the Properties of Molding Sand In order to determine the properties of molding sands, tests were carried out in which the following properties were determined: flowability (PD), permeability (Pss), tensile strength (Rms), and bending strength (Rgs) after dryingandwear resistance (Sss), also on samples in the dry state. Flowability (PD) was determined by the method of H.W. Dieterta and F. Valtiera [3]. It makes use of compacting molding sand with a laboratory compactor. In the test, a cylindrical sample is compacted by hitting a rammer five times. Flowability is measured by testing the loss in height of the sample measured between the fourth and fifth impact of the rammer. The value of flowability was calculated from the Formula (1):(1) PD=100−40x,% where: x—difference in height of the sample between the fourth and fifth impact of the rammer [mm]. Permeability (Pss) was determined by an accelerated method on a digital apparatus designed for determining the permeability of LPiR1 (Multiserw-Morek, Marcyporęba, Poland). Determination of parameters concerning tensile strength (Rms) and bending strength (Rgs) was carried out on the universal testing machine LRu-2e (Multiserw-Morek, Marcyporęba, Poland). Wear resistance (Sss) was determined according to BN-77/4024-02 standard at an ambient temperature for 3 min. The wear resistance (Sss) was calculated from the Formula (2):(2) Ss=a-ba × 100, % where: a—sample mass before testing (g), b—sample mass after testing (g). 2.7. SEM Analysis of Molding Sand In order to assess the structure of molding sand, microscopic observations were carried out by the SEM (Scanning Electron Microscope) TM 3030 (Hitachi, Tokio, Japan), cooperating with the EDX detector (Energy Dispersive X-ray Spectroscopy) TM3000 MICSF+ (Oxford Instruments, Oxford, UK). 2.8. Test Castings Casting molds were prepared by hand using the model shown in the photo—Figure 1, from molding sands with the addition of various binders. After obtaining the castings, we were able to assess their surface roughness depending on the molding sands used. The castings were made of gray cast iron EN-GJL-250. This alloy was chosen because it is a commonly used type of cast iron to evaluate the influence of molding sands on the solidification process of cast irons. Moreover, it is a typical, commonly used cast iron with good technological properties and a significant share in world production. Cast iron melts for the tests were carried out in a medium-frequency induction furnace, Type PI 30 (ELKON, Rybnik, Poland), using a crucible with a capacity of 6 kg. Gray cast iron (3.52% C; 1.80% Si; 0.76% Mn; 0.19% P; 0.01% S), ferrosilicon (FeSi75T) and a carburizer were used as charge materials. The liquid cast iron was overheated to temperature of 1350 ÷ 1400 °C.The slag was pulled off, and carburizer and ferrosilicon in the amount of about 1.3% in relation to the weight of the cast iron were introduced. Next, liquid alloy was cast by gravity into previously prepared molds made of various molding masses. After solidification and cooling, the castings were knocked out of the molds and subjected to further tests. The chemical composition of the castings was determined using the S1 Mini Lab spectrometer (GNR Analytical Instruments Group, Milan, Italy), and the measurement results are presented in Table 2. 2.9. Roughness of Test Casts Measurements of surface roughness of the test castings were carried out using the SV-3200 surface roughness tester (MITUTOYO, Kawasaki, Japan)—Figure 2. The device was scaled as follows: measurement length (X) 15.0000 mm, measurement step 0.0005 mm, measurement speed 1.00 mm/s, axis range (Z) 0.800 mm. Measurements were made in accordance with EN ISO 4287: 1998/AC: 2008 and PN-EN ISO 4288: 2011. 3. Results and Discussions During the pouring process, organic components of molding and core sands, mainly binders, as well as hardeners and protective coatings, usually burn quickly and evaporate, the remainder is ash [29]. During thermal decomposition of barley malt, casting defects may appear as a result of gasification. Therefore, research attempts were made to determine the content of organic compounds in barley malt and to analyze processes of thermal decomposition of the binder. 3.1. Tests of Barley Malt as a Binder for Molding Sand 3.1.1. Determination of Starch Content As a result of polarimetric tests, the starch content in barley malt was determined, which was 58.30%. Obtained results confirmed that starch—a carbon biopolymer composed of glucose units—is the main component of the tested binder. 3.1.2. Determination of Protein Content Using the Kjeldahl method, the total white content of binder was determined. For this purpose, the content of protein nitrogen was determined, which was 1.7312%. The amount of nitrogen was then multiplied by the nitrogen-to-protein conversion factor, appropriate for barley malt—6.25. The protein content was therefore 10.82%. The obtained results show that the total content of starch and protein substances constitutes nearly 70% of the ingredients contained in binder. The rest consists of the remaining sugars, including dextrins, cellulose, and fats. 3.1.3. TGA/TG Analysis The results of TG and DTG tests were analyzed in order to determine the content of substances that underwent thermal decomposition. The analysis was performed by estimating the parameters of the Gauss equation for the curves visible on the graph, shown in Figure 3. Calculations were made using the OriginPro 2019b software (OriginLab, Northampton, MA, USA). With a furnace heating rate of 12.5 °C/min, the main thermal decomposition of the sample took place when the temperature reached 440 °C and progressed to 510 °C (initial thermal degradation was 92.6 °C)—Figure 3a. At the heating rate of 25 °C/min, the weight loss of the sample could be observed in the temperature range of 500–620 °C (the initial thermal degradation was 172.0 °C)—Figure 3b, which indicates an intensive decomposition of organic matter. On heating at 50 °C/min, the weight loss of the sample could be observed when the temperature reached 650 °C (the initial thermal degradation was 218.0 °C). The decomposition of the organic matter lasted up to 800 °C—Figure 3c. Regardless of the furnace heating rate, the tested material was characterized by a final weight of ~25% of the initial value, which was mainly ash. Thus, the loss of sample mass of ~75% was organic matter, which was mainly composed of bound carbon. This material, undergoing thermal decomposition, can cause gasification by combustion products. As is commonly known, carbon oxides (CO2, CO) are products of coal combustion (the main component of binder), depending on the oxygen access to the reaction system. Detailed analysis of DTG results shows that the temperature range of thermal decomposition of tested binder depends significantly on heating rate and that the tested binder may contain from 2 to 4 main compounds. These substances are thermally decomposed. This is indicated by the test results in Table 3 and Figure 4. At a furnace heating rate of 12.5 °C/min, 4 curves appeared—Figure 4a, the maximum decomposition of which took place at temperatures of 318.2 °C, 438.7 °C, 438.9 °C, and 525.8 °C. This indicates thermal decomposition of malt organic components at these temperatures, and their content in the binder amounted to 8.5%, 20.8%, 33.8%, and 36.9%, respectively—Table 3. With the increase of the heating rate to 25 and 50 °C/min, the maxima of curves decreased, lowering and moving together toward higher temperatures. This may indicate the fact that the compounds included in the binder have similar thermal decomposition characteristics—Figure 4b,c. Similar to the TG/DTG plot, a detailed analysis of binder thermal decomposition curves showed a significant shift in organic compound decomposition curves with an increase in heating rate. This may also be due to the large mass of the sample (2 g) used in the tests. Therefore, at a high heating rate, the tested material was not able to heat up in its entire volume. The matter distribution of the outer layers could overlap with the matter distribution of the inner layers, which reached the set value with a delay. The effect of limited heat conduction deep into the sample may be proven by the research conducted by Seruga et al. [30]. They performed a barley TGA at an oven heating rate of 10 °C/min, achieving a similar final TG ~30%, but with major malt decomposition at much lower temperatures of 200–500 °C and three decomposition curves. 3.2. Microscopic Evaluation of Molding Sands In order to determine the morphology of the surface of grains and bonding bridges, microscopic observations were made using SEM with a BSE (Back-Scattered Electrons) detector. To obtain a high-resolution image, samples were carbon sprayed using a Q150T high-vacuum sputtering machine (Quorum, Laughton, UK). The surface morphology of the grains and the bonding bridges for the tested masses is presented using photos in Figure 5. The SEM image of molding sand (by SE imaging technique) with malted barley is shown in Figure 5a,b. The layer of malted barley binder creating bonding bridges between high-silica sand grains is smooth. Periodically, flakes of malted barley binder have been observed. Slowly heating the molding sand results in slow water flow into the airstream, which floats the dry load and ensures the quality of the bridges. The binder, dried in the form of smooth and mild bridges between the grains, allows very good strength properties of the molds and cores. The mass with the water glass (Figure 5c) has a very similar characteristic of bonding bridges as in the case of masses based on barley malt. However, the presence of irregularly shaped breaches was additionally noticed on the surface of the grains. The SEM image of bentonite molding sand is shown in Figure 5d. The flake structure characteristic of bentonite is visible on the grain surface. Small cracks in the bonding bridges were also observed, which translates into lower strength of these masses. Figure 5e shows topography of bridges connecting masses from sand surrounded by resin. The image shows a layer of resin evenly distributed over the matrix grains with embedded fine-dispersion spherical gas bubbles. Also, numerous “drops” of unbound binder were observed on the grain surface. Moreover, SEM observations of samples after testing the bending strength of masses were carried out. Figure 6 shows photos of fractures of these samples, showing destruction of bonding bridges. Masses based on barley malt (Figure 6a,b) and a water glass (Figure 6c) are characterized by the destruction of the cohesive type, i.e., the system in which the adhesive forces are greater than the cohesive forces [31,32]. The lower strength of the molding mass with the addition of 2% of barley malt binder compared to 5% is due to the so-called droplet distribution of binder, which is characteristic of the lower binder content and its high viscosity. The SEM image of bentonite mass (Figure 6d) is an example in which forces of adhesion and cohesion balance each other, and therefore destruction takes place on the grain surface of sand matrix and inside the layer of binding material [31]. In the case of a sand mass surrounded by resin (Figure 6e), there is a typical adhesive destruction. 3.3. Properties of Molding Sand The results of the flowability PD of masses are shown in Figure 7. Comparing this with the results for binder from barley malt, there is a noticeable tendency toward a decrease in flowability with an increase in the binder content. If a mass of B 8% is added to this conclusion, the above-mentioned trend is confirmed. Comparing the obtained results, it can be seen that all masses obtained good flowability, because the flowability exceeded or was close to 80%. It allows us to conclude that the tested masses can be compacted by all known methods, from manual forming to various machine forming methods. Increasing the binder content, resulting in a reduction in flowability, is related to a significant difference in the size of sand and binder grains, resulting in an increase in the value of the internal friction coefficient [33]. Results of the permeability Pss of molding sand are shown in Figure 8. A trend was observed that with increasing binder content, the permeability decreased. This is related to the greater number of voids between sand grains in the masses with a lower binding material content [6,8,9]. During pouring with liquid metal, significant amounts of gases can be released from molds with organic binders. Therefore, a lower malt content (2%) is preferred. The results of Rms from the tensile test are shown in Figure 9. Increasing the binder content in MB masses resulted in a significant increase in tensile strength. The strength Rms of the mass with the content of 5% of barley malt was highest. Results of measurements of bending strength Rgs of masses are shown in Figure 10. It can be seen, as with the tensile strength results, that increasing binder content increases the bending strength of the mass with malt binder. According to data from the manufacturer, bending strength for molding sand made of coated sand is close to 8 MPa. The manufacturer does not carry out other tests that were performed in this work. Figure 11 shows the results of wear resistance Sss. Increasing binder content reduces weight loss during the measurement. Reducing the value of Sss may mean a smaller number of raw casting defects. 3.4. Macroscopic Evaluation of Test Castings 3.4.1. Visual Assessment Preliminary visual analysis showed that the casting process was carried out correctly and all the obtained castings showed no signs of damage. Figure 12 shows a photo of the crude casting obtained in a mold made of a MB 2% mass (it is confirmed by a correctly conducted casting process). The thickness of the burnt layer of MB 2% and MB 5% was also assessed. A clear difference in the thickness of the burnt molding sand layer was observed. In the case of MB 2%, the thickness was ~14.5 mm and MB 5%~20 mm.This difference is related to higher malt content in the 5% MB mass, which translates into a greater depth of the scorch zone. In mass with a binder content of 2%, a greater amount of quartz sand eliminates the possibility of such a deep combustion of malt during the casting process. 3.4.2. Surface Roughness of the Castings In order to determine the surface roughness, castings were sandblasted. Even a preliminary visual assessment showed differences in roughness between individual castings. Roughness measurements were made along each “stair” from 1 to 4 of all the test casts. Values of Ra—arithmetic mean deviation of profile from the mean line and value of Rz—sum of height of highest elevation and deepest depression were determined. Averaged measurement results are presented in Table 4. The highest roughness was observed on the surface of the casting obtained in the mass of B 8%—Figure 13. In turn, the lowest roughness was characteristic of the cast made of RCS mass—Figure 14. The results of the roughness measurements of castings cast into molds from masses containing MB in the amount of 2% (Figure 15) and 5% (Figure 16) turned out to be satisfactory and comparable with results obtained for casting made in sand molds with WG (Figure 17). This applied in particular to casting made in a mold based on mass with the addition of 5% MB. In this case, surface defects in the form of cracks were observed on the surface of “stair” (Figure 16). The relatively high content of barley malt in the mass meant that during the solidification of the liquid metal, the binder burned intensively and a large amount of gases was released. The gasification probably caused the appearance of defects on the surface of solidified casting. 4. Conclusions The addition of MB binder in the amount of both 2% and 5% to mass guarantees relatively good strength properties, required from commonly used molding sands. Increasing the content of this binder increases the strength of the mass. It is related to the formation of bonding bridges between the binder and the molding sand, similar to the case of sand with a water glass. High flowability allows us to draw the conclusion that molding sand with MB binder can be compacted by standard methods. The mass containing MB 5% was characterized by lower fluidity compared to the mass with an MB 2% addition. This is due to the higher content of sticky material, resulting in a higher viscosity. The results of the TG/DTG analysis show that the loss of ~75% of the mass of the MB binder is organic matter. Over 50% of this is starch, which, undergoing thermal decomposition and gasification, may cause surface defects in castings. This dependence was observed for the mass containing 5% MB. There were no such defects in the case of the MB 2% mass. Thus, it is possible to eliminate this problem by introducing a smaller amount of binder into the molding sand or by using appropriate venting solutions. In addition, the greater content of binder in the mass resulted in its lower permeability. Gas activity of the binder will be the subject of further studies. Moreover, the mass of MB 2% was characterized by formation of a burnt layer 25% smaller than the mass of MB 5%. It is extremely important from the point of view of regeneration of the molding sand. This is due to the exothermic reaction when more binder is burnt. On the basis of the obtained results, it can be concluded that barley malt (compared to reference binders) can be a binder for molding sand. In addition, it is a natural resource, renewable through agricultural production. Therefore, in relation to conventional binding materials, it can be an alternative material that fits into the concept of sustainable development. Author Contributions Conceptualization, B.S. and D.M.; Data curation, B.S. and D.M.; Formal analysis, B.S., D.M., J.K.-R., K.Ś. and A.J.; Investigation, B.S., D.M., D.N., K.Ś., A.G. and A.J.; Methodology, B.S., D.M., D.N., K.Ś., A.G. and A.J.; Project administration, B.S. and D.M.; Supervision, D.M., J.K.-R. and A.J.; Writing—original draft, B.S., D.M. and D.N. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Model for molding foundry molds [27,28]. Figure 2 MITUTOYO CV-3200 profilometer and view of the analyzed stair cast (the roughness of each “stair” was individually measured using a needle measurement). Figure 3 Results of TG/DTG analysis of barley malt for the heating rate: (a) 12.5 °C/min, (b) 25 °C/min, (c) 50 °C/min. Figure 4 Detailed distribution of DTG diagram for furnace heating speed: (a) 12.5 °C/min, (b) 25 °C/min, (c) 50 °C/min. Figure 5 The surface morphology of molding sand for binders (SE imaging technique): (a) MB 2%, (b) MB 5%, (c) WG, (d) B 8%, (e) RCS. Figure 6 Destruction of bonding bridges after testing the bending strength of masses containing binders (SE imaging technique): (a) MB 2%, (b) MB 5%, (c) WG, (d) B 8%, (e) RSC. Figure 7 Results of flowability measurements of molding sands with different binders. Figure 8 Results of permeability measurements of different molding sands. Figure 9 Results of dry tensile strength measurements of different molding sands. Figure 10 Results of dry bending strength measurements of different molding sands. Figure 11 Results of dry wear resistance measurements of different molding sands. Figure 12 Open form after cooling, made of a mass containing 2% MB. Figure 13 Raw casting made in a mold with a mass of B 8%. A chart of an example of surface roughness for the 1st “stair” of the casting. 1–4 are numbers of stairs of the casting. Figure 14 Raw casting made in a mold with a mass of RCS. A chart of an example of surface roughness for the 1st “stair” of the casting. 1–4 are numbers of stairs of the casting. Figure 15 Raw casting made in a mold with a mass of MB 2%. A chart of an example of surface roughness for the 1st “stair” of the casting. 1–4 are numbers of stairs of the casting. Figure 16 Raw casting made in a mold with a mass of MB 5%. A chart of an example of surface roughness for the 1st “stair” of the casting. 1–4 are numbers of stairs of the casting. Figure 17 Raw casting made in a mold with a mass of WG 5%. A chart of an example of surface roughness for the 1st “stair” of the casting. 1–4 are numbers of stairs of the casting. materials-15-03375-t001_Table 1 Table 1 Composition of molding sands used in the research. Fine Silica Sand (%mass) Binder Type Binder (%mass) Distilled Water (%mass) 98 barley malt—MB 2% 2 2 95 barley malt—MB 5% 5 5 95 water glass—WG 5 - 92 bentonite—B 8% 8 0.8 97 resin coated sand—RCS 3 - materials-15-03375-t002_Table 2 Table 2 Chemical composition of the cast iron used for the test castings. Chemical Composition [%mass] Cast iron C Si Mn Ni Cu Mo Cr Al P S Fe 3.55 3.32 0.46 0.14 0.20 0.11 0.03 0.03 1.21 0.20 rest materials-15-03375-t003_Table 3 Table 3 Characteristic values of DTG curves for different heating rates. DTG, °C/min Characteristic Values Curve 1 Curve 2 Curve 3 Curve 4 12.5 Tmax, °C 318.2 438.7 468.9 525.8 Share, % 8.5 20.8 33.8 36.9 25 Tmax, °C 547.4 570.5 - - Share, % 49.4 50.6 - - 50 Tmax, °C 663.3 707.8 742.6 - Shar, % 35.4 48.1 16.4 - materials-15-03375-t004_Table 4 Table 4 Results of roughness measurements of castings made of molds from molding sands with the addition of various binders. Measurement for an Individual “Stair” of Test Castings Roughness Parameters for Castings Made in Molds from Masses with Various Binders MB 2% MB 5% WG 5% B 8% RCS 1 Raav 10.8130 8.2973 12.6998 18.8292 8.1663 Rzav 71.8486 58.5037 69.0820 87.2708 83.5917 2 Raav 13.8019 12.9511 14.8688 22.4032 6.4821 Rzav 79.7763 79.4316 81.4624 104.0351 92.6132 3 Raav 13.9996 16.004 12.3565 18.1374 10.633 Rzav 89.3288 90.9578 81.4827 92.1897 89.5778 4 Raav 11.5767 14.5426 11.8558 20.3440 10.5243 Rzav 79.5225 90.9488 75.8473 102.5992 86.0159 Average value from all “stair” Raav 12.55 12.95 12.95 21.43 8.95 Rzav 80.12 79.96 76.97 96.52 87.95 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Grabowska B. Holtzer M. World development tendencies in the field of moulding and core sands with regard to their environmental impact Arch. Foundry Eng. 2008 58 212 215 2. Major-Gabryś K. Environmentally Friendly Foundry Molding and Core Sands J. Mater. Eng. Perform. 2019 28 3905 3911 10.1007/s11665-019-03947-x 3. Lewandowski J.L. Moulding Materials Akapit Kraków, Poland 1997 (In Polish) 4. Czerwinski F. Mir M. Kasprzak W. Application of core sand binders in metal casting Int. J. Cast Met. Res. 2015 28 129 139 10.1179/1743133614Y.0000000140 5. Guilherme da Silva H. Espíndola Ferreira J.C. Vikas K. Garza-Reyes J.A. Benchmarking of cleaner production in sand mould casting companies Manag. Environ. Qual. Int. J. 2020 31 1407 1435 10.1108/MEQ-12-2019-0272 6. Ochlorl E.F. Ugboaja J.O. Olowomeye O.A. Performance of kaolin and cassava starch as replacements for bentonite in mouldings and used in thin wall ductile iron castings Niger. J. Technol. 2019 38 947 956 10.4314/njt.v38i4.18 7. Popoola A.P.I. Abdulwahab M. Fayomi O.S.I. Synergetic per formance of palm oil (Elaeisguineensis) and pine oil (Pinus sylvestris) as binders on foundry core strength Int. J. Phys. Sci. 2012 7 3062 3066 8. Atanda P.O. Akinlosotu O.C. Oluwole O.O. Effect of Some Polysaccharide Starch Extractson Binding Characteristics ofFoundry Moulding Sand Int. J. Sci. Eng. Res. 2014 5 362 367 9. Fayomi O.S.I. Hybrid Effect of Selected Local Binders on the Moulding Properties of River Niger Silica Sand for Industrial Application J. Nanosci. Adv. Technol. 2016 1 19 23 10.24218/jnat.2016.19 10. Yaro S.A. Suleiman M.U. Cassava/guinea corn starches and soybean oil as core binders in sand casting of aluminium silicon(Al-Si)lloy J. Eng. Technol. 2006 1 47 55 11. Grabowska B. Kaczmarska K. Cukrowicz S. Drożyński D. Żymankowska-Kumon S. Bobrowski A. Gawluk B. Influence of Carbon Fibers Addition on Selected Properties of Microwave-Cured Moulding Sand Bonded with BioCo2 Binder Arch. Foundry Eng. 2018 18 152 160 12. Chowdhury S.I. Rashid H. Mumtaz G.R. Comparison and CFD verification of binder effects in sand mould casting of aluminum alloy ANNALS Fac. Eng. Hunedoara Int. J. Eng. 2016 14 143 146 13. Manley D. Meals, grits, flours and starches (other than wheat), Technology of Biscuits, Crackers and Cookies 3rd ed. Woodhead Publishing Great Britain, UK 2000 104 111 14. Yu W. Quek W. Li C. Gilbert R. Fox G. Effects of the Starch Molecular Structures in Barley Malts and Rice Adjuncts on Brewing Performance Fermentation 2018 4 103 10.3390/fermentation4040103 15. Ukwenya J. Idoko F.A. Ameh J.E. Studies on the Effect of some Cereal Binders on the Hardness Values of Cast Aluminium Alloys (Al – 12%Si) Taraba J. Eng. Technol. (TAJET) 2019 1 1 3 16. Fox G.P. Chemical Composition in Barley Grains and Malt Quality Genetics and Improvement of Barley Malt Quality Zhang G. Li C. Springer Berlin/Heidelberg, Germany 2009 63 98 17. Czubaszek A. Wojciechowicz-Budzisz A. Spychaj R. Kawa-Rygielska J. Effect of AddedBrewer’s Spent Grain on the Baking Value of Flour and the Quality of Wheat Bread Molecules 2022 27 1624 10.3390/molecules27051624 35268724 18. Gałkowska D. Witczak T. Witczak M. Ancient Wheat and Quinoa Floursas Ingredients for Pasta Dough—Evaluation of Thermal and Rheological Properties Molecules 2021 26 7033 10.3390/molecules26227033 34834126 19. Czubaszek A. Wojciechowicz-Budzisz A. Spychaj R. Kawa-Rygielska J. Baking properties of flour and nutritional value of rye bread with brewer’s spent grain LWT 2021 150 111955 10.1016/j.lwt.2021.111955 20. Kotecka-Majchrzak K. Kasałka-Czarna N. Spychaj A. Mikołajczak B. Montowska M. The Effect of Hemp Cake (Cannabis sativa L.) on the Characteristics of Meatballs Stored in Refrigerated Conditions Molecules 2021 26 5284 10.3390/molecules26175284 34500717 21. Łucarz M. Grabowska B. Grabowski G. Determination of parameters of the moulding sand reclamation process, on the thermal analysis bases Arch. Metall. Mater. 2014 59 1023 1027 10.2478/amm-2014-0171 22. Bobrowski A. Drożyński D. Grabowska B. Kaczmarska K. Kurleto-Kozioł Ż. Brzeziński M. Studies on thermal decomposition of phenol binder using TG/DTG/DTA and FTIR-DRIFTS techniques in temperature range 20–500 °C China Foundry 2018 15 145 151 10.1007/s41230-018-7035-4 23. Żymankowska-Kumon S. Holtzer M. Grabowski G. Thermal analysis of foundry bentonites Arch. Foundry Eng. 2011 11 209 213 24. Kaczmarska K. Grabowska B. Grabowski G. Bobrowski A. Kurleto-Kozioł Ż. Thermal decomposition of binder based on etherified starch to use in foundry industry J. Therm. Anal. Calorim. 2017 130 285 290 10.1007/s10973-017-6451-9 25. Grabowska B. Żymankowska-Kumon S. Cukrowicz S. Kaczmarska K. Bobrowski A. Tyliszczak B. Thermoanalytical tests (TG–DTG–DSC, Py-GC/MS) of foundry binders on the example of polymer composition of poly(acrylic acid)–sodium carboxymethylcellulose J. Therm. Anal. Calorim. 2019 138 4427 4436 10.1007/s10973-019-08883-5 26. Carnin R.L.P. Folgueras M.V. Luvizão R.R. Correia S.L. da Cunha C.J. Dungan R.S. Use of an integrated approach to characterize the physicochemical properties of foundry green sands Thermochim. Acta 2012 543 150 155 10.1016/j.tca.2012.05.018 27. Bartocha D. Melting of two-phase cast steel using the recovery method Arch. Foundry Eng. 2014 14 3 8 (In Polish) 28. Cholewa M. Baron C. Kozakiewicz Ł. Influence of heat-insulating molding sand on the microstructure of gray cast iron Arch. Foundry Eng. 2015 15 119 123 29. Chen S. Zhang J. Xu K. Xu Q. Thermal Decomposition Behaviour of Foundry Sand for Cast Steel in Nitrogen and Air Atmospheres Math. Probl. Eng. 2020 2020 8121276 10.1155/2020/8121276 30. Jackowski M. Niedzwiecki L. Lech M. Wnukowski M. Arora A. Tkaczuk-Serafin M. Baranowski M. Krochmalny K. Veetil V.K. Seruga P. 2HTC of Wet Residues of the Brewing Process: Comprehensive Characterization of Produced Beer, Spent Grain and Valorized Residues J. Energ. 2020 13 31. Dobosz S.M. Water in Molding and Core Sand AKAPIT Kraków, Poland 2006 32. Jelinek P. Pojivové Soustavy Slévárenských Formovacích Směsí P. Jelínek Ostrava, Czech Republic 2004 (In Czech) 33. Szymański A. Soil Mechanics SGGW Warsow, Poland 2007 (In Polish)

PMC009xxxxxx/PMC9099644.txt

==== Front Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods11091313 foods-11-01313 Article Modifying the Quality of Pig Carcasses, Meat, and Dry Fermented Sausage from Black Slavonian Pigs by Selecting the Final Body Weight and Nutrition Samac Danijela * Senčić Đuro Antunović Zvonko https://orcid.org/0000-0001-9763-3522 Novoselec Josip Prakatur Ivana Steiner Zvonimir https://orcid.org/0000-0003-4078-6864 Klir Šalavardić Željka Ronta Mario Kovačić Đurđica Domínguez Rubén Academic Editor Faculty of Agrobiotechnical Sciences Osijek, Josip Juraj Strossmayer University of Osijek, V. Preloga 1, 31000 Osijek, Croatia; dsencic@fazos.hr (Đ.S.); zvonko.antunovic@fazos.hr (Z.A.); josip.novoselec@fazos.hr (J.N.); ivana.prakatur@fazos.hr (I.P.); zvonimir.steiner@fazos.hr (Z.S.); zeljka.klir@fazos.hr (Ž.K.Š.); mario.ronta@fazos.hr (M.R.); djkovacic@fazos.hr (Đ.K.) * Correspondence: dsamac@fazos.hr 30 4 2022 5 2022 11 9 131307 3 2022 28 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). A total of 96 Black Slavonian pigs were subjected to the research, in which they were split into 6 groups. Three groups (100, 120, and 130 kg) were fed a higher level (HL) of crude protein in fodder mixtures (CPFM), and three groups (100, 120, and 130 kg) were fed a lower level (LL) of CPFM. After the pigs were slaughtered, pig carcasses were dissected and the meat and halves quality indicators were determined. According to the influence of the final body weight (BW) and nutrition of pigs on the quality of their halves, meat, and dry fermented sausages (kulens), it was concluded that feeding an HL of CPFM increased the proportion of loin, belly rib part, and chin and increased the muscle tissue in the ham, loin, shoulder, neck, and belly rib parts. However, the chemical composition of the meat and the sensory properties of the kulen were not significantly affected by feeding the pigs an HL of CPFM. It was concluded that, by selecting the final BW and adjusting the feeding strategies for pigs, it is possible to modify the conformation and composition of pig carcasses and the quality of meat and kulens produced from the Black Slavonian pig, which is important because consumers prefer products with certain characteristics and of a standard quality and are ready to pay for them. Black Slavonian pigs feeding final body weight kulen meat quality This research received no external funding. ==== Body pmc1. Introduction The Black Slavonian pig, also known as “Fajferica”, is an autochthonous Croatian pig breed of the meaty and fatty type created at the end of the 19th century on the estate of Count Karl Pfeiffer called Orlovnjak, near Osijek. Count Pfeiffer mated Mangalica sows with Berkshire boars and then mated the offspring with Poland China boars [1]. From the end of the 19th century until the middle of the 20th century, the Black Slavonian pig was the most widespread pig breed in Croatia, but after the Second World War, because of modern, highly productive white-meat breeds from abroad, the breeding of this pig was neglected to the point where it almost became extinct [2]. However, due to good meat quality, resistance [3], longevity, and adaptability to extensive housing conditions [4,5], this breed was again recognized by Croatian producers, who realized that the Black Slavonian pig should be protected for biodiversity in a “bank of genes” that will permanently preserve genetic materials from rare and endangered breeds or individuals of the population [6]. This would encourage agricultural production, contributing to environmental protection, the preservation of biodiversity, and protection of rural areas [7]. Financial incentives provided by the government for the biological preservation of the breed cannot be used as a permanent solution. Therefore, this breed should be evaluated through adequate use for economic purposes. To date, this breed has been grown extensively without systematic research on its production properties under the influence of various paragenetic factors, particularly nutrition. Taking similar European Mediterranean breeds (the Basque, the Gascon, and the Corsican pig in France; the Nero Siciliano and the Cinta Senese in Italy; and the Iberian pig in Spain) as a model, the Black Slavonian pig needs to be used for economic purposes by producing specific and protected products (fresh meat, dry fermented sausage, ham, greaves, etc.) with a trademark (brand) [8]. To preserve this breed, it is important to increase the production of traditional standardized meat products with a higher added value. Dry fermented sausage, referred to as kulen, is the product with the highest degree of valorization, price, and quality [9]. As a result of measures to halt the breed’s decline, research related to the Black Slavonian pig intensified at the beginning of the 21st century [4]. The effective population in 1996 was only 20 pigs, when the survival of the breed was endangered [10]. Croatia, therefore, signed the Biodiversity Treaty [11], and “A Survey of the State of Biological and Environmental Diversity of Croatia with Strategy and Protection Plan Action” was elaborated [12], as well as “A Program for Breeding up of the Black Slavonian Breed” [13]. Even today, the Black Slavonian pig is studied by few scientists from the region of its origin. Some research has been conducted on evaluating different housing and breeding systems [5,14], whereas several research groups have conducted various genetic studies [15,16,17,18] on breed determination and conservation of local breeds. Other studies have been related to the fattening of the Black Slavonian pig to obtain high-quality meat and meat products [19]. Mostly, the quality of pig halves, meat, kulen [20,21,22,23], ham [24], and sausages [25] has been studied. The aim of this study is to show how selecting the final body weight (BW) of Black Slavonian pigs and feeding with different levels of crude protein in fodder mixtures (CPFM) influence the composition of pork halves (muscle, fat, and bone tissue), the conformation of pork halves (share of basic parts), and the composition of muscle tissue, particularly the content of intramuscular fat. Furthermore, the aim is to show that the modification of muscle tissue (meat) composition can improve the nutritional, physicochemical, and sensory properties of meat and kulen. 2. Materials and Methods 2.1. Animal Menagement The study was conducted in compliance with the legal regulations set by the Animal Protection Act of Croatia (NN 133/06, NN 37/13, and NN 125/13) and the European Union Directive 2010/63/UE regarding animal protection, as approved by the Bioethics Committee for Research on Animals of the Faculty of Agrobiotechnical Sciences, Osijek (2158-94-02-22-01). We used 96 Black Slavonian pigs, 32 of which were fattened up to approximately 100 kg final BW (groups A and B), 32 up to approximately 120 kg final BW (groups C and D), and 32 up to approximately 130 kg final BW (groups E and F). During the fattening, the pigs were fed fodder mixtures of grains (corn and barley) and a superconcentrate (Table 1). Groups A, C, and E were fed 14% CPFM during the first fattening period and 12% CPFM during the second fattening period, while groups B, D, and F were fed 12% CPFM during the first fattening period and 10% CPFM during the second fattening period. During the fattening, the pigs also consumed green alfalfa ad libitum. The sex ratio in each experimental group was equal (50:50%). During the fattening, all pigs were kept in the same housing conditions, in a semi-outdoor system. After the fattening was completed, the pigs were transported to a slaughterhouse in a truck 24 h prior to being slaughtered. At the time of slaughter, the pigs were between 16 and 18 months old. 2.2. Carcass Measurements After the pigs were slaughtered, pig carcasses were left to chill at +4 °C at a relative moisture of 90% for 24 h. After that, the right-half carcasses were dissected according to the modified method of [26]. The length of cold pig halves was measured by means of a measuring tape from os pubis to atlas and from os pubis to the 1st rib. Backfat thickness (S) was measured on the small of the back, at the point where musculus gluteus medius pars piriformis grows into bacon most intensively. The thickness of the longissimus lumborum in mm (M) was measured as the shortest connection between the front end of musculus gluteus medius and the upper edge of the vertebral canal. The surface area of the cross section (cm2) of muscullus longissimus dorsi (MLD) was measured between the 13th and 14th ribs, as per the method provided in [27], by means of a digital planimeter Haff Digiplan 305/306. The volume of the leg was measured at the widest part of the leg. A measuring tape was used to measure ham circumference and length. 2.3. Physico-Chemical Analysis of Meat Meat quality indicators were investigated on a sample taken from the long back muscle (MLD), from between the 13th and 14th ribs. The pH1 value (45 min postmortem) was determined at the meat temperature of 35 °C, and the pH2 value (24 h postmortem) was measured after the meat was chilled at +4 °C. A contact pH meter (Mettler Toledo) was used for this purpose. The water-holding capacity (WHC) was determined by means of compression as per the method provided in [28], and consistency was presented as the surface area in cm2 of muscular tissue that was compressed on a filter paper in the process of measuring the WHC. Meat color was measured with a mobile device colorimeter Minolta Chromameter CR-410 (Minolta Camera Co., Ltd. Japan) Illuminant D65, Observer 2 degrees Closely matches CIE 1931Standard Observers: x¯2λ, y¯λ, z·λ, according to the standard CIE L* a* b* color system [29], according to the reference method in Honikel (1998), approved by the International Commission on Illumination CIE (1976). Meat color was measured on the long back muscle (MLD), taken from between the 13th and 14th ribs, cooled at +4 °C, after color stabilization (after 10 min), with 3 repetitions. The CPFM content in meat was determined according to the Kjeldahl method [30]. Intramuscular fat was determined according to the Soxhlet method [31]. The water content was defined as the loss sample of mass after the sample had been dried at 105 °C until constant mass was achieved. The ash content was determined by burning the organic matter at 550 °C until constant mass was achieved, and the ash content was shown as the percentage remains of the sample mass. 2.4. Kulen Menagement After the carcass was dissected, meat from ham, loin, and shoulder, without fat and joint tissue, was separated for processing into kulens. The proportion (%) of the meat from ham and loin (80%) and shoulder (16.20%) was the same in all the tested groups. This “pure” muscular tissue was ground in a grinding machine (matrix diameter 6 mm and 8 mm), and the kulen mixture was prepared with 80% meat from ham and loin, 16.20% meat from shoulder, 2% salt, 1% sweet ground pepper, and 0.8% minced garlic. The mixture for each tested group of pigs was prepared separately and then stuffed into a pig cecum by means of a stuffing device. Thus prepared, the kulens were weighed and spread on drying racks in a drying room. For the first 15 days, the kulens were dried over smoke obtained by burning ash wood (Fraxinus excelsior). During these 15 days, fire was made 7 times. Kulens were then left to dry and mature for the next 9 months. 2.5. Kulen Analysis After reaching maturity, the kulens were submitted for organoleptic evaluation to a taste panel consisting of 5 members with previous experience in assessment. The panelists have consented to participation in the study. Assessment was carried out at the Faculty of Agrobiotechnical Sciences Osijek. Each kulen sample was cut in half on a white glass plate with a knife. The assessor was provided one half to assess the appearance, the cross-section appearance, and the structure via visual inspection and touching. The assessor was provided the other half for tasting, for which each assessor was served a slice (around 0.5 cm thick) immediately after the kulen was cut. After every 6 evaluated kulens, the assessors took a break for an hour. During kulen assessment, the assessors were also served cheese, bread, apples, and water (at room temperature) to eliminate (neutralize) the traces of taste from the mouth between tasting individual samples. Kulen properties were scored as follows: appearance (1–5), structure (1–3), cross-section appearance (1–3), smell (1–5), taste (1–10), and general impression (1–5). A total of 8 kulens from each of the tested groups were assessed. The assessment of the sensory properties of the kulens was followed by an investigation of their physical and chemical properties. The pH values of the kulen were measured with a contact pH meter (Mettler Toledo) applied to the center of the cross section. The kulen color was measured with a mobile device Minolta Chromameter CR-410 (Minolta Camera Co., Ltd. Japan) according to the standard CIE L* a* b* color system [29]. Water activity (aw) in the kulen was measured using a HygroLab 3 (Rotronic) by applying the Aw Quick model of operation on samples prepared by chopping and homogenizing 100 g of the central part of the kulen. The NaCl content in the kulen was determined by the titrimetric method [32], and the CPFM content of the kulen was obtained using the Kjeldahl method. Intramuscular fat was determined according to the Soxhlet method. The water content was defined as the loss of sample mass after the sample had been dried at 105 °C until constant mass was achieved. The ash content was determined by burning the organic matter at 550 °C until constant mass was achieved, and the ash content was shown as the percentage remains of the sample mass. 2.6. Statistical Analysis Statistical data processing was performed by means of a Stat. Soft. Inc. (2012) computer software. Significance testing between and within groups was determined by an analysis of variance (ANOVA), and the calculated F value was compared with the theoretical F value. The significance of differences between mean values was determined using Fischer’s LSD test. 3. Results and Discussion 3.1. Indicators of the Quality of Pig Halves The basic indicators of pig carcass quality are presented in Table 2. As the weight of the pig carcasses increased by weight groups, the length of the pig carcasses from os pubis to atlas also increased. The carcass length was similar to that of the Iberian pig [33] and the Majorcan Black pig [34]. For this property, a significant (p < 0.05) and a very significant difference (p < 0.01) were determined between the weight groups of pigs fed with a lower level (LL) of CPFM, whereas a very significant difference was found at a higher level (HL) of CPFM (p < 0.01) only between the 100 and 130 kg pig groups. Similar significant differences were determined also in terms of the carcass length from os pubis to the 1st rib. As the pig carcass weight increased, so did the ham circumference. Therefore, very significant differences (p < 0.01) were determined among all weight groups of pigs fed with both lower and higher CPFM levels. The level of CPFM affected the ham volume only within the weight group of 100 kg pigs. Therefore, the hams of pigs fed with higher CPFM levels were very significantly (p < 0.01) higher in volume. The increase in the CPFM level in fodder mixtures affected the increase in ham length very significantly (p < 0.01) between the 120 and 130 kg pig groups, while in the 100 kg pig group, no significant difference (p > 0.005) was determined. The increase in the level of CPFM very significantly affected (p < 0.01) the increase in the ham index in the 120 and 130 kg pig groups. The increase in the BWs of pigs within groups fed with fodder mixtures with an LL of CP affected the increase in the ham index. An HL of CPFM affected the reduction in backfat thickness between the 120 and 130 kg pig groups. The increase in the BWs within pig groups fed with both LL and HL of CPFM led to a significant (p < 0.05) and a very significant (p < 0.01) increase in backfat thickness. Backfat thickness in the Black Slavonian pig is not excessive, unlike some other rural breeds. For example, in the research in [33], Iberian pigs of approximately 130 kg had a backfat thickness of 6.00 cm, whereas in the research in [35], even higher values were recorded for the Majorcan Black pig. The level of CPFM did not significantly (p > 0.05) affect the cross-sectional area of MLD. However, this property was very significantly (p < 0.01) affected by the BWs of pigs, so that higher BW pigs had a very significantly (p < 0.01) smaller MLD area. 3.2. Conformation of Pig Carcasses As can be seen from Table 3, no significant (p > 0.05) influence of the level of CPFM on the absolute proportion of hams in pig halves was determined, though a very significant (p < 0.01) or significant (p < 0.05) influence of the final BWs of pigs on the absolute proportion of hams was determined. In addition, [36] determined that the CPFM level in fodder mixtures had no effect on the absolute proportion of hams in pig halves. As the BWs increased within pig groups, the absolute proportion of hams increased. The level of CPFM did not affect the proportion of less valuable parts in the halves, but it affected very significantly (p < 0.01) the proportion of the loin, shoulder, neck, belly rib part, chin, and fat in some weight groups. As the level of CPFM decreased, the loin proportion increased very significantly (p < 0.01) in all the weight groups, the shoulder proportion increased in the 120 kg pig group, and the neck and fat proportions increased in the 130 kg pig group. The proportion of belly rib part was significantly (p < 0.01) higher in all weight groups fed with fodder mixtures containing an HL of CP, which corresponds to previous studies by [36]. The relative proportions of basic parts in the halves are shown in Table 4. No significant (p > 0.05) impact of the level of CPFM or pig BW on the proportion of hams in the pig halves was found. In the research by [37], it was determined that pigs with lower final BWs produce carcasses with a significantly (p < 0.05) higher share of hams in halves. With the increased level of CPFM, the proportion of the belly rib part increased very significantly (p < 0.01) in all the weight groups. The level of CPFM very significantly (p < 0.01) or significantly (p < 0.05) affected the proportion of other basic parts in pig halves in some weight groups. Thus, the neck and fat proportions were very significantly (p < 0.01) higher at an LL of CPFM in weight groups ranging from 100 to 130 kg and in the 130 kg pig group, respectively. As the BW increased, the relative proportion of loin also increased. Therefore, significant differences were determined between weight groups ranging from 100 to 130 kg at an HL of CPFM, between weight groups ranging from 100 to 130 kg at an LL of CPFM, as well as between weight groups ranging from 120 to 130 kg at an LL of CPFM. As the BW increased, the proportion of shoulder decreased significantly (p < 0.05) or very significantly (p < 0.01) at an HL, but not at an LL of CPFM. This was in accordance with the results obtained by [38], who investigated the effects of four dietary treatments characterized by ranging from 0 to 20% progressive reduction in the dietary CPFM and indispensable amino acid contents on the carcass quality and uniformity of pigs. In the research on the influence of the final BW on the quality of pig carcasses of the Black Slavonian pigs, [19] observed that, as the final BW increased, the relative proportion of the shoulder in the carcasses decreased. The proportion of the belly rib part decreased significantly (p < 0.05) or very significantly (p < 0.01) with an increase in the pig BW within groups. Very significant differences (p < 0.01) were found between the 100 and 120 kg pig weight groups and between the 100 and 130 kg pig weight groups at an HL of CPFM, as well as between the 100 and 130 kg pig weight groups and between the 120 and 130 kg pig weight groups at an LL of CPFM. The chin proportion increased with increasing pig BW. Significant differences (p < 0.05) were determined between the 120 and 130 kg weight groups at an HL of CPFM and between the 100 and 120 kg pig weight groups, between the 100 and 130 kg pig weight groups, and between the 120 and 130 kg pig weight groups at an LL of CPFM. The fat proportion in the pig halves also increased with increasing pig BW, but very significant differences (p < 0.01) were found only between the 100 and 130 kg pig weight groups and between the 120 and 130 kg pig weight groups at an LL of CPFM. This was in accordance with the study in [37]. 3.2.1. The Tissue Proportion in Carcasses As shown in Table 5, the increased level of CPFM affected the increase in the absolute and relative proportions of muscle tissue in halves in the same pig weight groups very significantly (p < 0.01), while no significant differences were found in terms of fat proportion (p > 0.05). In the study on the influence of CPFM levels on the composition of pig carcasses of Iberian pigs, [39] concluded that pigs fed a meal with an LL of CPFM significantly deposited fat in the carcass compared to pigs fed a meal with an HL of CP. Additionally, refs. [40,41] confirmed these conclusions with their research. Analogously, refs. [42,43] concluded that, when pigs are fed low-protein diets, fatter carcasses are produced compared with pigs fed with high-protein diets. However, they suggested the adoption of the net energy system and balanced amino acids as a means to still achieve acceptable performance, carcass characteristics, and meat quality. As stated in the review paper by [44], heavier pigs are associated with a greater backfat thickness and a decreased percentage of fat-free or lean meat. This research also confirmed that an increase in the BWs of pigs had a very significant (p < 0.01) influence on the proportion of muscular and adipose tissue in pig carcasses among different weight groups. The absolute proportion of muscular tissue grew very significantly (p < 0.01) as the BW increased, whereas the relative proportion of muscular tissue decreased very significantly (p < 0.0.1) with the increase in BW within both levels of CPFM. Accordingly, the proportion of adipose tissue in carcasses grew very significantly (p < 0.01) as the BW increased within both levels of CPFM. 3.2.2. Proportion of Muscular Tissue of the Basic Parts in the Weight of Pig Carcasses As can be seen from Table 6, the increase in the level of CPFM had no significant (p > 0.05) influence on the increase in the muscular tissue proportion of ham in the 100 and 130 kg pig weight groups, whereas in the 120 kg pig weight group, a significantly (p < 0.05) higher proportion of the muscular tissue of ham was determined at an HL of CPFM. As the BW of the pigs increased from 100 to 130 kg, a tendency was recorded toward a decrease in the proportion of the muscular tissue of a ham, but very significant differences (p < 0.01) were determined only between the 100 and 130 kg pig groups and between the 120 and 130 kg pig groups at an HL of CPFM. A significant difference (p < 0.05) was determined between the 100 and 130 kg pig groups at an LL of CPFM. A tendency toward an increase in the proportion of the muscular tissue of the loin was recorded with an increase in the CPFM content. However, a significant difference was determined only in the 130 kg pig weight group. An increase in the BW of pigs had no significant (p > 0.05) influence on the proportion of muscular tissue of the loin. Considering the proportion of shoulder and neck muscle tissue, no regular upward or downward trend was detected depending on the level of CPFM and the pig slaughter mass. A decrease in the level of CPFM led to a decrease in the proportion of muscular tissue in the belly rib part. Very significant (p < 0.01) differences in the part of this muscular tissue within the 120 and 130 kg pig weight groups were determined. An increase in the weight of the pig halves within the weight groups did not significantly (p > 0.05) affect the proportion of muscular tissue in the belly rib part at an HL of CPFM, but a very significant (p < 0.01) influence of the pig BW on this indicator at an LL of CPFM was determined. 3.2.3. Proportion of Adipose Tissue of Basic Parts in the Weight of Pig Carcasses As shown in Table 7, the proportion of CPFM did not significantly (p > 0.05) affect the proportion of adipose tissue of the ham in the weights of pig halves in the 100 and 120 kg pig weight groups, but a significant (p < 0.05) influence in the 130 kg pig weight group was determined. An increase in the BW of pigs by weight groups led to the tendency for an increase in the proportion of adipose tissue of ham, but very significant differences (p < 0.01) were detected only between the 100 and 130 kg pig groups and between the 120 and 130 kg pig groups at an HL of CPFM, and significant differences (p < 0.05) were detected between the 100 and 120 kg weight groups at an LL of CPFM. The proportion of adipose tissue of the loin was higher at an LL of CPFM in all pig weight groups, but very significant differences (p < 0.01) were detected only within the 100 and 130 kg weight groups. The increase in the BWs of pigs led to the tendency of increase in the proportion of adipose tissue of the loin in all weight groups at both LL and HL of CPFM. Very significant differences (p < 0.01) were determined between the 100 and 120 kg weight groups and between the 100 and 130 kg weight groups at an HL of CPFM and also between the 120 and 130 kg weight groups at an LL of CPFM. The proportion of adipose tissue of the neck was higher at an LL of CPFM in all pig weight groups, but a significant difference (p < 0.05) was determined within the 100 kg weight group. As the pig carcass weight increased, the relative proportion of adipose tissue of the neck in all pig weight groups at both LL and HL of CPFM also increased, but a very significant difference (p < 0.01) was determined only between the 100 and 120 kg weight groups and a significant difference (p < 0.05) was determined between the 100 and 130 kg weight groups at an HL of CPFM. The proportion of adipose tissue of the belly rib part was higher at an HL of CPFM in all the pig weight groups, but very significant differences (p < 0.01) were detected only in the 100 kg weight group. With the increase in pig BW, the relative proportion of adipose tissue of the belly rib part decreased at an HL of CPFM, while at an LL of CPFM, it increased. Very significant differences (p < 0.01) were detected between the 100 and 120 kg pig weight groups at an HL of CPFM, between the 100 and 130 kg pig weight groups at an HL of CPFM, and between the 100 and 130 kg weight groups at an LL of CPFM. The proportion of adipose tissue of the chin was higher in the 100 kg weight group at an HL of CPFM. The proportion of adipose tissue of fat was higher at an LL of CPFM in the 130 kg weight group, whereas in all the other weight groups, no significant differences (p > 0.05) were detected in terms of the CPFM level. Moreover, no influence of pig BW on the relative proportion of adipose tissue of fat was determined. 3.2.4. Proportion of Bone Tissue of the Basic Parts in the Weight of Pig Carcasses The proportion of bone tissue of the ham in the weight of pig carcasses did not significantly (p > 0.05) differ between pig groups, as well as in terms of the level of CPFM and pig BW (Table 8). 3.3. Pork Quality Indicators 3.3.1. Physico-Chemical Properties of Meat The physical properties of the meat are shown in Table 9. The pH value of the meat is an important characteristic of meat quality, as the transformation of muscles into meat changes its pH value. A neutral pH value shifts toward an acidic pH value. According to the classification of [45,46], the preferred values for pH1 are above 6.0 (a value between 5.8 and 6.0 is suspicious), while for pH2, values below 5.7 indicate PSE meat (i.e., pale, soft, and exudative meat) and values above 6.0 indicate DFD meat (i.e., dark, firm, and dry meat) [47]. The data obtained for the pH1 and pH2 values, WHC, and meat consistency point to meat of normal quality, and no significant differences (p > 0.05) were determined among the groups of pigs when considering their final BWs and the level of CPFM. According to [48], pigs fed with an LL of CPFM have meat with a lower pH1 value and WHC, while [41] reported the limited influence of the level of CPFM on the meat’s pH value. In [49], a decrease in WHC was observed when pigs were fed reduced CPFM levels. As per [50], as the BWs of pigs increased from 100 to 160 kg, the pH values of the meat decreased. This correlates with the results obtained by [51], who determined higher pH values of meat in pigs of lower weight. For white modern breeds, [3] determined that an increase in the BWs of pigs was also followed by an increase in the pH values, WHC, and marbling. Surface color is an important visual quality indicator of meat, and meat color measurements are usually expressed on the L* (lightness), a* (redness), and b* (yellowness) scale [52]. An increase in the level of CPFM did not significantly (p > 0.05) influence differences in terms of the L* value for meat color in the 100 and 120 kg weight groups, but a significant difference (p < 0.05) in the 130 kg weight group was detected. In this research, in most of the investigated groups of pigs, the L* values for meat somewhat exceeded the desired values (43–50) reported by [53]. Higher L* values (HL of lightness) are a consequence of a higher proportion of fat in the meat, which ranged from 6.97% in group 1 to as much as 16.98% in group 6. According to [48], pigs fed with lower-CPFM diets have a lighter (L*) meat color. In terms of the level of CPFM, [41] determined a limited effect of diet on meat color. As the BWs increased within weight groups at an HL of CPFM, no significant differences (p < 0.05) were determined for this meat quality between the 120 and 130 kg pig groups and a very significant difference (p < 0.01) was determined between the 100 and 130 kg pig groups at an LL of CPFM. In [54], it was indicated that the higher final BW of pigs very significantly (p < 0.01) influenced the L* value for meat color, whereas [51] indicated that an increase in the L* value for meat color corresponds to a final BW increase. Many authors [55,56,57] have not recorded any statistically significant differences (p > 0.05) in the L* values for meat color in relation to pig BW. Very significant differences (p < 0.01) were determined between pig groups with regard to the level of CPFM for a* values of meat color. With an increase in pig BW by weight groups at an HL of CP, no significant differences (p > 0.05) for a* values were detected, while very significant differences (p < 0.01) were determined for this parameter between the 100 and 130 kg pig weight groups and between the 120 and 130 kg pig weight groups at an LL of CPFM. In [54], it was indicated that higher final BWs of pigs very significantly (p < 0.01) influenced the a* value for meat color. As per [58] also, there were significant differences (p < 0.05) in the level of meat redness between the 100 and 130 kg pigs. In the research carried out by [59], higher a* values for meat color were determined in heavier pigs, whereas [56,57] did not determine statistically significant differences in terms of the a* value for meat color. Increased a* values for meat color, which were recorded within pigs with higher final BWs, are in correlation with a higher content of muscular pigment in older pigs. In terms of the b* values of meat color, no significant differences (p > 0.05) were detected between the 100 and 120 kg pig groups and with an increased level of CPFM. As the BWs increased by weight groups at an HL of CPFM, no significant differences (p > 0.05) between the 100 and 130 kg pig weight groups and between the 120 and 130 kg pig weight groups were detected, whereas significant differences (p < 0.05) were detected between the 100 and 130 kg pig weight groups and between the 120 and 130 kg pig weight groups at an LL of CPFM by weight groups. In our previous research [37], higher b* values were determined for the meat color of heavier animals. While [54] determined that higher final BW of pigs very significantly (p < 0.01) influenced the b* value for color, [55,56,57] did not find any statistically significant differences in terms of the b* value for meat color. 3.3.2. Basic Chemical Properties of Meat As can be seen from Table 10, with an increased level of CPFM, the water and protein content in meat also increased within the weight groups of pigs, but not statistically significantly (p > 0.05). As expected, the water and protein content in the meat decreased very significantly (p < 0.01) as the weight of the carcasses increased by weight groups between the 100 and 120 kg weight groups and between the 100 and 130 kg weight groups at both LL and HL of CP. This is in agreement with our previous studies [19,37], in which we determined that pigs with lower BWs produce meat with a higher water and protein content. However, as the level of CPFM increased, the fat content in the meat decreased within all pig weight groups, but not in a statistically significant way (p > 0.05). With the increase in the weights of pig halves within weight groups, the fat content also increased very significantly (p < 0.01) between the 100 and 120 kg pig weight groups and between the 100 and 130 kg pig weight groups at both LL and HL of CPFM, whereas between the 120 and 130 kg weight groups, no significant differences (p > 0.05) were detected. With the increase in the level of CPFM, no statistically significant differences (p > 0.05) in the ash content were detected within weight groups. As the weight of carcasses increased by weight groups, very significant differences (p < 0.01) were detected in the ash content of meat between the 100 and 120 kg pig weight groups and between the 100 and 130 kg pig weight groups at both LL and HL of CP, whereas between the 120 and 130 kg weight groups, no significant differences (p > 0.05) were detected. As assumed, an increase in the final BWs of pigs caused an increase in the content of intramuscular fat and a decrease in the content of CPFM and water in meat. 3.4. Indicators of Kulen Quality 3.4.1. Physical and Chemical Properties of Kulen In terms of the pH value of kulens, kulen color (L*, a*, and b* values), and the content of NaCl and water in kulens, no significant differences (p > 0.05) were detected among groups when taking into consideration the level of CPFM and the BWs of pigs. Kulens with slightly lower pH values were found in the research of [60] (5.35) and [61] (5.07–5.75) if compared to the pH values of kulens obtained in this research (5.86–5.97), which can be attributed to the influence of different pig genotypes, different production technologies, and different phases of kulen ripening. These studies were carried out on Slavonian kulens produced by various producers and originating from different areas of Slavonia. As expected, with an increase in the level of CPFM within weight groups, the protein content in kulens also increased, but not in a statistically significant way (p > 0.05). As the weights of pig halves by weight groups increased, the protein content in kulens decreasing very significantly (p < 0.01) between the 100 and 120 kg pig groups and between the 100 and 130 kg pig groups at both HL and LL of CP. No significant differences were detected (p > 0.05) between the 120 and 130 kg pig weight groups in terms of the protein content in kulens. The research results for the protein content in kulens obtained by [60] (22.92%), [61] (30.3–39.6), and [20] (40.99%) as well as the results of this research (43.59–45.94%) indicate that kulens, compared to some other traditional sausages researched by other authors [62,63], have a higher average protein content. An increased level of CPFM did not significantly (p > 0.05) affect the fat content in kulens within the 120 and 130 kg pig weight groups, whereas in the 100 kg weight group, a very significant difference (p < 0.01) was determined, because pigs fed with an HL of CPFM gave kulens with less fat. The final BWs of pigs had a very significant (p < 0.01) influence on the fat content in kulens between the 100 and 120 kg pig groups and between the 100 and 130 kg pig groups with an HL of CP and between the 100 and 120 kg pig groups with an LL of CP. Increased BWs of pigs resulted in increased crude fat in kulens. Compared to the results obtained by [60] (24.23–60.34%), [61] (16.40–31.00%), and [20] (23.03%), in this research, a lower fat content in kulens was determined (16.24–18.51%). This can be explained by differences in the production of kulens in their research (fatter meat, backfat added, etc.). No statistically significant differences (p > 0.05) were detected in terms of ash, water, and NaCl content, as well as L*, a*, b*, and aw values in kulens among the groups of pigs of different weights. The results of the research on the physico-chemical properties of kulens are in line with expectations and the results of the chemical analysis of meat from Black Slavonian pig (Table 11), which indicate that, with increasing levels of CPFM, the crude fat content decreases, while the CPFM content in meat increases among pig weight groups, but not in a statistically significant way (p > 0.05). 3.4.2. Sensory Properties of Kulens The sensory properties of kulens are provided in Table 12. No significant differences (p > 0.05) were recorded in terms of appearance (it was even), structure (firm, but not too hard), cross-section appearance (it was uniform in all groups), and smell (it was nice) between the groups when considering the level of CPFM and the final BWs of pigs. An increase in the level of CPFM did not significantly (p > 0.05) affect the taste of the kulens. However, the BW before slaughter significantly influenced the taste of the kulens. Pigs with higher BWs provided better-tasting kulens. Very significant differences (p < 0.01) were detected in terms of the taste between kulens produced from pigs with BWs of 100 and 120 kg and from pigs with BWs of 100 and 130 kg at an HL of CPFM and between the 100 and 120 kg pig groups and between the 100 and 130 kg pig groups at an LL of CPFM. It should be emphasized that the Black Slavonian pigs from all the investigated weight groups (100, 120, and 130 kg) resulted in the production of good-quality kulens. The reason for this is the already “mature” meat in pigs with BWs of 100 kg, as they have a high fat content, i.e., a higher dry matter content. Another reason is that, compared to modern high-meat genotype pigs, the Black Slavonian pigs reach that BW when they are older and they also have an adequate enzyme composition of meat. 4. Conclusions An HL of CPFM significantly increased the absolute and relative proportions of the loin, the chin, and the belly rib part and decreased the less valuable part proportion in all weight groups. If compared to pigs fed with an LL of CPFM, a significantly higher proportion of muscular tissue was determined in the hams, loins, shoulders, necks, and belly rib parts of pigs fed a meal with an HL of CPFM. An increase in the final BWs led to an increase in the absolute proportions of the ham, loin, chin, adipose tissue, and less valuable parts as well as to an increase in the relative proportions of the loin and adipose tissue within all weight groups. No significant differences were detected in terms of pH value, kulen color values (L*, a*, and b*), and the content of NaCl, water, and ash in kulens, considering the final BWs of pigs, but it was determined that, as the final BWs of pigs increased, the CPFM content in kulens decreased very significantly. The level of CPFM did not significantly affect the sensory properties of kulens, but the increase in pig BW very significantly improved kulen taste. In general, by selecting the final BWs of pigs and specific feeding strategies for pigs, it is possible to modify the conformation and composition of pig carcasses and the quality of meat and kulens produced from the Black Slavonian pig. The obtained results are important because consumers prefer meat and meat products of specific characteristics and a standard quality and are ready to pay for them and, by selecting preferred meat characteristics, producers can ensure this. Author Contributions Conceptualization, D.S.; methodology, D.S. and Đ.S.; formal analysis, D.S., Z.A., Đ.S. and J.N.; data curation, D.S.; writing—original draft preparation, D.S.; writing—review and editing, Đ.K., I.P., Z.S., Ž.K.Š. and M.R.; visualization, Đ.K. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in compliance with the legal regulations set by the Animal Protection Act of Croatia (NN 133/06, NN 37/13, and NN 125/13) and the Euro-pean Union Directive 2010/63/UE regarding animal protection, as approved by the Bioethics Committee for Research on Animals of the Faculty of Agrobiotechnical Sciences, Osijek (2158-94-02-22-01). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The data presented in this study are available on request from the corresponding author. Conflicts of Interest The authors declare no conflict of interest. foods-11-01313-t001_Table 1 Table 1 The mixing ratio of grains and superconcentrate “35” in the experimental fattening. Fattening Period Fodder Mixture Proportion (%) CPFM Proportion (%) Metabolic Energy (MJ/kg) HL of CP 1st (25−60 kg) Corn 78.00 6.31 Super “35” 22.00 7.70 100.00 14.01 13.37 2nd (60−100, 120, or 130 kg) Corn 68.00 5.54 Barley 20.00 2.14 Super “35” 12.00 4.20 100.00 11.88 13.34 LL of CP 1st (25−60 kg) Corn 85.00 6.88 Super “35” 15.00 5.25 100.00 12.13 13.30 2nd (60−100, 120, or 130 kg) Corn 70.00 5.67 Barley 25.00 2.67 Super “35” 5.00 1.75 100.00 10.09 13.25 HL of CP—high level of crude protein; LL of CP—lower level of crude protein. foods-11-01313-t002_Table 2 Table 2 Basic indicators of the quality of pig carcasses. Indicators A B C D E F x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s Pig BW (kg) 100.62 A ± 1.36 100.75 a ± 1.24 120.37 B ± 1.36 121.12 b ± 2.02 131.00 C ± 1.63 131.37 c ± 0.88 Cold carcass mass (kg) 38.30 A ± 1.12 37.07 a ± 0.56 48.06 B ± 0.95 47.85 b ± 0.58 53.43 C ± 1.11 52.93 c ± 0.36 Half length (cm) Os pubis−atlas 97.19 A ± 3.45 95.25 a ± 1.73 101.12 AB ± 1.20 100.25 b ± 1.87 104.37 B ± 1.15 106.62 c ± 9.32 Half length (cm), Os pubis−1st rib 78.25 A ± 2.74 78.19 a ± 2.71 85.31 B ± 1.40 95.65 b ± 13.45 84.31 xAB ± 1.96 93.69 yb ± 4.44 Volume of leg (cm) 64.87 xA ± 1.09 62.56 ya ± 2.34 67.12 B ± 1.20 68.25 b ± 2.32 70.94 C ± 1.06 72.12 c ± 1.84 Leg length (cm) 27.19 A ± 2.34 26.50 a ± 1.86 28.50 xAB ± 0.89 34.94 yb ± 5.30 30.19 xB ± 1.90 40.69 yc ± 0.48 Fat thickness (cm) 3.30 A ± 0.37 3.00 a ± 0.53 3.66 xA ± 0.50 4.26 yb ± 0.37 4.85 xB ± 0.21 5.53 yc ± 0.43 Sectional area MLD-a (cm2) 34.62 A ± 0.72 34.50 a ± 1.03 33.12 B ± 0.96 32.87 b ± 0.96 32.62 B ± 1.86 32.25 b ± 1.24 x,y—different letters differ significantly (p < 0.05) within the final body weight level group (100 kg, 120 kg, and 130 kg). A,B,C—different letters differ significantly (p < 0.05) between 1, 3, and 5 groups (fed with 14% CPFM during the first fattening period and with 12% CPFM during the second fattening period). a,b,c—different letters differ significantly (p < 0.05) between 2, 4, and 6 groups (fed with 12% CPFM during the first fattening period and with 10% CPFM during the second fattening period). foods-11-01313-t003_Table 3 Table 3 Shares of the basic parts of the pig carcasses (kg). Part of Half Groups of Pigs A B C D E F x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s Ham (kg) 9.75 A ± 0.35 9.69 a ± 1.23 12.82 B ± 0.93 12.52 b ± 0.61 13.64 C ± 0.44 13.39 c ± 0.47 Loin (kg) 5.16 xA ± 0.30 6.27 ya ± 0.76 7.60 xB ± 0.78 8.58 yb ± 1.11 8.90 xC ± 1.37 11.23 yc ± 0.39 Shoulder (kg) 5.87 xA ± 0.27 4.82 ya ± 0.35 5.53 xA ± 0.65 6.45 yb ± 0.66 6.78 B ± 0.57 6.78 b ± 0.23 Neck (kg) 2.78 A ± 0.25 4.27 a ± 1.48 5.52 B ± 1.85 5.44 a ± 1.35 4.78 xC ± 1.27 6.79 yb ± 0.32 Belly rib part (kg) 9.36 xA ± 0.41 6.72 yab ± 1.37 9.96 xA ± 0.47 7.57 ya ± 2.68 11.32 xB ± 0.44 6.11 yb ± 0.15 Chin (kg) 1.12 xA ± 0.18 0.79 ya ± 0.24 1.13 A ± 0.31 1.36 b ± 0.34 1.57 B ± 0.32 1.48 b ± 0.16 Fat (kg) 1.28 A ± 0.16 1.08 a ± 0.23 1.63 B ± 0.45 1.51 b ± 0.38 2.11 xC ± 0.45 2.59 yc ± 0.17 Less valuable parts (kg) 2.96 A ± 0.11 3.43 a ± 0.69 3.87 B ± 0.40 4.42 b ± 0.96 4.33 B ± 0.59 4.56 b ± 0.36 x,y—different letters differ significantly (p < 0.05) within the final body weight level group (100 kg, 120 kg, and 130 kg). A,B,C—different letters differ significantly (p < 0.05) between 1, 3, and 5 groups (fed with 14% CPFM during the first fattening period and with 12% CPFM during the second fattening period). a,b,c—different letters differ significantly (p < 0.05) between 2, 4, and 6 groups (fed with 12% CPFM during the first fattening period and with 10% CPFM during the second fattening period). foods-11-01313-t004_Table 4 Table 4 Shares of the basic parts of the pig carcasses (%). Part of Half Groups of Pigs A B C D E F x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s Ham (%) 25.45 ± 0.60 26.11 ± 1.41 26.66 ± 1.70 26.17 ± 1.09 25.53 ± 0.84 25.30 ± 0.95 Loin (%) 13.46 A ± 0.69 15.50 a ± 6.09 15.82 AB ± 1.69 17.92 a ± 2.18 16.66 B ± 2.58 21.22 b ± 0.73 Shoulder (%) 15.33 xA ± 0.62 13.10 y ± 1.02 11.53 xB ± 1.49 13.48 y ± 1.35 12.67 C ± 0.84 12.80 ± 0.35 Neck (%) 7.26 xA ± 0.65 12.29 y ± 2.72 11.45 B ± 3.75 11.38 ± 2.86 8.97 xAB ± 2.47 12.83 y ± 0.52 Belly rib part (%) 24.45 xA ± 0.99 18.09 ya ± 2.76 20.74 xB ± 1.09 15.83 ya ± 5.63 21.19 xB ± 0.81 11.53 yb ± 0.22 Chin (%) 2.95 xAB ± 0.38 2.17 ya ± 0.70 2.34 A ± 0.65 2.84 b ± 0.71 2.95 B ± 0.55 2.81 b ± 0.32 Fat (%) 3.36 ± 0.35 2.90 a ± 0.47 3.39 ± 0.96 3.14 a ± 0.80 3.94 x ± 0.79 4.90 yb ± 0.33 Less valuable parts (%) 7.74 x ± 0.26 9.84 y ± 1.03 8.07 x ± 0.78 9.24 y ± 0.07 8.09 ± 1.20 8.61 ± 0.66 x,y—different letters differ significantly (p < 0.05) within the final body weight level group (100 kg, 120 kg, and 130 kg). A,B,C—different letters differ significantly (p < 0.05) between 1, 3, and 5 groups (fed with 14% CPFM during the first fattening period and with 12% CPFM during the second fattening period). a,b,c—different letters differ significantly (p < 0.05) between 2, 4, and 6 groups (fed with 12% CPFM during the first fattening period and with 10% CPFM during the second fattening period). foods-11-01313-t005_Table 5 Table 5 Shares of tissue in the carcass. Type of Tissue Groups of Pigs A B C D E F x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s Muscle tissue kg 18.55 xA ± 0.90 17.14 ya ± 1.77 22.78 xB ± 1.29 20.71 yb ± 1.36 22.65 xB ± 0.64 20.53 yb ± 1.30 % 48.47 A ± 2.57 46.29 a ± 1.88 47.37 xA ± 2.52 43.29 yb ± 2.93 42.39 xB ± 1.76 38.76 yc ± 2.21 Adipose tissue kg 13.68 A ± 1.23 12.87 a ± 1.76 17.08 B ± 1.85 18.29 b ± 1.80 22.19 C ± 2.05 23.08 c ± 1.52 % 35.69 A ± 2.62 34.69 a ± 2.09 35.55 A ± 4.00 38.21 b ± 3.55 41.49 B ± 3.27 43.64 c ± 3.16 Bone tissue kg 3.09 A ± 0.12 3.63 ± 0.60 4.34 B ± 0.53 4.43 ± 0.67 4.26 ± 0.47 4.76 B ± 0.66 % 8.10 x ± 0.26 9.82 y ± 0.96 9.01 ± 0.99 9.26 ± 1.40 8.03 ± 0.97 8.99 ± 1.21 x,y—different letters differ significantly (p < 0.05) within the final body weight level group (100 kg, 120 kg, and 130 kg). A,B,C—different letters differ significantly (p < 0.05) between 1, 3, and 5 groups (fed with 14% CPFM during the first fattening period and with 12% CPFM during the second fattening period). a,b,c—different letters differ significantly (p < 0.05) between 2, 4, and 6 groups (fed with 12% CPFM during the first fattening period and with 10% CPFM during the second fattening period). foods-11-01313-t006_Table 6 Table 6 Shares of muscle tissue of the basic parts in the weight of the pig carcasses. Muscle Tissue of the Basic Parts of Halves Groups of Pigs A B C D E F x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s Ham (%) 15.57 A ± 0.97 15.19 a ± 1.31 15.58 xA ± 1.95 14.11 yab ± 1.40 12.85 xB ± 1.05 13.62 yb ± 0.72 Loin (%) 6.93 ± 0.87 7.15 ± 1.61 6.70 ± 0.62 7.36 ± 1.35 6.38 x ± 0.86 7.94 y ± 0.37 Shoulder (%) 9.58 xA ± 0.38 7.68 ya ± 0.85 6.66 xB ± 1.07 7.93 ya ± 0.62 7.08 B ± 0.38 6.87 b ± 0.63 Neck (%) 4.92 A ± 0.36 6.31 a ± 1.42 7.15 B ± 2.10 6.87 ab ± 1.65 5.28 xA ± 1.23 7.79 yb ± 0.74 Belly rib part (%) 11.47 ± 0.93 9.94 a ± 1.33 11.28 x ± 1.02 7.00 yb ± 4.29 10.79 x ± 0.20 2.55 yc ± 0.08 x,y—different letters differ significantly (p < 0.05) within the final body weight level group (100 kg, 120 kg, and 130 kg). A,B,C—different letters differ significantly (p < 0.05) between 1, 3, and 5 groups (fed with 14% CPFM during the first fattening period and with 12% CPFM during the second fattening period). a,b,c—different letters differ significantly (p < 0.05) between 2, 4, and 6 groups (fed with 12% CPFM during the first fattening period and with 10% CPFM during the second fattening period). foods-11-01313-t007_Table 7 Table 7 Shares of adipose tissue of the basic parts in the weight of the pig carcasses. Adipose Tissue of the Basic Parts of Halves Groups of Pigs A B C D E F x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s Ham (%) 7.47 A ± 0.96 8.12 a ± 0.54 8.40 A ± 0.95 9.44 b ± 1.19 10.33 xB ± 1.27 9.03 yb ± 1.40 Loin (%) 4.93 xA ± 0.33 7.75 ya ± 1.40 7.29 B ± 1.58 8.32 a ± 1.19 8.51 xB ± 2.06 10.98 yb ± 0.58 Shoulder (%) 4.17 A ± 0.52 3.68 ± 0.51 3.24 B ± 0.86 3.78 ± 0.94 4.11 A ± 0.54 4.25 ± 0.40 Neck (%) 1.30 xA ± 0.27 3.20 y ± 1.23 2.72 B ± 1.34 2.86 ± 0.78 2.53 ± 0.98 3.42 B ± 0.63 Belly rib part (%) 11.51 xA ± 1.04 6.85 ya ± 1.44 8.16 B ± 0.68 7.81 ab ± 1.35 9.14 B ± 0.70 8.26 b ± 0.15 Chin (%) 2.94 xAB ± 0.38 2.17 ya ± 0.70 2.34 A ± 0.65 2.85 b ± 0.71 2.95 B ± 0.55 2.81 b ± 0.32 Fat (%) 3.36 ± 0.35 2.90 a ± 0.47 3.39 ± 0.96 3.14 a ± 0.80 3.94 x ± 0.79 4.89 yb ± 0.33 x,y—different letters differ significantly (p < 0.05) within the final body weight level group (100 kg, 120 kg, and 130 kg). A,B—different letters differ significantly (p < 0.05) between 1, 3, and 5 groups (fed with 14% CPFM during the first fattening period and with 12% CPFM during the second fattening period). a,b—different letters differ significantly (p < 0.05) between 2, 4, and 6 groups (fed with 12% CPFM during the first fattening period and with 10% CPFM during the second fattening period). foods-11-01313-t008_Table 8 Table 8 Shares of bone tissue of the basic parts in the weight of pig carcasses. Bone Tissue of the Basic Parts of Halves Groups of Pigs A B C D E F x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s Ham (%) 2.41 ± 0.08 2.80 ± 0.40 2.68 ± 0.36 2.61 ± 0.45 2.34 ± 0.41 2.65 ± 0.51 Loin (%) 1.60 x ± 0.22 2.22 y ± 0.29 1.82 x ± 0.36 2.24 y ± 0.46 1.77 x ± 0.08 2.30 y ± 0.06 Shoulder (%) 1.57 ± 0.07 1.73 ± 0.13 1.63 ± 0.18 1.75 ± 0.27 1.49 ± 0.10 1.69 y ± 0.27 Neck (%) 1.04 x ± 0.10 1.78 y ± 0.52 1.58 ± 0.46 1.65 ± 0.52 1.16 ± 0.35 1.62 ± 0.54 Belly rib part (%) 1.47 A ± 0.23 1.29 a ± 0.23 1.30 AB ± 0.18 1.00 b ± 0.24 1.24 B ± 0.31 0.72 c ± 0.04 x,y—different letters differ significantly (p < 0.05) within the final body weight level group (100 kg, 120 kg, and 130 kg). A,B,C—different letters differ significantly (p < 0.05) between 1, 3, and 5 groups (fed with 14% CPFM during the first fattening period and with 12% CPFM during the second fattening period). a,b,c—different letters differ significantly (p < 0.05) between 2, 4, and 6 groups (fed with 12% CPFM during the first fattening period and with 10% CPFM during the second fattening period). foods-11-01313-t009_Table 9 Table 9 Physical properties of meat. Indicators Groups of Pigs A B C D E F x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s pH1 6.72 ± 0.03 6.74 ± 0.03 6.75 ± 0.03 6.72 ± 0.03 6.71 ± 0.04 6.71 ± 0.04 pH2 5.81 ± 0.05 5.79 ± 0.05 5.80 ± 0.04 5.79 ± 0.06 5.79 ± 0.04 5.78 ± 0.03 Water-holding capacity (cm2) 4.34 ± 0.04 4.34 ± 0.02 4.35 ± 0.03 4.33 ± 0.04 4.34 ± 0.02 4.33 ± 0.03 Consistency (cm2) 2.47 ± 0.12 2.51 ± 0.14 2.49 ± 0.14 2.46 ± 0.10 2.50 ± 0.10 2.45 ± 0.11 Color (CIE L*) 51.74 ± 1.96 55.10 a ± 3.74 53.75 ± 3.25 52.75 a ± 5.08 52.80 x ± 1.74 48.99 yb ± 0.76 Color (CIE a*) 17.30 x ± 0.70 18.90 ya ± 0.77 18.43 x ± 1.36 19.91 ya ± 1.68 18.28 x ± 0.80 23.02 yb ± 1.22 Color (CIE b*) 4.58 A ± 0.80 5.74 a ± 0.91 5.80 B ± 1.06 5.67 a ± 1.08 5.08 xAB ± 0.87 6.86 yb ± 1.04 x,y—different letters differ significantly (p < 0.05) within the final body weight level group (100 kg, 120 kg, and 130 kg). A,B—different letters differ significantly (p < 0.05) between 1, 3, and 5 groups (fed with 14% CPFM during the first fattening period and with 12% CPFM during the second fattening period). a,b—different letters differ significantly (p < 0.05) between 2, 4, and 6 groups (fed with 12% CPFM during the first fattening period and with 10% CPFM during the second fattening period). foods-11-01313-t010_Table 10 Table 10 Chemical composition of the meat. Indicators Groups of Pigs A B C D E F x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s Water (%) 67.78 A ± 1.50 67.51 a ± 1.54 63.05 B ± 3.78 62.80 b ± 3.45 62.31 B ± 3.09 62.04 b ± 3.74 Crude protein (%) 24.18 A ± 1.03 23.76 a ± 1.19 20.21 B ± 1.43 20.00 b ± 1.31 20.15 B ± 1.27 20.07 b ± 1.35 Crude fat (%) 6.97 A ± 1.37 7.70 a ± 1.93 15.82 B ± 5.32 16.30 b ± 4.72 16.62 B ± 4.54 16.98 b ± 4.88 Ash (%) 1.07 A ± 0.02 1.03 a ± 0.03 0.92 B ± 0.06 0.90 b ± 0.06 0.92 B ± 0.05 0.91 b ± 0.06 A,B—different letters differ significantly (p < 0.05) between 1, 3, and 5 groups (fed with 14% CPFM during the first fattening period and with 12% CPFM during the second fattening period). a,b—different letters differ significantly (p < 0.05) between 2, 4, and 6 groups (fed with 12% CPFM during the first fattening period and with 10% CPFM during the second fattening period). foods-11-01313-t011_Table 11 Table 11 Physical and chemical properties of kulens. Indicators Groups of Pigs A B C D E F x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s pH 5.92 ± 0.11 5.96 ± 0.08 5.96 ± 0.16 5.92 ± 0.05 5.97 ± 0.15 5.86 ± 0.05 Water activity (aw) 0.87 A ± 0.01 0.84 a ± 0.03 0.81 B ± 0.03 0.79 b ± 0.02 0.79 B ± 0.03 0.83 ab ± 0.06 Color (CIE L*) 35.38 ± 1.25 35.65 ± 1.95 36.66 ± 1.67 36.18 ± 1.51 36.72 ± 1.86 36.21 ± 1.39 Color (CIE a*) 17.08 ± 0.80 17.27 ± 0.16 17.10 ± 0.99 17.57 ± 0.74 17.36 ± 1.06 17.60 ± 0.66 Color (CIE b*) 9.27 ± 1.01 9.42 ± 0.62 10.00 ± 1.23 9.80 ± 0.85 10.11 ± 1.23 9.87 ± 0.69 NaCl (%) 5.22 ± 0.12 5.18 ± 0.09 5.01 ± 0.47 5.31 ± 0.29 5.22 ± 0.43 5.42 ± 0.20 Water (%) 31.45 ± 0.56 31.01 ± 0.86 31.39 ± 3.07 32.05 ± 2.65 32.10 ± 3.20 32.61 ± 2.64 Crude protein (%) 45.94 A ± 0.19 45.74 a ± 0.71 43.89 B ± 1.63 43.39 b ± 1.15 43.59 B ± 1.73 43.00 b ± 1.09 Crude fat (%) 16.24 xA ± 0.75 17.47 ya ± 0.33 18.51 B ± 0.99 18.61 b ± 1.03 18.24 B ± 0.97 18.33 ab ± 1.04 Ash (%) 6.37 x ± 0.56 5.78 y ± 0.14 6.21 ± 0.60 5.95 ± 0.54 6.07 ± 0.62 5.97 ± 0.52 x,y—different letters differ significantly (p < 0.05) within the final body weight level group (100 kg, 120 kg, and 130 kg). A,B—different letters differ significantly (p < 0.05) between 1, 3, and 5 groups (fed with 14% CPFM during the first fattening period and with 12% CPFM during the second fattening period). a,b—different letters differ significantly (p < 0.05) between 2, 4, and 6 groups (fed with 12% CPFM during the first fattening period and with 10% CPFM during the second fattening period). foods-11-01313-t012_Table 12 Table 12 Sensory properties of kulens. Indicators Groups of Pigs A B C D E F x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s x¯ ± s Appearance (1–5) 4.42 ± 0.36 4.55 ± 0.28 4.50 ± 0.32 4.65 ± 0.20 4.52 ± 0.38 4.40 ± 0.18 Structure (1–3) 2.67 ± 0.21 2.80 ± 0.21 2.75 ± 0.23 2.92 ± 0.24 2.76 ± 0.26 2.70 ± 0.26 Cross-section appearance (1−10) 8.30 ± 0.32 8.60 ± 0.34 8.50 ± 0.32 8.62 ± 0.25 8.42 ± 0.31 8.45 ± 0.21 Odor (1−5) 4.65 ± 0.09 4.55 ± 0.35 4.60 ± 0.30 4.65 ± 0.14 4.75 ± 0.23 4.67 ± 0.21 Taste (1−10) 8.75 A ± 0.18 8.80 a ± 0.15 9.32 B ± 0.34 9.30 b ± 0.28 9.52 B ± 0.21 9.37 b ± 0.17 General impression (1−5) 4.20 ± 0.21 4.30 ± 0.24 4.37 ± 0.31 4.45 ± 0.18 4.42 ± 1.20 4.35 ± 0.26 A,B—different letters differ significantly (p < 0.05) between 1, 3, and 5 groups (fed with 14% CPFM during the first fattening period and with 12% CPFM during the second fattening period). a,b—different letters differ significantly (p < 0.05) between 2, 4, and 6 groups (fed with 12% CPFM during the first fattening period and with 10% CPFM during the second fattening period). Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Latin K. Petrić T. Raguž N. Karolyi D. Klišanić V. Menčik S. Mahnet Ž. Lukić B. Body measurements of Black Slavonian pig Stočarstvo 2021 75 3 12 10.33128/s.75.1-2.1 2. FAO The State of the World’s Animal Genetic Resources for Food and Agriculture FAO Rome, Italy 2007 3. Senčić Đ. Bukvić Ž. Antunović Z. Šperanda M. Slaughter quality of Black Slavonian pig—Endangered breed and its cross-breeds with Swedish landrace while keeping them outdoor Poljoprivreda/Agriculture 2005 11 43 48 4. Karolyi D. Luković Z. Salajpal K. Đikić M. Black Slavonian pig—A breed for extensive husbandry (A review) Acta Agrar. Kaposvariensis 2010 14 221 227 5. Budimir K. Margeta V. Margeta P. Breeding of the Black Slavonian Pig in extensive conditions Acta Agric. Slov. 2013 4 93 96 6. Čačić M. Orehovački V. Vukobratović M. Dražić M.M. Smetko A. Pavlešić T. Čubrić Čurik V. Čurik I. The role of gene banks to conserve indigenous breeds of farm animals Stočarstvo 2015 69 85 92 7. Dupont F. Effect of geographic indications Proceedings of the Workshop on “Geographic Indications for Middle Eastern and North African Agri-Food Products“ Montpellier, France 7–10 June 2004 8. Budimir K. Djurkin Kušec I. Lukić B. Džijan S. Margeta V. Kušec G. Possibilities of branding the pork in Croatia—Review Acta Agrar. Kaposvariensis 2014 18 115 121 9. Kovačić D. Radman M. Kolega A. Markovina J. Karolyi D. Export Marketing of Slavonian Kulen: The Results of Market Research Agric. Conspec. Sci. 2003 68 185 191 10. Uremović M. Uremović Z. Luković Z. Production properties of the Black Slavonian pig breed. Zb. Biotehniške fak. Univ. v Ljubljani. Kmetijstvo Zootehnika 2000 76 131 134 11. The Convention on Biological Diversity (CBD) 1992 Available online: https://www.cbd.int/ (accessed on 20 January 2022) 12. Pregled Stanja Biološke i Krajobrazne Raznolikosti Hrvatske sa Strategijom i Akcijskim Planom Zaštite Državna Uprava za Zaštitu Prirode RH (DUZZP RH) Zagreb, Croatia 1999 65 117 13. Uremović M. Janeš M. Uzgojni Program Crne Slavonske Pasmine Svinja Hrvatski Stočarski Selekcijski Centar Zagreb, Croatia 2000 1 10 14. Djurkin Kušec I. Buha I. Margeta V. Gvozdanović K. Radišić Ž. Komlenić M. Kušec G. Carcass Composition and Meat Quality of Crna Slavonska Pigs from Two Different Housing Conditions Agric. Conspec. Sci. 2017 82 221 225 15. Druml T. Salajpal K. Dikić M. Urosevic M. Grilz-Seger G. Baumung R. Genetic diversity, population structure and subdivision of local Balkan pig breeds in Austria, Croatia, Serbia and Bosnia-Herzegovina and its practical value in conservation programs Genet. Sel. Evol. 2012 44 5 10.1186/1297-9686-44-5 22376364 16. Škorput D. Špehar M. Luković Z. Connectedness between contemporary groups in Black Slavonian pig Livest. Sci. 2018 216 6 8 10.1016/j.livsci.2018.06.012 17. Gvozdanović K. Margeta V. Margeta P. Djurkin Kušec I. Galović D. Dovč P. Kušec G. Genetic diversity of autochthonous pig breeds analyzed by microsatellite markers and mitochondrial DNA D-loop sequence polymorphism Anim. Biotechnol. 2019 30 242 251 10.1080/10495398.2018.1478847 29938587 18. Gvozdanović K. Djurkin Kušec I. Margeta P. Salajpal K. Džijan S. Bošnjak Z. Kušec G. Multiallelic marker system for traceability of Black Slavonian pig meat Food Control 2020 109 106917 10.1016/j.foodcont.2019.106917 19. Senčić Đ. Butko D. Antunović Z. Novoselec J. Influence of Pig Body Mass on Meat and Carcass Quality of Black Slavonian Pig Proceedings of the 7th International Symposium of Animal Biology and Nutrition Bucharest, Romania 25–26 September 2008 20. Pleadin J. Kovačević D. Perši N. Ochratoxin A contamination of the autochthonous dry-cured meat product “Slavonski Kulen” during a six-month production process Food Control 2015 57 377 384 10.1016/j.foodcont.2015.05.013 21. Pleadin J. Zadravec M. Brnić D. Perković I. Škrivanko M. Kovačević D. Moulds and mycotoxins detected in the regional speciality fermented sausage ‘slavonski kulen’ during a 1-year production period Food Addit. Contam. Part A 2016 34 282 290 10.1080/19440049.2016.1266395 27899057 22. Mastanjević K. Kovačević D. Frece J. Markov K. Pleadin J. The Effect of Autochthonous Starter Culture, Sugars and Temperature on the Fermentation of Slavonian Kulen Food Technol. Biotechnol. 2017 55 67 76 10.17113/ftb.55.01.17.4688 28559735 23. Lešić T. Zadravec M. Zdolec N. Vulić A. Perković I. Škrivanko M. Kudumija N. Jakopović Ž. Pleadin J. Mycobiota and Mycotoxin Contamination of Traditional and Industrial Dry-Fermented Sausage Kulen Toxins 2021 13 798 10.3390/toxins13110798 34822582 24. Marušić Radovčić N. Poljanec I. Vidinski P. Novina K. Medić H. Influence of different pig genotype on aroma, colour and fatty acid composition of smoked dry-cured ham Meso 2019 21 548 561 10.31727/m.21.6.4 25. Lešić T. Vahčić N. Kos I. Zadravec M. Sinčić Pulić B. Bogdanović T. Petričević S. Listeš E. Škrivanko M. Pleadin J. Characterization of Traditional Croatian Household-Produced Dry-Fermented Sausages Foods 2020 9 990 10.3390/foods9080990 26. Weniger H.I. Steinhauf D. Pahl G. Topography of Carcasses BLV Verlagsgesellschaft Munich, Germany 1963 27. Comberg G. Behrens H. Bollwahn W. Fielder E. Glodek P. Kallweit E. Meyer H. Stephan E. Schweinezucht Eugen Ulmer Verlag Stuttgart, Germany 1978 28. Grau R. Hamm R. Eine einfache Methode zur Bestimmung der Wasser bildung in Fleisch Fleischwirtschaft 1952 4 295 297 29. Commision International de l’Eclairage Colorimetry 3rd ed. Publication CIE 15:2004 CIE Vienna, Austria 1976 30. American Public Health Association (APHA) Standard Methods for the Examination of Water and Wastewater 29th ed. American Public Health Association Washington, DC, USA 1998 31. HRN ISO 1443:1999 Meat and Meat Products—Determination of Total Fat Content ISO Geneva, Switzerland 1973 32. V-3-32 Internal Methods for Examination of NaCl Veterinary Institute Vinkovci, Croatia 33. Benito J. Vázquez C. Menaya C. Ferrera J.L. García Casco J.M. Silió L. Rodrigáñez J. Rodríguez M.C. Evaluation of the productive parameters in different strains of Iberian pigs Tradition and Innovation in Mediterranean Pig Production Afonso J.A. Tirapicos J.L. CIHEAM Zaragoza, Spain 2000 113 121 34. Gonzalez J. Gispert M. Rodríguez P. Gil M. Jaume J. Tibau J. Oliver M.A. Carcass and meat quality of porc negre mallorquí (Majorcan Black Pig) Proceedings of the 6th International Symposium on the Mediterranean Pig Messin-Capo d’Orlando, Italy 11–13 October 2007 35. Ramírez R. Cava R. Carcass composition and meat quality of three different Iberian x Duroc genotype pigs Meat Sci. 2007 75 388 396 10.1016/j.meatsci.2006.08.003 22063794 36. García-Valverde R. Barea R. Lara L. Nieto R. Aguilera J.F. The effect of feeding level upon protein and fat deposition in Iberian heavy pigs Livest. Sci. 2008 114 263 273 10.1016/j.livsci.2007.05.005 37. Samac D. Senčić Đ. Antunović Z. Steiner Z. Novoselec J. Klarić I. Kvaliteta svježe svinjetine od crne slavonske svinje u odnosu na tjelesnu masu Proceedings of the 45th Croatian & 5th International Symposium of Agriculture Opatija, Croatia 15–19 February 2010 231 232 38. Gallo L. Dalla Montà G. Carraro L. Cecchinato A. Carnier P. Schiavon S. Carcass quality and uniformity of heavy pigs fed restrictive diets with progressive reductions in crude protein and indispensable amino acids Livest. Sci. 2015 172 50 58 10.1016/j.livsci.2014.11.014 39. Barea R. Nieto R. Lara L. Garcia M.A. Vichez M.A. Aguiler J.F. Effects of dietary protein content and feeding level on carcass characteristics and weights of Iberian pigs growing between 50 and 100 kg live weight Anim. Sci. 2008 82 405 413 10.1079/ASC200645 40. Teye G.A. Sheard P.R. Whittington F.M. Nute G.R. Steward A. Wood J.D. Influence of dietary oils and protein level on pork quality. 1. Effect on muscle fatty acid composition, carcass, meat and eating quality Meat Sci. 2006 73 157 165 10.1016/j.meatsci.2005.11.010 22062065 41. Millet S. Ongenae E. Hesta M. Seynaeve M. De Smet S. Janssens G.P.J. The feeding of ad libitum dietary protein to organic growing-finishing pigs Vet. J. 2006 7 483 490 10.1016/j.tvjl.2005.01.002 42. Wang Y. Zhou J. Wang G. Cai S. Zeng X. Qiao S. Advances in low-protein diets for swine J. Anim. Sci. Biotechnol. 2018 9 60 10.1186/s40104-018-0276-7 30034802 43. Tous N. Lizardo R. Vilà B. Gispert M. Font-i-Furnols M. Esteve-Garcia E. Effect of reducing dietary protein and lysine on growth performance, carcass characteristics, intramuscular fat, and fatty acid profile of finishing barrows J. Anim. Sci. 2014 92 129 140 10.2527/jas.2012-6222 24352966 44. Wu F. Vierck K.R. DeRouchey J.M. O’Quinn T.G. Tokach M.D. Goodband R.D. Dritz S.S. Woodworth J.C. A review of heavy weight market pigs: Status of knowledge and future needs assessment Transl. Anim. Sci. 2017 1 1 15 10.2527/tas2016.0004 45. Hofmann K. What is quality? Measurement and evaluation of meat quality Meat Focus Int. 1994 3 73 82 46. Virgili R. Schivazappa C. Muscle traits for long natured dried meats Meat Sci. 2002 62 331 343 10.1016/S0309-1740(02)00118-3 22061609 47. Karolyi D. Promjene u kvaliteti mesa Meso 2004 6 18 20 48. Russunen M. Partanen K. Pösö R. Puolanne E. The effect of dietary protein supply on carcass composition, size of organs, muscle properties and meat quality of pigs Livest. Sci. 2007 107 170 181 10.1016/j.livsci.2006.09.021 49. Hong J.S. Lee G.I. Jin X.H. Kim Y.Y. Effect of dietary energy levels and phase feeding by protein levels on growth performance, blood profiles and carcass characteristics in growing-finishing pigs J. Anim. Sci. Technol. 2016 58 37 10.1186/s40781-016-0119-z 27795835 50. Cisneros F. Ellis M. McKeith F.K. McCaw J. Fernando R.L. Influence of slaughter weight on growth and carcass characteristics, commercial cutting and curing yields, and meat quality of barrows and gilts from two genotypes J. Anim. Sci. 1996 74 925 933 10.2527/1996.745925x 8726722 51. Beattie V.E. Weatherup R.N. Moss B.W. Walker N. The effect of increasing carcass weight of finishing boars and gilts on joint composition and meat quality Meat Sci. 1999 52 205 211 10.1016/S0309-1740(98)00169-7 22062373 52. Xiong Y. Chen M. Warner R.D. Fang Z. Incorporating nisin and grape seed extract in chitosan-gelatine edible coating and its effect on cold storage of fresh pork Food Control 2020 110 107018 10.1016/j.foodcont.2019.107018 53. Joo S.T. Kaufman R.G. Kim B.C. Pork G.B. The relationship of sarcoplasmic and myofibrillar protein solubility to colour and water-holding capacity in porcine longissimus muscle Meat Sci. 1999 52 291 297 10.1016/S0309-1740(99)00005-4 22062578 54. Galián M. Poto A. Peinado B. Carcass and meat quality traits of the Chato Murciano pig slaughtered at different weights Livest. Sci. 2009 124 314 320 10.1016/j.livsci.2009.02.012 55. Piao J.R. Tian J.Z. Kim B.G. Choi Y.I. Kim Y.Y. Han I.K. Effects of sex and market weight on growth performance carcass characteristics and pork quality of market hogs Asian-Australas. J. Anim. Sci. 2004 10 1452 1458 10.5713/ajas.2004.1452 56. Lebret B. Effect of feeding and rearing systems on growth, carcass composition and meat quality in pigs Animal 2008 2 1548 1558 10.1017/S1751731108002796 22443914 57. Alonso V. Mar Campo M. Provincial L. Roncalés P. Beltrán J.A. Effect of protein level in commercial diets on pork meat quality Meat Sci. 2010 85 7 14 10.1016/j.meatsci.2009.11.015 20374857 58. Čandek-Potokar M. Zlender B. Bonneau M. Effects of breed and slaughter weight on longissimus muscle biochemical traits and sensory quality in pigs Ann. Zootech. 1998 47 3 16 10.1051/animres:19980101 59. Djurkin I. Dadić M. Brkić D. Lukić B. Kušec G. Mikolin M. Jerković I. Influence of gender and slaughter weight on meat quality traits of heavy pigs Acta Agric. Slov. 2012 3 211 214 60. Kovačević D. Suman K. Šubarić D. Mastanjević K. Vidaček S. Investigation of homogeneity and physicochemical characterization of the Homemade Slavonian Sausage Meso 2010 6 338 344 61. Karolyi D. Fizikalno-kemijska, higijenska i organoleptička karakterizacija slavonskog kulena Meso 2011 6 423 429 62. Ambrosiadis J. Soultos N. Abrahim A. Bloukas J.G. Physicochemical, microbiological and sensory attributes for the characterization of Greek traditional sausages Meat Sci. 2004 66 279 287 10.1016/S0309-1740(03)00100-1 22064129 63. Moretti V.M. Madonia G. Diaferia C. Mentasti T. Paleari M.A. Panseri S. Pirone G. Gandini G. Chemical and microbiological parameters and sensory attributes of a typical Sicilian salami ripened in different conditions Meat Sci. 2004 66 845 854 10.1016/j.meatsci.2003.08.006 22061017

PMC009xxxxxx/PMC9099645.txt

==== Front Int J Mol Sci Int J Mol Sci ijms International Journal of Molecular Sciences 1422-0067 MDPI 10.3390/ijms23095094 ijms-23-05094 Article Acidic and Alkaline Conditions Affect the Growth of Tree Peony Plants via Altering Photosynthetic Characteristics, Limiting Nutrient Assimilation, and Impairing ROS Balance https://orcid.org/0000-0002-8796-0148 Aung Theint Thinzar 1† Shi Fengrui 1† Zhai Yanning 1 Xue Jingqi 1 https://orcid.org/0000-0002-2700-009X Wang Shunli 1 https://orcid.org/0000-0002-7429-6743 Ren Xiuxia 1* Zhang Xiuxin 12* Skriver Karen Academic Editor 1 Key Laboratory of Biology and Genetic Improvement of Horticultural Crops, Ministry of Agriculture and Rural Affairs, Institute of Vegetables and Flowers, Chinese Academy of Agricultural Sciences, Beijing 100081, China; theintthinzaraungyau.hc@gmail.com (T.T.A.); shifengrui1023@126.com (F.S.); zhaiyanning@caas.cn (Y.Z.); xuejingqi@caas.cn (J.X.); wangshunli@caas.cn (S.W.) 2 National Agricultural Science and Technology Center, Chengdu 610213, China * Correspondence: renxiuxia@caas.cn (X.R.); zhangxiuxin@caas.cn (X.Z.); Tel.: +86-010-8210-5944 (X.Z.) † These authors contributed equally to this work. 03 5 2022 5 2022 23 9 509406 4 2022 02 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Exposure to acidic and alkaline conditions were found to cause the excess accumulation of reactive oxygen species in tree peony, thereby causing damage and inhibiting plant growth and development. The activities of antioxidant enzymes were also found to be significantly up-regulated, especially under alkaline conditions; this explained why tree peony is better adapted to alkaline than to acidic conditions. Through pairwise comparisons, 144 differentially expressed genes (DEGs) associated with plant growth, photosynthesis, and stress were identified. The DEGs related to stress were up-regulated, whereas the remaining DEGs were almost all down-regulated after acid and alkaline treatments. The nutrient assimilation was greatly inhibited. Chlorophyll synthesis genes were suppressed, and chlorophyll content was reduced. The development and structures of stomata and chloroplasts and the transcription of related genes were also influenced. Among photosynthesis-related DEGs, electron transport chains were the most sensitive. The suppressed expression of photosynthesis genes and the reduced light-harvesting capacity, together with the impairment of chloroplasts and stomata, finally led to a sharp decrease in the net photosynthetic rate. Carbohydrate accumulation and plant biomass were also reduced. The present study provides a theoretical basis for the response mechanisms of tree peony to adverse pH conditions and enriches knowledge of plant adaptation to alkaline conditions. pH stress responses plant adaptability transcriptome analysis regulation network National Key R&D Program of China2020YFD1000500 eijing Municipal Natural Science Foundation6214045 National Key R&D Program of China2018YFD1000401 China Agriculture Research SystemCARS-21 This research was supported by the National Key R&D Program of China (2020YFD1000500), Beijing Municipal Natural Science Foundation (No. 6214045), the National Key R&D Program of China (2018YFD1000401), and China Agriculture Research System (CARS-21). ==== Body pmc1. Introduction Tree peony (Paeonia suffruticosa Andr.) is a famous Chinese traditional flowering plant referred to as ‘the king of flowers’, with more than 2000 years of cultivation history [1]. It is also famous worldwide due to its ornamental features and economic value [2]. The tree peony has been used as a medicinal plant since ancient times and at present has gained attention as an emerging oil plant [1,3]. Moreover, tree peony is widely used in landscaping, gardening, potted flower culturing, forcing culture, and oriental flower arranging. The rise in soil pH is one of the factors restricting the vegetative growth and development of tree peony. Hence, a systematic study of pH as it affects plant growth is urgently required to improve cultivation techniques for tree peony. Soil acidification is a major limiting factor for worldwide sustainable agricultural production. Acidic soil covers approximately 30–40% of the world’s arable land [4]. Soil alkalization is also a significant problem in China [5]. These adverse pH conditions induce the production of reactive oxygen species (ROS) in plant cells. This can cause damage to the plant (in the form of protein oxidization, destroyed nucleic acids, and lipid peroxidation). However, ROS are also involved in various cellular processes, including stress resistance, as signal molecules [6,7]. It has been reported that pH significantly affects nutrient uptake, root growth, flower quality, and other cellular processes [6,8]. To reduce the damage caused by stress, plants will activate several antioxidant enzymes to eliminate the excess ROS [9]. The activities of superoxide dismutase (SOD), peroxidase (POD), and catalase (CAT) are significantly increased when plants are exposed to stress for a long period of time [10], and those species or cultivars with stronger antioxidant enzyme systems exhibit higher tolerance to stress [11]. The process of photosynthesis in leaves is fundamental to biomass production and crop yields [12]. Like most stresses, adverse pH stress is toxic to the photosystem because it inhibits chlorophyll synthesis, destroys thylakoid membrane and chloroplast structure, and hinders photosynthetic electron transport chains (ETCs) [13,14,15]. Electrons become oversupplied when photosynthesis is inhibited under stress conditions, in turn causing excess ROS production and affecting plant growth [10]. In tea plants, a pH of 2.5 was shown to reduce the number of chloroplasts, alter stomatal density and size, and suppress gene expression related to photosynthesis and carbohydrate metabolites [16]. In soybeans, soil acidity caused the disorganization of grana lamellae and decreased the number of thylakoids [17]. Stomatal closure was enhanced and chlorophyll content was reduced under salt and alkali stress in alkaline grasslands, and the two factors together caused a reduction in the net photosynthesis rate (Pn) [18]. Moreover, an extremely high or low pH leads to nutrient element unavailability, ion imbalances, damage to plant membranes, and osmotic stress, thereby inhibiting nutrient absorption and thus affecting plant growth, photosynthesis, and plant disease resistance [14,15,19]. Transcriptome technology is an important method that can be used to reveal the molecular mechanisms of plant responses and adaptability to stress. Using transcriptome techniques, a total of 855 differentially expressed genes (DEGs) associated with photosynthesis, cell walls, and phenylpropanoid metabolism have been found in woody plants, among which the majority of the DEGs related to photosynthesis are up-regulated under optimal pH conditions, including PSI and PSII reaction centers, ATP synthase, the cytochrome b6/f complex, and photosynthetic ETCs [13]. Although research on pH and plant growth has made some progress, there has been systematic studies and even less research on how plants adapt to adverse pH conditions. However, these gaps can be filled through the present research. Two representative pH stress conditions, pH 4.0 and 10.0, were selected to study the effects of acidic and alkaline conditions on a series of physiological processes and the related gene expression patterns in tree peony. The responses included plant growth, flowering, chlorophyll synthesis, photosynthesis, stomatal development and movement, nutrient assimilation and transport, hormone synthesis and signal transduction, and ROS signaling, and metabolism elimination. The results are expected to reveal the mechanism of acidic and alkaline toxicity to tree peony plants and the mechanism of plant adaptation to pH stress. 2. Results 2.1. Morphological Parameters and Anthocyanin Content in Flowers The morphological flower parameters were observed at the four stages from the bud initiation stage to the flowering stages (Table 1 and Figure 1). The results showed that the developmental process of tree peony was delayed under acidic and alkaline treatments, especially under acidic conditions (Figure 1a). Flower growth parameters, including flower diameter and stalk length, gradually increased from bud initiation until the flowering stage (Figure 1b,c). Flower diameter and stalk length had no significant differences among the three treatments during the first three stages (S1–S3), while alkaline and acidic treatments greatly decreased the growth rates of flower diameter and stalk length at the flowering stage (S4) compared to the controls (pH 7.0). Flower diameter, flower height, stalk length, stalk diameter, flower biomass, and the number of petals were reduced under alkaline and acidic conditions compared to the controls (Table 1). Alkaline and acidic treatments also caused a rapid increase in the percentage of abnormal flowering, approximately 3.25 times and 4.13 times the percentage of controls in the acidic and alkaline stress groups, respectively (Table 1). Flower quality, especially the petal color, was also significantly affected. Color indices showed that under respective acidic and alkaline conditions, the L* value was enhanced by 14.5% and 8.34%, the a* value was decreased by 12.47% and 6.8%, and the C* value was reduced by 11.1% and 5.4%; additionally, the −b* value was increased by 11.6% compared to the controls (Figure 1e). The H° value showed no significant difference among treatments (Figure 1f). Anthocyanin content was significantly reduced under acidic conditions (15.44%) and not significantly under alkaline conditions (9.67%) (Figure 1d). The higher L* value and lower a* and C* values were responsible for the color fading of tree peony petals under acidic and alkaline conditions, and the effect was more serious under acidic conditions. 2.2. Morphological Parameters in Leaves and Roots As shown in Figure 2a, leaves were significantly smaller under acidic and alkaline treatments, and leaf color turned yellow under acidic stress conditions. The chlorophyll level slowly increased from S3 to S5 and then dramatically decreased (from S5 to S6) in treated leaves of tree peony, especially under acidic conditions (Figure 2b). Compared to the control group, the chlorophyll level was significantly lower under acidic and alkaline treatments from the S4 to S6 stages. It appeared that chlorophyll was degraded in the later growth stages, and degradation was more severe under acidic conditions. Total chlorophyll (a + b) content was decreased under acidic conditions, while chlorophyll a, b, and carotenoid had no significant differences among treatments at stage S4. The ratio of chlorophyll a/b was reduced under both the acidic and alkaline treatments (Figure 2c). The leaf area and biomass all gradually increased under the three treatments (Figure 2d,e), and compared to the control group, there were significant reductions in both measures in the acidic and alkaline groups at stage S4. As shown in Figure 2f, many fibrous roots rotted and the color turned abnormally dark brown in treated plants. The biomass of adult roots was slightly reduced, while the biomass of fibrous roots was significantly lower under acidic and alkaline conditions compared to the control group (Figure 2g,h): the fresh and dry weights of the roots were decreased by 23.6% and 7.25% under acidic treatment and by 13.9% and 15.16% under alkaline treatment compared to controls, respectively; the fresh and dry weights of new fibrous roots were decreased by 56.25% and 49.12% under acidic conditions and by 39.28% and 43.85%, respectively, under alkaline conditions compared to the controls. The root lengths in the acidic and alkaline groups were concentrated at around 1–3 cm (35% and 33.3%, respectively) and 4–5 cm (30% and 22.2%, respectively), while the root lengths in the control group were concentrated at around 9–10 cm (34.8%) and greater than 10 cm (30.4%) (Figure 2i). The diameter of adult roots was concentrated at around 4–6 cm (27.3%) and 6–8 cm (31.8%) in the control group. Root diameter was concentrated at around 6–8 cm (20%) and 8–10 cm (30%) under acidic conditions and was concentrated at around 6–8 cm (33.3%) under alkaline conditions (Figure 2j). 2.3. De novo Transcriptome Assembly, Identification of DEGs, and KEGG Pathway Analysis of DEGs After the removal of ambiguous reads, adapter sequences, and low-quality reads, a total of 132,167 unigenes were assembled with average sequence length of 698 bp, an N50 length of 1317 bp, and a GC percentage of 39.02%. Among these assembled unigenes, 44,376 and 41,017 were annotated by the NR and NT databases, respectively; 31,208 could be annotated in Swiss-Prot, 33,675 were annotated in KEGG, 33,428 (25.29%) were annotated using KOG, 30,503 (23.08%) were annotated in the Pfam database, and 33,482 could be annotated in GO. A total of 4574 DEGs were identified in tree peony leaves following exposure to acidic conditions (pH 4.0) compared to the control (pH 7.0), of which 2858 were up-regulated and 1716 were down-regulated, and a total of 5006 DEGs were identified in tree peony leaves following exposure to alkaline conditions (pH 10.0) compared to the control, of which 1327 were up-regulated and 3679 were down-regulated (Figure S1a). The two groups (pH 4.0 vs 7.0 and pH 10.0 vs 7.0) shared 1510 DEGs (Figure S1b). A total of 33,428 unigenes were assigned to 25 KOG functional classifications, among which ‘General function prediction only’ was the largest category (26.00%), followed by ‘Signal transduction mechanisms’ (10.53%), ‘Posttranslational modification’ (7.96%), ‘Transcription’ (5.70%), and ‘Translation’ (4.24%) (Figure 3a). Additionally, 1352 unigenes (4.04%) were assigned to ‘Carbohydrate transport and metabolism’, 976 (2.89%) unigenes were assigned to ‘Energy production and conversion’, and 631 (1.89%) unigenes were assigned to ‘Inorganic ion transport and metabolism’, a class that shared a relatively high percentage of genes among the categories. In addition, 32,377 unigenes were assigned to KEGG pathways (Figure 3b), among which ‘Metabolism pathways’ comprised 61.17%, followed by ‘Genetic information processing’ (23.66%), ‘Environmental information processing’ (6.61%), and ‘Cellular processes’ (4.58%). Further analyses showed that ‘Global and overview maps’ (23.26%), ‘Carbohydrate metabolism’ (10.85%), and ‘Translation pathway’ (9.53%) accounted for the greatest proportions. Furthermore, 173,061 unigenes were successfully annotated by GO assignments; these were classified into 44 functional groups belonging to three GO categories (Figure 3c): ‘Biological process’ (68,425; 39.54%), ‘Molecular function’ (52,331; 30.24%), and ‘Cellular component’ (52,305; 30.22%) (Figure 3c). Among ‘Biological process’, the top two categories were ‘Cellular process’ (23,930; 13.83%) and ‘Metabolic process’ (18,570; 10.73%). Among ‘Cellular component’, ‘Cellular anatomical entity’ (30,418; 17.58%) was the largest category. Within the ‘Molecular function’, the greatest numbers were involved in ‘Binding’ (24,511; 14.16%), ‘Catalytic activity’ (22,017; 12.72%), and ‘Transporter activity’ (2631; 1.52%). A pathway enrichment analysis of DEGs based on the KEGG database with p < 0.05 as the threshold was performed to identify the functional consequences of gene expression changes associated with plant growth, flowering, photosynthesis, and stress. The results revealed that the most enriched pathways were ‘Metabolism pathways’, ‘Environmental information processing’, ‘Cellular processes’, ‘Carbohydrate metabolism’, ‘Energy production and conversion’, and ‘Inorganic ion transport and metabolism’. A total of 144 DEGs—including 35 DEGs related to growth, flowering, and its regulatory metabolism (Table 2); 75 DEGs related to photosynthesis (Table 3); 22 DEGs related to stress signal and tolerance (Table 4); and 12 DEGs related to iron transport (Table 4)—were identified after further analyses of top DEGs. Of the DEGs related to plant growth, flowering, and regulatory metabolism, five DEGs were found to be related to flowering, seven DEGs were found to be involved in plant growth, four DEGs were found to be related to hormone metabolism, three DEGs were found to be involved in signal transduction, and 16 DEGs were found to be related to the regulatory processes of transcription and translation. Photosynthesis-related gene expression was significantly affected by acidic and alkaline treatments; seven DEGs were shown to be involved in PSI, seven DEGs were shown to be involved in PSII, six DEGs were shown to be involved in light harvesting, 17 DEGs were shown to be involved in photosynthetic ETCs, seven DEGs were shown to be involved in thylakoid formation, six DEGs were shown to be involved in chlorophyll biosynthesis, 10 DEGs were shown to be involved in ATP synthesis, 11 DEGs were shown to be involved in carbon fixation, and three DEGs were shown to be involved in stomatal development and movement, indicating that acidity or alkalinity could affect both the light reactions and carbon fixation of photosynthesis, especially for photosynthetic ETCs. 2.4. Expression Profiles Analysis of Important DEGs Related to Growth and Flowering The expression levels of flowering-related DEGs were significantly down-regulated under acidic and alkaline conditions (Figure 4). The expression levels of PAUSED (PSD, Unigene8_All) and two-component response regulator-like APRR1 (CL5436.Contig4_All) were significantly reduced by the acid and alkaline treatments. JMJ18 (CL1104.Contig4_All) and WUSCHEL-related homeobox 1 (WOX1, CL8062.Contig3_All) were both down-regulated under acidic conditions, while the flowering time control protein FY (CL8788.Contig3_All) was only down-regulated under alkaline conditions. Accordingly, cell-division- and plant-growth-related genes were also suppressed under acidic and alkaline conditions. The expression levels of TOUGH (TGH, CL845.Contig2_All) and MIZU-KUSSEI 1 (MIZ1, CL9598.Contig2_All) were significantly reduced under alkaline conditions (Figure 4). B-box zinc finger protein 19 (BBX19) acts as a negative regulator of seedling photomorphogenesis [20]. Acidity and alkalinity increased the expression of BBX19 (CL856.Contig1_All), and exocyst complex component SEC10 (CL14716.Contig6_All) and callose synthase 7 (CALS7, CL542.Contig25_All) were down-regulated under both treatments. The expression level of cell division cycle protein 27 homolog B (CDC27B, CL7886.Contig1_All) was highly reduced under alkaline conditions, and tubulin alpha-5 chain (TUBA5, CL4204.Contig2_All) was decreased under acidic conditions. One auxin biosynthesis (YUC, Unigene53925_All) and two signal transduction genes (TMK1, CL2731.Contig4_All; AFB2, CL789.Contig8_All) were significantly down-regulated under both treatments. As a negative regulator of the ethylene response pathway, CTR1 (CL7684.Contig1_All) was also down-regulated under both conditions. Three signal transduction pathway genes—CML38-like (Unigene2096_All), SK5 (CL5696.Contig3_All), and IQD14 (Unigene40182_All)—were significantly down-regulated under acidic and alkaline conditions, especially under acidic conditions. A total of 14 genes involved in transcription and translation regulation were down-regulated under acidic and alkaline conditions: BRASSINAZOLE INSENSITIVE PALE GREEN 2 (BPG2, CL6966.Contig2_All), transcription termination factor MTERF4 (CL4833.Contig7_All), shaggy-related protein kinase ASK1 (Unigene19526_All), BLH1 (CL1863.Contig5_All), FAR1-RELATED SEQUENCE 5-like (FRS, Unigene39555_All), and pre-mRNA-processing protein 40A (PRP40A) had extremely low expression levels under alkaline conditions, and methyl-CpG-binding domain (MBD, Unigene46268_All), pre-mRNA-processing-splicing factor 8A (PRP8A, Unigene23105_All), eukaryotic translation initiation factor 3 subunit F (eIF3f, Unigene39851_All), and squamosa-binding protein-like 39 (SPL39, CL6274.Contig5_All) were not expressed under acidic conditions; MBD7 (CL7329.Contig2_All), YRDC (CL3635.Contig6_All), splicing factor U2af large subunit B (U2AF65B, CL12897.Contig1_All), and G-box-binding factor 3 (GBF3, CL1061.Contig4_All) had no detectable expression under both treatments. Acidic and alkaline conditions increased the expression levels of argonaute 1 (AGO1, CL4080.Contig4_All) and replication factor C subunit 1 (RFC1, CL4788.Contig6_All), two genes involved in gene silencing regulation. 2.5. Effect of Irrigation pH Treatment on the Photosynthetic Characteristics The results regarding photosynthetic characteristics showed that the Pn initially increased and then decreased in all three treatments, with a peak at S5 (Figure 5a). The Pn was significantly weakened under acidic and alkaline conditions compared to the control in all stages, and this effect was much stronger under acidic conditions. The Tr, Gs, and WUE were also reduced by the treatments compared to the control group, with the lowest levels under acidic conditions (Figure 5b,c,e). Nevertheless, the change trend of Ci showed no significant differences among the three treatments (Figure 5d), which suggested that acidic and alkaline conditions could substantially reduce the Pn, Gs, and WUE while affecting Ci, and acidity could apparently do more damage to tree peony photosynthesis than alkaline conditions. Accordingly, soluble sugar was significantly decreased under acidic conditions (20.15%), even though the starch content had no significant differences among treatments (Figure 5f). The soluble protein was significantly reduced under acidic (23.46%) and alkaline (8.46%) treatments compared to the controls. 2.6. Stomata Characteristics and Leaf Structure Stomata play critical roles in photosynthesis. The results showed that stomata size (including guard cell length, guard cell pair width, stomata length, and stomata width) increased with the growth of the tree peony, but there were no significant differences among treatments (Figure 6a,b and Table S3). Stomatal number and density were significantly reduced under acidic and alkaline conditions compared to those of the control from the S4 to S6 stages: the stomatal density was reduced by 18.84% and 7.82% at S3, 30.32% and 27.04% at S4, and 37.64% and 41.87% at S6, respectively, under acidic and alkaline conditions (Figure 6a,c). Moreover, in all three stages, the pore aperture declined under the acidic and alkaline treatment conditions by 55.47% and 55.68% at S3, 62.13% and 46.01% at S4, and 81.64% and 66.6% at S6, respectively, compared to that of the control (Figure 6d). There were clear differences in internal leaf structure among the three treatments: the results showed that the leaves of the plants grown under acidic and alkaline conditions were thinner, with loose palisade tissue and irregularly arranged spongy mesophyll cells; the leaves in the control group showed the most compact leaf palisade parenchyma, and cell wall thickness was reduced under alkaline conditions. In addition, the shape of palisade mesophyll cells was also affected by the treatments, tending to be round instead of elliptical (Figure 6f,g). The number, size, shape, and the ultrastructure of the chloroplast were influenced by both the acidic and alkaline conditions (Figure 6f). As shown in Figure 5g, despite the cell size having no significant differences among the three treatments, the numbers of chloroplasts per cell were significantly decreased by 69.9% and 65.03% under the acidic and alkaline conditions, respectively, in comparison to the control group. Chloroplast size was also reduced under the acidic and alkaline conditions. The chloroplasts in leaves from the control group had a highly organized inner membrane system; many grana thylakoids were regularly distributed with plentiful grana lamellae, and osmiophilic granules were dispersed and fewer in number. In contrast, the stacks of grana disappeared from the chloroplasts in the yellow leaves grown under the acidic and alkaline conditions. These chloroplasts had only a few stromal thylakoid membranes remaining, along with clusters of osmiophilic granules. The structures of thylakoid membranes in these chloroplasts were extremely disordered. The structures of the stromal lamella and basal lamella in chlorophyll under acidic and alkaline stress were unclear, and the starch granules were not tightly arranged. Under the acidic and alkaline conditions, the numbers of lipid droplets, basal lamellae, and osmophilic granules were much lower in leaf cells grown and the matrix lamella was looser. Starch grain size was decreased under the alkaline conditions. Stomatal development and movement genes were affected by both treatments. Beta carbonic anhydrase (BCA) is involved in the CO2 signaling pathway that controls gas-exchange between plants and the atmosphere by modulating stomatal development and movement [21,22]. Serine/threonine/tyrosine-protein kinase HT1 is involved in the control of stomatal movement in response to CO2 and functions as a major negative regulator of CO2-induced stomatal closing [23]. Exposure to acidity and alkalinity were found to result in the significant down-regulation of BCA (Unigene56632_All) and HT1 (Unigene2658_All). Translationally-controlled tumor protein (TCTP) is involved in the regulation of abscisic acid- and calcium-mediated stomatal closure, and acidity and alkalinity enhanced the expression of TCTP (Unigene79230_All). 2.7. Expression Profiles Analysis of Important DEGs in Photosynthesis Acidic and alkaline conditions significantly reduced gene expression related to the light reactions of photosynthesis (Figure 7), including 5 PSⅠ genes (Photosystem I iron-sulfur center (PSAC), CL5623.Contig3_All; photosystem I reaction center subunit IV (PSAE), Unigene26014_All; photosystem I reaction center subunit VI (PSAH1), CL553.Contig2_All; photosystem I reaction center subunit N (PSAN), Unigene16611_All; and PHOTOSYSTEM I ASSEMBLY 2 (PSA2), Unigene40033_All), seven PSⅡ genes (Photosystem II protein D2 protein (PSBB), CL2899.Contig8_All; serine/threonine-protein kinase STN8, CL5574.Contig1_All; photosystem II phosphoprotein PSBH, CL2899.Contig10_All; oxygen-evolving enhancer protein 1 (PSBO1), CL7453.Contig2_All; psbP domain-containing protein 3 (PPD3), CL2018.Contig1_All; PPD7, UniGene 11530_All; photosystem II 22 kDa (PSBS), CL2734.Contig2_All), 5 light-harvest genes (serine/threonine-protein phosphatase 5 (PAPP5), CL5682.Contig2_All; chlorophyll a-b binding protein (CAB6A), CL7283.Contig4_All; CAB7, CL349.Contig1_All; CAB, UniGene 44832_All, CL1191.Contig3_All; and CL6697.Contig2_All)), 17 ETC genes (electron transfer flavoprotein-ubiquinone oxidoreductase (ETFQO, CL8017.Contig1_All; cytochrome b5 reductase 1 (CBR1), CL13977.Contig4_All; ATP-dependent NAD(P)H-hydrate dehydratase (NAXD), CL2990.Contig3_All; cytochrome f (CYTF), Unigene46676_All; chlorophyllide a oxygenase (CAO), CL1939.Contig9_All; 3Fe-4S ferredoxin (FDXA), CL647.Contig1_All; NADPH: adrenodoxin oxidoreductase MFDR, Unigene11788_All; plastocyanin-like (PETE), Unigene20162_All; NADH-Ubiquinone/plastoquinone complex I (MNHD), CL9837.Contig2_All; NADH-plastoquinone oxidoreductase subunit 2 (NDHB2), CL12489.Contig1_All; quinone oxidoreductase (NDH2), CL2308.Contig4_All; NAD(P)H-quinone oxidoreductase subunit O (NDHO), CL11886.Contig4_All; fibrillin-5 (FBN5), CL269.Contig4_All; photosynthetic NDH subunit of lumenal location 1 (PNSL1), Unigene11863_All, Unigene27888_All; PNSL3, UniGene 10120_All; dihydrodipicolinate reductase-like DAPB3, CL1744.Contig1_All)), and 5 thylakoid membrane formation genes (THYLAKOID FORMATION 1 (THF1), CL14413.Contig2_All; thylakoid membrane TERC, CL2696.Contig4_All; 50S ribosomal protein L24 (RPL24), CL1839.Contig2_All; ALBINO3-like protein 2 (ALB3L2), Unigene39908_All, CL1726.Contig1_All; OBG-like GTPase (OBGL), CL588.Contig4_All); see Figure 6. In addition, several ATP synthase genes (ATP synthase delta chain (ATPD), CL80.Contig9-15_All; ATP synthase CF1 epsilon chain (ATPE), CL8509.Contig1_All; and CL633.Contig10_All) and chlorophyll biosynthesis genes (magnesium chelatase subunit I (CHLI1), CL10274.Contig3_All; pheophytinase (PPH), CL6625.Contig1_All; chlorophyll a oxygenase (CAO), Unigene50987_All) were highly repressed under acidic and alkaline conditions. Adenylate kinase (ADK, CL5310.Contig4_All) was reduced under alkaline conditions. Chlorophyllase-1-like (CLH1, CL7533.Contig2_All) was suppressed under acidic conditions, while the expression levels of 7-hydroxymethyl chlorophyll a reductase (HCAR, UniGene 52328_All) and chlorophyll(ide) b reductase (NOL, CL5229.Contig4_All) were slightly decreased under alkaline conditions. The expression levels of ATP synthase CF1 alpha subunit (ATPA, CL633.Contig6) and adenylate kinase 4 (AK4, CL9808.Contig2_All) were enhanced under the alkaline conditions, and the levels of ATPA (Contig 12_All) were enhanced under both treatments. Both the acidic and alkaline conditions also suppressed the expression of 11 genes in the Calvin cycle of photosynthesis, including bifunctional riboflavin kinase/FMN phosphatase CBBY (CL3370.Contig1_All), 2-carboxy-D-arabinitol-1-phosphatase CA1P (CL13509.Contig2_All), phosphoglycerate mutase 2 (PGM2, CL3429.Contig1_All), ribulose bisphosphate carboxylase small chain (RBCS, CL4922.Contig3, Contig11, and Contig12), phosphoglycerate kinase (PGK, CL3299.Contig2_All), fructose-bisphosphate aldolase (FBA, CL12763.Contig2_All), fructose-1,6-bisphosphatase (FBP, CL1239.Contig5_All), 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (PFKFB, UniGene 23926_All), and transketolase (TKL, CL7000.Contig2_All) (Figure 8). 2.8. The Activities of Enzymes, Nutrient Assimilation, and Expression Profiles of Key Genes in ROS Scavenging Pathway and Nutrient Transport The level of hydrogen peroxide (H2O2) was significantly increased by 8.19% and 6.78% at S4 and by 16.53% and 10.59% at S6 under the acidic and alkaline conditions, respectively, compared to that of the control group (Figure 9a). Most of DEGs involved in ROS signaling cascades, including one EXECUTER 1 (EX1, CL5897.Contig2_All), one Serine/threonine-protein kinase 2 (SAPK2, CL600.Contig6_All), one Lysine-rich arabinogalactan protein 19 (AGP19, Unigene6180_All), two Transcription factor ORG2 (CL2470.Contig4_All, CL2470.Contig8_All), one LRR receptor-like serine/threonine-protein kinase (RPK, CL13877.Contig3_All), two Mitogen-activated protein kinase kinase 3 (MKK3, CL2322.Contig1_All, CL10999.Contig2_All), one Protein phosphatase 2C 50 (PP2C50, CL13633.Contig1_All), and one Receptor-like protein 51 (RLP51, Unigene1326_All) genes were enhanced under both the acidic and alkaline conditions. The activities of superoxide dismutase (SOD), peroxidase (POD), and catalase (CAT) were significantly enhanced under the alkaline conditions at S4 and being enhanced under both the acidic and alkaline conditions at S6, with a higher level in the latter group: SOD activity increased by 23.97% and 48.60%, POD activity increased by 15.16% and 27.00%, and CAT activity increased by 24.49% and 32.42% at S6 under the acidic and alkaline conditions, respectively (Figure 9b–d). Acidity and alkalinity resulted in the significant up-regulation of the expression levels of SOD (CL12133.Contig2_All), CAT (CL8683.Contig2_All), and POD (Unigene2231_All, Unigene79660_All, Unigene6219_All) genes. One SOD (CL11553.Contig1_All) gene, three POD (Unigene63828_All, Unigene26867_All, and Unigene82390_All) genes, one peroxisome biogenesis protein 12 (PEX12, CL7714.Contig3_All) gene, and one transmembrane ascorbate-dependent reductase CYB561A (CL11875.Contig1_All) gene were up-regulated only under the acidic conditions (Figure 9e). The results of the nutrient analysis showed that the uptake of P was reduced by 22.22% and 7.4% while the uptake of magnesium (Mg) was decreased 11.11% and 3.7%, respectively, under the acidic and alkaline conditions compared to the controls (Figure 9f). The uptake of K had no significant differences among treatments, but calcium uptake was enhanced in leaves grown under both the acidic and alkaline conditions. The uptake rates of B and Mn were significantly decreased in the adverse pH groups compared to the control group, especially in the acidity group with reductions of 24.81% and 28.88%, respectively (Figure 9g). The assimilation rate of Fe was only reduced under the acidic conditions, with a reduction of 5.33%. The uptake of Si was also inhibited under the acidic conditions in comparison to the control group, but it was increased under the alkaline conditions. Thus, nutrient assimilation, especially for those elements related to photosynthesis and flowering, was reduced under both treatment conditions. The effect was more serious under acidity compared to that under alkalinity. Nutrient transporter genes including one calcium-transporting ATPase 12 (ACA12, Unigene17975_All) gene, three boron transporter (BOR1, Unigene48447_All; BOR4, Unigene41688_All, and CL3601.Contig5_All) genes, six phosphate transporter (PPT2, CL4107.Contig2_All; PHO1, Unigene39529_All, CL1982.Contig2_All, Unigene40561_All, and CL14859.Contig1_All; and glycerol-3-phosphate transporter 1, glpT1, CL8237.Contig4_All) genes, one potassium transporter 17 (POT17, Unigene56892_All) gene, and one potassium channel SKOR (Unigene48839_All) gene were down-regulated under both the acidic and alkaline conditions (Figure 9h). 3. Discussion 3.1. Inhibition of the Growth and Development of Tree Peony Plants Exposed to Acidic and Alkaline Stresses Acidic and alkaline stresses limit plant growth and development by disturbing numerous physiological processes, including photosynthesis, ionic homeostasis, ROS balance, and the antioxidant system [24]. Acidity stress significantly suppresses root growth, reduces root diameter, and decreases the biomass of rice seedlings [8,25]. Alkaline stress markedly reduces survival percentage and total biomass and inhibits root growth [26,27]. In this study, we observed the severe inhibition of tree peony growth under both the acidic and alkaline conditions, and tree peony was found to be better adapted to alkaline conditions than acidic conditions. Flowering was inhibited and petal color became faded under acidic and alkaline conditions in Ipomoea nil [28] and Paeonia lactiflora [6], consistent with our results. To determine the internal molecular mechanism, we analyzed gene expression related to plant growth and flowering. PSD and WOX1 are required for shoot apical meristem growth [29]. APRR1 controls the photoperiodic flowering response [30]. JMJ18 and FY are involved in the control of flowering time [31,32]. The expression of most genes related to flowering, including PSD, APRR1, JMJ18, WOX1, and FY, were found to be reduced in tree peony grown under acidic and alkaline conditions, explaining why flowering was delayed and the flower quality was greatly reduced. TOUGH (TGH) and MIZU-KUSSEI 1 (MIZ1) are required for plant growth and development [33,34]. Consistent with the flowering genes, TGH and MIZ1 were also significantly down-regulated under adverse pH conditions. The cell-division-related genes SEC10, CDC27B, and CALS7 [35,36] were also down-regulated in tree peony under both the acidic and alkaline conditions, explaining the reason for the suppression of plant growth as affected by adverse pH conditions. TUBA5 is important for the synthesis of microtubules, and microtubules play crucial roles in plant adaptation to stressful environments [37]. The expression of the microtubule synthesis gene TUBA5 was much higher under alkaline conditions than under acidic conditions, a result that may explain why the tree peony grown under alkaline conditions displayed stronger tolerance. Auxin plays an important role in controlling various aspects of plant growth and development [38]. The genes in plant hormone signal transduction pathways are significantly inhibited by highly acidic conditions in tea plants [16]. In tree peony leaves, the expression levels of genes related to auxin biosynthesis and transduction, including TUC, TMK1, and AFB2, were also significantly reduced under acidic and alkaline treatments. Transcription and translation are signs of cellular activity. High levels of transcription and translation accelerate plant growth, development, and metabolism. The present study showed that most of the transcription and translation regulation genes were also suppressed under both the acidic and alkaline conditions. In contrast, the adverse pH conditions raised the possibility of gene silencing through transcriptional and post-transcriptional regulations. Transcription and translation profiles are also affected in response to environmental stresses in other species [39]. These results show that acidic and alkaline conditions affected the plant growth and flowering of tree peony through regulating gene transcription and translation. 3.2. Exposure to Acidity and Alkalinity Reduces Photosynthesis via Weakening Light Capture, Photosynthetic ETCs, ATP Synthesis, Carbon Fixation, and the Development of Stomata The structure of leaf and chloroplasts, as well as chloroplast movement, are vital for photosynthesis [40,41]. The leaves of tree peony plants grown under acidic and alkaline conditions were thinner than controls, with loose palisade tissue and irregularly arranged spongy mesophyll cells. The number of chloroplasts was significantly reduced in both treatment groups. Moreover, the structures of thylakoid membranes in the chloroplasts of the yellow leaves grown under acidic and alkaline conditions were extremely disordered, and the amounts were also decreased. These changes were similar to the disorganization of thylakoid membranes observed in Ocimum basilicum under stress conditions [42]. Thylakoid-membrane-related genes were found to be down-regulated under abiotic stress at both transcriptome and proteome levels [43]. Six thylakoid membrane formation genes were also down-regulated under acidic and alkaline conditions in tree peony. Changes in chloroplast ultrastructure and quantity were one of the most important reasons for the decrease in the chlorophyll content, and chlorophyll content is also a critical determinant for the Pn [44]. In Puccinellia tenuiflora, photosynthesis is remarkably reduced under alkaline stress due to stomata closure and decreases in chlorophyll content [18]. We also found that the changes in chloroplast ultrastructure and quantity in the acidic and alkaline treatment groups were positively related to chlorophyll content and Pn. Consistently, chlorophyll biosynthesis genes were also highly down-regulated under acidic and alkaline conditions. Significantly lowered chlorophyll contents have been reported in tomato and maize subjected to alkaline stress [45,46]. Chlorophyll and carotenoid contents were found to be significantly lower in plants grown under low pH treatment than in the control groups [47]. Abiotic stress causes the breakdown of chlorophyll and reductions in photosynthetic pigments in rice [48]. In tree peony leaves, it seems that chlorophyll may be degraded in the late growth stage and that the degradation is more severe under acidic conditions, a result that could explain the leaf chlorosis in the acidic group, as has been also reported in quince, pear, and olive [49]. A decrease in chlorophyll content can directly affect light energy absorption capacity [50]. A previous study showed that light energy absorption capacity decreased when white willow was subjected to stress [51]. This study has shown that light-harvesting-related genes were down-regulated under acidic and alkaline conditions. Therefore, the adverse pH level directly affected the number and structure of chloroplasts, the production of chlorophyll molecules, and the expression of LHC-related genes, resulting in a significant reduction in light-harvesting capacity. Stomata, formed by a pair of guard cells, play an important role as a regulatory gate for the exchange of CO2 between plants and the environment; thus, they regulate stomatal conductance (Gs) and the Pn by changing their aperture and/or density [52]. Our results showed that acidic and alkaline treatments led to stomatal closure. The decrease in Pn under stressful conditions is normally attributed to the suppression of mesophyll conductance and to stomatal closure under moderate and severe stress [53]. When plants are exposed to changing environmental conditions for a short period, stomatal aperture may be the main factor influencing Gs, whereas changes in Gs may be determined by the alteration of both stomatal aperture and stomatal density in response to a changing environment over a longer period [54,55,56]. A low pH (pH 2.5) was found to greatly alter stomatal density and size in tea leaves [16]. The stomatal density of tree peony leaves in our study was also decreased under acidic and alkaline conditions for a long period of treatment, especially under acidic conditions. Accordingly, two stomatal development and movement genes, BCA and HT1, were down-regulated under acidic and alkaline conditions, while one stomatal-closure-inducing gene, TCTP, was enhanced under both conditions. Moreover, the increase in stomatal density was positively correlated with WUE in Leymus chinensis [57], in agreement with our results. Hence, acidity and alkalinity reduced stomatal aperture and density to modulate gas diffusion, thereby affecting photosynthesis. Photosynthesis is one of the most sensitive processes to stress [10,43]. Previous research has shown that acidic and alkaline stress significantly reduces photosynthesis and productivity [16,18]. In the present study, Pn was significantly decreased under acidic and alkaline treatments compared to the control group. Consistent with this result, the Gs, Tr, and WUE were also significantly reduced under acidic and alkaline conditions. Photosynthesis includes two major stages: light-dependent reactions and light-independent reactions. The light-dependent reactions take place in the thylakoid membrane via two photosystems called PSI and PSII, in which electrons are transferred and the light energy is converted into chemical energy in the form of the ATP and NADPH molecules. In tea leaves, the expression levels of multiple genes related to photosynthesis, including one light-harvesting complex, two PSII subunits, one PSI subunit, and one ferredoxin-NADP(+) reductase (FNR) were found to be reduced under pH 2.5 [16]. Rhododendron prefers acidic soils with a pH of 5.0 or below; a transcriptome comparison showed that photosynthesis-related genes, including LHC genes and petC, petE, petH, and ATP synthesis genes, were all down-regulated under high pH [13]. The transcriptome analysis of tree peony leaves following exposure to acidic and alkaline conditions showed that the most significant DEGs in the light-dependent photosynthesis pathway were concentrated in photosynthetic ETCs, the PSI or PSII reaction-center complex, and ATP synthesis. A total of 17 ETCs, 5 PSI, 7 PSII, and 4 ATP synthase genes were down-regulated under acidic and alkaline conditions, and photosynthetic ETCs were most sensitive. The light-independent stage, also known as the Calvin cycle, takes place in the stroma of chloroplasts. It uses the stored chemical energy from the light-dependent reactions to ‘fix’ CO2 and then creates a product that can be converted into glucose. Adverse stress has been shown to markedly reduce the expression of Calvin cycle genes in cucumbers [58,59]. Based on the transcriptome gene expression, we found that acidic and alkaline conditions suppressed the expression of 11 Calvin cycle genes. In conclusion, our results indicated that acidic and alkaline conditions inhibit tree peony photosynthesis by repressing photosynthetic ETCs, diminishing light-harvesting capacity, decreasing stomatal density and aperture, and weakening enzyme activities in the Calvin Cycle. 3.3. Acidic and Alkaline Conditions Interfere with Nutrient Assimilation and Transport in Tree Peony Leaves Soil pH can affect nutrient availability and assimilation [18]. It has been reported that soil acidity stress causes a decreased uptake of nutrients (i.e., nitrogen, phosphorus, potassium, calcium, and magnesium) [8] and that alkaline conditions lead to the deficiency of nutrient minerals that in turn limits plant growth and agricultural productivity [60,61]. P is a main component of nucleic acids, proteins, and phospholipids [62]. P deficiency affects protein synthesis, energy metabolism, and signal transduction; decreases chlorophyll content and CO2 assimilation; and impairs photosynthetic ETCs [63]. In maize, alkaline conditions were found to lead to the deficiency of P [64]. Similar results were found in tree peony leaves. B is closely related to flowering and yield [65]. B becomes less available with increasing solution pH [66]. Here, we found a significantly lower level of B in tree peony leaves exposed to both the acidic and alkaline conditions. A previous study showed that B deficiency affects photosynthetic capacity and the transport of photosynthesis products in woody plants [67], a result that may explain why tree peony photosynthesis was inhibited when B absorption was reduced. Fe is required for the synthesis of the heme structure and is an essential component of chlorophyll [68]. In addition, Fe is involved in photosynthetic ETCs in the form of ferritin and ferredoxin. Mg and Mn are not only components of chlorophyll but also activators of Calvin cycle enzymes such as RuBP carboxylase and ribulokinase 5-phosphate [69]. Fe availability for plants depends on the physico-chemical properties of the soil. High pH decreases the availability of Mn [70]. In lettuce, the content of Mg was found to be decreased under low and high pH, but Fe and Mn levels were decreased at higher pH [68]. In tea leaves, the level of Mg decreased with increasing acidity, thereby causing the inhibition of chlorophyll biosynthesis [71,72]. The chlorophyll synthesis and photosynthetic capacity of Carya illinoinensis were found to be reduced when plants were deficient in Mn [73]. Our results showed that the assimilation of Mg, and Mn were all inhibited in the tree peony plants grown under both the acidic and alkaline conditions, as well as that Fe assimilation was only reduced under acidic conditions; this was one of the most important reasons for the decrease in chlorophyll content and the impairment of light-harvesting capacity and ETCs. Therefore, the reductions in P, B, Fe Mg, and Mn assimilation caused the inhibition of chlorophyll biosynthesis, the impairment of light harvesting, and the obstruction of ETCs under acidic and alkaline conditions, consequently suppressing photosynthesis. Si is widely considered to possess significant potential as a substance that can ameliorate the negative effects of abiotic stresses and improve plant growth and biomass accumulation [74]. The accumulation of Si in tree peony plants was shown to be enhanced under alkaline conditions; this explained why the injury and negative effects on tree peony plants grown under acidic conditions were much greater than those caused by alkaline conditions. The lower pH was shown to increase the absorption of Si in rice due to species adaptation [75]. Ca also has a stimulating effect on plant tolerance to different stresses by regulating antioxidant metabolism [76]. The enhanced antioxidant activities of tree peony plants grown under acidic and alkaline conditions were consistent with the higher accumulation of Ca2+. Mineral nutrient transport genes, including GmPTs, GmZIPs, and GmHKT1, were shown to be significantly down-regulated by acidity in soybeans [15]. Here, we also found that nutrient transporter genes including one calcium-transporting gene, three boron transporting genes, six phosphate transporting genes, one potassium transporting gene, and one potassium channel gene were greatly down-regulated in tree peony leaves under acidic and alkaline conditions. Therefore, the absorption and transport of nutrients were affected by the stress, and these in turn influenced plant growth, flowering, photosynthetic capacity, and plant resistance. 3.4. Redox Homeostasis and the Activities of Antioxidant Enzymes in the Response to Acidic and Alkaline Treatment At low or moderate levels, ROS are implicated as second messengers in signaling cascades that mediate most biological processes, including programmed cell death (PCD), stomatal closure, and tolerance to different stresses [77]. Meanwhile, a high level of reactive ROS leads to direct oxidative damage for plants and ultimately results in cell death [77,78]. Adverse pH conditions have been shown to lead to increased ROS levels [18,79], and similar results were obtained in this study. In addition, we found that the ROS level was lower under alkaline conditions than acidic conditions. Plants perceive abiotic and biotic stresses and adapt to these stresses by a series of signal transduction factors, including ROS [77,80]. EX1 enables plants to perceive singlet oxygen as a stress signal, activating a nuclear stress response program, triggering a PCD, and impeding PSII without causing photooxidative damage to the plant [81]. AGP19 also regulates PCD [82]. SAPK2, RPK, and PP2C50 are involved in the ABA signal transduction pathway when plants are subjected to stress [83,84,85]. MKK3 is one important component of the ABA signaling pathway; it negatively regulates ROS accumulation [86,87]. RLP51 takes part in plant defense responses [88]. ORG2 plays an important role in iron deficiency-mediated stress regulation [89]. Consistent with the ROS level, the expression levels of these ROS-signaling genes were found to be enhanced in tree peony following exposure to adverse pH conditions, a result that may show the possibility of ROS as a signaling molecule involved in plant responses and adaptions to pH stress. ROS signaling was also shown to be enhanced in Arabidopsis under stress conditions [78]. In response to excess ROS accumulation under stress conditions, plants activate a set of ROS-scavenging enzymes (SOD, POD, CAT, and ascorbate peroxidase (APX)) and non-enzymatic antioxidants (ascorbate, glutathione, carotenoids, and phenolic compounds) to restore cellular ROS homeostasis [18]. SOD is the first antioxidant enzyme that can catalyze O2− to H2O2, and, as such, it plays a central role in plant defense against oxidative stress. The increased activity of SOD directly results in enhanced stress tolerance in plants [90]. Subsequently, POD, CAT, APX, and GPX catalyze the conversion of H2O2 to water. The ability of plants to control oxidant levels under stressful conditions is highly correlated with their stress tolerance [91]. Alkaline stress was found to stimulate the activities of ROS-scavenging enzymes and increase the gene expression of SOD, CAT, POD, and APX in Puccinellia tenuiflora [18]. In this study, the activities of SOD, POD, and CAT were more enhanced under alkaline conditions than under acidic conditions. Moreover, acidic and alkaline treatments resulted in the significant up-regulation of the expression of SOD, POD, and CAT genes. The increasing ROS-scavenging capability of tree peony plants observed in the alkaline group is critical for ROS homeostasis and alkaline tolerance. Accordingly, many tree peony roots were found to be damaged and rotten under acidic conditions, while roots grown under alkaline conditions were in much better condition. Root damage under adverse pH stress is also associated with ROS accumulation in rice [26,27]. ROS can damage the photosynthetic apparatus, particularly PSII, and inhibit the translation of photosynthetic genes, resulting in reductions in Pn and the inhibition of PSII repair [92,93]. In this research, we observed negative correlations of ROS and Pn, which were also confirmed by Sharma et al. (2012) [77]. Adverse pH, particular acidic toxicity, can directly damage citrus roots, thus affecting the uptake of mineral nutrients [8]. The high ROS level in tree peony plants may cause damage to plant roots and negatively regulate the global transcription level, thereby affecting nutrient assimilation and cell metabolism. In addition, the adverse pH conditions were shown to stimulate stomatal closure and regulate a series of genes related to plant growth, photosynthesis, flowering, hormone and signal transduction, transcription, and translation (Figure 9). Ultimately, plant growth, flowering, photosynthesis, nutrient assimilation and transport, and ROS production and elimination were all shown to be influenced by acidic and alkaline conditions (Figure 10). Adverse pH affected the availability of important nutrients such as P, Fe, Mg, Mn, B, and Si and caused the excess production of ROS that may damage root cells and reduce nutrient assimilation. This could in turn affect chlorophyll synthesis, photosynthesis, and stress tolerance. Furthermore, the activities of ROS-eliminating enzymes (including SOD, CAT, and POD) and the expression of the associated coding genes were enhanced to alleviate the damage caused by adverse pH stress. These adverse pH conditions suppressed the expression of a series of genes related to plant growth, flowering, signal transduction, transcription, translation, light harvesting, photosynthetic ETCs, thylakoid membranes, carbon fixation enzymes, chloroplast development, chlorophyll synthesis, stomatal development, and stomatal aperture through a series of signaling cascades. As a result, plant growth was inhibited, flowering quality was reduced, photosynthesis was impaired, and plant biomass was decreased. 4. Materials and Methods 4.1. Plant Materials and Forcing Culture Conditions Five-year-old adult plants of P. suffruticosa Andr. ‘Luo Yang Hong’ were collected from the experimental field of the Department of Peony, Chinese Academy of Agricultural Sciences, Beijing, China, and they were potted in plastic flowerpots with pH-balanced media. Sixty plants free of pests or disease and with similar growth conditions were randomly selected in this study, with 20 plants for each treatment. The plants were treated with different pH levels (pH 4.0, 7.0, and 10.0) for five months. The plants were thoroughly irrigated using a glycine buffer solution (pH 4.0 and 10.0) once a week, and distilled water (pH 7.0) was used as the control treatment. The growth parameters, photosynthesis, and samples for the physiological indices and section observation of leaves were collected during the treatment time (S0, bud sprouting stage; S1, hard bud stage; S2, loose bud stage; S3, half open stage; S4, fully opened stage; S5, two weeks after the fully opened stage; and S6, four weeks after the fully opened stage). Flowers were sampled to study flower quality and biomass. Root samples for morphological characteristics and secondary metabolites were collected after five months. All samples were immediately frozen in liquid nitrogen and then stored at −80 °C until analysis. 4.2. Color Indexes and Pigment Estimation The flower and leaf color indexes were measured with a Minolta CR-300 Chroma Meter (Konica Minolta Optic Inc., Tokyo, Japan). For L* values from 0 to 100, the darkness gradually decreases; the green tone gradually decreases and the red tone becomes clearer from −a* to +a*. The shade of blue gradually decreases and the shade of yellow increases from −b* to +b*. The hue angle was calculated as follows: H° = arctangent (b*/a*). Chroma (C*) and hue angle (h) were calculated according to the following equations: C* = (a*2 + b*2)1/2 and h = tan − 1 (b*/a*). C* is the distance perpendicular from the lightness axis (more distance leads to more chroma). The anthocyanin content was measured according to the method of Brown [94] at 530 nm. Chlorophyll a and b contents were determined at 663 and 645 nm, respectively, and calculated based on the method of Li [95]. The relative chlorophyll index was constructed using a portable chlorophyll SPAD 502 m (Konica Minolta Optic Inc., Tokyo, Japan). For carotenoid determination, fresh leaves were homogenized in 80% acetone with a homogenizer. Homogenates were centrifuged at 4 °C for 15 min (3000 rpm). Supernatants were used for the analysis of pigments. Absorbances were determined at 470 nm. 4.3. Measurement of Photosynthetic Indexes The net photosynthesis rate (Pn), transpiration rate (Tr), intercellular carbon dioxide (CO2) concentration (Ci), stomatal conduction (Gs), and photosynthetic water use efficiency (WUE) of leaves were determined between 08:30 and 10:30 from fully expanded third blades using a portable open flow gas exchange system (CIRAS-3 Portable Photosynthesis machine, Amesbury, MA, USA) under ambient CO2 concentrations (chemicals removed). The Pn, Gs, Ci, Tr, and WUE were recorded once the rate of CO2 uptake had stabilized. WUE is given by the ratio of the net CO2 assimilation rate to the transpiration rate. 4.4. Stomata Observation by Light Microscope and Leaf Ultrastructure Observation by TEM Nail polish imprints were taken from the abaxial surface of mature leaves from plants and examined immediately with an optical microscope (Olympus CX31RTSF, Tokyo, Japan). Stomatal properties were analyzed using the ImageJ software (version 1.8.0). Stomatal density, aperture, and size were calculated as previous described [96]. Samples for transmission electron microscopy (TEM) were prepared according to standard TEM sample preparation protocols. Ultrathin tissue sections were mounted on nickel grids and observed using a transmission electron microscope (Hitachi HT7500, Tokyo, Japan). 4.5. Estimation of Contents of Macro and Micro-Nutrients, Soluble Sugar, Protein, and Starch Leaf samples were dried in an oven at 80 °C for 72 h and then ground to a fine powder using a mortar and pestle. About 30 mg was reduced to ashes at 550 °C in a muffle furnace for 5 h and then digested with 2 mL of 20% HCl (6N) for 5 min at 60 °C using a heating block. This hot water extract was cooled and filtered using Whatman no. 42 filter paper and finally diluted to a volume of 50 mL with distilled deionized water. Macro and micro nutrient concentrations were determined using an inductively coupled plasma optical emission spectrometer (ICP-OES, Agilent 725, Beijing, China). Soluble sugar, protein, and starch were determined following the work of Ren et al. (2018) [96]. 4.6. Determination of Hydrogen Peroxide Level and Activities of Antioxidant Enzymes The activities of CAT, SOD, and POD were determined using an antioxidant assay kit (Sigma-Aldrich KGT 00150-1, St. Louis, MO, USA) according to the manufacturer’s protocols, and the absorbances were measured at 405, 550, and 420 nm, respectively. The H2O2 level was evaluated using a hydrogen peroxide assay kit (Sigma-Aldrich KGT018, St. Louis, MO, USA) by comparing its absorbance at 405 nm to a standard calibration curve according to its manufacturer’s protocols. 4.7. RNA-seq, Library Construction, Sequence Assembly and Annotation Total RNA was isolated from the leaves of tree peony plants grown under pH 4.0, 7.0, and 10.0 at S4 using a Trizol extraction kit (Invitrogen, Carlsbad, CA, USA), and DNA was removed using RNase-free DNase I (Takara Biotechnology, Dalian, China) according to the manufacturer’s instructions. Three leaves were used in each sample. The quality of RNA was detected using a NanoDrop 2000 UV/Visible Spectrophotometer (Thermo Scientific, Waltham, MA, USA) and an electrophoresis apparatus (Thermo Scientific, Waltham, MA, USA). High-quality RNA from samples of three treatments was used for cDNA library construction and Illumina sequencing. Each sample, including leaves from the pH 4.0, 7.0, and 10.0 samples, was used to construct one cDNA library, so three cDNA libraries were constructed in our study. The cDNA library was constructed using the Truseq RNA sample preparation kit (Illumina, San Diego, CA, USA) according to the Illumina manufacturer’s instructions. Briefly, the poly (A) mRNA was isolated using oligo-dT beads (QIAGEN, Hilden, Germany). Subsequently, 200-nt-long mRNA fragments were generated using a fragmentation buffer and first-strand cDNA was synthesized with the addition of random hexamer primers. The second-strand cDNA was synthesized with a SuperScript double-stranded cDNA synthesis kit (Invitrogen, Carlsbad, CA, USA) and purified using a QiaQuick PCR extraction kit (QIAGEN, Hilden, Germany) according to the manufacturer’s instructions. The double-stranded cDNA of the above-mentioned three samples was sequenced using an Illumina HiSeq™4000 platform at the Beijing Genomics Institute Company (Shenzhen, China). All transcription sequencing data are available at the NCBI Short Read Archive (Accession number: SRR19039925-SRR19039927). All raw reads were initially processed by passing them through quality control (QC) filters to remove adapter sequences, low-quality reads (phred score < 20), unknown nucleotides (Ns), relatively short reads (<50 nt), and terminal nucleotides in both 3′ and 5′ ends to produce clean reads. The clean reads were de novo assembled using the Trinity software to construct unique consensus sequences with no extension on either end. All assembled unigenes were searched and annotated against the publicly available NCBI non-redundant nucleic acid sequence database (NT) using BLASTn analysis and protein databases including the Swiss-Prot protein sequence database (Swiss-Prot), NCBI non-redundant protein database (NR), Clusters of EuKaryotic Orthologous Groups (KOG), and Kyoto Encyclopedia of Genes and Genomes (KEGG) using BLASTx analysis with an E-value cut-off of 1 × 10−5. To understand the functional classification of the unigenes, Gene Ontology (GO) analysis was conducted on the annotated sequences using the Blast2GO Program and NR annotation results. The potential coding sequences (CDS) of unigenes were identified from NCBI (Open Reading Frame Finder, https://www.ncbi.nlm.nih.gov/orffinder/, accessed on 4 April 2022) and confirmed by BLAST in Swiss-Prot and Pfam protein sequences with Hmmscan. 4.8. Analysis of Differentially Expressed Genes (DEGs), GO and KEGG Enrichment Differential expression analysis was performed using DESeq2 on the identified DEGs. The DEGs were evaluated based on the genes with FPKM of >1 in at least one sample, and this parameter was set to the p-adjust of <0.05 and |log2FC| of ≥2. GO and KEGG enrichment analyses were performed using Fisher’s exact test for the elucidation of the biological functions of the genes. The false discovery rate was <0.001. K-mean clustering was performed on log2-transformed FPKM values with the Euclidean correlation as a similarity metric for the visualization of genes with similar expression patterns and the exploration of their functions. A heatmap of DEGs was drawn with the Heml software (version 1.0). 4.9. Statistical Analysis The design of the experiment was completely randomized with twenty replications. The experimental data are expressed as mean ± SE and were analyzed by one-way ANOVA, followed by Duncan’s multiple range test at p < 0.05 to find the statistical significance among treatments using the SPSS Statistics software (version 20.0). 5. Conclusions Plant growth, flowering, photosynthesis, and the associated regulatory genes of tree peony were found to be affected by acidic and alkaline conditions, and acidity was more toxic than alkalinity to plants because the ROS-scavenging capability of tree peony was enhanced under alkaline conditions. Acidic and alkaline conditions produced excess ROS that caused damage to root, chloroplasts, and photosynthetic systems. In addition, a series of genes related to plant growth, cell division, flowering, auxin biosynthesis and signal transduction, transcription, translation, photosynthesis, chloroplast development, chlorophyll synthesis, and stomatal development was also significantly down-regulated. The DEGs related to photosynthesis were concentrated in light-harvesting capacity, ETCs, the reaction centers of PSII and PSI, ATP synthesis, and carbon fixation, among which ETCs were the most sensitive to the adverse pH. Nutrient assimilation was also affected by acidic and alkaline conditions. The reduced chlorophyll content and low expression of photosynthetic antenna genes synergistically suppressed the light-harvesting capacity. These results jointly led to the reduction in Pn. Accordingly, sugar accumulation and plant biomass were decreased. Acknowledgments We thank LetPub (www.letpub.com, accessed on 4 April 2022) for its linguistic assistance during the preparation of this manuscript. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ijms23095094/s1. Click here for additional data file. Author Contributions Conceptualization and Methodology, X.Z. and X.R.; Resources, X.Z., S.W., J.X. and Y.Z.; Data Curation, T.T.A., X.R. and F.S.; Writing—Original Draft Preparation, T.T.A.; Writing—Review and Editing, X.R.; Visualization and Supervision, X.R.; Project Administration, X.R.; Funding Acquisition, X.Z. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare that they have no competing interests. Figure 1 Influence of different pH treatments on the flowering features of P. suffruticosa ‘Luoyanghong’. (a) Morphology of flowers, (b) flower diameter, (c) flower stalk length, (d) anthocyanin accumulation in petal, and (e,f) flower color indexes. S0, bud sprouting stage; S1, hard bud stage; S2, loose bud stage; S3, half open stage; S4, fully opened stage. Scale bar is 1 cm. Asterisks and different lowercase letters indicate significant differences among different treatments in leaves (Duncan’s test at p < 0.05 after analysis of variance; data are shown as mean ± SE). Figure 2 Influence of different pH treatments on the leaf and root growth of P. suffruticosa ‘Luoyanghong’. (a) Morphological changes of leaf, bar = 5 cm; (b,c) chlorophyll and carotenoid contents; (d) leaf area; (e) leaf biomass; (f) morphological changes of root, bar = 1 cm; (g,h) biomass of adult roots and fibrous roots; and (i,j) the distribution of root. S0, bud sprouting stage; S1, hard bud stage; S2, loose bud stage; S3, half open stage; S4, fully opened stage; S5, two weeks after the fully opened stage; S6, four weeks after the fully opened stage. Asterisks and different lowercase letters indicate significant differences among different treatments in leaves (Duncan’s test at p < 0.05 after analysis of variance; data are shown as mean ± SE). Figure 3 DEG analysis via (a) KOG functional classification, (b) KEGG pathways distribution, and (c) Gene Ontology (GO) assignments for tree peony transcriptome unigenes. Figure 4 Gene expression patterns related to the flowering, growth, hormone, signal, and regulation of transcription and translation by FPKM analysis in three samples. Figure 5 Influence of different pH treatments on the photosynthetic indexes of P. suffruticosa ‘Luoyanghong’ at four consecutive weekly stages from the bud initiation stage to 4 weeks after flowering stage. (a) Net photosynthetic rate (Pn); (b) transpiration rate (Tr); (c) stomatal conductance (Gs); (d) intercellular CO2 levels (Ci); (e) water use efficiency (WUE) values; and (f) the accumulation of carbohydrates and protein. S0, bud sprouting stage; S1, hard bud stage; S2, loose bud stage; S3, half open stage; S4, fully opened stage; S5, two weeks after the fully opened stage; S6, four weeks after the fully opened stage. Asterisks and different lowercase letters indicate significant differences among different treatments in leaves (Duncan’s test at p < 0.05 after analysis of variance; data are shown as mean ± SE). Figure 6 Influence of different pH treatments on the characteristics of stomata and chloroplast of P. suffruticosa ‘Luoyanghong’. (a) The morphology of stomata, bar = 50 µm; (b) stomata size; (c) stomata density; (d) stomata aperture; (e) expression patterns of genes involved in stomatal development and movement; (f) the structure of leaf and chloroplast with different magnification of 700× (left), 3000× (middle), and 50,000× (right); and (g) the characteristics of chloroplast. UE, upper epidermis; PL, palisade mesophyll; CW, cell wall; CP, chloroplast; GL, grana lamella; M, mitochondria; O, osmiophilic granule; SG, starch grain. S0, bud sprouting stage; S1, hard bud stage; S2, loose bud stage; S3, half open stage; S4, fully opened stage; S5, two weeks after the fully opened stage; S6, four weeks after the fully opened stage. Asterisks and different letters indicate significant differences from the control group (one-way ANOVA, p < 0.05). Figure 7 Expression pattern of DEGs involved in the light reactions of photosynthesis situated in the KEGG pathway by FPKM analysis. Figure 8 Expression pattern of DEGs shown to be involved in the Calvin cycle of photosynthesis by FPKM analysis. Figure 9 Influence of different pH treatments on H2O2 content, the activities of antioxidant enzymes, and nutrient assimilation. (a) H2O2 content, (b) SOD activity, (c) POD activity, (d) CAT activity, (e) heatmap of gene expression of ROS signal cascades and scavenging enzymes, (f,g) nutrient uptake, and (h) heatmap of gene expression of nutrient transporter. Different lowercase letters indicate significant differences among different treatments in leaves (Duncan’s test at p < 0.05 after analysis of variance; data are shown as mean ± SE). Figure 10 An overview of the effects of acidity and alkaline stress on flowering, plant growth, photosynthesis, oxidative stress, nutrient relation, and the regulation pathways in the tree peony plant. The upward arrow means increase, and the downward arrow means decrease. SAM is short for shoot apical meristem (SAM) and AM is short for axillary meristem, which are important for flowering and branching patterns, respectively. ijms-23-05094-t001_Table 1 Table 1 Influence of different pH treatments on the morphological characteristics of flower quality at the full flowering stage. Different lowercase letters indicate significant differences among different treatments in leaves (Duncan’s test at p < 0.05 after analysis of variance; data are shown as mean ± SE). Treatment Flower Diameter (cm) Flower Height (cm) Flower Stalk Length (cm) Flower Stalk Diameter (cm) Fresh Weight (g) Dry Weight (g) No. of Petals Abnormal Flowering Percentage (%) pH 4.0 10.77 ± 0.38b 3.84 ± 0.17b 20.83 ± 0.77b 6.9 ± 0.26b 15.24 ± 0.61b 3.12 ± 0.05a 75.66 ± 4.33b 26 pH 7.0 12.14 ± 0.28a 6.16 ± 0.29a 23.44 ± 0.62a 7.82 ± 0.28a 21.75 ± 0.82a 3.48 ± 0.13a 90.16 ± 2.18a 8 pH 10.0 10.27 ± 0.43b 4.05 ± 0.28b 21.94 ± 0.3ab 7.08 ± 0.18b 16.22 ± 0.73b 3.03 ± 0.08a 81.83 ± 2.3ab 33 ijms-23-05094-t002_Table 2 Table 2 Identification of the candidate genes involved in photosynthesis. No. Gene Name UniGene ID Sequence Length (bp) Coding Sequence Length (bp) Gene Name Homology Species and GenBank Number CDS Length of Homology Species (bp) PSI 1 PsaC CL5623.Contig3 All 1455 945 Photosystem I iron-sulfur center Actinidia chinensis, NKQK01000029.1 1212 2 PsaC CL553.Contig3 All 1288 213 Photosystem I iron-sulfur center Nicotiana tomentosiformis, XM 009612956.3 438 3 PsaF CL15133.Contig3 All 651 237 Photosystem I reaction center subunit III Abrus precatorius, XM 027497800.1 669 4 PsaE Unigene26014 All 723 501 photosystem I reaction center subunit IV Juglans regia, XM 018994420.2 573 5 PsaH1 CL553.Contig2 All 1316 216 photosystem I reaction center subunit VI-1 Nicotiana tomentosiformis, XM 009612956.3 438 6 PsaN Unigene16611 All 396 162 Photosystem I reaction center subunit N Cephalotus follicularis, BDDD01006702.1 429 7 PSA2 Unigene40033 All 1497 501 PHOTOSYSTEM I ASSEMBLY 2 Morus notabilis, XM 024170885.1 363 PSII 8 PsbB CL2899.Contig8 All 9368 1527 Photosystem II protein D2 protein Paeonia obovata, YP 009114474.1 1527 9 STN8 CL5574.Contig1 All 1222 489 Serine/threonine-protein kinase STN8 Pistacia vera, XM 031391586.1 1500 10 PsbH CL2899.Contig10 All 2735 345 Photosystem II phosphoprotein Paeonia obovata, NC 026076.1 222 11 PSBO1 CL7453.Contig2 All 1521 786 Oxygen-evolving enhancer protein 1 Nicotiana attenuata, XM 019379586.1 999 12 PPD3 CL2018.Contig1 All 1082 555 psbP domain-containing protein 3 Carica papaya, XM 022046797.1 762 13 PPD7 Unigene11530 All 1296 390 psbP domain-containing protein 7 Populus euphratica, XM 011038539.1 858 14 PSBS CL2734.Contig2 All 381 282 photosystem II 22 kDa Tanacetum cinerariifolium, BKCJ010081569.1 550 Light-harvesting complex 15 PAPP5 CL5682.Contig2 All 718 615 partial serine/threonine-protein phosphatase 5 Hibiscus syriacus, XM 039206886.1 1623 16 CAB6A CL7283.Contig4 All 683 234 chlorophyll a-b binding protein 6A Benincasa hispida, XM 039035318.1 741 17 CAB7 CL349.Contig1 All 399 399 chlorophyll a-b binding protein 7 Tripterygium wilfordii, XM 038846989.1 939 18 CAB5 Unigene44832 All 372 255 chlorophyll a-b binding protein 5 Amborella trichopoda, XM 006836691.3 795 19 CAB CL1191.Contig3 All 508 321 chlorophyll a-b binding protein Vitis vinifera, XM 010657584.1 816 20 CAB5 CL6697.Contig2 All 222 166 chlorophyll a-b binding protein 5 Telopea speciosissima, XM 043835320.1 825 Photosynthetic electron transport chain 21 ETFQO CL8017.Contig1 All 1732 732 partial electron transfer flavoprotein-ubiquinone oxidoreductase Cannabis sativa, XM 030652288.1 1329 22 CBR1 CL13977.Contig4 All 1536 738 NADH--cytochrome b5 reductase 1-like Vitis riparia, XM 034832700.1 837 23 NAXD CL2990.Contig3 All 1485 996 ATP-dependent NAD(P)H-hydrate dehydratase Juglans microcarpa x Juglans regia, XM 041145087.1 1137 24 Cytf Unigene46676 All 336 219 Cytochrome f Eurycoma longifolia, MH751519.1 963 25 CAO CL1939.Contig9 All 2422 1215 Chlorophyllide a oxygenase Juglans regia, XM 018952223.2 1605 26 fdxA CL647.Contig1 All 1425 468 3Fe-4S ferredoxin Lupinus albus, WOCE01000019.1 1014 27 MFDR Unigene11788 All 1231 957 NADPH:adrenodoxin oxidoreductase Carya illinoinensis, XM 043121482.1 1176 28 petE Unigene20162 All 747 555 Plastocyanin-like protein Corchorus olitorius, AWUE01016532.1 534 29 MNHD CL9837.Contig2 All 2528 324 NADH-Ubiquinone/plastoquinone complex I protein Prunus dulcis, AP021287.1 1944 30 ndhB2 CL12489.Contig1 All 243 165 NADH-plastoquinone oxidoreductase subunit 2 Iseilema macratherum, NC 030611.1 1533 31 ndh2 CL2308.Contig4 All 1627 213 Quinone oxidoreductase Vitis vinifera, QGNW01001796.1 1089 32 ndhO CL11886.Contig4 All 681 189 NAD(P)H-quinone oxidoreductase subunit O Carica papaya, XM 022044181.1 501 33 FBN5 CL269.Contig4 All 1159 792 Fibrillin-5 Prunus dulcis, XM 034364758.1 816 34 PNSL1 Unigene11863 All 697 357 Photosynthetic NDH subunit of lumenal location 1 Vitis vinifera, QGNW01000023.1 708 35 PNSL1 Unigene27888 All 1097 342 Photosynthetic NDH subunit of lumenal location 1 Vitis vinifera, QGNW01000023.1 708 36 PNSL3 Unigene10120 All 1308 633 Photosynthetic NDH subunit of lumenal location 3 Camellia sinensis, XM 028241493.1 666 37 DAPB3 CL1744.Contig1 All 1424 605 partial Dihydrodipicolinate reductase-like protein CRR1 Juglans regia, XM 018993624.2 903 Thylakoid formation and chloroplast development 38 THF1 CL14413.Contig2 All 971 759 THYLAKOID FORMATION 1 Senna tora, JAAIUW010000012.1 897 39 TERC CL2696.Contig4 All 993 804 thylakoid membrane protein TERC Camellia sinensis, XM 028237267.1 1077 40 RPL24 CL1839.Contig2 All 1030 480 50S ribosomal protein L24 Telopea speciosissima, XM 043834893.1 480 41 ALB3L2 Unigene39908 All 1042 666 ALBINO3-like protein 2 Quercus suber, XM 024024971.1 726 42 ALB3L2 CL1726.Contig1 All 1396 567 ALBINO3-like protein 2 Vitis riparia, XM 034854167.1 921 43 OBGL CL588.Contig4 All 1305 996 GTP-binding protein Nelumbo nucifera, XM 010251428.2 990 44 CSP41B CL2622.Contig3 All 1339 192 chloroplast stem-loop binding protein of 41 kDa b Cannabis sativa, XM 030649812.1 1146 Chlorophyll biosynthesis 45 CHLI1 CL10274.Contig3 All 3153 1083 magnesium chelatase subunit I (CHLI) Ziziphus jujuba, XM 016018749.2 1266 46 PPH CL6625.Contig1 All 1231 636 pheophytinase Manihot esculenta, XM 021744990.2 1116 47 CLH1 CL7533.Contig2 All 1010 690 chlorophyllase-1-like Vitis riparia, XM 034834294.1 960 48 HCAR Unigene52328 All 1076 585 7-hydroxymethyl chlorophyll a reductase Vitis vinifera, XM 019220129.1 1128 49 NOL CL5229.Contig4 All 1303 852 chlorophyll(ide) b reductase NOL Prunus mume, XM 008233562.2 1059 50 CAO Unigene50987 All 211 162 chlorophyll a oxygenase Capsicum annuum, DQ423120.1 299 Chlorophyll catabolism 51 PAO CL12667.Contig2 All 2501 444 partial pheophorbide a oxygenase Parasponia andersonii, JXTB01000035.1 1626 ATP synthase 52 ADK CL5310.Contig4 All 1007 714 adenylate kinase Tripterygium wilfordii, KAF5731549.1 897 53 ATPD CL80.Contig9 All 2583 267 ATP synthase delta chain Fragaria vesca, XM 004290397.2 753 54 ATPD CL80.Contig10 All 2488 267 ATP synthase delta chain Fragaria vesca, XM 004290397.2 753 55 ATPE CL8509.Contig1 All 1125 402 ATP synthase CF1 epsilon chain Paeonia ludlowii, NC 035623.1 402 56 ATPE CL633.Contig10 All 14,020 1527 ATP synthase CF1 alpha subunit Paeonia obovata, YP 009114434.1 1527 57 ATPD CL80.Contig15 All 2028 276 ATP synthase delta chain Fragaria vesca subsp. Vesca, XM 004290397.2 753 58 ATPD CL80.Contig14 All 1868 270 ATP synthase delta chain Fragaria vesca, XM 004290397.2 753 59 ATPA CL633.Contig6 All 13,886 1527 ATP synthase CF1 alpha subunit Paeonia obovata, YP 009114434.1 1527 60 ATPA CL633.Contig12 All 13,308 1527 ATP synthase CF1 alpha subunit Paeonia obovata, YP 009114434.1 1527 61 AK4 CL9808.Contig2 All 806 669 Adenylate kinase 4 Populus alba, XM 035053161.1 741 Carbon fixation 62 CBBY-like CL3370.Contig1 All 936 708 riboflavin kinase Bifunctional riboflavin kinase/FMN phosphatase Camellia sinensis, XM 028260557.1 891 63 CA1P CL13509.Contig2 All 1869 900 2-carboxy-D-arabinitol-1-phosphatase Camellia sinensis, XM 028230931.1 1605 64 PGM2 CL3429.Contig1 All 1293 243 phosphoglycerate mutase, 2,3-bisphosphoglycerate-independent Actinidia rufa, GFZ05492.1 258 65 RBCS-F1 CL4922.Contig11 All 700 279 ribulose-1,5-bisphosphate carboxylase small chain F1 Lupinus angustifolius, XM 019587259.1 531 66 RBCS CL4922.Contig12 All 290 261 ribulose bisphosphate carboxylase small chain Tanacetum cinerariifolium, BKCJ010045582.1 1899 67 RBCS CL4922.Contig3 All 310 251 ribulose bisphosphate carboxylase small subunit Carya illinoinensis, XM 043105128.1 549 68 PGK CL3299.Contig2 All 1682 1368 phosphoglycerate kinase Herrania umbratica, XM 021443140.1 1206 69 FBA CL12763.Contig2 All 754 216 Fructose-bisphosphate aldolase Apostasia shenzhenica, KZ451885.1 1110 70 Fbp CL1239.Contig5 All 770 219 fructose-1,6-bisphosphatase Vigna angularis, XM 017556201.1 477 71 PFKFB Unigene23926 All 2108 1788 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase Vitis riparia, XM 034831796.1 2271 72 TKL CL7000.Contig2 All 846 321 transketolase Fragaria vesca, FJ887833.1 363 Stomatal development and movement 73 BCA Unigene56632 All 1532 414 carbonic anhydrase Vitis riparia, XM 034820072.1 984 74 HT1 Unigene2658 All 2008 1122 Serine/threonine-protein kinase HT1 Vitis vinifera, XM 002270717.3 1125 ijms-23-05094-t003_Table 3 Table 3 Identification of the candidate genes involved in stress and nutrient transport. No. Gene Name UniGene ID Sequence Length (bp) Coding Sequence Length (bp) Gene Name Homology Species and GenBank Number CDS Length of Homology Species (bp) Stress-related genes 1 SOD CL12133.Contig2 All 752 663 Superoxide dismutase [Fe] Quercus suber, XM 024033160.1 936 2 SOD CL11553.Contig1_All 235 114 superoxide dismutase Vitis riparia, XM_034849451.1 687 3 CAT1 CL8683.Contig2_All 241 150 Catalase isozyme 1 Cocos nucifera, CM017878.1 2833 4 CAT CL8683.Contig3_All 231 129 Catalase Thalictrum thalictroides, ABWDY010037995.1 1464 5 POD20 Unigene63828_All 233 233 Peroxidase 20 Populus trichocarpa, XM006368377.2 1017 6 POD20 Unigene2231_All 264 264 Peroxidase 20 Vitis vinifera, QGNW01000083.1 855 7 POD45 Unigene79660_All 336 336 Peroxidase 45 Gossypium hirsutum, XM016865128.2 996 8 POD Unigene6219_All 895 838 Peroxidase Thalictrum thalictroides, JABWDY010016849.1 972 9 POD48 Unigene82390_All 305 305 Peroxidase 48 Malus domestica, XM029108818.1 699 10 PEX12 CL7714.Contig3 All 1173 840 Peroxisome biogenesis protein 12 Ziziphus jujuba, XM 016020053.2 1182 11 CYB561A CL11875.Contig1 All 873 669 Transmembrane ascorbate ferrireductase 3 Coffea arabica, XM 027240953.1 663 12 TCTP Unigene79230 All 375 300 partial Translationally-controlled tumor protein Capra hircus, XM 018056766.1 660 13 AGO1 CL4080.Contig4 All 3683 3255 Protein argonaute 1 Vitis vinifera, XM 002271189.3 3258 14 RFC1 CL4788.Contig6 All 1558 1167 Replication factor C subunit 1-like Populus euphratica, XM 011015553.1 1302 15 EX1 CL5897.Contig2_All 2799 1380 EXECUTER 1 Camellia sinensis, XM_028237292.1 1488 16 EX1 CL5897.Contig4_All 2822 1095 protein EXECUTER 1 Camellia sinensis, XM_028237292.1 1488 Ion transport 17 ACA12 Unigene17975_All 611 534 Calcium-transporting ATPase 12 Glycine soja, XM_028369732.1 3162 18 BOR1 Unigene48447_All 566 174 boron transporter 1 Zingiber officinale, XM_042585939.1 2139 19 BOR4 Unigene41688_All 263 212 boron transporter 4-like Ricinus communis, XM_025157068.1 1923 20 BOR4 CL3601.Contig5_All 1903 1062 boron transporter 4-like Camellia sinensis, XM_028235763.1 2136 21 PPT2 CL4107.Contig2_All 2381 855 phosphoenolpyruvate/phosphate translocator 2 Prunus avium, XM_021948621.1 1014 22 PHO1 Unigene39529_All 1188 609 Phosphate transporter PHO1-like 3 Vitis vinifera, QGNW01000145.1 2394 23 TDT Unigene43599_All 977 588 tonoplast dicarboxylate transporter Camellia sinensis, XM_028254388.1 1644 24 PHO1 CL1982.Contig2_All 5028 774 phosphate transporter PHO1 homolog 3 Cannabis sativa, XM_030628536.1 2424 25 PHO1 Unigene40561_All 727 354 phosphate transporter PHO1 homolog 3-like Vitis vinifera, XM_019218049.1 614 26 glpT CL8237.Contig4_All 2509 1053 glycerol-3-phosphate transporter 1 Prunus avium, XM_021967732.1 1563 27 PHO1 CL14859.Contig1_All 1305 945 phosphate transporter PHO1 homolog 3-like Carya illinoinensis, XM_043134065.1 2406 28 SKOR Unigene48839_All 228 226 potassium channel SKOR-like Herrania umbratica, XM_021436349.1 2448 29 POT17 Unigene56892_All 1885 1593 potassium transporter 17 Rosa chinensis, XM_024336662.2 1785 ijms-23-05094-t004_Table 4 Table 4 Identification of the candidate genes involved in growth, flowering, and related regulation metabolism. No. Gene Name UniGene ID Sequence Length (bp) Coding Sequence Length (bp) Gene Name Homology Species and GenBank Number CDS Length of Homology Species (bp) Flowering-related genes 1 PSD Unigene8 All 1646 216 partial Exportin-T-like isoform X2 Populus alba, XM 035040257.1 2892 2 APRR1 CL5436.Contig4 All 1643 813 partial Two-component response regulator-like APRR1 Vitis vinifera, QGNW01000154.1 1662 3 JMJ18 CL1104.Contig4 All 3551 2940 Lysine-specific demethylase JMJ18-like isoform X1 Vitis riparia, XM 034842267.1 3261 4 WOX1 CL8062.Contig3 All 1081 645 WUSCHEL-related homeobox 1-like, partial Macadamia integrifolia, XM 042620797.1 825 5 FY CL8788.Contig3 All 2714 2313 Flowering time control protein FY Vitis vinifera, XM 010647342.2 2346 6 AAO CL5090.Contig6 All 1541 723 F-box/FBD/LRR-repeat protein Sesamum indicum, XM 011078410.2 1311 Plant-growth-related genes 7 TGH CL845.Contig2 All 999 843 G-patch domain-containing protein TGH Actinidia rufa, BJWL01000006.1 822 8 MIZ1 CL9598.Contig2 All 895 705 Protein MIZU-KUSSEI 1 Ricinus communis, XM 002510226.3 690 9 BBX19 CL856.Contig1 All 1081 672 B-box zinc finger protein 19 isoform X1 Ziziphus jujuba, XM 016036814.2 639 Genes involved in cell growth and division 10 SEC10 CL14716.Contig6 All 3072 2505 Exocyst complex component Sec10-like Parasponia andersonii, JXTB01000066.1 2538 11 CDC27B CL7886.Contig1 All 1665 735 partial Cell division cycle protein 27 homolog B isoform X2 Manihot esculenta, XM 021741898.2 2190 12 CALS7 CL542.Contig25 All 2923 360 partial Callose synthase 7-like Quercus suber, XM 024021939.1 2802 13 TUBA5 CL4204.Contig2 All 1741 1350 Tubulin alpha-5 chain Macadamia integrifolia, XM 042625060.1 1353 14 TSS CL4166.Contig4 All 6317 5649 TSS Vitis vinifera, XM 002278334.4 5592 15 EXPA1 CL12917.Contig4 All 1172 753 Expansin-A1 Fragaria vesca, XM 004297244.2 756 Hormone-related genes 16 TMK1 CL2731.Contig4 All 3327 2901 Receptor protein kinase TMK1-like Vitis vinifera, XM 002274874.4 2883 17 YUC Unigene53925 All 418 315 partial Indole-3-pyruvate monooxygenase YUCCA3 Cajanus cajan, XM 020366995.2 1293 18 AFB2 CL789.Contig8 All 2719 1719 AUXIN SIGNALING F-BOX 2 Vitis vinifera, XM 019225111.1 1719 19 CTR1 CL7684.Contig1 All 4854 4242 Serine/threonine-protein kinase CTR1 Vitis vinifera, QGNW01001367.1 4296 Signal transduction 20 CML38 Unigene2096 All 371 345 Calcium-binding protein CML38-like Mangifera indica, XM 044647080.1 423 21 SK5 CL5696.Contig3 All 1165 1050 Calcium-dependent protein kinase SK5-like Hevea brasiliensis, XM 021816377.1 1629 22 IQD14 Unigene40182 All 1308 480 partial IQ-DOMAIN 14 isoform X1 Senna tora, JAAIUW010000012.1 1536 Regulation of transcription and translation 23 BPG2 CL6966.Contig2 All 2327 1671 BRASSINAZOLE INSENSITIVE PALE GREEN 2 Vitis vinifera, XM 010655630.2 1983 24 MTERF4 CL4833.Contig7 All 2113 1578 Transcription termination factor MTERF4 Carya illinoinensis, XM 043135353.1 1575 25 ASK1 Unigene19526 All 2064 1230 Shaggy-related protein kinase alpha Quercus lobata, XM 031102093.1 1230 26 BLH1 CL1863.Contig5 All 3450 2073 BEL1-like homeodomain protein 1 Morella rubra, RXIC02000092.1 2091 27 FRS Unigene39555 All 1222 399 partial FAR1-RELATED SEQUENCE 5-like Rosa chinensis, XM 040508409.1 1290 28 MBD Unigene46268 All 905 816 Methyl-CpG-binding domain-containing protein 10-like isoform X2 Juglans microcarpa, XM 041162513.1 822 29 MBD7 CL7329.Contig2 All 958 885 Methyl-CpG-binding domain-containing protein 5 Carica papaya, XM 022038861.1 738 30 YRDC CL3635.Contig6 All 1361 873 YRDC domain-containing protein Theobroma cacao, XM 007016659.2 837 31 PRP40A CL379.Contig13 All 3502 3117 Pre-mRNA-processing protein 40A Nelumbo nucifera, XM 010275221.2 3141 32 PRP8A Unigene23105 All 3667 2472 partial Pre-mRNA-processing-splicing factor 8A Vitis vinifera, XM 003632714.3 7044 33 eIF3f Unigene39851 All 1004 597 Eukaryotic translation initiation factor 3 subunit F-like Telopea speciosissima, XM 043846840.1 858 34 SPL39 CL6274.Contig5 All 1738 1020 Squamosa-binding protein-like 39 Paeonia suffruticosa, MT239473.1 2700 35 U2AF65B CL12897.Contig1 All 942 537 partial Splicing factor U2af large subunit B-like isoform X1 Telopea speciosissima, XM 043832750.1 1665 36 GBF3 CL1061.Contig4 All 1254 909 G-box-binding factor 3 Benincasa hispida, XM 039043265.1 1254 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Guo L. Guo S. Xu J. He L. Carlson J.E. Hou X. Phylogenetic analysis based on chloroplast genome uncover evolutionary relationship of all the nine species and six cultivars of tree peony Ind. Crops Prod. 2020 153 112567 10.1016/j.indcrop.2020.112567 2. Yang Y. Sun M. Li S.S. Chen Q.H. da Silva J.A.T. Wang A.J. Yu X.N. Wang L.S. Germplasm resources and genetic breeding of Paeonia: A systematic review Hortic. Res. 2020 7 107 10.1038/s41438-020-0332-2 32637135 3. Zhang Y.C. Ge M. Wang X.H. Zeng Y. Liao Y.X. Wang Y.P. Zeng Z.L. Discrimination of Moutan Cortex from different sources and geographical origins for quality evaluation using microsatellite markers coupled with chemical analysis Biochem. Syst. Ecol. 2020 89 104005 10.1016/j.bse.2020.104005 4. Kochian L.V. Hoekenga O.A. Pineros M.A. How do crop plants tolerate acid soils? Mechanisms of aluminum tolerance and phosphorous efficiency Annu. Rev. Plant Biol. 2004 55 459 493 10.1146/annurev.arplant.55.031903.141655 15377228 5. Wang L. Seki K. Miyazaki T. Ishihama Y. The causes of soil alkalinization in the Songnen Plain of Northeast China Paddy Water Environ. 2009 7 259 270 10.1007/s10333-009-0166-x 6. Zhao D.Q. Hao Z.J. Wang J. Tao J. Effects of pH in irrigation water on plant growth and flower quality in herbaceous peony (Paeonia lactiflora Pall.) Sci. Hortic. 2013 154 45 53 10.1016/j.scienta.2013.02.023 7. Herb M. Gluschko A. Schramm M. Reactive Oxygen Species: Not omnipresent but important in many locations Front. Cell Dev. Biol. 2021 9 716406 10.3389/fcell.2021.716406 34557488 8. Long A. Zhang J. Yang L.T. Ye X. Lai N.W. Tan L.L. Lin D. Chen L.S. Effects of low pH on photosynthesis, related physiological parameters, and nutrient profiles of citrus Front. Plant Sci. 2017 8 185 10.3389/fpls.2017.00185 28270819 9. Zhang H.H. Li X. Xu Z.S. Wang Y. Teng Z.Y. An M.J. Zhang Y.H. Zhu W.X. Xu N. Sun G.Y. Toxic effects of heavy metals Pb and Cd on mulberry (Morus alba L.) seedling leaves: Photosynthetic function and reactive oxygen species (ROS) metabolism responses Ecotoxicol. Environ. Saf. 2020 195 110469 10.1016/j.ecoenv.2020.110469 32179235 10. Wang Y. Yu Y. Zhang H. Huo Y. Liu X. Che Y. Wang J. Sun G. Zhang H. The phytotoxicity of exposure to two polybrominated diphenyl ethers (BDE47 and BDE209) on photosynthesis and the response of the hormone signaling and ROS scavenging system in tobacco leaves J. Hazard. Mater. 2022 426 128012 10.1016/j.jhazmat.2021.128012 34923383 11. Wang Y. Zhu H. Tam N.F.Y. Effect of a polybrominated diphenyl ether congener (BDE-47) on growth and antioxidative enzymes of two mangrove plant species, Kandelia obovata and Avicennia marina, in South China Mar. Pollut. Bull. 2014 85 376 384 10.1016/j.marpolbul.2014.02.012 24631399 12. Johnson M.P. Photosynthesis Essays Biochem. 2016 60 255 273 10.1042/EBC20160016 27784776 13. Zhou X.L. Wang Y.H. Shen S.K. Transcriptomic comparison reveals modifications in gene expression, photosynthesis, and cell wall in woody plant as responses to external pH changes Ecotoxicol. Environ. Saf. 2020 203 111007 10.1016/j.ecoenv.2020.111007 32888586 14. Guo R. Yang Z. Li F. Yan C. Zhong X. Liu Q. Xia X. Li H. Zhao L. Comparative metabolic responses and adaptive strategies of wheat (Triticum aestivum) to salt and alkali stress BMC Plant Biol. 2015 15 170 10.1186/s12870-015-0546-x 26149720 15. Chen Q. Wu W. Zhao T. Tan W. Tian J. Liang C. Complex gene regulation underlying mineral nutrient homeostasis in soybean root response to acidity stress Genes 2019 10 402 10.3390/genes10050402 31137896 16. Zhang C. Yi X. Zhou F. Gao X. Wang M. Chen J. Huang J. Shen C. Comprehensive transcriptome profiling of tea leaves (Camellia sinensis) in response to simulated acid rain Sci. Hortic. 2020 272 109491 10.1016/j.scienta.2020.109491 17. Sun J. Hu H. Li Y. Wang L. Zhou Q. Huang X. Effects and mechanism of acid rain on plant chloroplast ATP synthase Environ. Sci. Pollut. Res. Int. 2016 23 18296 18306 10.1007/s11356-016-7016-3 27278067 18. Yin Z. Zhang H. Zhao Q. Yoo M.-J. Zhu N. Yu J. Yu J. Guo S. Miao Y. Chen S. Physiological and comparative proteomic analyses of saline-alkali NaHCO3-responses in leaves of halophyte Puccinellia tenuiflora Plant Soil 2019 437 137 158 10.1007/s11104-019-03955-9 19. Moreau H. Zimmermann S.D. Gaillard I. Paris N. pH biosensing in the plant apoplast—A focus on root cell elongation J. Plant Physiol. 2021 187 504 514 10.1093/plphys/kiab313 20. Kumagai T. Ito S. Nakamichi N. Niwa Y. Murakami M. Yamashino T. Mizuno T. The Common Function of a Novel Subfamily of B-Box Zinc Finger Proteins with Reference to Circadian-Associated Events in Arabidopsis thaliana Biosci. Biotechnol. Biochem. 2008 72 1539 1549 10.1271/bbb.80041 18540109 21. Hu H. Boisson-Dernier A. Israelsson-Nordström M. Böhmer M. Xue S. Ries A. Godoski J. Kuhn J.M. Schroeder J.I. Carbonic anhydrases are upstream regulators of CO2-controlled stomatal movements in guard cells Nat. Cell Biol. 2010 12 87 93 10.1038/ncb2009 20010812 22. Engineer C.B. Ghassemian M. Anderson J.C. Peck S.C. Hu H. Schroeder J.I. Carbonic anhydrases, EPF2 and a novel protease mediate CO2 control of stomatal development Nature 2014 513 246 250 10.1038/nature13452 25043023 23. Hashimoto M. Negi J. Young J. Israelsson M. Schroeder J.I. Iba K. Arabidopsis HT1 kinase controls stomatal movements in response to CO2 Nat. Cell Biol. 2006 8 391 397 10.1038/ncb1387 16518390 24. Geng G. Wang G. Stevanato P. Lv C. Wang Q. Yu L. Wang Y. Physiological and Proteomic Analysis of Different Molecular Mechanisms of Sugar Beet Response to Acidic and Alkaline pH Environment Front. Plant Sci. 2021 12 682799 10.3389/fpls.2021.682799 34178001 25. Zhang Y.K. Zhu D.F. Zhang Y.P. Chen H.Z. Xiang J. Lin X.Q. Low pH-induced changes of antioxidant enzyme and ATPase activities in the roots of rice (Oryza sativa L.) seedlings PLoS ONE 2015 10 e0116971 10.1371/journal.pone.0116971 25719552 26. Lv B.-S. Li X.-W. Ma H.-Y. Sun Y. Wei L.-X. Jiang C.-J. Liang Z.-W. Differences in Growth and Physiology of Rice in Response to Different Saline-Alkaline Stress Factors J. Agron. 2013 105 1119 1128 10.2134/agronj2013.0017 27. Wei L.X. Lv B.S. Wang M.M. Ma H.Y. Yang H.Y. Liu X.L. Jiang C.J. Liang Z.W. Priming effect of abscisic acid on alkaline stress tolerance in rice (Oryza sativa L.) seedlings Plant Physiol. Biochem. 2015 90 50 57 10.1016/j.plaphy.2015.03.002 25780993 28. Fukada-Tanaka S. Inagaki Y. Yamaguchi T. Saito N. Iida S. Colour-enhancing protein in blue petals Nature 2000 407 581 10.1038/35036683 29. Hunter C.A. Aukerman M.J. Sun H. Fokina M. Poethig R.S. PAUSED Encodes the Arabidopsis Exportin-t Ortholog J. Plant Physiol. 2003 132 2135 2143 10.1104/pp.103.023309 30. Matsushika A. Makino S. Kojima M. Mizuno T. Circadian Waves of Expression of the APRR1/TOC1 Family of Pseudo-Response Regulators in Arabidopsis thaliana: Insight into the Plant Circadian Clock Plant Cell Physiol. 2000 41 1002 1012 10.1093/pcp/pcd043 11100772 31. Yang H. Han Z. Cao Y. Fan D. Li H. Mo H. Feng Y. Liu L. Wang Z. Yue Y. A companion cell-dominant and developmentally regulated H3K4 demethylase controls flowering time in Arabidopsis via the repression of FLC expression PLoS Genet. 2012 8 e1002664 10.1371/journal.pgen.1002664 22536163 32. Henderson I.R. Liu F. Drea S. Simpson G.G. Dean C. An allelic series reveals essential roles for FY in plant development in addition to flowering-time control Development 2005 132 3597 3607 10.1242/dev.01924 16033802 33. Calderon-Villalobos L.I.A. Kuhnle C. Dohmann E.M.N. Li H. Bevan M. Schwechheimer C. The Evolutionarily Conserved TOUGH Protein Is Required for Proper Development of Arabidopsis thaliana Plant Cell 2005 17 2473 2485 10.1105/tpc.105.031302 16024589 34. Moriwaki T. Miyazawa Y. Kobayashi A. Uchida M. Watanabe C. Fujii N. Takahashi H. Hormonal Regulation of Lateral Root Development in Arabidopsis Modulated by MIZ1 and Requirement of GNOM Activity for MIZ1 Function J. Plant Physiol. 2011 157 1209 1220 10.1104/pp.111.186270 21940997 35. Fendrych M. Synek L. Pečenková T. Toupalová H. Cole R. Drdová E. Nebesářová J. Šedinová M. Hála M. Fowler J.E. The Arabidopsis Exocyst Complex Is Involved in Cytokinesis and Cell Plate Maturation Plant Cell 2010 22 3053 3065 10.1105/tpc.110.074351 20870962 36. Blilou I. Frugier F. Folmer S. Serralbo O. Willemsen V. Wolkenfelt H. Eloy N.B. Ferreira P.C. Weisbeek P. Scheres B. The Arabidopsis HOBBIT gene encodes a CDC27 homolog that links the plant cell cycle to progression of cell differentiation Genes Dev. 2002 16 2566 2575 10.1101/gad.237302 12368267 37. Chun H.J. Baek D. Jin B.J. Cho H.M. Park M.S. Lee S.H. Lim L.H. Cha Y.J. Bae D.W. Kim S.T. Microtubule Dynamics Plays a Vital Role in Plant Adaptation and Tolerance to Salt Stress Int. J. Mol. Sci. 2021 22 5957 10.3390/ijms22115957 34073070 38. Lv B. Yan Z. Tian H. Zhang X. Ding Z. Local Auxin Biosynthesis Mediates Plant Growth and Development Trends Plant Sci. 2019 24 6 9 10.1016/j.tplants.2018.10.014 30448230 39. Lackner D.H. Schmidt M.W. Wu S. Wolf D.A. Bähler J. Regulation of transcriptome, translation, and proteome in response to environmental stress in fission yeast Genome Biol. 2012 13 R25 10.1186/gb-2012-13-4-r25 22512868 40. Lauterbach M. van der Merwe P.d.W. Keßler L. Pirie M.D. Bellstedt D.U. Kadereit G. Evolution of leaf anatomy in arid environments—A case study in southern African Tetraena and Roepera (Zygophyllaceae) Mol. Phylogenetics Evol. 2016 97 129 144 10.1016/j.ympev.2016.01.002 41. Gotoh E. Suetsugu N. Higa T. Matsushita T. Tsukaya H. Wada M. Palisade cell shape affects the light-induced chloroplast movements and leaf photosynthesis Sci. Rep. 2018 8 1472 10.1038/s41598-018-19896-9 29367686 42. Bishehkolaei R. Fahimi H. Saadatmand S. Nejadsattari T. Lahouti M. Yazdi F.T. Ultrastructural localisation of chromium in Ocimum basilicum Turk. J. Bot. 2011 35 261 268 43. Zhang J. Wang S. Song S. Xu F. Pan Y. Wang H. Transcriptomic and proteomic analyses reveal new insight into chlorophyll synthesis and chloroplast structure of maize leaves under zinc deficiency stress J. Proteom. 2019 199 123 134 10.1016/j.jprot.2019.03.001 44. Garty J. Tamir O. Hassid I. Eshel A. Cohen Y. Karnieli A. Orlovsky L. Photosynthesis, Chlorophyll Integrity, and Spectral Reflectance in Lichens Exposed to Air Pollution J. Environ. Qual. 2001 30 884 893 10.2134/jeq2001.303884x 11401277 45. Khan A. Kamran M. Imran M. Al-Harrasi A. Al-Rawahi A. Al-Amri I. Lee I.-J. Khan A.L. Silicon and salicylic acid confer high-pH stress tolerance in tomato seedlings Sci. Rep. 2019 9 19788 10.1038/s41598-019-55651-4 31874969 46. Mir M.A. John R. Alyemeni M.N. Alam P. Ahmad P. Jasmonic acid ameliorates alkaline stress by improving growth performance, ascorbate glutathione cycle and glyoxylase system in maize seedlings Sci. Rep. 2018 8 2831 10.1038/s41598-018-21097-3 29434207 47. Liu Y. Wang Q. Zhang Y. Cui J. Chen G. Xie B. Wu C. Liu H. Synergistic and antagonistic effects of salinity and pH on germination in switchgrass (Panicum virgatum L.) PLoS ONE 2014 9 e85282 10.1371/journal.pone.0085282 24454834 48. Wang H. Wu Z. Chen Y. Yang C. Shi D. Effects of salt and alkali stresses on growth and ion balance in rice (Oryza sativa L.) Plant Soil Environ. 2011 57 286 294 10.17221/36/2011-PSE 49. De la Guardia M.D. Alcántara E. A comparison of ferric-chelate reductase and chlorophyll and growth ratios as indices of selection of quince, pear and olive genotypes under iron deficiency stress Plant Soil 2002 241 49 56 10.1023/A:1016083512158 50. Lu Y. Gan Q. Iwai M. Alboresi A. Burlacot A. Dautermann O. Takahashi H. Crisanto T. Peltier G. Morosinotto T. Role of an ancient light-harvesting protein of PSI in light absorption and photoprotection Nat. Commun. 2021 12 679 10.1038/s41467-021-20967-1 33514722 51. Ran X. Wang X. Gao X. Liang H. Liu B. Huang X. Effects of salt stress on the photosynthetic physiology and mineral ion absorption and distribution in white willow (Salix alba L.) PLoS ONE 2021 16 e0260086 10.1371/journal.pone.0260086 34793527 52. Vialet-Chabrand S.R.M. Matthews J.S.A. McAusland L. Blatt M.R. Griffiths H. Lawson T. Temporal Dynamics of Stomatal Behavior: Modeling and Implications for Photosynthesis and Water Use J. Plant Physiol. 2017 174 603 613 10.1104/pp.17.00125 28363993 53. Chaves M.M. Flexas J. Pinheiro C. Photosynthesis under drought and salt stress: Regulation mechanisms from whole plant to cell Ann. Bot. 2009 103 551 560 10.1093/aob/mcn125 18662937 54. Franks P.J. Farquhar G.D. The mechanical diversity of stomata and its significance in gas-exchange control J. Plant Physiol. 2007 143 78 87 10.1104/pp.106.089367 55. Yan F. Sun Y. Song F. Liu F. Differential responses of stomatal morphology to partial root-zone drying and deficit irrigation in potato leaves under varied nitrogen rates Sci. Hortic. 2012 145 76 83 10.1016/j.scienta.2012.07.026 56. Büssis D. von Groll U. Fisahn J. Altmann T. Stomatal aperture can compensate altered stomatal density in Arabidopsis thaliana at growth light conditions Funct. Plant Biol. 2006 33 1037 1043 10.1071/FP06078 32689314 57. Limin Y. Mei H. Guangsheng Z. Jiandong L. The changes in water-use efficiency and stoma density of Leymus chinensis along Northeast China Transect Acta Ecologica. Sinica. 2007 27 16 23 10.1016/S1872-2032(07)60006-7 58. Shu S. Chen L. Lu W. Sun J. Guo S. Yuan Y. Li J. Effects of exogenous spermidine on photosynthetic capacity and expression of Calvin cycle genes in salt-stressed cucumber seedlings J. Plant Res. 2014 127 763 773 10.1007/s10265-014-0653-z 25069716 59. Sayed A.I.H. El-hamahmy M.A.M. Rafudeen M.S. Ebrahim M.K.H. Exogenous spermidine enhances expression of Calvin cycle genes and photosynthetic efficiency in sweet sorghum seedlings under salt stress Biol. Plant. 2019 63 511 518 60. Artiola J. Rasmussen C. Freitas R. Effects of a Biochar-Amended Alkaline Soil on the Growth of Romaine Lettuce and Bermudagrass Soil Sci. 2012 177 561 570 10.1097/SS.0b013e31826ba908 61. Gao Z. Han J. Mu C. Lin J. Li X. Lin L. Sun S. Effects of Saline and Alkaline Stresses on Growth and Physiological Changes in Oat (Avena sativa L.) Seedlings Not. Bot. Horti. Agrobot. Cluj-Napoca 2014 42 357 362 10.15835/nbha4229441 62. Dotaniya M.L. Pingoliya K.K. Lata M. Verma R. Regar K.L. Role of phosphorus in chickpea (Cicer arietinum L.) production Afr. J. Agric. Res. 2014 9 3736 3743 63. Meng X. Chen W.W. Wang Y.Y. Huang Z.R. Ye X. Chen L.S. Yang L.T. Effects of phosphorus deficiency on the absorption of mineral nutrients, photosynthetic system performance and antioxidant metabolism in Citrus grandis PLoS ONE 2021 16 e0246944 10.1371/journal.pone.0246944 33596244 64. Sirisuntornlak N. Ullah H. Sonjaroon W. Anusontpornperm S. Arirob W. Datta A. Interactive Effects of Silicon and Soil pH on Growth, Yield and Nutrient Uptake of Maize Silicon 2021 13 289 299 10.1007/s12633-020-00427-z 65. Alimuddin S. Musa Y. Azrai M. Asrul L. Effect of boron on flowering, yield components and grain yield of two prolific maize (Zea mays L.) hybrids IOP Conference Series: Environ. Earth Sci. 2020 484 012076 10.1088/1755-1315/484/1/012076 66. Goldberg S. Lesch S.M. Suarez D.L. Predicting selenite adsorption by soils using soil chemical parameters in the constant capacitance model Geochim. Et Cosmochim. Acta 2007 71 5750 5762 10.1016/j.gca.2007.04.036 67. Wang N. Yang C. Pan Z. Liu Y. Peng S. Boron deficiency in woody plants: Various responses and tolerance mechanisms Front. Plant Sci. 2015 6 916 10.3389/fpls.2015.00916 26579163 68. Roosta H.R. Interaction between water alkalinity and nutrient solution ph on the vegetative growth, chlorophyll fluorescence and leaf magnesium, iron, manganese, and zinc concentrations in lettuce J. Plant Nutr. 2011 34 717 731 10.1080/01904167.2011.540687 69. Millaleo R. Reyes- Diaz M. Ivanov A.G. Mora M.L. Alberdi M. Manganese as essential and toxic element for plants: Transport, accumulation and resistance mechanisms Soil Sci. Plant Nutr. 2010 10 470 481 10.4067/S0718-95162010000200008 70. George E. Horst W.J. Neumann E. Chapter 17—Adaptation of Plants to Adverse Chemical Soil Conditions Marschner’s Mineral Nutrition of Higher Plants 3rd ed. Marschner P. Academic Press San Diego, CA, USA 2012 409 472 10.1016/B978-0-12-384905-2.00017-0 71. Hu X.F. Wu A.Q. Wang F.C. Chen F.S. The effects of simulated acid rain on internal nutrient cycling and the ratios of Mg, Al, Ca, N, and P in tea plants of a subtropical plantation Environ. Monit. Assess. 2019 191 99 10.1007/s10661-019-7248-z 30680466 72. Duan X. Hu X. Chen F. Deng Z. Effects of Simulated Acid Rain and Aluminum Enrichment on Growth and Photosynthesis of Tea Seedlings Adv. Mat. Res. 2012 610–613 181 185 10.4028/www.scientific.net/AMR.610-613.181 73. Henriques F.S. Gas exchange, chlorophyll a fluorescence kinetics and lipid peroxidation of pecan leaves with varying manganese concentrations Plant Sci. J. 2003 165 239 244 10.1016/S0168-9452(03)00163-8 74. Verma K. Zeng X.-P. Zeng Y. Li D.-M. Guo D.-J. Rajput V. Chen G.-L. Barakhov A. Minkina T. Li Y. Characteristics of leaf stomata and their relationship with photosynthesis in Saccharum officinarum under drought and silicon application ACS Omega 2020 5 24145 24153 10.1021/acsomega.0c03820 32984737 75. Tavakkoli E. Lyons G. English P. Guppy C.N. Silicon nutrition of rice is affected by soil pH, weathering and silicon fertilization Soil Sci. Plant Nutr. 2011 174 437 446 10.1002/jpln.201000023 76. Jiang Y. Huang B. Effects of calcium on antioxidant activities and water relations associated with heat tolerance in two cool-season grasses J. Exp. Bot. 2001 52 341 349 10.1093/jexbot/52.355.341 11283179 77. Sharma P. Jha A.B. Dubey R.S. Pessarakli M. Reactive oxygen species, oxidative damage, and antioxidative defense mechanism in plants under stressful conditions J. Bot. 2012 2012 217037 10.1155/2012/217037 78. Castro B. Citterico M. Kimura S. Stevens D.M. Wrzaczek M. Coaker G. Stress-induced reactive oxygen species compartmentalization, perception and signalling Nat. Plants 2021 7 403 412 10.1038/s41477-021-00887-0 33846592 79. Noctor G. Mhamdi A. Foyer C.H. The roles of reactive oxygen metabolism in drought: Not so cut and dried J. Plant Physiol. 2014 164 1636 1648 10.1104/pp.113.233478 80. Wang L. Sadeghnezhad E. Guan P. Gong P. Review: Microtubules monitor calcium and reactive oxygen species signatures in signal transduction Plant Sci. J. 2021 304 110589 10.1016/j.plantsci.2020.110589 81. Kim C. Meskauskiene R. Zhang S. Lee K.P. Lakshmanan Ashok M. Blajecka K. Herrfurth C. Feussner I. Apel K. Chloroplasts of Arabidopsis are the source and a primary target of a plant-specific programmed cell death signaling pathway Plant Cell 2012 24 3026 3039 10.1105/tpc.112.100479 22797473 82. Yang J. Sardar H.S. McGovern K.R. Zhang Y. Showalter A.M. A lysine-rich arabinogalactan protein in Arabidopsis is essential for plant growth and development, including cell division and expansion Plant J. 2007 49 629 640 10.1111/j.1365-313X.2006.02985.x 17217456 83. Kim H. Hwang H. Hong J.W. Lee Y.N. Ahn I.P. Yoon I.S. Yoo S.D. Lee S. Lee S.C. Kim B.G. A rice orthologue of the ABA receptor, OsPYL/RCAR5, is a positive regulator of the ABA signal transduction pathway in seed germination and early seedling growth J. Exp. Bot. 2012 63 1013 1024 10.1093/jxb/err338 22071266 84. Osakabe Y. Maruyama K. Seki M. Satou M. Shinozaki K. Yamaguchi-Shinozaki K. Leucine-rich repeat receptor-like kinase1 is a key membrane-bound regulator of abscisic acid early signaling in Arabidopsis Plant Cell 2005 17 1105 1119 10.1105/tpc.104.027474 15772289 85. Han S. Min M.K. Lee S.Y. Lim C.W. Bhatnagar N. Lee Y. Shin D. Chung K.Y. Lee S.C. Kim B.G. Modulation of ABA signaling by altering motif of PP2Cs in Oryza sativa Mol. Plant 2017 10 1190 1205 10.1016/j.molp.2017.08.003 28827170 86. Takahashi F. Mizoguchi T. Yoshida R. Ichimura K. Shinozaki K. Calmodulin-dependent activation of MAP kinase for ROS homeostasis in Arabidopsis Mol. Cell 2011 41 649 660 10.1016/j.molcel.2011.02.029 21419340 87. Danquah A. de Zélicourt A. Boudsocq M. Neubauer J. Frei dit Frey N. Leonhardt N. Pateyron S. Gwinner F. Tamby J.P. Ortiz-Masia D. Identification and characterization of an ABA-activated MAP kinase cascade in Arabidopsis thaliana Plant J. 2015 82 232 244 10.1111/tpj.12808 25720833 88. Steinbrenner Adam D. Muñoz-Amatriaín M. Chaparro Antonio F. Aguilar-Venegas Jessica M. Lo S. Okuda S. Glauser G. Dongiovanni J. Shi D. Hall M. A receptor-like protein mediates plant immune responses to herbivore-associated molecular patterns Proc. Natl. Acad. Sci. USA 2020 117 31510 31518 10.1073/pnas.2018415117 33229576 89. Wang H.-Y. Klatte M. Jakoby M. Bäumlein H. Weisshaar B. Bauer P. Iron deficiency-mediated stress regulation of four subgroup Ib BHLH genes in Arabidopsis thaliana Planta 2007 226 897 908 10.1007/s00425-007-0535-x 17516080 90. Gupta A.S. Heinen J.L. Holaday A.S. Burke J.J. Allen R.D. Increased resistance to oxidative stress in transgenic plants that overexpress chloroplastic Cu/Zn superoxide dismutase Proc. Natl. Acad. Sci. USA 1993 90 1629 1633 10.1073/pnas.90.4.1629 8434026 91. Kusvuran S. Kiran S. Ellialtioglu S. Antioxidant enzyme activities and abiotic stress tolerance relationship in vegetable crops Book Abiotic and Biotic Stress in Plants-Recent Advances and Future Perspectives Shanker A.K. Shanker C. IntechOpen London, UK 2016 481 506 10.5772/62235 92. Nishiyama Y. Allakhverdiev S.I. Yamamoto H. Hayashi H. Murata N. Singlet oxygen inhibits the repair of photosystem II by suppressing the translation elongation of the D1 protein in Synechocystis sp. PCC 6803 Biochemistry 2004 43 11321 11330 10.1021/bi036178q 15366942 93. Nishiyama Y. Murata N. Revised scheme for the mechanism of photoinhibition and its application to enhance the abiotic stress tolerance of the photosynthetic machinery Appl. Microbiol. Biotechnol. APPL 2014 98 8777 8796 10.1007/s00253-014-6020-0 94. Brown P.N. Shipley P.R. Determination of anthocyanins in cranberry fruit and cranberry fruit products by high-performance liquid chromatography with ultraviolet detection: Single-laboratory validation J. AOAC Int. 2011 94 459 466 10.1093/jaoac/94.2.459 21563679 95. Li Y. He N. Hou J. Xu L. Liu C. Zhang J. Wang Q. Zhang X. Wu X. Factors Influencing Leaf Chlorophyll Content in Natural Forests at the Biome Scale Front. Ecol. Evol. 2018 6 64 10.3389/fevo.2018.00064 96. Ren X.X. Liu Y. Jeong H.K. Jeong B.R. Supplementary Light Source Affects the Growth and Development of Codonopsis lanceolata Seedlings Int. J. Mol. Sci. 2018 19 3074 10.3390/ijms19103074

PMC009xxxxxx/PMC9099646.txt

==== Front Sensors (Basel) Sensors (Basel) sensors Sensors (Basel, Switzerland) 1424-8220 MDPI 10.3390/s22093558 sensors-22-03558 Article High-Resolution Doppler and Azimuth Estimation and Target Detection in HFSWR: Experimental Study Golubović Dragan 12* Erić Miljko 12 Vukmirović Nenad 13 Bhutani Akanksha Academic Editor Pauli Mario Academic Editor 1 University of Belgrade, School of Electrical Engineering, 11120 Belgrade, Serbia; miljko.eric@vlatacom.com (M.E.); nenad.vukmirovic@ic.etf.bg.ac.rs (N.V.) 2 Vlatacom Institute, 11070 Belgrade, Serbia 3 University of Belgrade, Innovation Center of the School of Electrical Engineering, 11120 Belgrade, Serbia * Correspondence: dragan.golubovic@vlatacom.com 07 5 2022 5 2022 22 9 355816 3 2022 03 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). In this paper, we present a new high-resolution algorithm for primary signal processing in High Frequency Surface Wave Radar (HFSWR). The algorithm has been developed to achieve and improve primary signal processing performance in existing HFSWR radars in terms of radar target detection. The proposed algorithm is based on a high-resolution estimate of the Range–Doppler (RD-HR) map using given number of frames in the selected integration period. RD-HR maps are formed at every antenna in receive antenna array. Target detection is based on an RD-HR map averaged across all the antennas. Azimuth estimation is performed by a high-resolution MUSIC-type algorithm that is executed for all detections we found in the RD-HR map. The existence of strong Bragg’s lines in the RD-HR map complicates the detection process but the contrast of the RD-HR map as well as the detectability of targets on the RD-HR map is significantly better compared to the RD-FFT map used by many existing radars, such as WERA. HFSWR OTHR high-resolution methods Range–Doppler map ship detection WERA Vlatacom InstituteSerbian Ministry of Education, Science and Technological DevelopmentThis research, as a part of P.148 Project, was funded by Vlatacom Institute. The APC was funded by Vlatacom Institute. The research was also supported by the Serbian Ministry of Education, Science and Technological Development. ==== Body pmc1. Introduction Recently, in many countries which have direct access to the sea, great importance has been attached to the monitoring of the Exclusive Economic Zone (EEZ), which is defined in accordance with the United Nations Convention on the Law of the Sea. This is very important because many illegal activities can be carried out beyond the horizon, which are under the jurisdiction of the EEZ of a particular country. In this regard, High Frequency Surface Wave Radars (HFSWR) are widely used in the past for maritime surveillance of ships at ranges up to 200 nautical miles [1,2,3]. Various theoretical as well as practical implementation aspects of HFSWR radars are highlighted in numerous published works [4,5,6,7,8,9,10,11,12]. Despite the fact that the principles of HFSWR radars are theoretically well studied and clarified, and practically implemented and verified [13,14,15,16], HFSWR radars are still the subject of intensive research and development with the aim to improve detectability and resolution performance of targets with small radar cross-section (small boats, UAV, drones, etc.). The motivation for those research and development efforts are related to the application of HFSWR radars for monitoring of illegal crime activities, drug trafficking, attack on petrol platforms and strategic objects, etc. Therefore, the focus of this paper is to develop and propose new high resolution algorithm and new detection scheme in order to improve detection and resolution performance of HFSWR radars based on FMCW principle. We have shown here that multidimensional signal at the output of dechirper of FMCW HFSWR radar (with neglected coupling between domains) can be modeled as superposition of ionospheric interference, sea clutter, additive noise and attenuated sinusoids (cissoids) in 3D space (fast time domain, slow time domain and spatial domain), each of them correspond to the range, Doppler/radial velocity and azimuth of one target in multi-target scenario, typical for HFSWR radars. Therefore, the task of signal processing in FMCW HFSWR radar is to solve detection/estimation problem:to detect the number of superposed cisoids, what is equivalent to the detection targets number in multi-target scenario; to estimate frequencies of those cisoids in fast time, slow time and spatial domain, what is equivalent to the estimation of their parameters (range, Doppler/radial velocity and azimuth). If the receiving antenna array is linear and uniform, the estimation problem is usually solved by 3D Fourier transform. Such solution is applied in many practical implementations of FMCW HFSWR radars. As a result, so called range/Doppler/azimuth (RDA map) is provided. It is usually followed by CFAR detection of targets in RDA map. That is conventional approach of the signal processing in FMCW HFSWR radar. It is well known that the target detection performance of HFSWR radars are basically limited by the presence of sea clutter, ionospheric and external interference and additive noise. Two main groups for target detection methods in FMCW HFSWR radars are proposed so far: Constant False Alarm Ratio (CFAR) based methods and image processing based methods. CFAR detection methods with its various variants are usually used in FMCW HFSWR radars [17,18,19,20]. However, some other detection methods like [21,22,23,24,25,26] are proposed. In this paper, we proposed new detection algorithm based on [27] which is primarily used in Image Processing, not in HFSWR. The key problem of detection process is how to mitigate external interference. Various schemes for interference mitigation as a preprocessing step are proposed so far. In regard of this, target detection is still a challenging research problem, especially in the context of the application of high-resolution methods for range/Doppler/azimuth estimation, as presented in [28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45]. If RDA map is estimated by classical 3D FFT, resolution in Range, Doppler/Radial velocity and Azimuth domain are limited by system parameters such as chirp signal bandwidth (range), the period of integration (Doppler/radial velocity) and the number of antennas in the receiving antenna arrays. Increasing of these parameters leads to better detectability and better resolution properties of radar in each domain. Increasing the integration period improves the resolution of the Doppler estimate, but this increase is limited by the Coherent integration time related to the physics of wave propagation. The integration period must be shorter than the Coherent integration time. On the other hand, it is implicitly assumed that the target within the integration period does not change the radial velocity, which is not a valid assumption for large integration periods. Increasing the number of antennas in the antenna array improves the resolution properties, but this results in an increase in the physical space for placing the antenna array. The RD-HR map is the key to the high-resolution method proposed. In that case, criterion functions are calculated on all antennas for a set of discrete values of normalized Doppler frequencies, for the range of Doppler frequencies of interest, and with a grid resolution that is many times better than the FFT Doppler resolution, thus obtaining a zoomed high-resolution RD-HR map. The high-resolution properties of the obtained RD-HR maps and better detectability and contrast of targets in the RD-HR map in relation to the RD-FFT map are clearly observed in rest of the paper. MUSIC-type high-resolution algorithm is used to form the covariance matrix that is the basis for the application of most high-resolution methods. The main contributions of this paper are as follows. The high-resolution method for estimating the RD-HR map (uniform and more computationally efficient non-uniform variant) is proposed. Then the Doppler shift was compensated before high-resolution azimuth estimation, which is also newly proposed. The contribution is also in a novel detection algorithm that comes from the field of Image processing whose kernel function is more convenient to the morphologies of peaks of criterion functions of high-resolution methods than the classic CFAR, which is more suitable for use in 3D FFT RDA map estimation. The rest of this paper is organized as follows. Section 2 introduces the system and signal model used to generate results and to test the proposed method. In Section 3, we presented detailed algorithm description. Here, we explained 2 variants of the algorithm: High-resolution Range–Doppler map estimation (both for uniform and non-uniform in slow time domain). We explained also the detection of targets on the Range–Doppler map, and a novel method for azimuth detection with improved accuracy and numerical complexity. We discuss some of the experimental results in Section 4 and provide concluding remarks in Section 5. The derivation of dechirped signal was presented in the Appendix A. 2. System and Signal Model 2.1. System Model The High Frequency Surface Wave radar (HFSW) to be analyzed in this paper uses Frequency Modulated Continuous Waves (FMCW) which are vertically polarized. Specific propagation of waves is the key for over-the horizon-coverage. This is the most important difference between HF radar and standard microwave surveillance radars. They operate in HF spectrum 3–30 MHz. Exact frequency selection depends on specific requirements. A typical configuration is given in Figure 1. Generally, the radar consists of 3 subsystems: transmitter (Tx) antenna array, receiver (Rx) antenna array and transceiver that can be in one place or physically separated. Positions of Tx antennas are known. One of the antennas in the Tx antenna array is marked with A(Tx). The transmitter should provide a signal at a specific frequency and a sufficient power level to allow detection of targets at great distances. The output amplifier should provide a variable level of RF signal to optimize the system in different operating conditions and environmental influences. The following requirements are set when designing the Tx antenna system: the radiated energy should be maximally directed in the direction of the sea and the system must radiate as little as possible towards the Rx antenna array. A problem in FMCW systems is the isolation between the Tx and Rx antenna array. The signal level of the direct wave from the Tx antenna to the Rx antennas must not saturate the Rx. The Rx antenna array is linear, but other geometries of antenna arrays can be used. Antennas are installed parallel to the sea shore. The center of the Tx array should be in line with the Rx antenna array. Positions of Rx antennas are known, too. They are connected to N collocated receiving channels by calibrated cables (coaxial or fiber optic). The n-th antenna in the Rx antenna array is marked with An(Rx). The Rx must provide mutual coherence for all receiving channels (one channel for each antenna). In the Rx, instead of performing classical demodulation, the received signals are multiplied by a conjugated replica of the Tx waveform (generated in Chirp Generator) in a dechirper. Finally, the dechirped signals are sampled and sent to the block for signal processing. HFSWR radar WERA located on Ibeju Lekki, Nigeria, was used as a reference system for signal acquisition and performance comparison. On this specific test site, an Rx antenna array consists of 16 monopole antennas, where the distance between antennas in the antenna array is 0.45λc (λc is the wavelength of the used carrier frequency) and Rx antenna array aperture is 6.75λc. Also, the Tx planar array consisting of four antennas is used (two active and two passive antennas as reflectors). Since it is a continuous radar, the critical parameter of such a system is the isolation between Tx and Rx, which is systematically provided by careful design of Tx and Rx antenna array geometry and careful receiver design. Due to strong Tx-Rx wave, the reduction of the dynamic range of the analog signal at the input to the A/D converter is needed. In practical implementations, Rx and Tx arrays have to be separated as much as possible. The selected distance is limited by the size of the site on which the radar is placed. In the specific installation, the distance between the Tx and Rx antenna arrays is 1200 m. This length is sufficient for all frequencies of interest, and this is taken into account when designing the site. Another solution to improve the isolation between Tx and Rx is to install a notch filter at the dechirper output which suppresses signals around 0 Hz (DC), in order to reduce the impact of transmitter leakage. It is important to note that this filter does not influence the stationary targets located away from the radar, but only close targets. The bandwidth of the transmitted signal is 100 kHz and the chirp duration is 0.260022 s. The Tx signal is amplified to the desired level using a power amplifier and fed to the transmit antenna array. In existing HFSWRs, it is common to use power levels 500–1000 W. At the test site, we used a power level of 500 W. From each of the 16 Rx antennas, the signal is first filtered in order to suppress out-of-band components, then it is amplified to the level needed for A/D conversion and further processing. 2.2. Signal Model The Tx and the Rx array are synchronized in time and phase. The Tx transmits a periodic sequence of chirps at a carrier frequency fc. In Figure 2 the proposed signal model was shown. Continuous time variable t˜ spans the entire time-axis, whereas t is the time elapsed from the beginning of a chirp. (1) −∞<t˜<+∞ (2) t≡t˜modT,t∈0,T For convenience, besides the continuous time variable t˜, we also define (3) m=t˜T (4) t˜=t+mT,m∈Z, where m is the index of a chirp (also called “slow time”) and t is the time variable within a chirp (also called “fast time”). Note that t˜, t, and m are mutually dependent variables and that t˜ can be equivalently expressed by the pair m,t. A single chirp, denoted by c(t), is modeled as (5) c(t)=ej2πfc−B2t+B2Tt2. The transmitted signal, denoted by rt˜, is a periodic sequence of chirps with period T, defined as (6) rt=c(t),0≤t<T (7) rt˜=rt˜+T,−∞<t˜<+∞. Time delay of the transmitted chirp signal of the q-th target on the range R(q) on the n-th antenna is modeled as (8) τn(q)(t˜)=2Rm(q)c+2vm(q)ct+τAn(q), where c is the wave propagation velocity, τAn(q) is the relative delay at the n-th receive antenna w.r.t. the referent point of the receiving array, Rm(q) is the range to the target relative to the referent point of the Rx array and vm(q) is the radial velocity of the target during the m-th chirp (and is assumed constant during each chirp). Signal model of the received signal reflected from the q-th target on n-th antenna is modeled as (9) xn(q)(t˜)=a(q)rt−τn(q)(t˜),forτn(q)(t˜)<t<Ta(q)rt−τn(q)(t˜)+T,for0<t<τn(q)(t˜) where τ is the two-way propagation time for the target and a∈R is an attenuation factor. The model is continous in fast time, and discrete in slow time and antenna domain. The signal at the output of the dechirper is given by (10) yn(q)(t˜)=xn(q)(t˜)r(t˜)*=a(q)rt−τn(q)(t˜)r(t˜)*,forτn(q)(t˜)<t<Ta(q)rt−τn(q)(t˜)+Tr(t˜)*,for0<t<τn(q)(t˜) The complete derivation of the equations is given in the Appendix A, whereas the approximated equations, under some assumptions, are given in the Appendix B. So, we have an approximated model of the dechirped signal, as shown in the following 2 equations. In the first case, when τm,n(q)<t<T, the dechirped signal is modeled as (11) yn(q)(t˜)=a(q)ej2πR0(q)−2tBcT−2fc−Bc×ej2πv(q)−2fc−BcmT−2mtBc×ej2πτAn(q)−BtT−fc+B2×ej2πτAn(q)R0(q)2BcT×ej2πτAn(q)v(q)2Bmc×ej2πτAn(q)2B2T. In the second case, when 0<t<τm,n(q), the dechirped signal is modeled as:(12) yn(q)(t˜)=a(q)ej2πR0(q)−2tBcT−2fc+Bc×ej2πv(q)−2fc+BcmT−2mtBc×ej2πτAn(q)−BtT−fc−B2×ej2πτAn(q)R0(q)2BcT×ej2πτAn(q)v(q)2Bmc×ej2πτAn(q)2B2T×ej2πBt+fcT. We can notice that mixed terms appear in the expression for the dechirped signal, which clearly shows that there is a certain coupling between range, Doppler and azimuth domain. The terms standing next to the R0(q) in the previous two equations correspond to the range of the q-th target. The terms standing next to the v(q) correspond to Doppler effect because Doppler frequency fd(q) can be expressed using radial velocity as fd(q)=2v(q)fc/c, and normalized Doppler frequency is μq=(2πT)2v(q)fc/c. In addition, the terms standing next to the τAn(q) correspond to the azimuth of the q-th target, because the delay at the n-th receive antenna w.r.t. the referent point of the receiving array can be expressed as τAn(q)=(n−1)dsinθ(q)/c, where d represents the distance between antenna elements of the receiving array and θ(q) is the azimuth of the target. In the Rx, the received signal is given by (13) ynt˜=ηnt˜+∑qyn(q)t˜, where ionospheric interference, sea clutter and additive noise are modeled by ηnt˜, and the sum is over all the targets. Finally, the dechirped signal is sampled at a rate fs in the A/D converter. 3. Detailed Algorithm Description A three-dimensional matrix with acquired complex time samples of IQ branch signals at the receiving channels is denoted by Y∈CM×P×N, whose elements are (14) ym,p,n=yn(m−1)T+(p−1)/fs, for 1≤m≤M, 1≤p≤P, 1≤n≤N. In practical situations, P and N are predefined values and M is the value to be chosen and it should correspond to the integration period in which signal coherence is preserved. The developed algorithms were tested for the length of the segment of M=256, where the successive segments overlap in 128 frames. This ensures that the results are ejected every 128 frames. RAW data were recorded with 2048 frames per file. In the analysis, the acquired frames are divided into 15 overlapping segments, each with 256 overlapping frames. A vector with P complex signal samples at the n-th antenna in the m-th frame is defined as ym,n=ym,1,n,…,ym,p,n,…,ym,P,n∈C1×P, where ym,p,n denotes the p-th complex signal sample at the n-th antenna in the m-th frame (segment). Figure 3 shows Y matrix. The first step in classical processing in many radars, such is WERA radar, is the implementation of FFT algorithm of vectors ym,n for all frames m=1,2,…,M and all antennas n=1,2,…,N, and thus obtaining a three-dimensional matrix S∈CM×P×N with complex samples of the spectrum whose rows sm,n∈C1×P represent vectors with the spectrum samples in the m-th frame and at the n-th antenna, which are obtained as:(15) sm,n=(wP⊙ym,n)FP, where wP=w1,w2,…,wP∈R1×P denotes the vector with the samples of the applied Blackman–Harris window function, the operator (⊙) denotes Hadamard’s (Shur’s) product and FP∈CP×P denotes the Fourier matrix of P×P dimensions. The second step in the classic primary processing of many radars is the implementation of the FFT algorithm per matrix S columns tp,n=s1,p,n,…,sm,p,n,…,sM,p,n⊤∈CM×1 for all antennas n=1,2,…,N and all samples p=1,2,…,P per frame as follows:(16) hp,n⊤=(wM⊙tp,n⊤)FM∈C1×M, where wM=w1′,w2′,…,wM′∈R(1×M) denotes samples of the applied Blackman–Harris window function, and (·)⊤ is transpose operation. The vectors hp,n are calculated for p=P−R+1,P−R+2,…,P, where R corresponds to the maximum projected radar range. The matrix FM∈CM×M denotes the Fourier matrix of M×M dimensions. Figure 4 shows S matrix with selected vectors used in the previous analysis. A matrix Hn∈CM×R is then formed as follows:(17) Hn=hP−R+1,n,hP−R+2,n,…,hP,n∈CM×R. The matrix Hn represents the Range–Doppler matrix at the n-th antenna. This matrix is obtained using the FFT algorithm and we’ll call it the RD-FFT map. A further procedure of primary signal processing in some existing OTHR radars, such is WERA, is CFAR detection of targets in the RD-FFT map, followed by the estimation of signal arrival direction for detected targets using classical single snapshot beamforming. 3.1. High-Resolution Range–Doppler (RD-HR) Map Estimation (Uniform Sampling Method in Slow Time Domain) The high-resolution algorithm for creating the RD-HR map starts from the formation of the matrix S∈CM×P×N using Fourier transform. By adding zeros to the vectors with signal samples, better computational (but not actual) resolution can be achieved when applying FFT. Thus, the proposed algorithm for creating the RD-HR map is computationally high-resolution in the range domain and essentially computationally high-resolution in the Doppler domain. From the original matrix S we form an extended matrix SE∈C(M+r(L−1))×P×N by adding r(L−1) additional frames. Then new matrices Qp,n∈CM×L are formed for 1≤p≤P and 1≤n≤N, whose columns are vectors ql,p,n, 1≤l≤L. A total of M+r(L−1) frames are used to form this matrix. The index r denotes the number of frames that do not overlap in adjacent vectors ql,p,n. The developed algorithm was tested for r=1. The reason for this is a better estimate of the covariance matrix later. Figure 5 shows the creation of Qp,n matrix from SE matrix. Then the covariance matrices Cp,n∈CM×M are formed for n=1,2,…,N and p=P−R+1,P−R+2,…,P as follows:(18) Cp,n=1LQp,nQp,nH∈CM×M. The matrix Cp,n is a complex square Hermitian (conjugate symmetric) positive definite matrix, which means that the eigenvalues of that matrix are positive quantities. The formation of the covariance matrix Cp,n is a key step in the formulation of the high-resolution algorithm for the creation of the RD-HR map. Based on the covariance matrix, it is possible to formulate several criterion functions of high-resolution algorithms for creating an RD-HR map. In this case, we use MUSIC-based algorithm, as follows:(19) PMUSRD(μ,p,n)=1∥aμ(μ)HEp,n∥. The matrix Ep,n∈CM×(M−K) is a matrix of noise subspace of the covariance matrix Cp,n whose columns are eigenvectors of the covariance matrix Cp,n which correspond to M−K of the smallest eigenvalues of the covariance matrix Cp,n, where K represents a parameter of the MUSIC-based algorithm. As is well known, the MUSIC algorithm requires the knowledge of the dimensionality of the signal subspace (parameter K). There are several algorithms in the literature that can be used to estimate K, such as the Minimum Description Length algorithm (MDL) [41] or the Akaike Information Criteria (AIC) [46]. Those algorithms work well in controlable signal scenario, such as a simulation. However, from our previous experience, we know that in real coherent systems (such as direction finder we developped) these algorithms tend to overestimate the value of K. It is known from the theory of subspace algorithms, that when the eigenvalues are sorted in descending order, there is an inflection point in the curve whose position indicates the dimensionality of the signal and the complementary noise subspace. The eigenvalues of the covariance matrix, whose values are approximately equal to the variance of the noise on receiving channels, determine the dimensionality of the noise subspace. So, we analyzed empirically the position of that inflection point and observed that the K values 5 and 10 produce satisfactory results compared with AIS. However, an adaptive algorithm should be applied for the estimation of K. Also, the term coherence we use in two contexts. The first is related to a multipath environment where multiple delayed replicas of the same signal are superposed on the antenna array. This is known to be a problem for most algorithms in array processing. The second is related to the period of integration. The choice of the integration period implicitly assumes that in this period of time the Doppler shift of the signal due to the movement of the target is constant. The vector aμ(μ)∈CM×1 is an equivalent steering vector formulated in the normalized Doppler domain as:(20) aμ(μ)=1,e−jμ,…,e−jμ(M−1)⊤, where the parameter μ denotes the normalized Doppler frequency in radians per frame. The criterion functions are calculated for a set of discrete values of normalized Doppler frequencies, for the range of Doppler frequencies of interest and with a resolution that is many times better than FFT resolution per Doppler, thus obtaining a zoomed high-resolution RD-HR map. With RD-HR maps obtained by MUSIC method, the high-resolution properties and better detectability and contrast of targets in the RD-HR map in relation to the RD-FFT map are clearly observed. Range and Doppler frequencies are estimated by detecting peaks in the RD-HR map. The arguments of the maxima of these peaks are (μq,pq), where 1≤q≤Nd, Nd is the total number of detected peaks and μq and pq correspond to Doppler and range domain, respectively. For each of the detected Nd peaks, the direction (azimuth) of the target is then estimated. For high-resolution azimuth estimation, the same type of criterion function is used as for RD-HR map estimation, with the difference that steering vectors and covariance matrices are formed in the spatial domain by n and for the detected indexes μq and pq in Doppler and range domain, respectively. 3.2. High-Resolution Range–Doppler (RD-HR) Map Estimation (Non-Uniform Sampling Method in Slow Time Domain) It was found that the covariance matrix Cp,n. For all n and p has a conditional number of order 1019 which means that it is close to the singular matrix which can in some cases result in problems of numerical instability during inversion and eigenvalue decomposition of the matrix. Known procedures for reducing the conditional number are analyzed (such as adding a small scalar value on the diagonal of the matrix Cp,n. The motivation for formulating a high-resolution algorithm for estimating the RD-HR map with non-uniform sampling is related to reducing the numerical complexity of the matrix algorithm and to reducing the conditional number of the matrix of that matrix. In uniform sampling method, the dimensionality of the covariance matrix is M×M. The idea of non-uniform sampling is that dimensionality of the covariance matrix be smaller (J×J) where J<M. The problem of non-uniform sampling method in this case is analogous to the problem formulated in the field of antenna arrays - how to replace a linear uniform antenna array with a non-uniform antenna array with the same aperture and a smaller number of antennas without significant degradation of the antenna array factor? This problem in antenna array theory is known as the problem of minimally redundant linear antenna arrays. High-resolution algorithm for the RD-HR matrix creation therefore has its theoretical foundation in the theory of antenna arrays or array processing. The idea is to select a subset of J rows of the matrix Qp,n by choosing an appropriate mapping ℓ:1,2,…,J→1,2,…,M, J<M. We then form Qp,n(ℓ) as (21) Qp,n(ℓ)j,l=Qp,nℓ(j),l, where Aj,l denotes the j,l-th element of A. Thus we form a new matrix Qp,n(ℓ)∈C(J×L) on the principle of non-uniform selection of the elements of the column vector of the matrix Qp,n. The same mapping ℓ is used to form the steering vector aμ(ℓ)(μ) by non-uniform selection of the elements of the vector aμ(μ). Thus, we get a much smaller covariance matrix. Covariance matrices Cp,n(ℓ)∈C(J×J) are formed for n=1,2,…,N and p=P−R+1,P−R+2,…,P as follows:(22) Cp,n(ℓ)=1LQp,n(ℓ)Qp,nH(ℓ). The criterion function of the high-resolution MUSIC-type algorithm for creating RD-HR map with non-uniform sampling has the same form as the criterion function for the variant with uniform sampling (with the noise subspace matrix and steering vector in the Doppler domain being formed as described above) and defined as:(23) PMUSRD(ℓ)(μ,p,n)=1∥aμ(ℓ)(μ)HEp,n(ℓ)∥. Ep,n(ℓ) is obtained from Cp,n(ℓ) in the same way Ep,n is obtained from Cp,n in the uniform sampling method by selecting the elements according to the same mapping ℓ. The procedure for detection and high-resolution estimation of direction (azimuth) is further identical as in the case of the variant of the algorithm with uniform sampling. 3.3. Detection of Targets on the RD-HR Map The complete procedure, presented in the previous part of this chapter, is used to evaluate the high-resolution RD-HR (Range–Doppler High Resolution) map. The Range–Doppler map in this document is defined as a numerical two-dimensional image consisting of a finite number of points, which provide accurate information on activity at sea. It is formed at all antennas and for all data segments. The high-resolution properties of the algorithm contribute to better ship detectability, as well as the ability to detect some ships, which are not visible at all using the currently used primary signal processing algorithms. So, the precondition for detection is the formation of RD-HR map, which is the most computationally demanding part of this algorithm. Comparing the RD-FFT and the RD-HR map clearly shows the advantages of the RD-HR map in terms of resolution properties, map contrast and the ability to detect targets, which fully justifies the use of the algorithm. It was also noticed that successive detections of ships of interest are chained, passing through all segments, and that there is an overlap of contours of criterion functions of successive detections, which is the basis for formulating criteria for detection consistency and a step towards an improved version of tracking. In this regard, we will describe the procedure for detecting targets (ships) from the RD-HR map using a newly developed algorithm for Image Processing [27]. The algorithm is innovative and adapted to the specifics of this type of Range–Doppler maps (characteristics of peaks in their criterion functions). Therefore, a joint estimation of the ship’s distance and its Doppler frequency is performed, based on finding peaks at the RD-HR map according to a precisely defined criterion. At the beginning of the detection process, we have to determine arithmetic mean of RD-HR maps at all antennas. Thus, we’ll have only one RD-HR map (criterion function) that is used to find detections:(24) P¯MUSRD(μ,p)=1N∑n=1NP(μ,p,n), where P(μ,p,n) can be uniform or non-uniform criterion function obtained by choosing appropriate mapping ℓ. From the criterion function P¯MUSRD(μ,p) we form the matrix PMUSRD which consists of elements of interest (1≤μ≤MP,1≤p≤R). MP and R are lengths of RD-HR map by Doppler and range dimension, respectively. The grid resolution is 4 times better for range and 2 times better for Doppler than in the case of FFT map. The next step in the detection procedure is to analyze the RD-HR map (data matrix) and numerically present it using 16 bits. This actually means that the RD-HR map (2D image) is represented by 216 different values, and that the minimum value of the RD-HR map criterion function is zero and the maximum value is 65,535. The reason for this conversion is the higher speed of ship detection, during the practical implementation of this algorithm, without losing the image quality at all in terms of poorer detection. The next procedure is to remove unwanted noise from the image (RD-HR map) which is presented as a relatively high value in the vicinity of which are relatively low values. This type of noise would manifest itself in the picture as a single point (of course, unwanted in this case). In practice, this would mean that this point would represent a peak in the criterion function, and later also the detection, which we already know in advance is not a real detection, but a false alarm. In the literature, this type of forest is known as “Salt and Pepper” and its elimination is required. The reason why these values not be considered as detection lies in the analysis of peaks at the RD-HR map. It has been experimentally observed that the peaks are not point values, but that they have some shape everywhere, whose width is smaller or larger, and their values of the criterion function cannot drastically fall at all neighboring points around one peak. The elimination of unwanted values is done by Median Filtering algorithm used in Image Processing. This filtering method is executed for each element of the input matrix (each pixel of the input map), analyzes its 8 adjacent values, and calculates the median of these nine values. The procedure is as follows:A window (matrix) of size 3×3 points is placed around the observed image element (at the beginning it starts from the upper leftmost point in the image, ie the first row and the first column); Then all the elements in the window are placed in an array; The array is sorted in ascending order; Select the mean element (5-th element in the array); The corresponding image element, which has been filtered, is written to the appropriate element of the output matrix (the pixel of the output map); The procedure steps 1–5 are repeated for each element of the output matrix. It is necessary to emphasize that the filtering window can be of different lengths, but in this particular case it is necessary to remove only unwanted single-point peaks from the image, because we know for sure that they do not represent detection. Setting up a larger window would not be practical, as there is a possibility that the criterion function of the Range–Doppler map would be very narrow, and in that case it would be unreasonably rejected, which would lead to poorer detection results. Also, it should be emphasized that the algorithm is applied to points located on the edges of the image, and all points of the window function, which can not include the nearest neighbors of the filtered point, are simply supplemented by zeros and the median is sought, as which has been explained before. There are several cases when zeros are filled in the window function, and they are: first row, first column, last row and last column of the Range–Doppler map. Because of this, there will certainly be zeros at all ends of the filtered image. In Figure 6, the procedure of the Median Filtering algorithm was shown. In other words, for a given map P, we select a 3×3 submatrix M(i,j) centered at (i,j) as (25) M(i,j)=Pi−1:i+1,j−1:j+1∀i,j, where each edge of the map is padded with zeros, or more formally, Pi,j=0, for all i,j where i∈0,MP+1 or j∈0,R+1. P can be PMUSRD in this case. MP and R are lengths of RD-HR map by range and Doppler dimension, respectively. Then the submatrix is rearranged into a vector, m(i,j)=vecM(i,j)∈R9×1, the vector is sorted which produces the vector ms(i,j), and the resulting pixel is PFi,j=ms(i,j)5. These steps are executed for each pixel (i,j) of the RD-HR map in order to eliminate spurious single-pixel peaks. Next, the value of the threshold above which detections are taken is defined. So, the threshold value is a parameter that determines whether some peaks in the criterion function will be considered as detections or noise (such detections are rejected). The selection of the threshold value is an important procedure during detection and it is necessary to pay special attention to it. Its value can vary depending on whether the detection of close or distant ships is desired, and it can also be an adaptive threshold value depending on the distance. Figure 7 shows the ship detection procedure, whose values of the criterion function are above the threshold. From a numerical point of view, we wonder what that threshold value is. If a low threshold is chosen, the number of detections will be higher, and all ships will be detected, both those that are close, but also those that are very far from the radar. In that case, the number of detections will be large, but this increases the complexity of the calculation later, when we are detecting the azimuth. It has been noticed that there are a large number of false alarms among these detections, so the question is how the threshold can be set. To choose the appropriate value for the threshold, the amplitude of the targets have to be taken into account. The greater the amplitude (on average), the higher the threshold value should be, otherwise the false alarm rate would be increased. If, on the other hand, the amplitude is low and we do not decrease the threshold adequately, the probability of misdetection would increase. The amplitude of the signal generally decreases with the range. Additionally, it increases with the increase of the target RCS. All in all, an adaptive threshold strategy should be implemented, so that the threshold value is nonuniform along the range dimension. Also, the threshold value should be increased in the immediate vicinity of a target with a large RCS, to keep the false alarm rate down. Based on experimental tests of data obtained from radar in operational work, it was concluded that the threshold value should be in the range of 0.1×216 and 0.2×216, while the number of successful detections should be maximum. In the implementation process of the algorithm, the values 0.1 and 0.2 are used, thinking about the previously defined values 0.1×216 and 0.2×216. In this study, the threshold values were chosen according to a large set of real radar data (large statistical sample) and applied within the algorithm to obtain the target detections. In this paper we present a small subset of these results. Of course, it should be noted that not all points that exceed the threshold value will be detections, but they will become candidates to be detections, which will be discussed later. The next step is the 2D convolution procedure. Here, the RD-HR map (2D image) is additionally filtered, but there must be another 2D filter matrix of smaller dimensions (the so-called kernel matrix).The main goal is to obtain a smooth image Range–Doppler map. The filter matrix in this case is a Gaussian 2D filter whose dimensions are 7×7. The dimensions of this filter can be different, but it would be best to choose them based on the characteristics of the RD-HR map peaks. The center of the filter matrix must be positioned on the pixel to be filtered. This operation in which we summarize the products of the elements of two 2D functions, where it is allowed for one of the two functions to move over the elements of the other function is actually a convolution. The 2D convolution operation is quite computationally demanding, so it is not very fast to execute unless small kernel filters are used. Their dimension should be odd, so that they have a center, for example 3×3, 5×5 and 7×7. In Figure 8 it is shown a Gaussian 2D filter measuring 7×7, as well as its criterion function. Figure 9 illustrates the 2D convolution procedure with a 3×3 kernel matrix for one pixel at the RD-HR map. As explained earlier, the same is true here if the dot is found on the edges of the RD-HR map, then the corresponding values of the kernel matrix are filled with zeros. Figure 10 illustrates the 2D convolution procedure with a kernel matrix, which clearly shows how to process one pixel at a time from the image. The last step is to determine the actual detections, based on the matrix obtained after the procedure of all the above filtering and with a defined value of the detection threshold. It should be noticed that not all values obtained in this way. First, non-zero elements are determined (which are significantly smaller than in the original image), and they represent candidates for detection. Then, around each point, a criterion function is observed with 2 points on all sides. In other words, we want to apply a 2D linear FIR (Finite Impulse Response) filter to the map PF to obtain PFF. Its impulse response (or kernel) can be thought of as a 7×7 matrix and is given by (26) κ(i,j)=1σ2exp−i2+j22σ2;∀i,j∈−3,−2,…,3. The result is the convolution (27) PFFi,j=∑ζ=−33∑ξ=−33κ(ζ,ξ)PFi−ζ,j−ξ where the edges of the map PF are appropriately padded with zeros, i.e., PFi,j=0, for all i,j where i∈−2,−1,0,MP+1,MP+2,MP+3 or j∈−2,−1,0,R+1,R+2,R+3. Kernels of different sizes, such as 9×9, 5×5, or 3×3, can be used instead, but filtering with large kernels can be computationally demanding. Figure 11 shows the layout of different kernel functions. There are 2 ways to search for the appropriate detection. The first way is the method of local maximum of the criterion function, more precisely, if the point, which is located in the center of the part of the criterion function of dimensions 5×5, has a maximum value in that window, then it represents detection. The second way involves the analysis of the same part of the criterion function, but also the sigma by distance and Doppler frequency. Sigma, as it was defined in WERA radar, means 2nd moments of the distance and Doppler estimation, while the distance and Doppler estimates are centroids (centers of mass) of this part of the criterion function. The result of the execution of the algorithm are detections, whose x and y coordinates represent estimates of the distance and Doppler frequency. Figure 12 shows all steps in detection process. Finally, we obtain detections on Range–Doppler map and we can continue with the last step in the detecion process which is azimuth detection. Figure 13 shows detections on RD-HR map. 3.4. Azimuth Detection of Targets Detected in RD-HR Map The idea is to first evaluate the high-resolution RD-HR map, to detect targets on it using a newly developed algorithm for Image Processing, and to estimate the azimuth, using another algorithm. Azimuth detection performs only for distances and Doppler frequencies of such detected targets. In this way, the numerical complexity is significantly reduced and the execution process of the algorithm is accelerated. For the calculation of azimuth, a similar procedure will be used as before. For detected distances, a criterion function will be required, but only for detected Doppler frequencies. The MUSIC method will give an accurate azimuth estimation. We will assume that the number of detections found on the Range–Doppler map is equal to Nd. Therefore it is necessary to determine the unknown parameters μq,pq and θq or the unknown vessel Doppler frequencies (radial velocities), ranges and directions of arrival (azimuths) for each of q=1,2,…,Nd targets, respectively. Unlike the RD-FFT map, when RD-HR map is formed, the phase data is lost and therefore we have to return to the matrix SE. The first step in azimuth detection is to select column vectors ql,pq,n for 1≤l≤L, 1≤n≤N and for pq-th FFT sample. These columns correspond to the detected range from the RD-HR map. Thus, for each of q detections we find the appropriate column vectors ql,pq,n, as shown in Figure 14. To compensate for the Doppler effect in matrix Qpq,n for all antennas n=1,2,…,N, we use the steering vector aμ(μq) and we get:(28) rn(q)=wM⊙aμ(μq)HQpq,n∈C1×L. Then we form appropriate matrix (29) R(q)=r1(q)⊤,r2(q)⊤,…,rN(q)⊤⊤∈CN×L. We get the covariance matrix by averaging over the L shifts as follows (30) CA(q)=1LR(q)R(q)H∈CN×N. The same snapshots/frames (Qp,n from Figure 5) that are used for the estimation of the covariance matrix (18) of the RD-HR map are also used for the estimation of the covariance matrix in (30), after pre-processing according to (28). When selecting the parameter L, a compromise must be made here because on one hand, to allow a more statistically stable estimate of the covariance matrix, a large enough number of snapshots should be taken, but on the other hand it increases the numerical complexity, so the signal processing cannot be performed in real-time. Another negative effect of increasing L too much is that it can become longer than the coherence interval and, so, will blur the RD-HR map. In the practical implementation, L=64 is the value that satisfies all the above requirements. The steering vector in this case is (31) aθ(θ)=1,e−jν,…,e−jν(N−1)⊤. for a ULA, where ν=2πfcdsinθ/c, θ is the azimuth, and d is the distance between adjacent antennas. The criterion function for the azimuth is (32) PMUSA(θ,q)=1∥aθH(θ)E(q)∥, where E(q) is the noise subspace matrix calculated for the q-th detected target. The matrix E(q)∈CN×(N−KA) is a matrix of noise subspace of the covariance matrix CA(q) whose columns are eigenvectors of the covariance matrix CA(q) which correspond to N−KA of the smallest eigenvalues of the covariance matrix. KA represents a parameter of MUSIC-based algorithm. The final estimate of the azimuth is determined by:(33) θ^(q)=arg maxθ|PMUSA(θ,q)|. 4. Experimental Results The results presented in this section are based on the measured radar data, and their verification was made using AIS data. A set of real signals (RAW data) acquired on April 19, 2020 from the OTHR radar located on Ibeju Lekki, Nigeria, in a time interval of 5 h was used for testing. The section is divided in 2 parts. The first part of this section shows Range–Doppler maps obtained by the proposed HR algorithm, and one example of azimuth estimation based on detections from RD-HR maps. In this part, the basic system parameters are explained too. The second part analyzes the performance of the proposed algorithm. Because we have real data and also AIS data, we can made one experiment to see the numerical results of detecting vessels. P = 1536 and N = 16 are predefined values and M is the value to be selected and it should correspond to the integration period in which the coherence of the signal is preserved. The developed algorithm was tested for the length of the segment M = 256, where the successive segments overlap with 128 frames, which ensures the results are ejected every 128 frames. First step in the proposed algorithm is forming of High Resolution Range–Doppler map. We used uniform sampling method. Here, we can see properties of Range–Doppler maps and their main advantages and differences. In Figure 15 a comparative view of the RD-FFT map, obtained by the FFT algorithm, and a high-resolution RD-HR map, obtained by the new high-resolution algorithm, was presented. Both maps were calculated for an integration period of 256 frames with a duration of 0.260022 s each (integration interval of 66.5656 s). At this point, it is worth noting the difference between two notions related to the performance of a position estimation algorithm. The first one is the target resolvability, which quantifies the ability of the algorithm to perceive two targets close to each other as separate targets and not a single larger target, depending on the distance between them. The second notion is the grid resolution, which represents the density of the points at which the criterion function of the algorithm is calculated. Even though it can be chosen arbitrarily, care must be taken not to make it too coarse, otherwise the accuracy and target resolvability of the algorithm would be degraded. On the other hand, choosing the grid resolution to be too fine directly increases the computational complexity of the algorithm. The resolution of the RD-FFT map according to Doppler is basically determined by the resolution properties of the Fourier transform and in this case it is 0.0150 Hz. The range of Doppler frequencies obtained by the FFT algorithm is from −1.9229 Hz to +1.9229 Hz. The RD-HR map is calculated using a new high-resolution algorithm when, and unlike the FFT algorithm, there is a possibility to choose the bandwidth and grid resolution with which the RD-HR map is calculated. In this particular case, the RD-HR map for the integration period of 256 frames was calculated in the range −0.4804 Hz to +0.4804 Hz with a grid resolution of 0.0019 Hz (RD-HR map is calculated in 513 points as opposed to RD-FFT map calculated in 256 points). It follows from the above that the grid resolution of the RD-HR Doppler map is 7.8947 times better than Doppler resolution of the RD-FFT map. The range in which the RD-HR map is calculated from −0.4804 Hz to +0.4804 Hz is chosen so that in this case it includes the Doppler frequencies of ships of interest, including Bragg’s lines. This range can be set arbitrarily. The calculation of the RD-HR map was performed with a range grid resolution of 375 m. It is 4 times better than the range resolution used in the calculation of the RD-FFT map (1.5 km). Comparing the RD-FFT and the new RD-HR map clearly shows the advantages of the RD-HR map in terms of target resolvability in Doppler domain, map contrast and the ability to detect targets. It was also noticed that successive detections of ships of interest are chained and that there is an overlap of contours of criterion functions of successive detections, which is the basis for formulating criteria for detection consistency. RD-HR maps are formed for each antenna individually and independently. Target detection is performed on an averaged RD-HR map. A 5×5 Gaussian kernel function was chosen based on the analysis of the shape of the lobes of the RD-HR map in order to improve detection performance. In the second step, the azimuth is estimated using a high-resolution MUSIC type algorithm that is executed only for all detections in the RD-HR map. Because of that, numerical complexity and the time of algorithm execution is reduced. Angle grid resolution had chosen to be 0.2 degrees, and it is much better resolution then the resolution used in many algorithms which are currently in use (typically 1 degree). In monostatic radars of this type, the accuracy in the azimuth dimension is usually the lowest of all the three dimensions. Because of this, it is significant to improve the azimuth estimation accuracy, if the estimator and the useful information embedded in the received signals allow, as long as this increase in the accuracy does not increase the numerical complexity so much that the algorithm can no longer run in real-time. In this particular case, the increase in the numerical complexity of the entire algorithm when the azimuth grid is refined from 0.2∘ to 0.1∘ is negligible. There are a few more reasons why it is important to choose a finer azimuth resolution. Firstly, the lobes in the MUSIC-based criterion function are very narrow, so the grid resolution should be chosen such that there are enough points on each lobe to enable us to recover the location of its maximum correctly. Furthermore, the MUSIC-based algorithm has high target resolvability, which is better than the FFT resolution cell. This allows us to detect vessels at the same range and radial speed, but at slightly different azimuths (within one FFT resolution cell), but this can be done only if the azimuth grid resolution is fine enough.Also, refining the grid without bound will not improve the performance of the estimator arbitrarily, but, instead, the performance improvement would experience a saturation effect. This is because the amount of useful information in the raw signals is limited, and thus determines the performance bound. However, by refining the resolution in this case from 0.2∘ to 0.1∘ does increase the performance noticeably (the saturation still does not occur). Figure 16 shows complete detection process. As mentioned earlier, in a time interval of approximately 5 h, we want to detect vessels using the proposed algorithm, and then compare the results with AIS data. Because of that, we made an experiment by randomly selected 10 vessels and monitored the detections throughout the time interval. The complete AIS data in a time interval of approximately 5 h is plotted in Figure 17. First, a contour is formed around the AIS data, where the width of the contour is equal to the size of the initial resolution cell of 1.5 km. Then the criterion was made so that the detections and AIS data are monitored for one hour, hour by hour. According to this criterion, if the detection is within the contour we will consider that the real detection and not a false alarm. Figure 18 illustrate the forming of the criterion contour. We made an experiment by randomly selected 10 vessels whose basic information is available on websites (accessed date: 10 January 2022): www.vesseltracker.com, https://www.myshiptracking.com, https://www.marinetraffic.com and https://maritimeoptima.com. We will monitor the vessels by MMSI (A Maritime Mobile Service Identity) number. MMSI is a nine-digit number that uniquely identify vessel stations and it is sent over a radio frequency channel. Figure 19 shows all vessel’s detections in a time interval of 5 h for different detection parameters of the proposed algorithm. It can be clearly seen that as the order of the model K increases, the number of detections increases too, which is an expected result. Also with increasing the value of the normalized detection threshold, the number of detections decreases, which is a consequence of poorer detection of peaks in the Range–Doppler map. Figure 19a also shows the markings of some individual tracks because we want to explain the nature of these tracks. Note that a small part of the area in Figure 16 was zoomed in and shown in Figure 19 so that the details, especially the categorization of the detections, could be clearly seen. Tracks marked by 1 indicate the traces for which AIS data exists. Light blue tracks represent the AIS data and we use this as the benchmark (as real vessel trajectories), where available. Note that AIS tracks are partially visible because they are overlayed (covered) by the target detections (yellow markers). The tracks marked by 2 indicate traces for which there is no AIS data, but from the aspect of the detection algorithm, this vessel can be considered as detectable. We know that these tracks correspond to real vessels, because this was confirmed by the tracker that was executed after the proposed detection algorithm (the tracker is outside the scope of the paper). The track marked with the number 3 is a consequence of ionospheric interference. Other scattered detections, marked with the number 5, are a consequence of the sea clutter. Note that the first-order Bragg’s lines in Figure 12 were suppressed in preprocessing, so that they do not appear in Figure 19. Additionally, even though Bragg’s lines are (more or less) concentrated in two regions on an RD-HR map, they appear dispersed on a geographical map. What can be noticed is that in all cases the detections match the AIS data very well, and in the next step an accuracy analysis will be made depending on the selected algorithm parameters (K and threshold). Since we observe detections hour by hour, Figure 20 shows an example of the appearance of detections on a geographic map from 17 pm to 18 pm for different values of algorithm parameters. Graphs of this type are very useful and can be used to monitor changes by the hour, for example if there is a change in climatic conditions, or if there is a change from day to night, etc. Note that the detections in Figure 20 were intentionally shown only for the interval 17:33–18:00, whereas the AIS data was shown for the entire hour (17:00–18:00), so that the beginnings (roughly the first 50%) of the AIS tracks would be visible (not overlayed by detections). In the following analyses, the results of the detection of arbitrarily selected vessels will be presented in order to see the impact of certain parameters. Figure 21 shows detections for a vessel with MMSI = 636014619 in a time interval of 1 h (18–19 h) for different detection parameters of the proposed algorithm. Note that there is only one vessel in the selected area—the one inside the vessel contour (this was verified by the tracker). Tracks can be clearly seen and detections successfully follow the AIS data. The highest number of detections is in the case when the model number is higher and the detection threshold is lower. This certainly increases the detectability of vessels, but also increases the number of false alarms. Figure 22 shows detections for a vessel with MMSI = 657,199,400 in a time interval of 1 h (18–19 h) for different detection parameters of the proposed algorithm when ionospheric interference is present. It interferes with the detection process but the tracks are still visible and the vessel is detectable for the whole period of time. Similarly, there is only one vessel in the selected area (verified by the tracker) and there is also strong ionospheric interference. The following example which is shown in Figure 23 will show how stationary targets are detected. One of the stationary targets, which is also shown on the map, will be taken. This target is labeled as G-132km. The model number and the choice of the detection threshold do not significantly affect the detectability of such a target and it is detectable in all cases. The following analyses will provide numerical data of the proposed algorithm for different parameters K and threshold in order to obtain their optimal values. Figure 24 shows the total number of detections inside the selected contour for all vessels. From this picture, you can clearly see which ships are detectable and in what period of time. Based on all previous analyses, the assessment of vessel detectability was made. Table 1 and Table 2 show the final results. As can be seen from the previous figure, the detection success is high in all cases. In order to determine the optimal value of the algorithm parameters, in practical situations it is necessary to determine the percentage of detection success. Figure 25 shows the percentage of detection success of the proposed algorithm for different parameters K and threshold. In this case, the best performance is 76.67% for chosen algorithm parameters K = 10 and threshold 0.1. This clearly shows that the order of the model must be higher and the detection threshold lower in order for this percentage to be higher. But in addition to this, an important parameter can be the ratio of true detections and total number of detections, and it is desirable that this number be as small as possible so that there are not too many false alarms. The ratio of total number of detections and number of detections within contours for all vessels and for all algorithm parameters is shown in Table 3. To provide more useful information about the performance of the algorithm than the cumulative number of detections and false alarms inside vessel contours (Table 3), we approximately determined the probability of detection and the probability of a false alarm (Pd and Pfa). In order to determine Pd and Pfa, we made a criterion. Suppose there is a single vessel in an area of interest in a given time interval. Let G be the set of points in the search grid (finite set). Also define T=t1,t2,…,tn, as the set of timestamps in the given interval, A as the selected area of interest, B(t), t∈T as the ball centered at the benchmark location of the vessel at time t, and D(t) as the set of locations of the detections obtained by the algorithm from the segment with timestamp t. Then, we can define the area inside the contour around the vessel’s trajectory as (34) C=⋃t∈TB(t). One possible estimate of the false alarm rate (FAR) is then (35) Pfa=∑t∈TD(t)\CA\C∩G·T, where · denotes the number of elements of a set. Another way of defining the FAR estimate, which also counts the detections inside the contour, but outside the ball for a given t (it includes an extended area), is (36) Pfae=∑t∈TD(t)\B(t)∑t∈TA\B(t)∩G. We also estimate the probability of detection, Pd, as the ratio of the number of timestamps t in which there is at least one detection inside the ball around the vessel, i.e., B(t)∩D(t)≠∅, and the total number of timestamps in which the vessel is present (in the given interval). The ball is centered at the AIS location of the vessel at the given timestamp (the timestamp of the given signal segment) and its radius is equal to the size of the initial resolution cell of 1.5 km. Total number of time stamps in the selected period of time (18–19 h) is 105. Figure 26 shows the proposed criteria used to determine the probability of detection and the probability of false alarm. Table 4 and Table 5 show the numerical results for two vessels presented in the paper. We have chosen these two vessels so that the detection of one is quite bothered by ionospheric interference, and the other is not. The results for the same vessels are presented in the previous part of the paper. The selected estimation area is the same as the area shown in the paper. In the cases in which the sample was too small to have at least 100 events of each of the two types (presence/absence of detection), we gave the estimate of the probability in the form of the ratio of the number of events of the given type and the total number of events (the sample size). In those cases, the probability estimate is not considered stable enough, but still provides some useful information about the behavior of the algorithm. Increasing K from 5 to 10 for the vessel in Table 4 increases Pd by 0.1143 and 0.0286 for threshold values of 0.1 and 0.2, respectively, while the FAR remains very low (in the order of 10−4). Similarly, for the vessel in Table 5 the increase in Pd is 0.0762 and 0.1143. In addition, one of the most important advantages of the high-resolution algorithm presented in this paper is target resolvability. As it is well known, target resolvability in HFSWR is limited by the size of Range–Doppler-azimuth cells in 3D RDA map. The size of Range–Doppler-azimuth cell provided by DFT based primary signal processing (usually applied in HFSWR) is primarily limited by the resolution properties of DFT (the number of samples in range and Doppler domain and the number of antennas in linear antenna array). In the paper we proposed MUSIC-based methods for high-resolution Doppler and azimuth estimation. It is also known that high-resolution methods can resolve signal targets which are inside Rayleigh resolution bandwidth (in this case in Doppler and azimuth domain). So, it is clear that the size of Range–Doppler-azimuth cells provided by high-resolution methods are smaller then those provided by DFT methods. It means that the application of high-resolution methods in HFSWR can improve its target resolvability. Practically, it means, for example, that two targets at the same range and Doppler (this scenario is possible in practice) can be resolved in spatial (azimuth) domain. In the presented signal scenario, we did not have two close targets representative for the illustration of better resolvability. So, we found a representative close target scenario from many other scenarios for which we have acquired signals. In Figure 27, we show an example with two vessels that were very close to each other. According to the detections shown in the figure, the algorithm has a high target resolvability rate, even for vessels at the same range and Doppler shift. And what is very important, the proposed high-resolution algorithm presented in the paper achieves real-time processing. The selected chirp duration is 0.260022 s, and since we want to output the results after every 128 frames, the real-time requirement is 33.28 s. Thus, the processing time of one segment must be shorter than 33.28 s. Also, the parallelization of the program was made, so that it is executed on several CPU cores and runs on on a PC computer with i7 CPU, and on a better CPU (AMD Ryzen) in real time. It should be emphasized that the parallelization was done for the most complex part of the algorithm (calculation of the RD-HR map) where CPU utilization is 100%, while in other less computationally demanding parts it is not. Table 6 shows the average segment processing time for a sample of 200 segments. Figure 28 shows logical processors usage in the multithread software. 5. Conclusions In this paper, we presented new developed algorithm for FFT-based range and MUSIC-based high-resolution Doppler and Azimuth estimation in HFSWRs. The performance of the proposed algorithm based on experimental study are also presented. We analyzed numerical results of the proposed algorithm for different parameters K and threshold values in order to obtain their optimal values. Based on the analyses, the assessment of vessel detectability was made. We improved the detectability in the selected time interval.The highest number of detections is in the case when the model number is higher and the detection threshold is lower, and what is more important, the ratio of total number of detections and true detections is not too high, and the percentage of detection success is high. Based on the experimental results, the assessment of vessel detectability was made. The properties of RD-HR maps are presented and also their main advantages and differences, related to RD-FFT maps, in order to have better target detectability. An RD-HR map has a higher contrast and it is more suitable for use in HFSWRs then RD-FFT maps. The contribution is also the compensation of the Doppler shift before high-resolution azimuth estimation which is performed using a high-resolution MUSIC-type algorithm, that is executed for each detection in the RD-HR map. Because of that, numerical complexity and the algorithm execution time are reduced. We do not need to process all the points from the RD-HR map, but only detections. We also form the criterion to compare detections and AIS data. It can be noticed that in all cases the detections match the AIS data very well, and this comparison can be very useful to verify empirically obtained results. Tracks can be clearly seen and detections successfully follow the AIS data. We have also made rough estimates of the probability of detection and the probability of a false alarm, which show the advantages of the proposed algorithm. In HFSWRs, CFAR detectors are usually used. The detection algorithm that we proposed, which comes from the field of Image processing, eliminates spurious single-pixel peaks in the criterion function and it can improve detectability by adapting the shape of the kernel to the shape of the lobes in the RD-HR map. The high-resolution MUSIC-based algorithm was proposed for both the RD-HR map and the azimuth estimation. It increases the target resolvability of the radar both in Doppler and azimuth domain, as well as the detectability of the targets. The MUSIC-based algorithm was improved by using non-uniform sampling across the signal frames (in Doppler domain), to drastically reduce numerical complexity, while sacrificing little performance. These results are obtained in a real environment within an experimental study. Therefore, the algorithm may have practical application in the future and have great potential because the surveillance with HFSWR radar can be done continuously, 24 h a day, 7 days a week at the significantly lower costs. It is also possible to define potentially suspicious activities and immediately warn the user about it. An extremely large detection range is therefore possible without the so-called “Blind zones”. Also, high reliability is enabled and possible integration with an automatic identification system (AIS) that provides the information about ships that are currently in a zone of interest. Therefore, all of the above represent a challenge to many researchers around the world to further develop and define HFSWR radar algorithms. A special challenge is to compare the performance of new algorithms with existing algorithms for primary signal processing, as well as to define mathematical models that will lead to better ship detectability and better radar resolution properties, as well as the ability to detect some ships that are not detectable or separable by distance/azimuth using the currently used algorithms for primary signal processing. The algorithm is potentially applicable for other purposes, for example for surveillance and other targets of interest, such as detection of sea currents, sea winds, icebergs, and can also be used to search and rescue people in the case the ship is lost from AIS and if an accident occurs, in the fishing, in the exploitation of marine resources, as well as tsunami detection, which can lead to saving many human lives, etc. As for the future research directions, it will be important to explore a few key points that could improve the proposed method in practical situations further:Mathematical modeling and the simulation of the whole system, which includes the setting of all the parameters of the multi-target signal scenario (such as interference, propagation, see clutter and signal parameters, such as SNR and RCS). That way we would also be able to estimate the false alarm rate and the probability of detection in a fully controlable scenario. Automatic selection of detection thresholds dependent of the distance or weather conditions, sea clutter and external interference, that is better than the experimentally obtained threshold in terms of detection results; Suitable algorithm used to estimate parameter K, such as an adaptive determination method, or some custom variant. As we noticed, AIC and MDL give a higher value of K than the actual value; Automatic segment length selection; Algorithmic optimization for obtaining non-uniform frame selection based on which the RD map is calculated, which leads to a lower numerical complexity. Most known technical radar solutions such as WERA, ONERA, OSMAR use the same type of primary signal processing based on Fourier spectral analysis. In all the listed HFSWRs, in the first step, by applying the Fourier transform, an RD map is formed. Target detection is performed on an RD map using one or more CFAR detector variants. In the second step, the direction estimate is determined for the targets detected in the RD map using the classic single snapshot beamformer. High-resolution methods represent a new solution in relation to the state-of-the-art in the field of signal processing in HFSWRs and certainly lead to the improved detection and separability of close targets. Acknowledgments We thank Vlatacom Institute for overall supporting this research. We gratefully thank Petar M. Djurić (Stony Brook University, New York, NY, USA) for carefully reading the manuscript, and for his useful comments and suggestions. In addition, we also thank anonymous reviewers for their useful comments, remarks and suggestions. Author Contributions Conceptualization, D.G., M.E. In addition, N.V.; methodology, D.G., M.E. In addition, N.V.; software, D.G. In addition, M.E.; validation, M.E. In addition, N.V.; formal analysis, M.E.; investigation, D.G., M.E. In addition, N.V.; resources, D.G. In addition, M.E.; data curation, D.G. In addition, M.E.; writing—original draft preparation, D.G. In addition, M.E.; writing—review and editing, N.V. In addition, M.E.; visualization, D.G.; supervision, N.V.; project administration, D.G. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Abbreviations The following abbreviations are used in this manuscript: HFSWR High Frequency Surface Wave Radar FFT Fast Fourier Transform OTHR Over The Horizon Radar EEZ Exclusive Economic Zone DoA Direction of Arrival MUSIC Multiple Signal Classification A/D Analog to Digital AIS Automatic Identification System HR High Resolution RF Radio Frequency MMSI Maritime Mobile Service Identity RD Range–Doppler Appendix A. Derivation of Dechirped Signal Complete derivations of equations for the dechirped signal are given here. (A1) t−τn(q)(t˜)=t−2Rm(q)c−2vm(q)ct−τAn(q) (A2) t−τn(q)(t˜)2=t2−2tτn(q)(t˜)+τn(q)(t˜)2 (A3) τn(q)(t˜)2=2cRm(q)+vm(q)t+τAn(q)2=4c2Rm(q)+vm(q)t2+4cRm(q)+vm(q)tτAn(q)+τAn(q)2=4c2Rm(q)2+2Rm(q)vm(q)t+vm(q)2t2+4cRm(q)τAn(q)+4cvm(q)tτAn(q)+τAn(q)2=4c2Rm(q)2+8c2Rm(q)vm(q)t+4c2vm(q)2t2+4cRm(q)τAn(q)+4cvm(q)tτAn(q)+τAn(q)2. In the first case, for τn(q)(t˜)<t<T, the dechirped signal is (A4) xn(q)(t˜)r(t˜)*=a(q)rt−τn(q)(t˜)r(t)*=a(q)ej2πfc−B2t−τn(q)(t˜)+B2Tt−τn(q)(t˜)2e−j2πfc−B2t+B2Tt2=a(q)ej2πfc−B2t−τn(q)(t˜)+B2Tt−τn(q)(t˜)2−fc−B2t−B2Tt2=a(q)ej2πφm,n(q)(t). Thus, the phase term φm,n(q)(t) in the previous equation can be derived as follows:(A5) φm,n(q)(t)=fct−Bt2−fcτn(q)(t˜)+B2τn(q)(t˜)+B2Tt2−B2T2tτn(q)(t˜)+B2Tτn(q)(t˜)2−fct+Bt2−B2Tt2=τn(q)(t˜)−fc+B2−BtT+B2Tτn(q)(t˜)2=2Rm(q)c+2vm(q)ct+τAn(q)−fc+B2−BtT+B2Tτn(q)(t˜)2=−2fcRm(q)c−2fcvm(q)ct−fcτAn(q)+Rm(q)Bc+vm(q)Bct+B2τAn(q)−2Rm(q)BcTt−2vm(q)BcTt2−BTτAn(q)t+2BTc2Rm(q)2+4BTc2Rm(q)vm(q)t+2BTc2vm(q)2t2+2BTcRm(q)τAn(q)+2BTcvm(q)τAn(q)t+B2TτAn(q)2=t−2vm(q)cfc−B2−2BRm(q)cT+4BRm(q)vm(q)c2T+t2−2vm(q)BcT+2(vm(q))2Bc2T+tτAn(q)−BT+2Bvm(q)Tc+τAn(q)2B2T+2BTc2Rm(q)2−2Rm(q)cfc−B2+τAn(q)−fc+B2+2BRm(q)Tc. Finally, for τn(q)(t˜)<t<T, dechirped signal can be expressed as:(A6) xn(q)(t˜)r(t˜)*=a(q)ej2πt−vm(q)c2fc−B−2BRm(q)cT+4BRm(q)vm(q)c2T×ej2πtτAn(q)−BT+2Bvm(q)Tc×ej2πt2−2vm(q)BcT+2(vm(q))2Bc2T×ej2πτAn(q)−fc+B2+2BRm(q)cT×ej2πτAn(q)2B2T×ej2π2BRm(q)2Tc2−Rm(q)2fc−Bc. In the second case, for 0<t<τn(q)(t˜), the dechirped signal is (A7) xn(q)(t˜)r(t˜)*=a(q)rt−τn(q)(t˜)+Tr(t)*=a(q)ej2πfc−B2t−τn(q)(t˜)+T+B2Tt−τn(q)(t˜)+T2e−j2πfc−B2t+B2Tt2=a(q)ej2πfc−B2t−τn(q)(t˜)+T+B2Tt−τn(q)(t˜)+T2−fc−B2t−B2Tt2=a(q)ej2πφm,n(q)(t). In the previous equation, we found:(A8) t−τn(q)(t˜)+T2=t2−2tτn(q)(t˜)+τn(q)(t˜)2+2tT−2Tτn(q)(t˜)+T2. In the second case, the phase term φm,n(q)(t) from the first case has to be supplemented by following terms:(A9) B2T2tT−2Tτn(q)(t˜)+T2+fcT−BT2=tB−2Bcvm(q)+BT2−2BcRm(q)+fcT−BT2−τAn(q)B. Finally, in the second case, when 0<t<τn(q)(t˜), the dechirped signal is (A10) xn(q)(t˜)r(t˜)*=a(q)ej2πt−vm(q)c2fc+B−2BRm(q)cT+4BRm(q)vm(q)c2T+B×ej2πtτAn(q)−BT+2Bvm(q)Tc×ej2πt2−2vm(q)BcT+2(vm(q))2Bc2T×ej2πτAn(q)−fc−B2+2BRm(q)cT×ej2πτAn(q)2B2T×ej2π2BRm(q)2Tc2−Rm(q)c2fc+B+fcT. Appendix B. Approximation of the Dechirped Signal Under the assumption that vm(q)=v(q)=const and that vm(q)t is negligible (because we neglect Doppler effect during one frame), we have (A11) Rm(q)≈R0(q)+mTv(q) (A12) τn(q)(t˜)≈τm,n(q)=2Rm(q)c+τAn(q). Also the terms containing 1c2 are negligible, so we get an approximated model of the dechirped signal. In the first case, for τm,n(q)<t<T, the phase term φm,n(q)(t) from the Equation (A5) can be derived as follows: (A13) φm,n(q)(t)=−2tv(q)cfc−B20−2tBRm(q)cT+4tBRmqv(q)c2T0+tτAn(q)−BT+tτAn(q)2Bv(q)Tc0+t2−2v(q)BcT0+t22(v(q))2Bc2T0+τAn(q)−fc+B2+2BRm(q)cT+τAn(q)2B2T+2BTc2Rm(q)20−2Rm(q)cfc−B2=−2tBcTR0(q)+mTv(q)+tτAn(q)−BT+τAn(q)−fc+B2+τAn(q)2BcTR0(q)+mTv(q)+τAn(q)2B2T−2cfc−B2R0(q)+mTv(q)=R0(q)−2tBcT−2cfc−B2+v(q)−2tBcTmT+v(q)−2cfc−B2mT+τAn(q)−BtT−fc+B2+τAn(q)R0(q)2BcT+τAn(q)v(q)−2BcTmT+τAn(q)2B2T. Finally, for τm,n(q)<t<T, the approximation of the dechirped signal can be expressed as:(A14) yn(q)(t˜)=a(q)ej2πR0(q)−2tBcT−2fc−Bc×ej2πv(q)−2fc−BcmT−2mtBc×ej2πτAn(q)−BtT−fc+B2×ej2πτAn(q)R0(q)2BcT×ej2πτAn(q)v(q)2Bmc×ej2πτAn(q)2B2T. In the second case, for 0<t<τm,n(q), the phase term φm,n(q)(t) from the first case has to be supplemented by following terms: (A15) tB−t2Bcv(q)0−2BcRm(q)+fcT−τAn(q)B==tB−2BcR0(q)+mTv(q)+fcT−τAn(q)B So, for 0<t<τm,n(q), the approximation of the dechirped signal can be expressed as:(A16) yn(q)(t˜)=a(q)ej2πR0(q)−2tBcT−2fc+Bc×ej2πv(q)−2fc+BcmT−2mtBc×ej2πτAn(q)−BtT−fc−B2×ej2πτAn(q)R0(q)2BcT×ej2πτAn(q)v(q)2Bmc×ej2πτAn(q)2B2T×ej2πBt+fcT. Figure 1 System model. Figure 2 Transmitted chirp signal (solid blue line) and received chirp signal (dashed red line). Figure 3 Three-dimensional matrix Y with complex signal samples (M frames from N antennas) from the output of the dechirper. Figure 4 Matrix S formulation. Figure 5 The creation of Qp,n matrix from SE matrix. Figure 6 Median Filtering algorithm. Figure 7 Detection procedure using threshold. Figure 8 Gaussian 2D filter of dimensions 7×7. Figure 9 2D convolution process with a 3×3 kernel matrix for one pixel at the RD-HR map. Figure 10 2D convolution process. Figure 11 The layout of different kernel functions: (a) 7×7 (b) 5×5 (c) 3×3 (d) 9×9. Figure 12 RD-HR map at the beginning of the process (top), RD-HR map after median filtering and Bragg’s lines elimination (middle) and HD HR map after convolution process (bottom). Figure 13 Detections on RD-HR map (detections are denoted by “+” markers, and the blue contours represent the contours of the criterion function of the MUSIC-based algorithm). Figure 14 Finding of the appropriate column vector ql,pq,n(q) from FFT matrix for detected range pq and l=1. Figure 15 The comparison of RD-FFT map (left) and RD-HR map obtained by the proposed algorithm (right). Figure 16 Detection of a vessel using the proposed algorithm: Detections in RD-HR map denoted by “+” markers (upper left), Azimuth estimation (A-HR criterion function) for the selected vessel detected in RD-HR map (upper right) and display of that detection (red circle) on a geographical map (bottom) with AIS data as a benchmark (light blue tracks), where the A-HR function is drawn as a yellow line (in polar coordinates, with the origin at the radar station) and scaled so that its maximum is at the position of the detection. Figure 17 The complete AIS data in a time interval of approximately 5 h and randomly selected 10 vessels with their MMSI identifiers and 5 stationary located groups of vessels with G identifiers. Figure 18 A contour formed around the AIS data. Figure 19 The display of all vessel’s detections (yellow markers) in a small part of the area, with AIS data as a benchmark (light blue tracks) in a time interval of 5 h for different detection parameters of the proposed algorithm: (a) K=5 and normalized threshold = 0.2 (b) K=5 and normalized threshold = 0.1 (c) K=10 and normalized threshold = 0.2 (d) K=10 and normalized threshold = 0.1. Figure 20 The display of all vessel’s detections (yellow markers) with AIS data as a benchmark (light blue tracks) in a time interval of 1 h (17–18 pm) for different detection parameters of the proposed algorithm: (a) K=5 and normalized threshold = 0.2 (b) K=5 and normalized threshold = 0.1 (c) K=10 and normalized threshold = 0.2 (d) K=10 and normalized threshold = 0.1. Figure 21 The display of all detections of vessel with MMSI = 636014619 (yellow markers) with AIS data as a benchmark (light blue tracks) in a time interval of 1 h (18–19 h) for different detection parameters of the proposed algorithm: (a) K=5 and normalized threshold = 0.2 (b) K=5 and normalized threshold = 0.1 (c) K=10 and normalized threshold = 0.2 (d) K=10 and normalized threshold = 0.1. Figure 22 The display of all detections of vessel with MMSI = 657199400 (yellow markers) with AIS data as a benchmark (light blue tracks) in a time interval of 1 h (18–19 h) for different detection parameters of the proposed algorithm: (a) K=5 and normalized threshold = 0.2 (b) K=5 and normalized threshold = 0.1 (c) K=10 and normalized threshold = 0.2 (d) K=10 and normalized threshold = 0.1. Figure 23 The display of all detections of groups of vessels with with G-123km identifier (yellow markers) with AIS data as a benchmark (light blue tracks) in a time interval of 1 h (18–19 h) for different detection parameters of the proposed algorithm: (a) K=5 and normalized threshold = 0.2 (b) K=5 and normalized threshold = 0.1 (c) K=10 and normalized threshold = 0.2 (d) K=10 and normalized threshold = 0.1. Figure 24 Number of vessel detections inside selected contours. Figure 25 The percentage of detection success. Figure 26 The forming of the ball (left) and the selection of the estimation area around the vessel’s contour (right). Figure 27 The display of two vessels (rounded in red) that were very close to each other (left) and zoomed display of these vessels (right). Figure 28 Logical processors usage in multithread software. sensors-22-03558-t001_Table 1 Table 1 Detection results for K=5 and different threshold values in the selected period of time. Detection Parameter K = 5 Vessel Number Vessel ID Threshold = 0.1 Threshold = 0.2 17 h 18 h 19 h 20 h 21 h 22 h 17 h 18 h 19 h 20 h 21 h 22 h 1 229395000 ✕ ✕ ✕ ✕ ✓ ✓ ✕ ✕ ✕ ✕ ✓ ✓ 2 256609000 ✕ ✕ ✕ ✕ ✓ ✓ ✕ ✕ ✕ ✕ ✕ ✓ 3 352916000 ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ 4 511539000 ✓ ✕ ✓ ✓ ✕ ✕ ✓ ✕ ✓ ✓ ✕ ✕ 5 358008073 ✕ ✓ ✓ ✓ ✓ ✓ ✕ ✓ ✓ ✓ ✓ ✓ 6 636014619 ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✕ 7 352915000 ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ 8 657105000 ✕ ✕ ✕ ✕ ✓ ✓ ✕ ✕ ✕ ✕ ✕ ✕ 9 657162200 ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ 10 657199400 ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✕ ✕ ✕ ✕ ✓—detected, ✕—not detected or no AIS. sensors-22-03558-t002_Table 2 Table 2 Detection results for K=10 and different threshold values in the selected period of time. Detection Parameter K = 10 Vessel Number Vessel ID Threshold = 0.1 Threshold = 0.2 17 h 18 h 19 h 20 h 21 h 22 h 17 h 18 h 19 h 20 h 21 h 22 h 1 229395000 ✕ ✕ ✕ ✕ ✓ ✓ ✕ ✕ ✕ ✕ ✓ ✓ 2 256609000 ✕ ✓ ✕ ✓ ✕ ✓ ✕ ✕ ✕ ✕ ✕ ✓ 3 352916000 ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ 4 511539000 ✓ ✓ ✓ ✓ ✕ ✕ ✓ ✓ ✓ ✓ ✕ ✕ 5 358008073 ✕ ✓ ✓ ✓ ✓ ✓ ✕ ✓ ✓ ✓ ✓ ✓ 6 636014619 ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✕ 7 657129500 ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ 8 657105000 ✕ ✕ ✕ ✕ ✓ ✓ ✕ ✕ ✕ ✕ ✓ ✓ 9 657162200 ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ 10 657199400 ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✕ ✓ ✕ ✓—detected, ✕—not detected or no AIS. sensors-22-03558-t003_Table 3 Table 3 The ratio of total number of detections and detections within contours for all vessels. Algorithm Parameters Total Number of Detections within Contours for All Vessels Total Number of Non-Contour Detections for All Vessels The Ratio of Total Number of Detections and Number of Detections within Contours for All Vessels Threshold 0.1 Threshold 0.2 Threshold 0.1 Threshold 0.2 Threshold 0.1 Threshold 0.2 K = 5 17 h 976 780 4856 1294 5.97 2.66 18 h 1600 1315 18435 6849 12.52 6.21 19 h 952 669 6150 1400 7.46 3.09 20 h 1097 910 5295 1170 5.83 2.28 21 h 1968 1685 6024 1397 4.06 1.83 22 h 954 804 2426 622 3.54 1.77 K = 10 17 h 1390 1016 13298 5093 10.57 6.01 18 h 2091 1634 39678 18136 19.97 12.10 19 h 1363 869 18264 5878 14.40 7.76 20 h 1371 1052 17811 4946 13.99 5.70 21 h 2578 2031 21154 7191 9.20 4.54 22 h 1267 1008 8839 3254 7.98 4.23 sensors-22-03558-t004_Table 4 Table 4 Pd and Pf of the vessel with MMSI = 636014619 in a time interval of 1 h (18–19 h). Detection Parameters Pd Pfa Pfae K = 5, threshold = 0.2 0.7238 21/1,689,555 * 48/1,759,737 * K = 5, threshold = 0.1 0.7238 1.0417 ×10−4 1.1536 ×10−4 K = 10, threshold = 0.2 0.7524 4.5278 ×10−4 4.6257 ×10−4 K = 10, threshold = 0.1 0.8381 7.6470 ×10−4 7.7625 ×10−4 * too small a sample to estimate. sensors-22-03558-t005_Table 5 Table 5 Pd and Pf of the vessel with MMSI = 657199400 in a time interval of 1 h (18–19 h). Detection Parameters Pd Pfa Pfae K = 5, threshold = 0.2 0.2095 1.7061 ×10−4 1.7565 ×10−4 K = 5, threshold = 0.1 0.2952 5.6351 ×10−4 5.7442 ×10−4 K = 10, threshold = 0.2 0.3238 5.3808 ×10−4 5.4040 ×10−4 K = 10, threshold = 0.1 0.3714 13.3301 ×10−4 13.3301 ×10−4 sensors-22-03558-t006_Table 6 Table 6 The average segment processing time obtained by the implemented detection software. CPU Type Processing Time (1-Thread Software) Processing Time (Multi-Thread Software) Intel CORE i7 1075H 32.2 s 8.1 s AMD Ryzen 9 5900HX 22.9 s 4.4 s Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Headrick J.M. Skolnik M.I. Over-the-Horizon Radar in the HF Band Proc. IEEE 1974 62 664 673 10.1109/PROC.1974.9506 2. Georges T.M. Harlan J.A. New horizons for over-the-horizon radar? IEEE Antennas Propag. Mag. 1994 36 14 24 10.1109/74.317763 3. Ponsford A.M. Surveillance of the 200 nautical mile Exclusive Economic Zone (EEZ) using high frequency surface wave radar Can. J. Remote Sens. 2001 27 354 360 4. FM/CW Radar Signals and Digital Processing. Technical Report Available online: https://repository.library.noaa.gov/view/noaa/18645 (accessed on 20 February 2022) 5. Jankiraman M. FMCW Radar Design Artech House London, UK 2018 6. Dzvonkovskaya A. Gurgel K. Rohling H. Schlick T. Low Power High Frequency Surface Wave RadarApplication for Ship Detection and Tracking Proceedings of the International Conference on Radar Adelaide, Australia 2–5 September 2008 7. Gurgel K. Schlick T. Remarks on Signal Processing in HF Radars Using FMCW Modulation Proceedings of the International Radar Symposium IRS 2009 Hamburg, Germany 9–11 September 2009 8. The Effects of Sea Clutter on the Performance of HF Surface Wave Radar in Ship Detection and the Implication on the Radar Design. Defence Research and Development Canada–Ottawa Technical Memorandum DRDC Ottawa TM 2007-325 Available online: https://www.researchgate.net/publication/317290451 (accessed on 20 February 2022) 9. Menelle M. Auffray G. Jangal F. Bazin V. Urbani B. HF-Surface Wave Radar: First results for sea state studies Proceedings of the 7th WSEAS International Conference on Application of Electrical Engineering (AEE’08) Trondheim, Norway 2–4 July 2008 10. Ponsford A.M. Wang J. A review of high frequency surface wave radar for detection and tracking of ships Turk. J. Electr. Eng. Comput. Sci. 2010 18 409 428 11. Gurgel K.W. Essen H.H. Kingsley S.P. HF radars: Physical limitations and recent developments Coast. Eng. 1999 37 201 218 10.1016/S0378-3839(99)00026-5 12. Headrick J.M. Thomason J.F. HF radars: Applications of high-frequency radar Radio Sci. 1998 33 1045 1054 10.1029/98RS01013 13. Sevgi L. Ponsford A. Chan H.C. Integrated Maritime surveillance system based on high–frequency surface–wave radars. 1. Theoretical background and numerical simulations IEEE Antennas Propag. Mag. 2001 43 20 43 10.1109/74.951557 14. Ponsford A. Sevgi L. Chan H.C. An integrated Maritime surveillance system based on high–frequency surface–wave radars. 2. Operational status and system performance IEEE Antennas Propag. Mag. 2001 43 52 63 10.1109/74.979367 15. Menelle M. Auffray G. Jangal F. Full Digital High Frequency Surface Wave Radar: French Trials in the Biscay Bay Proceedings of the 2008 International Conference on Radar Adelaide, Australia 2–5 September 2008 16. Dzvonkovskaya A. Gurgel K. Rohling H. Schlick T. HF Radar WERA Application for Ship Detection and Tracking J. Navig. 2009 7 18 25 17. Rohling H. Radar cfar thresholding in clutter and multiple target situations IEEE Trans. Aerosp. Electron. Syst. 1983 19 608 621 10.1109/TAES.1983.309350 18. Turley M. Hybrid CFAR techniques for HF radar Proceedings of the Radar Systems (Conf. Publ. No. 449) Edinburgh, UK 14–16 October 1997 19. Dzvonkovskaya A.L. Rohling H. Adaptive thresholding for HF radar ship detection Proceedings of the Sixth International Radiowave Oceanography Workshop (ROW2006) Hamburg, Germany 15–18 May 2006 20. Hongbo L. Yiying S. Yongtan L. Estimation of detection threshold in multiple ship target situations with HF ground wave radar J. Syst. Eng. Electron. 2007 18 739 744 10.1016/S1004-4132(08)60013-4 21. Jangal F. Saillant S. Helier M. Wavelet contribution to remote sensing of the sea and target detection for a high-frequency surface wave radar IEEE Geosci. Remote Sens. Lett. 2008 5 552 556 10.1109/LGRS.2008.923211 22. Li Q. Zhang W. Li M. Niu J. Jonathan Wu Q.M. Automatic detection of ship targets based on wavelet transform for HF surface wavelet radar IEEE Geosci. Remote Sens. Lett. 2017 14 714 718 10.1109/LGRS.2017.2673806 23. Grosdidier S. Baussard A. Ship detection based on morphological component analysis of high-frequency surface wave radar images IET Radar Sonar Navig. 2012 8 48 53 10.1049/iet-rsn.2012.0062 24. Starck J.L. Elad M. Donoho D. Redundant multiscale transforms and their application for morphological component separation Adv. Imag. Electron Phys. 2004 132 288 348 25. Starck J.L. Moudden Y. Bobin J. Elad M. Donoho D. Morphological component analysis Proceedings of the SPIE Conference Wavelets San Diego, CA, USA 31 July 2005 26. Cai J. Zhou H. Huang W. Wen B. Ship Detection and Direction Finding Based on Time-Frequency Analysis for Compact HF Radar IEEE Geosci. Remote Sens. Lett. 2021 18 72 76 10.1109/LGRS.2020.2967387 27. Fast 2D Peak Finder Available online: https://github.com/adinatan/fastpeakfind/releases/tag/1.13.0.0 (accessed on 20 February 2022) 28. Schmidt R. Multiple emitter location and signal parameter estimation IEEE Trans. Antennas Propag. 1986 34 276 280 10.1109/TAP.1986.1143830 29. Capon J. High resolution frequency-wavenumber spectrum analysis Proc. IEEE 1969 57 1408 1418 10.1109/PROC.1969.7278 30. Roy R. Kailath T. ESPRIT-estimation of signal parameters via rotational invariance techniques IEEE Trans. Acoust. Speech Signal Process. 1989 37 984 995 10.1109/29.32276 31. Zhang L. Shi C. Niu J. Ji Y. Jonathan Wu Q.M. DOA Estimation for HFSWR Target Based on PSO-ELM IEEE Geosci. Remote Sens. Lett. 2022 19 1 5 10.1109/LGRS.2021.3062373 32. Kim B. Jin Y. Lee J. Kim S. High-Efficiency Super-Resolution FMCW Radar Algorithm Based on FFT Estimation Sensors 2021 21 4018 10.3390/s21124018 34200856 33. Kim B. Jin Y. Lee J. Kim S. FMCW Radar Estimation Algorithm with High Resolution and Low Complexity Based on Reduced Search Area Sensors 2022 22 1202 10.3390/s22031202 35161947 34. Li Y.-C. Choi B. Chong J.-W. Oh D. 3D Target Localization of Modified 3D MUSIC for a Triple-Channel K-Band Radar Sensors 2018 18 1634 10.3390/s18051634 29783782 35. Seo J. Lee J. Park J. Kim H. You S. Distributed Two-Dimensional MUSIC for Joint Range and Angle Estimation with Distributed FMCW MIMO Radars Sensors 2021 21 7618 10.3390/s21227618 34833690 36. Patole S. Torlak M. Wang D. Ali M. Automotive radars: A review of signal processing techniques IEEE Signal Process. Mag. 2017 34 22 35 10.1109/MSP.2016.2628914 37. Kim B. Kim S. Lee J. A novel DFT-based DOA estimation by a virtual array extension using simple multiplications for FMCW radar Sensors 2018 18 1560 10.3390/s18051560 38. Kim B. Kim S. Jin Y. Lee J. Low-complexity joint range and Doppler FMCW radar algorithm based on number of targets Sensors 2020 20 51 10.3390/s20010051 31861824 39. Akaike H. Fitting autoregressive models for prediction Ann. Inst. Stat. Math. 1969 21 243 247 10.1007/BF02532251 40. Rissanen J. Modeling by shortest data description Automatica 1978 14 465 471 10.1016/0005-1098(78)90005-5 41. Wax M. Ziskind I. Detection of the Number of Coherent Signals by the MDL Principle IEEE Trans. Acoust. Speech. Signal Process. 1989 37 1190 1196 10.1109/29.31267 42. Ji Y. Zhang J. Wang Y. Chang G. Sun W. Performance Analysis of Target Detection with Compact HFSWR Proceedings of the 2016 CIE International Conference on Radar (RADAR) Guangzhou, China 10–13 October 2016 43. Ji Y. Zhang J. Wang Y. Sun W. Li M. Target Monitoring Using Small-Aperture Compact High-Frequency Surface Wave Radar IEEE A E Syst. Mag. 2018 33 22 31 10.1109/MAES.2018.170023 44. Xie J. Yuan Y. Liu Y. Super-Resolution Processing for HF Surface Wave Radar Based on Pre-Whitened MUSIC IEEE J. Ocean. Eng. 1998 23 313 321 45. MoyWe D.E. Warrington E.M. Some super-resolution DF measurements within the HF band Proceedings of the 10th International Conference on Antennas and Propagation Edinburgh, UK 14–17 April 1997 46. Akaike H. A New Look at the Statistical Model Identification IEEE Trans. Automat. Contr. 1974 19 716 723 10.1109/TAC.1974.1100705

PMC009xxxxxx/PMC9099647.txt

==== Front J Clin Med J Clin Med jcm Journal of Clinical Medicine 2077-0383 MDPI 10.3390/jcm11092492 jcm-11-02492 Article Actual Associations between HLA Haplotype and Graves’ Disease Development Zawadzka-Starczewska Katarzyna 1 Tymoniuk Bogusław 2 Stasiak Bartłomiej 3 https://orcid.org/0000-0002-8748-3337 Lewiński Andrzej 14 https://orcid.org/0000-0002-2910-7691 Stasiak Magdalena 1* Lupoli Roberta Academic Editor 1 Department of Endocrinology and Metabolic Diseases, Polish Mother’s Memorial Hospital—Research Institute, 281/289 Rzgowska St., 93-338 Lodz, Poland; kasiula891@op.pl (K.Z.-S.); andrzej.lewinski@umed.lodz.pl (A.L.) 2 Department of Immunology, Rheumatology and Allergy, Medical University of Lodz, 251 Pomorska St., 92-213 Lodz, Poland; boguslaw.tymoniuk@umed.lodz.pl 3 Institute of Information Technology, Lodz University of Technology, 215 Wolczanska St., 90-924 Lodz, Poland; bartlomiej.stasiak@p.lodz.pl 4 Department of Endocrinology and Metabolic Diseases, Medical University of Lodz, 281/289 Rzgowska St., 93-338 Lodz, Poland * Correspondence: mstasiak33@gmail.com 29 4 2022 5 2022 11 9 249205 4 2022 26 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). The association between HLA and the risk of Graves’ disease (GD) has been analyzed for many years. However, the results were often inconsistent and mostly regarded Asian populations. The purpose of our study was to perform HLA genotyping using a next-generation sequencing (NGS) method in Caucasians, to find out which alleles are eventually correlated with GD morbidity as well as which of them can be considered protective. HLA-A, -B, -C, -DQB1, -DRB1 were genotyped using a next-generation sequencing method in 2376 persons, including 159 GD patients and 2217 healthy controls. We have demonstrated a significant association between the risk of GD and the following alleles: HLA-B*08:01, -B*39:06, -B*37:01, -C*07:01, -C*14:02, -C*03:02, -C*17:01, -DRB1*03:01, -DRB1*11:01, -DRB1*13:03, -DRB1*01:03, -DRB1*14:01, -DQB1*03:01, DQB1*02:01. The alleles HLA-B*39:06, -B*37:01, -C*14:02, -C*03:02, -C*17:01, -DRB1*14:01 are novel GD-associated, previously not-reported independent ones with no linkage disequilibrium with other high-risk alleles. On the other hand, the frequencies of HLA-B*07:02, -C*07:02, -C*03:04, DRB1*07:01, -DQB1*02:02, -DQB1*03:03 were significantly lower in GD compared to controls. This study demonstrated the actual relationships between HLA and GD based on the NGS method and provided a novel set of alleles as a reliable tool for an individual personalized risk assessment. Graves’ disease human leukocyte antigen HLA susceptibility alleles NATIONAL SCIENCE CENTRE2020/04/X/NZ5/00354 Polish Mother’s Memorial Hospital—Research instituteThis research was funded by the NATIONAL SCIENCE CENTRE of Poland, grant number 2020/04/X/NZ5/00354, and by the Polish Mother’s Memorial Hospital—Research institute, Łódź, Poland. ==== Body pmc1. Introduction Graves’ disease (GD) is an autoimmune thyroid disorder characterized by the production of specific antibodies against the thyrotropin (TSH) receptor. These TSH-receptor antibodies (TRAb) most frequently stimulate thyroid hormone production resulting in hyperthyroidism. However, TRAb may also block the TSH-receptor or have an ambivalent character [1]. The prevalence of GD in the Caucasian population is about 0.5–2.0% [1,2]. Similar to other autoimmune diseases, GD is usually triggered by environmental factors in genetically predisposed individuals [2,3]. Among genes associated with the immune response, human leukocyte antigen (HLA) genes have been found to be associated with autoimmune thyroid diseases (AITD), including GD [4]. Other genes such as cytotoxic T lymphocyte-associated factor 4 (CTLA-4), thyroglobulin (Tg) or CD40 genes can also be associated with an increased risk of GD [5,6]. However, taking into account the relevance of the major histocompatibility complex (MHC) for immune responses and high polymorphism of HLA region, it seems to play a prominent role as a molecular background of GD [4]. Previous studies on HLA-related susceptibility to GD indicated the existence of significant ethnic differences [2,3,7,8,9,10,11,12]. Furthermore, the obtained results in either Asian or Caucasian populations are not consistent [3,7,8,9,10,11,12]. Most of the previously published studies applied much older methods, especially serological ones, which had significantly lower accuracy than the high-resolution next-generation sequencing (NGS) method used in our study. Moreover, the symbols of particular alleles assessed by the previously applied methods differ from those currently used, and some antigens previously denoted by one symbol are separated as several individual alleles when assessed by the high-resolution method. This fact undoubtedly has an important impact on the accuracy and consistency of the already published results. Several studies conducted in the Asian population postulated the importance of the presence of HLA-B*46 in the development of GD [7,8,9], while others indicated a possible relationship with HLA-DRw8, -DQw4, -B5, Dw12 and -A11 antigens [13]. On the other hand, a Chinese study showed that the relationship between GD and HLA-B*46 concerned only men [14]. The same study found that the risk of GD was higher in patients with HLA-A*2, -Cw1, -DRB1*16:02, DRB1*03:01, DRB1*14:05, -DRB5*02, -DQB1*05:02, while the presence of -DRB1*15:01 and -DQB1*03:01 played a protective role [14]. Another study of Chinese patients observed the relationship between HLA-DR9 and -DQB1*03:03 in males [15]. Protective effects of HLA-DR12, previously described as an increased risk antigen, and of HLA-DQA1*04:01 were also postulated [15]. In turn, Japanese authors showed that the most important factor in the development of GD was the presence of HLA-DPB1*05:01 and/or HLA-A*2, with the risk being the highest in carriers of both of them [16]. In a Taiwanese study, HLA-A*02:07 was found to be a GD risk factor [17], while other Taiwanese authors showed that there was a correlation between GD and HLA-B*46:01, -DPB1*05:01, -DQB1*03:02, -DRB1*15:01 and -DRB1*16:02 [18]. Despite the apparent discrepancies, the results of a number of studies conducted on the Asian population are consistent in terms of the relationship between GD and the presence of alleles such as HLA-A*02:07, -B*46, -DRB1*08 or -DPB1*05:01 [6,17,18]. In Caucasians, studies are more scarce. Their results are consistent only with regard to the increased risk of GD in people with HLA-DRB1*03 and alleles remaining in linkage disequilibrium with HLA-DRB1*03 for this population, i.e., -DQA1*05: 01, -DQB1*02:01 [3,11,19,20]. However, it is well known that HLA-DRB1*03 is associated with an increased risk of all thyroid autoimmune diseases, not exclusively with GD. In regard to other alleles, Heward et al. postulated a possible role of HLA-DQB1*03:01/04 and -DQB1*02 in GD occurrence [3]. More recently, Vita et al. demonstrated significantly higher frequency of HLA-C*07, -C*17 and -DRB1*04 in patients with GD as compared to controls [2]. The small amount of data for the Caucasian population and the inconsistency between the results of individual studies, related mainly to different methods used and the size of the groups, left the HLA-related genetic basis of GD for the Caucasian population not satisfactorily explained. Therefore, there was a need to re-analyze and to compare HLA profiles in large groups of patients with GD and healthy controls using a modern high-resolution NGS method. By application of this method, our research group has recently demonstrated novel strong correlations between HLA and subacute thyroiditis (SAT), including not only the risk of SAT but also its clinical course [21,22,23,24,25]. It can be supposed that the relationship between GD and HLA is also much more complicated and includes more alleles than it was previously reported. The purpose of the study was to re-evaluate class I and class II MHC alleles in 159 patients with GD and 2217 healthy controls, and to clarify which HLA alleles are eventually associated with GD in the Caucasian population. Identification of an actual set of GD-related and GD-protective alleles would provide a novel reliable tool for the individual risk assessment and would constitute a great step in a development of personalized medicine. 2. Materials and Methods 2.1. GD Group and Control Group A total number of 2376 persons were included into the study, with 2217 healthy Polish hematopoietic stem cell donors with no medical history of thyroid disease, and 159 patients who were diagnosed with GD in the Department of Endocrinology and Metabolic Diseases, Polish Mother’s Memorial Hospital—Research Institute, Lodz, Poland, as well as in the Department-associated outpatient clinic. The size of the control group should be large enough to avoid any bias related to potential diseases which may appear in some members of this group in future, as well as to avoid any bias associated with random increased or decreased frequency of some alleles in a smaller control group. 2.2. HLA Typing Procedures HLA-A, -B, -C, -DQB1 and -DRB1 were genotyped using a high-resolution NGS method. DNA was isolated from whole blood collected to the anticoagulant (EDTA)-containing tubes. Further sample preparation consisted of several steps, including long-range PCR, genomic library preparation, and sequencing. Amplicons were quantified by fluorescence detection method, balanced, pooled and enzymatically fragmented. Afterwards, end repair and A-tailing of DNA fragments was performed followed by index adapter ligation. Genomic library was cleaned up and denatured prior to loading on NGS Illumina Platform (NextSeq). We analyzed sequencing data with NGS HLA Genotyping software MIA FORA. The data were considered sufficient whenever the coverage reached 40. 2.3. Statistical Analysis Allele frequencies were reported as absolute values and in percentages. The statistical significance of the differences between groups was evaluated by the chi-square test and by binomial logistic regression analysis, with p values ≤ 0.05 considered significant. The statistical analysis was carried out using Statistica v 13.1 software (Statsoft Polska, Kraków, Poland). 2.4. Inclusion Criteria In all patients included into the study, GD was diagnosed on the basis of standard criteria [1], including hyperthyroidism, elevated TRAb level, as well as typical ultrasound (US) pattern. 2.5. Biochemical and US Procedures Serum levels of TSH, free triiodothyronine (FT3), free thyroxine (FT4) and TRAb were measured by the electrochemiluminescence immunoassay (ECLIA) with Cobas e601 analyzer (Roche Diagnostics, Indianapolis, IN, USA). Ultrasound examinations (US) were performed in every patient, using a 7–14 MHz linear transducer (Toshiba Aplio XG; Toshiba, Japan). 2.6. Ethics Procedures Informed consent for all performed procedures was obtained from all of the patients after a full explanation of the purpose and nature of all procedures used. The study was approved by the Ethics Committee of the Polish Mother’s Memorial Hospital—Research Institute, Lodz, Poland (approval code—108/2018). 3. Results The mean age of patients at diagnosis of GD was 43.88 ± 17.44 years, with a male to female ratio of 1:4.48. The diagnosis of GD was based on the laboratory results (Table 1). Statistically significant differences in the frequency of HLA alleles between patients with GD and control group were found with several alleles having higher frequency and others having lower frequency in GD as compared to controls. 3.1. Alleles with Higher Frequencies in GD The alleles of higher frequency in GD as compared to controls were found in both MHC class I and class II. The differences were statistically significant for the following alleles of MHC class I: HLA-B*08:01 (12.5% vs. 9.0%), -B*39:06 (1.56% vs. 0.41%), -B*37:01 (2.19% vs. 0.83%) (Figure 1a), -C*07:01 (18.13% vs. 13.49%), -C*14:02 (2.19% vs. 0.95%), -C*03:02 (3.44% vs. 0.50%), -C*17:01 (2.50% vs. 0.50%) (Figure 1b). For the MHC class II, significant differences in the frequencies were found for the following alleles: -DRB1*03:01 (16.25% vs. 9.83%), -DRB1*11:01 (11.56% vs. 7.49%), -DRB1*13:03 (3.44% vs. 1.87%), -DRB1*01:03 (0.94% vs. 0.20%), -DRB1*14:01 (1.56% vs. 0.36%) (Figure 2a), -DQB1*03:01 (23.13% vs. 18.83%), DQB1*02:01 (16.25% vs. 9.72%) (Figure 2b). 3.2. Alleles with Lower Frequencies in GD On the other hand, the frequencies of the following alleles were significantly lower in GD as compared to controls: HLA-B*07:02 (5.94% vs. 10.53%), -C*07:02 (4.38% vs. 11.41%), -C*03:04 (1.56% vs. 5.19%), -DRB1*07:01 (8.75% vs. 12.97%), -DQB1*02:02 (5.63% vs. 9.47%), -DQB1*03:03 (2.19% vs. 4.74%) (Figure 3). 3.3. Significance of a Single High Risk Allele and of Co-Presence of Alleles In 26 patients (16.4%), only one of the alleles described above as correlated to a high risk of GD was found. Among this group, the following alleles were found: HLA-B*39:06, -C*03:02, -C*07:01, -C*14:02, -DRB1*14:01 and -DQB1*03:01 in two (7.7%), three (11.5%), six (23.1%), one (3.8%), one (3.8%) and thirteen (50%) patients, respectively. In 33 patients (20.8%), two of the high-risk alleles were present. Among this group the co-presence of HLA-DRB1*11:01 and -DQB1*03:01 was observed the most frequently (45%). The co-presence of alleles which are not in linkage disequilibrium was observed in 9 patients (27.3%). Among this group, the most frequent was the combination of HLA-C*14:02 with -DQB1*03:01 (22.2%). Interestingly, allele HLA-B*08:01 was most frequently present with -DRB1*03:01 and -DQB1*02:01. A combination of the three alleles, -B*08:01, -DRB1*03:01- and DQB1*02:01, occurred in 36 of 159 GD patients (22%), while the co-presence of these three alleles was found only in 5.87% of the control group. In only three patients (1.9%), the HLA-B*08:01 allele occurred with other alleles (Table 2). In none of the patients did -B*08:01 occur as the only high-risk allele. 4. Discussion In the recent years, it has become more and more clear that GD is triggered by environmental factors such as infections, stress, smoking, etc., in genetically predisposed individuals [6,7]. This genetic susceptibility plays a critical role in the pathogenesis of GD and has been previously demonstrated to be HLA-dependent. Moreover, in both Asian and Caucasian populations, this genetic background was demonstrated to include MCH class I and class II. Therefore, the identification of alleles specifically related to GD seems to be a crucial step in the development of personalized medicine in regard to thyroid disorders. However, the results that have been reported for the last two decades in both populations are not consistent. Significant discrepancies in the results obtained by various authors can undoubtedly depend on the applied method. The use of high-resolution methods can change the results obtained with older methods. Genotyping methods of resolution that allow the achievement of allelic specificity is currently a gold standard of research that is expected to demonstrate high reliability, and to avoid method-dependent errors. Older, less accurate methods provide results for the entire allelic group, not for a particular allele. This may result in erroneous conclusions and discrepancies in the results of the studies depending on the method. According to the results obtained in a strictly controlled group of HLA typing performed for the purposes of bone marrow transplantation between 1996 and 2011, 29.1% discrepancies were found between older methods and NGS method [26] which was applied in the present study. Another important example of the significance of the genotyping method is HLA-B*27 test commonly used to genetically confirm a diagnosis of ankylosing spondylitis. It has been recently demonstrated that HLA typing methods used so far gave insufficiently precise results, and two alleles, HLA-B*27:06 and HLA-B*27:09, are probably not associated with the disease [27]. These examples clearly present the importance of the method and a possible role of a method-dependent factor in the inconsistency of the previous results. To date, to the best of our knowledge, the present study included the largest Caucasian cohort to whom a modern high-resolution method was applied to analyze both MCH classes. Thus, this study can summarize and clarify the actual HLA-related genetic background of GD. The present study has confirmed a strong correlation between GD and HLA-B*08:01, DRB1*03:01, -DQB1*02:01 (Figure 1a and Figure 2a,b). Our observation is consistent with the previous reports [2,3,11,12,19,20]. Interestingly, in our study, HLA-B*08:01 was accompanied by these two alleles in most cases of its presence although it belongs to a different MCH class (Table 2). The co-presence of these three alleles in GD group was 4 times more frequent than in controls. Additionally, HLA-DRB1*03:01 and -DQB1*02:01 were rarely present without HLA-B*08:01 (Table 2). Such a close unexpected association between HLA-B*08:01 and MCH class II alleles was previously postulated in the Caucasian population [28], but the present study has confirmed it and demonstrated its strength for GD patients for the first time. This finding sheds a new light on the possible linkage disequilibrium between alleles from different MCH classes. Moreover, it indicates the existence of specific mechanism of impact augmentation between these alleles in GD. The currently discussed group of linkage disequilibrium includes also HLA-DQA1*05:01 [27,29,30], previously reported as GD high risk allele [2,3,19]. We did not perform comparison of frequencies of HLA-DQA1*05:01 between our groups because HLA-DQA1 alleles are not reported for transplantation purposes and results performed using NGS method are unavailable either for our control group or for any other representatively large cohort in Poland. Comparison with any population with available lower resolution results could introduce unacceptable bias, as the main strength of the present study is the application of the most reliable method. However, a strong linkage disequilibrium between HLA-DQA1*05:01 and DRB1*03:01, DQB1*02:01, as well as other alleles demonstrated as high-risk ones in the present study, i.e., DRB1*01:03, DQB1*03:01 and DRB1*13:03 (Figure 2) seems to confirm previous findings of HLA-DQA1*05:01 being a GD risk allele [2,3,19]. The potential role of HLA-DQB1*03:01 was previously postulated by Heward et al. [3]. Similar to our study, Martin et al. observed significantly higher frequency of HLA-DRB1*11:01 and -DRB1*13:03 [31]. Our study has confirmed the role of these alleles as well as has further supported the role of HLA-DQA1*05:01 because of linkage disequilibrium between either HLA-DRB1*11:01 or -DRB1*13:03 and HLA-DQA1*05:01 [29]. Both of these HLA-DRB1 alleles are also in linkage disequilibrium with HLA-DQB1*03:01 which has been confirmed in the present study as high-risk allele. To the best of our knowledge, this is the first report presenting strong susceptibility to GD related to HLA-DRB1*01:03—the allele that is also in linkage disequilibrium with HLA-DQA1*05:01 and -DQB1*03:01 [29]. Our study also demonstrated that the risk of GD is significantly increased in patients with a presence of HLA-C*07:01, being in linkage disequilibrium with previously discussed -B*08:01 [32,33]. It is worth mentioning that the importance of HLA-B*08:01 and -C*07:01 was previously postulated in the literature, mostly as -B*08 and -C*07 with application of lower resolution methods and two-digit results [2,34]. Our study has confirmed this association for HLA-B*08:01 and -C*07:01 (Figure 1a,b). When analyzing the above-described associations, one should keep in mind that susceptibility associated with alleles being in linkage disequilibrium cannot be considered as fully independent. However, the single presence of any of them can be correlated with the risk of GD. This study has confirmed the correlation of GD with HLA-C*17:01 postulated before by Vita et al. [2]. Our study based on NGS method has demonstrated even stronger statistical significance than previously reported (Figure 1b). No linkage disequilibrium has been reported between HLA-C*17:01 and other high-risk alleles [32,33], so HLA-C*17:01 should be considered an independent one. The most important finding of our present study is the significance of novel alleles which have been reported here as GD-related for the first time. This group includes HLA-B*37:01, -B*39:06, -C03:02, -C14:02, and -DRB1*14:01 (Figure 1a,b and Figure 2a). These alleles belong to both I and II MCH classes and are not in linkage disequilibrium either with each other or with any previously discussed GD-related alleles [27,32,33]. Thus, their significance for the pathogenesis of GD is particularly relevant. Similar to Vita et al. [2], we have not confirmed a higher frequency of HLA-DRB1*08 previously postulated in North American and British Caucasians [34,35]. The specificity of the applied method may have played the most important role in these differences. Our study has demonstrated strong association of GD with the presence of several alleles belonging to both MCH classes. Strong impact of any of them on the risk of GD can be additionally supported by our finding that the presence of a single allele from the high-risk group is sufficient to induce GD. In the present study, alleles HLA-B*39:06, -C*03:02, -C*07:01, -C*14:02, -DRB1*14:01 and -DQB1*03:01 were present as a single allele in GD patients. Most of them, i.e., HLA-B*39:06, -C*03:02, -C*14:02, -DRB1*14:01, have been reported for the first time as high-risk alleles in the present study. It is worth underlining that among patients with two high risk alleles, the co-presence of HLA-DRB1*11:01 and -DQB1*03:01, being in linkage disequilibrium was the most common. However, in 27.3% of the patients, the co-present alleles were totally independent, with HLA-C*14:02—newly reported in the present paper—being the most frequent independent allele. Predisposition for autoimmune disorders other than GD can also be HLA-related. Hashimoto’s thyroiditis (HT) and GD share a variety of common etiological and pathophysiological factors, including HLA-based predisposition, a trend to aggregate in the same families or even to coexist in the same gland [36]. Moreover, several reports suggested the existence of a continuum between HT and GD [37,38,39]. In our GD group, only four patients had preceding Hashimoto’s thyroiditis, and in all of them HLA-DQB1*02:01 was present. This allele was demonstrated in our results as one of the alleles related to high risk of GD and it is in linkage disequilibrium with -DRB1:03:01—an allele typical for thyroid autoimmunity. Interestingly, the co-presence of these two alleles was found in all patients with coexistence of GD and non-thyroidal autoimmunity (i.e., in two patients with Addison’s disease and two patients with diabetes type 1). These subgroups were too small to obtain any statistical results, but this phenomenon requires further analysis. It was previously postulated that the presence of some HLA alleles may play a protective role against GD. Similar to the results on high-risk alleles, most of the papers published so far regard the Asian population. Yin et al. postulated that HLA-B*33 can protect against GD [40], while Mehraji et al. demonstrated that HLA-DQB1*02:01 and -DQA1*02:01 play the protective role [41]. The results of studies in the Korean population did not confirm the previous findings and revealed that alleles HLA-DRB1*01:01, DRB1-*07:01, -DRB1*12:02 and -DRB1*13:02 were the protective ones [42] On the other hand, other Korean group confirmed the protective role of HLA-DQA1*02:01 and -DQB1*02:01 as well as HLA-DRB1*12, and additionally postulated the significance of -DQA1*06:01 and -DQB1*03:01 [43]. In the Chinese population, Cavan et al. postulated the significance of -DQA1*04:01 and confirmed previously reported protective role of HLA-DRB1*12 and -DQB1*03:01 [15]. The results obtained by the different research groups were inconsistent and it is difficult to unambiguously distinguish the actually protective alleles. The main candidates in the Asian population seem to be HLA-DRB1*12:02, -DQB1*02:01 and DQB1*03:01. A similar situation of inconsistent results regards the Caucasian population, and in addition, the results are more scarce. Therefore, our present study aimed also to compare GD and control cohorts with regard to potentially protective alleles. We confirmed the previously reported significantly less frequent presence of HLA-DRB1*07:01, [12,34,35] and -DQB1*02:02 [34] in patients with GD as compared to healthy controls (Figure 3). However, in the last study, the significance of -DQB1*02 was postulated in two-digit presentation [34]. We have refined and clarified this finding using the NGS method. Our study revealed also lower frequency of HLA-DQB1*03:03 in GD individuals (Figure 3). It has to be underlined that HLA-DQB1*02:02 and -DQB1*03:03 are in linkage disequilibrium with -DRB1*07:01, together with the previously postulated -DQA1*02:01 [12,27]. The presence of these protective alleles can play a very important role in GD development. Proteins related to HLA-DRB1*03:01 and its linkage disequilibrium alleles differ from those related to -DRB1*07:01 and its linkage disequilibrium-related alleles at position β74, a crucial site in the binding pocket of the HLA allele, the residue being arginine and glutamine, respectively [2,34]. It has been hypothesized that in patients with the co-presence of these two alleles, HLA-DRB1*07:01 can cancel out the GD-susceptibility associated with -DRB1*03:01 [2,34]. Similar to our findings on novel high-risk alleles, we have also demonstrated novel, previously not reported, protective associations. In the present study, the frequencies of HLA-B*07:02 and -C*07:02 were significantly lower in the GD cohort than in the control group (Figure 3). These alleles cannot be considered independent as there is also a linkage disequilibrium between them. However, as it was previously underlined, the presence of any of them can potentially be sufficient as a protective factor. Moreover, another novel independent allele was revealed as protective, i.e., -HLA-C*03:04, with a high statistical significance of p = 0.004 (Figure 3). This allele is not in linkage disequilibrium with any other potentially protective allele. The differences in the results between Caucasian and Asian populations with regard to the MCH class II alleles, which are in linkage disequilibrium with HLA-DQA1*05:01 can be considered unexpected. Among all of these alleles, only -DRB1*03:01 was proved to be GD-related in the Asian population [14]. However, all of the above-described linkage disequilibrium-based correlations are common not exclusively for the Caucasian population but for all analyzed populations, including Asians [29]. Therefore, the question arises why the correlations present in Caucasians are absent in Asians, in whom completely different alleles are considered as high risk of GD. Furthermore, taking into account our present results and the literature data, we can observe a phenomenon of entirely opposite roles of HLA-DQB1*02:01 and -DQB1*03:01 in these populations. These alleles have been demonstrated as high risk in Caucasians in the present study and in some previous ones [3]. However, they are reported as protective in Asian studies [15,42,43]. Therefore, although HLA-DRB1*03:01 has been demonstrated to be a high-risk allele in both populations, -DQB1*03:01 being in linkage disequilibrium with -DRB1*03:01 in both populations, is a high-risk allele in Caucasians and a protective allele in Asians. This fact demonstrates the impact of other factors influencing GD risk in the Asian population and points out the necessity for further analysis of this phenomenon. 5. Conclusions The present study has demonstrated the actual associations between HLA haplotype and GD. The application of the NGS method to genotype both MCH I and II classes in large groups of patients and controls has clarified many discrepancies present in the previous results possibly due to a lack of allelic specificity and/or the size of the groups being too small. A significant association was found between the risk of GD and the following alleles: HLA-B*08:01, -B*39:06, -B*37:01, -C*07:01, -C*14:02, -C*03:02, -C*17:01, -DRB1*03:01, -DRB1*11:01, -DRB1*13:03, -DRB1*01:03, -DRB1*14:01, -DQB1*03:01, DQB1*02:01. Among these alleles, -B*39:06, -B*37:01, -C*14:02, -C*03:02, -C*17:01, -DRB1*14:01 are novel, previously not-reported, independent alleles with no known linkage disequilibrium with other high-risk alleles. On the other hand, the frequencies of HLA-B*07:02, -C*07:02, -C*03:04, -DRB1*07:01, -DQB1*02:02, -DQB1*03:03 were significantly lower in GD compared to controls, with the first three alleles being reported as protective for the first time. These alleles can be considered protective. This study has provided a novel set of alleles as a reliable tool for individual risk assessment. The identification of alleles which in a particular population are associated with GD and which play a protective role is an essential step in the development of personalized medicine. Author Contributions Conceptualization, K.Z.-S. and M.S.; data curation, K.Z.-S. and M.S.; formal analysis, K.Z.-S., B.S. and M.S.; funding acquisition, A.L. and M.S.; investigation, K.Z.-S. and B.T.; methodology, K.Z.-S., B.S., B.T. and M.S.; project administration, M.S. and A.L. resources, K.Z.-S., B.S. and B.T.; software, B.T. and B.S. supervision, A.L. and M.S.; validation, B.S. and M.S. visualization, K.Z.-S., B.T. and M.S.; writing—original draft, K.Z.-S., B.S. and M.S.; writing—review and editing, A.L. and M.S. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Ethics Committee of the Polish Mother’s Memorial hospital—Research Institute, Łódź, Poland (protocol code 108/2018 date of approval 18 December 2018). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the patients to publish this paper. Data Availability Statement The source data are available on reasonable request from the corresponding author. Conflicts of Interest The authors declare no conflict of interest. Abbreviations AITD autoimmune thyroid diseases ATA American Thyroid Association CD40 cluster of differentiation 40 CTLA-4 cytotoxic T lymphocyte-associated factor 4 EDTA ethylenediaminetetraacetic acid (anticoagulant) FT3 free triiodothyronine FT4 free thyroxine GD Graves’ disease HLA human leukocyte antigens MHC major histocompatibility complex NGS next-generation sequencing SAT subacute thyroiditis Tg thyroglobulin TRAb TSH-receptor antibodies TSH thyroid stimulating hormone (thyrotropin) US ultrasound Figure 1 Frequencies (%) of human leukocyte antigen (HLA) over-represented alleles with statistically significant difference between control (open bars) and Graves’ disease (GD) patients (solid bars) for major histocompatibility complex (MHC) class I alleles: HLA-B (a) and HLA-C (b). Figure 2 Frequencies (%) of human leukocyte antigen (HLA_ over-represented alleles with statistically significant difference between control (open bars) and Graves’ disease (GD) patients (solid bars) for major histocompatibility complex (MHC) class II alleles: HLA-DRB1 (a) and HLA-DQB1 (b). Figure 3 Frequencies (%) of human leukocyte antigen (HLA) under-represented alleles with statistical difference between controls (open bars) and Graves’ disease (GD) patients (solid bars) for both major histocompatibility complex (MHC) class I and class II alleles. jcm-11-02492-t001_Table 1 Table 1 Laboratory characteristics of the Graves’ disease (GD) group. Parameter (Reference Range and Units) Mean ± SD Median TSH (0.27–4.2 µIU/mL) 0.14 ± 0.43 0.05 FT4 (0.9–1.7 ng/dL) 3.35 ± 2.39 2.33 FT3 (2.0–4.4 pg/mL) 11.07 ± 8.38 7.86 TRAb (<1.7 IU/L) 15.04 ± 13.62 10.12 Abbreviations: FT3, free triiodothyronine; FT4, free thyroxine; SD, standard deviation; TRAb, TSH receptor antibodies; TSH, thyrotropin. jcm-11-02492-t002_Table 2 Table 2 Frequencies and linkage disequilibrium of three-locus HLA-B-DRB1-DQB1 haplotypes in Graves’ disease (GD) patients depending on the presence of the HLA-B*08:01 allele. HLA Haplotype Haplotype Frequency B*08:01- DRB1*03:01- DQB1*02:01 22% [n = 36] B*XX:XX- DRB1*03:01- DQB1*02:01 6.6% [n = 11] B*08:01- DRB1*XX:XX- DQB1*XX:XX 1.9% [n = 3] B*XX:XX—allele other than -B*08:01; DRB1*XX:XX- DQB1*XX:XX—alleles other than -DRB1*03:01,- DQB1*02:01. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Ross D.S. Burch H.B. Cooper D.S. Greenlee M.C. Laurberg P. Maia A.L. Rivkees S.A. Samuels M. Sosa J.A. Stan M.N. 2016 American Thyroid Association Guidelines for Diagnosis and Management of Hyperthyroidism and Other Causes of Thyrotoxicosis Thyroid 2016 26 1343 1421 10.1089/thy.2016.0229 27521067 2. Vita R. Lapa D. Trimarchi F. Vita G. Fallahi P. Antonelli A. Benvenga S. Certain HLA alleles are associated with stress-triggered Graves’ disease and influence its course Endocrine 2017 55 93 100 10.1007/s12020-016-0909-6 26951052 3. Heward J.M. Allahabadia A. Daykin J. Carr-Smith J. Daly A. Armitage M. Dodson P.M. Sheppard M.C. Barnett A.H. Franklyn J.A. Linkage disequilibrium between the human leukocyte antigen class II region of the major histocompatibility complex and Graves’ disease: Replication using a population case control and family-based study J. Clin. Endocrinol. Metab. 1998 83 3394 3397 10.1210/jcem.83.10.5137 9768636 4. Shin D.H. Baek I.C. Kim H.J. Choi E.J. Ahn M. Jung M.H. Suh B.K. Cho W.K. Kim T.G. HLA alleles, especially amino-acid signatures of HLA-DPB1, might contribute to the molecular pathogenesis of early-onset autoimmune thyroid disease PLoS ONE 2019 14 e0216941 10.1371/journal.pone.0216941 31091281 5. Gu L.Q. Zhu W. Zhao S.X. Zhao L. Zhang M.J. Cui B. Song H.D. Ning G. Zhao Y.J. Clinical associations of the genetic variants of CTLA-4, Tg, TSHR, PTPN22, PTPN12 and FCRL3 in patients with Graves’ disease Clin. Endocrinol. 2010 72 248 255 10.1111/j.1365-2265.2009.03617.x 6. Wang P.W. Chen I.Y. Juo S.H. Hsi E. Liu R.T. Hsieh C.J. Genotype and phenotype predictors of relapse of graves’ disease after antithyroid drug withdrawal Eur. Thyroid J. 2013 1 251 258 10.1159/000342621 24783027 7. Li Y. Yao Y. Yang M. Shi L. Li X. Yang Y. Zhang Y. Xiao C. Association between HLA-B*46 allele and Graves disease in Asian populations: A meta-analysis Int. J. Med. Sci. 2013 10 164 170 10.7150/ijms.5158 23329888 8. Hawkins B.R. Ma J.T. Lam K.S. Wang C.C. Yeung R.T. Association of HLA antigens with thyrotoxic Graves’ disease and periodic paralysis in Hong Kong Chinese Clin. Endocrinol. 1985 23 245 252 10.1111/j.1365-2265.1985.tb00220.x 9. Naito S. Sasaki H. Arakawa K. Japanese Graves’ disease: Association with HLA-Bw46 Endocrinol. Jpn. 1987 34 685 688 10.1507/endocrj1954.34.685 3436299 10. Yeo P.P. Chan S.H. Thai A.C. Ng W.Y. Lui K.F. Wee G.B. Tan S.H. Lee B.W. Wong H.B. Cheah J.S. HLA Bw46 and DR9 associations in Graves’ disease of Chinese patients are age- and sex-related Tissue Antigens 1989 34 179 184 10.1111/j.1399-0039.1989.tb01734.x 2595722 11. Barlow A.B. Wheatcroft N. Watson P. Weetman A.P. Association of HLA-DQA1*0501 with Graves’ disease in English Caucasian men and women Clin. Endocrinol. 1996 44 73 77 10.1046/j.1365-2265.1996.634454.x 8706297 12. Zamani M. Spaepen M. Bex M. Bouillon R. Cassiman J.J. Primary role of the HLA class II DRB1*0301 allele in Graves disease Am. J. Med. Genet. 2000 95 432 437 10.1002/1096-8628(20001218)95:5<432::AID-AJMG5>3.0.CO;2-7 11146462 13. Inoue D. Sato K. Enomoto T. Sugawa H. Maeda M. Inoko H. Tsuji K. Mori T. Imura H. Correlation of HLA types and clinical findings in Japanese patients with hyperthyroid Graves’ disease: Evidence indicating the existence of four subpopulations Clin. Endocrinol. 1992 36 75 82 10.1111/j.1365-2265.1992.tb02905.x 1559302 14. Chan S.H. Lin Y.N. Wee G.B. Ren E.C. Lui K.F. Cheah J.S. Human leucocyte antigen DNA typing in Singaporean Chinese patients with Graves’ disease Ann. Acad. Med. Singap. 1993 22 576 579 8257062 15. Cavan D.A. Penny M.A. Jacobs K.H. Kelly M.A. Jenkins D. Mijovic C. Chow C. Cockram C.S. Hawkins B.R. Barnett A.H. The HLA association with Graves’ disease is sex-specific in Hong Kong Chinese subjects Clin. Endocrinol. 1994 40 63 66 10.1111/j.1365-2265.1994.tb02444.x 16. Dong R.P. Kimura A. Okubo R. Shinagawa H. Tamai H. Nishimura Y. Sasazuki T. HLA-A and DPB1 loci confer susceptibility to Graves’ disease Hum. Immunol. 1992 35 165 172 10.1016/0198-8859(92)90101-R 1363421 17. Huang S.M. Wu T.J. Lee T.D. Yang E.K. Shaw C.K. Yeh C.C. The association of HLA -A, -B, and -DRB1 genotypes with Graves’ disease in Taiwanese people Tissue Antigens 2003 61 154 158 10.1034/j.1399-0039.2003.00016.x 12694583 18. Chen P.L. Fann C.S. Chu C.C. Chang C.C. Chang S.W. Hsieh H.Y. Lin M. Yang W.S. Chang T.C. Comprehensive genotyping in two homogeneous Graves’ disease samples reveals major and novel HLA association alleles PLoS ONE 2011 6 e16635 10.1371/journal.pone.0016635 21307958 19. Yanagawa T. Mangklabruks A. Chang Y.B. Okamoto Y. Fisfalen M.E. Curran P.G. DeGroot L.J. Human histocompatibility leukocyte antigen-DQA1*0501 allele associated with genetic susceptibility to Graves’ disease in a Caucasian population J. Clin. Endocrinol. Metab. 1993 76 1569 1574 10.1210/jcem.76.6.8501164 8501164 20. Mangklabruks A. Cox N. DeGroot L.J. Genetic factors in autoimmune thyroid disease analyzed by restriction fragment length polymorphisms of candidate genes J. Clin. Endocrinol. Metab. 1991 73 236 244 10.1210/jcem-73-2-236 1677360 21. Stasiak M. Tymoniuk B. Michalak R. Stasiak B. Kowalski M.L. Lewinski A. Subacute Thyroiditis is Associated with HLA-B*18:01, -DRB1*01 and -C*04:01-The Significance of the New Molecular Background J. Clin. Med. 2020 9 534 10.3390/jcm9020534 32079059 22. Stasiak M. Tymoniuk B. Adamczewski Z. Stasiak B. Lewinski A. Sonographic Pattern of Subacute Thyroiditis Is HLA-Dependent Front. Endocrinol. 2019 10 3 10.3389/fendo.2019.00003 30728805 23. Stasiak M. Tymoniuk B. Stasiak B. Lewinski A. The Risk of Recurrence of Subacute Thyroiditis Is HLA-Dependent Int. J. Mol. Sci. 2019 20 1089 10.3390/ijms20051089 30832406 24. Stasiak M. Lewinski A. Strong Correlation between HLA and Clinical Course of Subacute Thyroiditis—A Report of the Three Siblings Genes 2020 11 1282 10.3390/genes11111282 25. Stasiak M. Zawadzka-Starczewska K. Lewinski A. Clinical Manifestation of Subacute Thyroiditis Triggered by SARS-CoV-2 Infection Can Be HLA-Dependent Viruses 2021 13 2447 10.3390/v13122447 34960716 26. Mayor N.P. Hayhurst J.D. Turner T.R. Szydlo R.M. Shaw B.E. Bultitude W.P. Sayno J.R. Tavarozzi F. Latham K. Anthias C. Recipients Receiving Better HLA-Matched Hematopoietic Cell Transplantation Grafts, Uncovered by a Novel HLA Typing Method, Have Superior Survival: A Retrospective Study Biol. Blood Marrow Transplant. 2019 25 443 450 10.1016/j.bbmt.2018.12.768 30935664 27. Khan M.A. Polymorphism of HLA-B27: 105 subtypes currently known Curr. Rheumatol. Rep. 2013 15 362 10.1007/s11926-013-0362-y 23990399 28. Ahmad T. Neville M. Marshall S.E. Armuzzi A. Mulcahy-Hawes K. Crawshaw J. Sato H. Ling K.L. Barnardo M. Goldthorpe S. Haplotype-specific linkage disequilibrium patterns define the genetic topography of the human MHC Hum. Mol. Genet. 2003 12 647 656 10.1093/hmg/ddg066 12620970 29. DR/DQ Associations Available online: http://www.ctht.info/Table%2013%20DRB1%20DQA1%20DQB1%20associations%20in%20various%20populations.pdf (accessed on 14 March 2022) 30. Płoski R. Szymański K. Bednarczuk T. The genetic basis of graves’ disease Curr Genom. 2011 12 542 563 10.2174/138920211798120772 31. Martin S. Dutescu M.I. Sirbu A. Barbu C. Albu A. Florea S. Fica S. The clinical value of human leukocyte antigen HLA-DRB1 subtypes associated to Graves’ disease in Romanian population Immunol Investig. 2014 43 479 490 10.3109/08820139.2014.886261 24661105 32. Common Associations of HLA-C Alleles with Alleles of HLA-B Available online: http://www.ctht.info/Table%209%20CB%20ASSOCIATIONS.pdf (accessed on 14 March 2022) 33. Common Associations of HLA-B Alleles with Alleles of HLA-C Available online: http://www.ctht.info/Table%208%20BC%20ASSOCIATIONS.pdf (accessed on 14 March 2022) 34. Simmonds M.J. Howson J.M. Heward J.M. Carr-Smith J. Franklyn J.A. Todd J.A. Gough S.C. A novel and major association of HLA-C in Graves’ disease that eclipses the classical HLA-DRB1 effect Hum. Mol. Genet. 2007 16 2149 2153 10.1093/hmg/ddm165 17597093 35. Chen Q.Y. Huang W. She J.X. Baxter F. Volpe R. Maclaren N.K. HLA-DRB1*08, DRB1*03/DRB3*0101, and DRB3*0202 are susceptibility genes for Graves’ disease in North American Caucasians, whereas DRB1*07 is protective J. Clin. Endocrinol. Metab. 1999 84 3182 3186 10.1210/jcem.84.9.5991 10487684 36. Desai M.P. Karandikar S. Autoimmune thyroid disease in childhood: A study of children and their families Indian Pediatr. 1999 36 659 668 10740302 37. Ludgate M. Emerson C.H. Metamorphic thyroid autoimmunity Thyroid 2008 18 1035 1037 10.1089/thy.2008.1551 18844474 38. Aversa T. Lombardo F. Corrias A. Salerno M. De Luca F. Wasniewska M. In young patients with Turner or Down syndrome, Graves’ disease presentation is often preceded by Hashimoto’s thyroiditis Thyroid 2014 24 744 747 10.1089/thy.2013.0452 24325626 39. Wasniewska M. Corrias A. Arrigo T. Lombardo F. Salerno M. Mussa A. Vigone M.C. De Luca F. Frequency of Hashimoto’s thyroiditis antecedents in the history of children and adolescents with graves’ disease Horm. Res. Paediatr. 2010 73 473 476 10.1159/000313395 20395669 40. Yin X. Latif R. Bahn R. Davies T.F. Genetic profiling in Graves’ disease: Further evidence for lack of a distinct genetic contribution to Graves’ ophthalmopathy Thyroid 2012 22 730 736 10.1089/thy.2012.0007 22663548 41. Mehraji Z. Farazmand A. Esteghamati A. Noshad S. Sadr M. Amirzargar S. Yekaninejad M.S. Amirzargar A. Association of Human Leukocyte Antigens Class I and II with Graves’ Disease in Iranian Population Iran. J. Immunol. 2017 14 223 230 28919585 42. Park M.H. Park Y.J. Song E.Y. Park H. Kim T.Y. Park D.J. Park K.S. Cho B.Y. Association of HLA-DR and -DQ genes with Graves disease in Koreans Hum. Immunol. 2005 66 741 747 10.1016/j.humimm.2005.03.001 15993720 43. Wongsurawat T. Nakkuntod J. Charoenwongse P. Snabboon T. Sridama V. Hirankarn N. The association between HLA class II haplotype with Graves’ disease in Thai population Tissue Antigens 2006 67 79 83 10.1111/j.1399-0039.2005.00498.x 16451208

PMC009xxxxxx/PMC9099648.txt

==== Front Int J Environ Res Public Health Int J Environ Res Public Health ijerph International Journal of Environmental Research and Public Health 1661-7827 1660-4601 MDPI 10.3390/ijerph19095313 ijerph-19-05313 Systematic Review Quantitative Methods to Detect Suicide and Self-Harm Clusters: A Systematic Review Benson Ruth 12* Rigby Jan 3 Brunsdon Christopher 3 Cully Grace 12 Too Lay San 4 Arensman Ella 125 Tchounwou Paul B. Academic Editor 1 School of Public Health, College of Medicine and Health, University College Cork, Western Gateway Building, T12 XF62 Cork, Ireland; grace.cully@ucc.ie (G.C.); ella.arensman@ucc.ie (E.A.) 2 National Suicide Research Foundation, University College Cork, 4.28 Western Gateway Building, T12 XF62 Cork, Ireland 3 National Centre for Geocomputation, Maynooth University, W23 F2H6 Maynooth, Ireland; jan.rigby@mu.ie (J.R.); christopher.brunsdon@mu.ie (C.B.) 4 Centre for Mental Health, Melbourne School of Population and Global Health, The University of Melbourne, Melbourne, VIC 3053, Australia; tiffany.too@unimelb.edu.au 5 Australian Institute for Suicide Research and Prevention, School of Applied Psychology, Griffith University, Brisbane, QLD 4122, Australia * Correspondence: ruth.benson@ucc.ie 27 4 2022 5 2022 19 9 531329 3 2022 23 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Suicide and self-harm clusters exist in various forms, including point, mass, and echo clusters. The early identification of clusters is important to mitigate contagion and allocate timely interventions. A systematic review was conducted to synthesize existing evidence of quantitative analyses of suicide and self-harm clusters. Electronic databases including Medline, Embase, Web of Science, and Scopus were searched from date of inception to December 2020 for studies that statistically analyzed the presence of suicide or self-harm clusters. Extracted data were narratively synthesized due to heterogeneity among the statistical methods applied. Of 7268 identified studies, 79 were eligible for narrative synthesis. Most studies quantitatively verified the presence of suicide and self-harm clusters based on the scale of the data and type of cluster. A Poisson-based scan statistical model was found to be effective in accurately detecting point and echo clusters. Mass clusters are typically detected by a time-series regression model, although limitations exist. Recently, the statistical analysis of suicide and self-harm clusters has progressed due to advances in quantitative methods and geospatial analytical techniques, most notably spatial scanning software. The application of such techniques to real-time surveillance data could effectively detect emerging clusters and provide timely intervention. systematic review suicide self-harm cluster detection contagion geospatial analysis Irish Health Research BoardIRRL-2015-1586 National Health and Medical Research Council Early Career FellowshipGNT1156849 This research was funded by Irish Health Research Board, grant number IRRL-2015-1586. L.S.T. was supported by a National Health and Medical Research Council Early Career Fellowship, grant number GNT1156849. ==== Body pmc1. Introduction Suicide clusters are commonly referred to as a higher number of suicide deaths, attempted suicides, or self-harm events that occur in a population, location, or period than usually expected, based on statistical probability or community expectancy [1]; however, due to a lack of consensus regarding an operational definition of suicide clusters, particularly relating to the minimal number of cases that constitute a cluster, definitions are typically determined on an ad hoc basis in terms of the spatial and temporal limits of a cluster [2]. Clusters are suggested to be a product of a phenomenon known as contagion, whereby direct or indirect exposure to suicide results in subsequent suicide cases [3]. Suicide clusters are mostly reported within the adolescent population, particularly 15–24-year-olds [4,5] and are estimated to account for between 1% and 5% of all adolescent deaths by suicide [6,7]. Previous evidence suggests that an increase in the incidence of suicide clusters in recent years is linked to the broadening of social connections through electronic communication systems and internet-based social sites [8], particularly in the form of suicide pacts [9]. Geographical remoteness, economic deprivation, and indigenous status are factors associated with suicide clusters [10,11,12]. Furthermore, suicide clusters are more likely to occur in areas inhabited by disadvantaged cohorts, as certain risk factors associated with suicide including unemployment, socio-economic deprivation, and substance abuse occur more often in this population [13]. The patterns of suicide and self-harm mainly researched and documented within the literature are of two main types: mass clusters and point clusters. Mass clusters (or temporal clusters) involve a temporary increase in the total number of suicides within a population relative to the period before and after the cluster, with a lack of spatial relevance typically observed in the aftermath of a real or fictional suicide documented in the media [9,10]. In contrast, point clusters (or spatiotemporal clusters) are those that occur close together in both space and time within a given community or institution, and clusters of this nature can occur without the presence of media coverage [11]. A third pattern of suicide (spatial clusters) has been identified in the literature [2,3,4,5,6,7,8,9,10,11,12], wherein deaths cluster by location but not time, and are known as ‘locations where people frequently take their lives’, often occurring at well-known public or historical sites; however, this pattern is not as extensively researched compared with mass and point clusters. A phenomenon known as echo clusters, wherein one or more successive suicide cluster occurs at a distinct point away from the initial cluster, has been statistically verified in indigenous populations in rural Australia, but there is a dearth of evidence of this phenomenon elsewhere [14,15]. In recent years, an increasing number of studies have addressed the identification and detection of suicide clusters at both a national and local level [16,17,18]. The detection of clusters enhances the knowledge on the aetiology of emerging suicide clustering by establishing links between confirmed or suspected suicide cases and identifying socioecological factors associated with the increased risk of clusters within the affected area or population [19]. Policy makers and public health officials also benefit from early detection of suicide clusters by means of implementing targeted and timely interventions. Significant advances in spatial cluster detection have emerged in recent decades with the development of computer mapping and its integration with robust statistical models [19]. Previous studies that investigated the presence of suicide clusters have applied different techniques that follow frequentist and Bayesian probability models, and they have incorporated spatial scanning software [16,17,18,19,20]. Nonetheless, a standardized and systematic approach to the statistical ascertainment of suicide and self-harm clusters is still lacking in contemporary research. To date, no systematic review of the quantitative methods that effectively detect suicide and self-harm clusters has been conducted. The main aim of this systematic review is to synthesize the existing evidence based on statistical techniques used in successfully detecting suicide and self-harm clusters. In this regard, this review seeks to determine an accurate and precise approach to quantitatively verify suicide and self-harm clusters within a population, and to ensure that clusters of suicide and self-harm are detected in a timely manner, hence mitigating further cases. 2. Materials and Methods In accordance with the PRISMA guidelines [21], a comprehensive search strategy was established, including MeSH terms where relevant (see supplementary material for completed PRISMA checklist). The review was registered with The International Prospective Register of Systematic Reviews (PROSPERO, registration number CRD42018100354) to avoid duplication. The search strategy was applied to four bibliography databases: Medline, Embase, Web of Science, and Scopus from their inception to August 2018, to identify as much relevant literature as possible. The lead author conducted an updated search, applying the same search strategy in December 2020. The search terms included ((suicide (MeSH) OR suicid*) OR (self-injurious behaviour (MeSH) OR (self-injur* OR self-poison* OR self-mutilat* OR self-harm*)] AND [(cluster* OR imitat* OR contagion OR copycat OR werther effect)] OR (spatiotemporal analysis OR time-space analysis OR geospatial analysis OR statistical analysis*)). Inclusion criteria included studies that (a) have been published in a scholarly journal, (b) have applied a statistical method to detect suicide or self-harm clusters in a population, and (c) have the full-text available in English. Exclusion criteria eliminated (a) narrative reports of suicide or self-harm clustering that were not statistically verified, (b) grey literature including media reports relating to potential suicide or self-harm clusters, and (c) non-English language articles. The title and abstracts of all references generated by the search were screened for relevance by three authors (RB, GC, LST) to avoid content bias. For those articles of which full texts were not available, the full text was requested from the lead author. Additional hand searches of reference lists of relevant systematic reviews were also conducted to identify other eligible studies. Only published scholarly articles were included to obtain the most robust methodological approach possible. Data extraction in table format was used to summarize study results. A meta-analysis was not considered due to the heterogeneity of statistical methodology applied in the included studies; hence, the data was narratively synthesized as a result. Subgroup analysis was conducted on four study groups based on commonalities in cluster type identified during preliminary analysis. For the purpose of the current research and to avoid misinterpretation, suicide clusters will henceforth refer to clusters of death by suicide, whereas self-harm clusters will describe the clustering of self-harm events including attempted suicide. 3. Results The electronic searches identified 7246 publications, excluding duplicates. Based on the screening of titles and abstracts generated by the database searches, 295 potentially relevant publications were identified. Of those publications selected for full text screening, 216 did not meet the eligibility criteria, resulting in 79 relevant articles applicable for review (Figure 1; full details of all relevant articles included as supplementary material). The relevant studies were sub-divided, based on their primary focus, into point suicide clusters (n = 51), point self-harm clusters (n = 8), mass suicide clusters (n = 19), and echo suicide clusters (n = 1). The literature in this area predominately originates from the Oceania continent, Europe, and the Americas. Considerably less research on the topic has been published in Asia and in the African region to date. In terms of the level of geographic samples analyzed, approximately half of all studies were based on a national sample (n = 39), almost a third involved a regional sample (n = 25), and the remaining studies focused on state, city and investigations into locations associated with frequently occurring suicides or self-harm acts. The statistical analysis of point suicide and self-harm clusters commenced in 1975, with over two thirds of studies published in the last 5 years (n = 35). Although mass suicide cluster statistical detection was first documented within the literature in 1986, almost two thirds of studies have been conducted within the last 5 years (n = 12). The majority of identified publications (n = 51) primarily focused on the statistical analysis of point suicide clusters (Table 1). Those studies with alternative primary objectives predominantly evaluated clusters in the context of their association with demographics, socio-economic factors, cultural variables, and risk factors associated with suicide clusters [7,22,23,24]. A Poisson model incorporating the Monte Carlo simulation is the most widely applied statistical model, applied in half of all point suicide cluster detection studies (n = 28) [16,17,22,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49]. This test determines the significance of likely clusters, most commonly at a 5% spatial parameter and three-month temporal parameter, as well as detecting the relative risk of cluster. Geospatial analysis was conducted in three quarters of point suicide cluster detection studies (75%, n = 38), by means of spatial scanning using SaTSCan and FleXScan software, as well as hotspot analysis, spherical trigonometry, and Nearest Neighbor Analysis Exploratory Spatial Data Analysis methods [16,17,18,22,23,25,26,27,28,29,30,31,32,36,37,38,40,41,42,43,44,45,46,47,50,51,52,53,54,55]. ArcGIS was the most predominantly employed data visualization tool, whereas a small number of studies utilized alternate tools such as QGis, GeoDa and Teraview. A minority of studies applied alternative statistical analyses, including a regression model (n = 10) [22,26,29,38,43,44,50,56,57,58,59], the Knox procedure (n = 1) [7], descriptive network analysis (n = 1) [47], Bayesian hierarchical modelling (n = 2) [24,25,48,49,59,60], the Kernel Density estimator (n = 2) [51,58], Nearest Neighbor Analysis (n = 1) [50], Ripley’s K function (n = 1) [52], Moran’s I (n = 4) [50,59,60,61], chi square (n = 5) [50,51,55,62,63], Fisher’s exact test (n = 1) [64], and the Anderson Darling test (n = 1) [65]. Specific numerical details of detected clusters were undocumented in the results of those studies that applied alternative statistical techniques without the application of geospatial analysis. Comparing different statistical methods, a discrete Poisson-based scan statistical approach will capture most parameters to identify point suicide clusters, since suicide mortality data tends to be in a Poisson distribution. A small number of studies (n = 8) focused on the detection of self-harm clusters within populations (Table 2). Almost a third of self-harm cluster detection studies were based on national samples (n = 3), with the remainder focusing on cluster detection at regional (n = 2), county (n = 1), and city levels (n = 2). Most studies reported a significant detection of self-harm clusters (n = 7), with over half of the studies (n = 5) indicating the specific number of self-harm clusters detected within the population, ranging from one to twenty-five clusters. Scan statistics were applied in over half of all self-harm cluster detection studies [66,67,68,69], with an alternative temporal scanning method applied in one investigation [70]. Those studies that excluded geospatial techniques from the statistical analysis applied a regression model or chi-squared test [71,72,73]; however, detailed information relating to identified clusters was not explicated from such analyses. Based on a comparison of the statistical methods applied, a regression-based scan statistical model will capture most parameters to detect point self-harm clusters. Within the identified studies, over one third (35%, n = 19) reported on mass suicide clusters (Table 3). Over two-thirds of mass suicide cluster research were based on national samples (68%, n = 13), relating to high-profile suicides reported within the media in their countries. The remaining studies investigated mass clustering with regional (n = 4), provincial (n = 1), and continental samples (n = 1). The primary aim of all the identified studies was the statistical verification of increased suicides within a population (i.e., the detection of mass clusters). The most employed statistical analyses include a time-series model such as the Seasonal Autoregressive Integrated Moving Average (SARIMA) model (n = 11) [74,75,76,77,78,79,80,81,82,83], a regression model (n = 8) [82,83,84,85,86,87,88,89], a Poisson model (n = 4) [84,87,88,89], and non-parametric tests (n = 3) [87,90,91]. When comparing statistical models to detect mass clusters, a time-series regression model will capture the parameters of mass clustering as accurately as possible, based on temporal data. One study based on the statistical analysis of echo clusters, conducted in Australia, was identified within the literature [92]. The application of a Poisson scan statistic method to data based on the same geographical area but from two different periods, effectively detected several clusters in each period. Although there are no additional studies of this kind to compare this methodological approach against, the identified literature applies the same methodology as point suicide clusters with an additional time dimension. 4. Discussion This systematic review provides unique insights into the scope of quantitative methods used to detect suicide and self-harm clusters. The findings of this review indicate that quantitative analysis of suicide and self-harm clusters continues to advance, in line with enhancements in statistical models of verification and spatial scanning methods. Developments in geographical cluster detection have coincided with a greater availability of spatial data [93]. Open-source Geographical Information System (GIS) software was applied in all but one identified study of spatial suicide clusters, offering strengths including cost-effectiveness, reproducibility, online support forums, and tutorials [94]. As corroborated by the results of the current review, the quality of geographical data captured in a GIS database is crucially important for geospatial analysis and depends on positional and attribute accuracy (e.g., latitude and longitude coordinates and health outcome), as well as completeness of data. An awareness of the specific criteria for what constitutes suicide and self-harm acts, the importance of data completeness, and the precision required in the measurement of geographical coordinates are all critical components of accurate data recording, and in turn, accurate cluster detection [95]. The vast majority of research involved retrospective ecological studies of suicide or self-harm clusters based on aggregated geographical, mortality, and census data. The implementation of active surveillance involving proactive contact with data providers to access, record, and complete, accurate and timely public health data, including geographical identifiers [95,96], is recommended to enhance the precision of cluster detection. To date, probabilistic model-based spatial scan statistics are the most widely applied and reliable methods employed in the detection of point suicide and self-harm clusters. Despite a dearth of literature within the area, research investigating echo clusters of suicide has followed the same quantitative methodology as point cluster detection (i.e., a Poisson based spatial scan statistic), integrating an additional time dimension to account for analysis of at least two different time periods. The Poisson approach models how many times the event is likely to occur within a specific period, whereas the Monte Carlo simulation is used to evaluate the statistical significance of the likelihood ratio for each circle. Based on the significance test, the scan statistic can identify the most likely cluster, as well as secondary clusters, for which the likelihood ratios are less, but are still of importance [97]. SaTScan, which is a type of software using a cylindrical scan statistic involving a moving circular geographical-based scan window and a time-based height dimension of continuously varying radii, appears to be the most used scan approach within the reviewed literature [20]. This tool evaluates the statistical significance of point clusters with no prior assumptions of the data. Although this software has been extensively applied within epidemiological studies, a limitation of SaTScan is its inability to detect non-circular shaped clusters or hotspots, such as the shapes of roads or rivers [98]. To detect irregularly shaped clusters, alternative approaches have been proposed and applied within the reviewed research [98,99,100]. FleXScan [53], based on an adjustable spatial scan window, is effective in detecting clusters that assume arbitrary shapes [100,101,102,103]; however, the efficacy of this software is limited to the detection of small to moderate clusters of approximately 30 cases [53]. Echelon scanning using EcheScan, also identified within the literature, is used to detect non-circular shaped hotspots based on their spatial hierarchal structure, visually represented by a dendrogram that is scanned from top to bottom [97,102]. Similar to the traditional spatial scan statistic, echelon scanning is based on the Poisson model with Monte Carlo simulation; however, the scan window is smaller. EcheScan software, developed in R, is easily accessible and incorporates open-source mapping tools; however, limitations exist in some instances wherein the shape of the detected hotspot may be too complex, or too large, to be easily interpreted [103]. The results of this review suggest that analysis window parameters of scan statistic algorithms should be manipulated to determine the appropriate population and duration thresholds, calibrating the optimal parameter combination, since the precision of results can be affected by scale. Future research should seek to compare the performance of the scan statistic algorithms via a simulation study and examine the spatial congruence and sensitivity of the models. Based on the unique purposes of the scan statistics, the robustness and sensitivity of a Poisson-based spatial scan hybrid approach should also be explored by future research. Mass cluster detection fundamentally concerns itself with an increase in cases during a specific period, irrespective of spatial relevance. Quasi-experimental research designs, such as time-series forecasting based on a regression model, measure how many future observations are predictable based on past behavior [85,104]. In mass cluster detection, media coverage of a fictional or real high-profile suicide is correlated with an increase in cases of suicide during the aftermath of the suicide, by means of comparing frequencies of suicide in an experimental time frame during and after the death was reported, against the frequency of suicide in a control period. Such studies involve a crucial limitation that must be considered when interpreting findings; that is, the difficulty to accept observed increases in suicide and self-harm rates in terms of being a direct link to the high-profile case with absolute confidence. 4.1. Strengths and Limitations This review sought to identify and synthesize literature relating to suicide and self-harm cluster detection, demonstrating inclusivity in systematically reviewing all published studies to date, and addressing all types of suicide and self-harm clusters. The primary focus of the review was to examine the most robust global evidence using statistical methods to detect suicide and self-harm clusters within a population as accurately as possible; therefore, non-peer reviewed reports have been excluded from the synthesis, which may limit the results. Due to study heterogeneity arising from methodological diversity, a full quality appraisal was not carried out, hence, possible biases must be considered in the context of limitations. Excluding non-English studies has not limited the review since most research in this area has been conducted in English speaking countries. 4.2. Implications for Suicide Prevention and Considerations for Future Research The findings of this review have implications for suicide prevention. More specifically, this review has synthesized all empirical studies of suicide and self-harm clusters in a population, arriving at the most comprehensive standardized approach to suicide and self-harm cluster detection currently available, in the absence of a gold-standard method. Innovatively, the conclusive approach of geospatial probabilistic modelling for point suicide cluster detection has been incorporated in the development and evaluation of a community response to a suicide cluster, demonstrating the utility of this technique for suicide prevention purposes [44]. The comprehensive study identified in the review applied spatiotemporal analysis to suicide mortality data and socioeconomic aggregated data by way of identifying suicide clusters and spatial variations of risk-factors in Hong Kong, for the purpose of informing the development of the targeted program, and evaluating its efficacy post-program, using changes in suicide incidence and cluster patterns as the outcome. The findings of the study emphasize the value of a temporal and spatial monitoring surveillance system based on the methodology described here in prioritizing suicide prevention measures. The outcome of the novel study further suggests a use for such techniques in the monitoring and evaluation of population-level interventions to be implemented as components in national suicide prevention strategies. Official suicide mortality records can take up to two years post-death to be released, due to delays resulting from prolonged medico-legal cause of death investigations, and late registered deaths [105]. The application of cluster detection methods identified in this review to provisional, real-time, suspected suicide data, would support the detection of emerging clusters, providing an advanced opportunity to effectively intervene and mitigate further contagion [106]. Early identification of emerging suspected clusters would also facilitate the acceleration of an evidence-based crisis response in vulnerable communities, wherein screening and referral of susceptible individuals to appropriate clinical and support services could occur in a timelier manner. Future research should consider the investigation of self-harm clusters and suicide clusters within a population, to determine whether clusters of self-harm precede clusters of suicide, thereby offering the opportunity for targeted clinical intervention in populations wherein emerging self-harm clusters are detected as a prevention strategy for possible subsequent suicide clustering. Real-time active surveillance of suicide and self-harm would facilitate prospective studies of suicide and self-harm clusters using prospective geospatial probabilistic modelling [107]. The findings of such prospective studies would subsequently inform suicide prevention strategies, action plans, policy planning, and service provision in a timely manner. Although unexplored in studies, including those in this review, temporal analysis of suicide data using a calendar approach based on date of death may detect temporal clusters relating to significant dates, such as the anniversary of the death of a loved one or high-profile individual, and seasonal trends when peaks are commonly observed. The detection of this phenomenon should be incorporated as a key objective of a real-time suicide surveillance system by way of indicating high-risk dates and periods that could require deployment of additional resources to respond to possible increases in imitative behavior. 5. Conclusions The synthesized results of this systematic review demonstrate advances made in epidemiological cluster detection, which is relevant to suicide and self-harm data, within the forty-five-year period since statistical investigations into clusters of suicide and self-harm were first published. Most notably, the evolvement of open-source GIS software, has effectively contributed to point cluster detection by means of geospatial probabilistic modelling. Mass suicide cluster detection traditionally employs a time-series regression analysis to verify temporal clustering within a population; however, the use of retrospective aggregated data in these studies compromises the accuracy and efficiency of cluster detection investigations. Acknowledgments We gratefully acknowledge the support of Joe Murphy, assistant librarian at the Mercy University Hospital Library, for his excellent advice that enhanced the search strategy applied. We also wish to extend our gratitude to John Browne who delivered a postgraduate module based on systematic reviews for the health sciences in University College Cork. The knowledge acquired from this training proved to be invaluable in conducting this study. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ijerph19095313/s1, Table S1: PRISMA 2009 Checklist; Table S2: Point Suicide Clusters; Table S3: Point Self-Harm Clusters; Table S4: Mass Clusters; Table S5: Echo Clusters. Click here for additional data file. Author Contributions Conceptualization, R.B., E.A. and J.R.; methodology, R.B., G.C. and L.S.T.; formal analysis, R.B.; investigation, R.B.; writing—original draft preparation, R.B.; writing—review and editing, R.B., J.R., C.B., G.C., L.S.T. and E.A., supervision, E.A., J.R. and C.B.; funding acquisition, E.A and L.S.T. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 PRISMA flow diagram. ijerph-19-05313-t001_Table 1 Table 1 Point suicide clusters. Number of Studies 51 Level of data used in the study Location City State Regional National 2 2 6 18 23 Type of analysis (Studies that performed multiple statistical analyses are counted twice.) Nearest Neighbour Statistic Kernel density estimator Spherical Trigonometry Descriptive network analysis Knox procedure Ripleys k function Chi-square Fishers Exact Test Morans I Bayesian hierarchical model Anderson Darling Regression model Poisson model 1 2 1 1 1 1 5 1 4 6 1 10 28 Geospatial analysis conducted No Yes 13 38 Clusters detected No Yes 3 48 Addressed analysed Not specified Location of death Residence 4 13 34 SaTScan spatial applied Not specified No Yes 7 24 20 Number of clusters reported Not specified 20+ 1–20 No clusters detected 7 3 28 3 ijerph-19-05313-t002_Table 2 Table 2 Point self-harm clusters. Number of Studies 8 Level Of Data Used in the Study National National Regional City Regional City National County Location of studies Sweden New Zealand New South Wales, Australia Edinburgh, Scotland Kwai Tsing, Hong Kong Hamadan, Iran Denmark Meru, Kenya Aggregated data used No No Yes No Yes No Yes No Type of statistical analysis conducted Logistical regression SaTScan and Knox Procedure SaTScan, Hotspot analysis (Getis-Ord Gi*) and ArcGIS for mapping A scan interval test proposed by Naus, 1966 Chi square and SaTScan Logistic regression and Chi-square, SaTScan and Monte Carlo simulation Multi-level regressions and log likelihood ratio tests Multiple logistic regression Geospatial analysis conducted No Yes Yes No Yes Yes No No Clusters detected Yes Yes Yes Yes Yes Yes No Yes Number of clusters reported Not specified Not specified Twenty-five spatial cluster regions identified 1 cluster Four spatial clusters, one spatiotemporal cluster 2 clusters N/A Not specified ijerph-19-05313-t003_Table 3 Table 3 Mass clusters. Number of Studies 19 Level of data used in the study Continental National Regional Provincial 1 13 4 1 Type of analysis conducted (Studies that performed multiple statistical analyses are counted twice.) Poisson model Regression analysis Non-parametric tests, i.e., Mann–Whitney U test, Kolmogorov-Smirnov test Time-series models, e.g., SARIMA 4 8 3 11 Geospatial analysis conducted Yes No 0 19 Mass cluster(s) detected Yes No 17 2 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. O’Carroll P.W. Mercy J.A. Steward J.A. CDC recommendations for a community plan for the prevention and containment of suicide clusters MMWR Supp. 1988 37 1 12 2. Niedzwiedz C. Haw C. Hawton K. Platt S. The definition and epidemiology of clusters of suicidal behavior: A systematic review Suicide Life Threat Behav. 2014 44 569 581 10.1111/sltb.12091 24702173 3. Bohanna I. Suicide “contagion”: What we know and what we need to find out Can. Med. Assoc. J. 2013 185 861 862 10.1503/cmaj.130678 23695602 4. Hazell P. Adolescent suicide clusters: Evidence, mechanisms, and prevention Aust. N. Z. J. Psychiatry 1993 27 653 665 10.3109/00048679309075828 8135690 5. Gould M.S. Wallenstein S. Kleinman M. Time-space clustering of teenage suicide Am. J. Epidemiol. 1990 131 71 78 10.1093/oxfordjournals.aje.a115487 2293755 6. Gould M. Wallenstein S. Kleinman N. A Study of Time-Space Clustering of Suicide: Final Report Centers for Disease Control and Prevention (CDC) Atlanta, GA, USA 1987 7. Gould M.S. Wallenstein S. Kleinman M.H. O’Carroll P. Mercy J. Suicide clusters: An examination of age-specific effects Am. J. Public Health 1990 80 211 212 10.2105/AJPH.80.2.211 2297071 8. Larkin G. Beautrais A. Geospatial Mapping of Suicide Clusters Te Pou o Te Whakaaro Nui Auckland, New Zealand 2012 9. Biddle L. Donovan J. Hawton K. Kapur N. Gunnell D. Suicide and the internet BMJ 2008 336 800 802 10.1136/bmj.39525.442674.AD 18403541 10. Haw C. Hawton K. Niedzwiedz C. Platt S. Suicide clusters: A review of risk factors and mechanisms Suicide Life Threat Behav. 2013 43 97 108 10.1111/j.1943-278X.2012.00130.x 23356785 11. Joiner T.E. Jr. The clustering and contagion of suicide Curr. Dir. Psychol. Sci. 1999 8 89 92 10.1111/1467-8721.00021 12. Cox G.R. Owens C. Robinson J. Nicholas A. Lockley A. Williamson M. Cheung Y.T.D. Pirkis J. Interventions to reduce suicides at suicide hotspots: A systematic review BMC Public Health 2013 13 214 10.1186/1471-2458-13-214 23496989 13. Hsu C.Y. Chang S.S. Yip P. Individual-, household-and neighbourhood-level characteristics associated with life satisfaction: A multilevel analysis of a population-based sample from Hong Kong Urban Stud. 2017 54 3700 3717 10.1177/0042098016677208 14. Hanssens L. ‘Echo Clusters’-Are they a Unique Phenomenon of Indigenous Attempted and Completed Suicide? Aborig. Isl. Health Work J. 2010 34 17 26 10.3316/ielapa.65107994950817915 15. Hanssens L. Suicide Echo Clusters: Are they Socially Determined, and the Result of a Pre-existing Vulnerability in Indigenous Communities in the Northern Territory? Aborig. Isl. Health Work J. 2011 35 14 16. Jones P. Gunnell D. Platt S. Scourfield J. Lloyd K. Huxley P. John A. Kamran B. Wells C. Dennis M. Identifying probable suicide clusters in Wales using national mortality data PLoS ONE 2013 8 e71713 10.1371/journal.pone.0071713 24015189 17. Bando D.H. Moreira R.S. Pereira J.C. Barrozo L.V. Spatial clusters of suicide in the municipality of Sao Paulo 1996–2005: An ecological study BMC Psychiatry 2012 12 124 10.1186/1471-244X-12-124 22913796 18. Larkin G. Beautrais A.L. Xu H. Geospatial mapping of suicide clusters in Cantebury New Zealand Inj. Prev. 2010 16 (Suppl. 1) A258 10.1136/ip.2010.029215.918 19. Mclafferty S. Disease cluster detection methods: Recent developments and public health implications Ann. GIS 2015 21 127 133 10.1080/19475683.2015.1008572 20. Kulldorff M. A spatial scan statistic Commun. Stat.-Theory Methods 1997 26 1481 1496 10.1080/03610929708831995 21. Moher D. Liberati A. Tetzlaff J. Altman D.G. Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement PLoS Med. 2009 6 e1000097 10.7326/0003-4819-151-4-200908180-00135 19621072 22. Bando D.H. Brunoni A.R. Bensenor I.M. Lotufo P.A. Suicide rates and income in Sao Paulo and Brazil: A temporal and spatial epidemiologic analysis from 1996 to 2008 BMC Psychiatry 2012 12 127 10.1186/1471-244X-12-127 22928689 23. Hill N.T. Spittal M.J. Pirkis J. Torok M. Robinson J. Risk factors associated with suicide clusters in Australian youth: Identifying who is at risk and the mechanisms associated with cluster membership EClinicalMedicine 2020 29 100631 10.1016/j.eclinm.2020.100631 33294825 24. Carcach C. A spatio-temporal analysis of suicide in El Salvador BMC Public Health 2017 17 339 10.1186/s12889-017-4251-6 28427363 25. Ngui A.N. Apparicio P. Moltchanova E. Vasiliadis H.M. Spatial analysis of suicide mortality in Québec: Spatial clustering and area factor correlates Psychiatry Res. 2014 220 20 30 10.1016/j.psychres.2014.07.033 25095757 26. Cheung Y.T.D. Spittal M.J. Williamson M.K. Tung S.J. Pirkis J. Application of scan statistics to detect suicide clusters in Australia PLoS ONE 2013 8 e54168 10.1371/journal.pone.0054168 23342098 27. Exeter D.J. Boyle P.J. Does young adult suicide cluster geographically in Scotland? J. Epidemiol. Community Health 2007 61 731 736 10.1136/jech.2006.052365 17630375 28. Johnson A.M. Woodside J.M. Johnson A. Pollack J.M. Spatial patterns and neighborhood characteristics of overall suicide clusters in Florida from 2001 to 2010 Am. J. Prev. Med. 2017 52 e1 e7 10.1016/j.amepre.2016.07.032 27692756 29. Robinson J. Too L.S. Pirkis J. Spittal M.J. Spatial suicide clusters in Australia between 2010 and 2012: A comparison of cluster and non-cluster among young people and adults BMC Psychiatry 2016 16 417 10.1186/s12888-016-1127-8 27876026 30. Núñez-González S. Lara-Vinueza A.G. Gault C. Delgado-Ron J.A. Trends and spatial patterns of suicide among adolescent in Ecuador, 1997–2016 Clin. Pract. Epidemiol. Ment. Health 2018 14 283 10.2174/1745017901814010283 30972127 31. Sy K.T.L. Shaman J. Kandula S. Pei S. Gould M.S. Keyes K.M. Spatiotemporal clustering of suicides in the US from 1999 to 2016: A spatial epidemiological approach Soc. Psychiatry Psychiatr. Epidemiol. 2019 54 1471 1482 10.1007/s00127-019-01736-4 31177308 32. Milner A. Too L.S. Spittal M.J. Cluster Suicides among Unemployed Persons in Australia Over the Period 2001–2013 Soc. Indic. Res. 2018 137 189 201 10.1007/s11205-017-1604-6 33. Gibbons R.D. Clark D.C. Fawcett J. A statistical method for evaluating suicide clusters and implementing cluster surveillance Am. J. Epidemiol. 1990 132 (Suppl. 1) S183 S191 10.1093/oxfordjournals.aje.a115781 2356830 34. Kirch M.R. Lester D. Suicide from the Golden Gate Bridge: Do they cluster over time? Psychol. Rep. 1986 59 1314 10.2466/pr0.1986.59.3.1314 3823330 35. Cox B. Skegg K. Contagious suicide in prisons and police cells J. Epidemiol. Community Health 1993 47 69 72 10.1136/jech.47.1.69 8436899 36. Fontanella C.A. Saman D.M. Campo J.V. Hiance-Steelesmith D.L. Bridge J.A. Sweeney H.A. Root E.D. Mapping suicide mortality in Ohio: A spatial epidemiological analysis of suicide clusters and area level correlates Prev. Med. 2018 106 177 184 10.1016/j.ypmed.2017.10.033 29133266 37. Qi X. Hu W. Page A. Tong S. Spatial clusters of suicide in Australia BMC Psychiatry 2012 12 86 10.1186/1471-244X-12-86 22824367 38. Qi X. Tong S. Hu W. Spatial distribution of suicide in Queensland, Australia BMC Psychiatry 2010 10 106 10.1186/1471-244X-10-106 21138551 39. Strauss M.J. Klimek P. Sonneck G. Niederkrotenthaler T. Suicides on the Austrian railway network: Hotspot analysis and effect of proximity to psychiatric institutions R. Soc. Open Sci. 2017 4 160711 10.1098/rsos.160711 28405359 40. Tomita M. Kubota T. Ishioka F. Spatial Clustering Properties in the Temporal Variation of Suicide Rates/Numbers among Japanese Citizens: A Comprehensive Comparison and Discussion PLoS ONE 2015 10 e0127358 10.1371/journal.pone.0127358 26161651 41. Too L.S. Pirkis J. Milner A. Bugeja L. Spittal M.J. Railway suicide clusters: How common are they and what predicts them? Inj. Prev. 2017 23 328 333 10.1136/injuryprev-2016-042029 27864309 42. Too L. Pirkis J. Milner A. Spittal M. Clusters of suicides and suicide attempts: Detection, proximity, and correlates Epidemiol. Psychiatr. Sci. 2017 26 491 500 10.1017/S2045796016000391 27278418 43. Sugg M.M. Woolard S. Lawrimore M. Michael K.D. Runkle J.D. Spatial Clustering of Suicides and Neighborhood Determinants in North Carolina, 2000 to 2017 Appl. Spat. Anal. Policy. 2021 14 395 413 10.1007/s12061-020-09364-1 44. Lai C. Law Y.W. Shum A.K. Ip F.W. Yip P.S. A community-based response to a suicide cluster: A Hong Kong experience Crisis 2020 41 163 171 10.1027/0227-5910/a000616 31418310 45. Too L.S. Spittal M.J. Suicide Clusters among Top 10 High-Risk Occupations: A Study From 2001 to 2016 in Australia J. Nerv. Ment. Dis. 2020 208 942 946 10.1097/NMD.0000000000001234 32910073 46. Hill N.T. Too L.S. Spittal M.J. Robinson J. Understanding the characteristics and mechanisms underlying suicide clusters in Australian youth: A comparison of cluster detection methods Epidemiol. Psychiatr. Sci. 2020 29 e151 10.1017/S2045796020000645 32758330 47. Yamaoka K. Suzuki M. Inoue M. Ishikawa H. Tango T. Spatial clustering of suicide mortality and associated community characteristics in Kanagawa prefecture, Japan, 2011–2017 BMC Psychiatry 2020 20 74 10.1186/s12888-020-2479-7 32070316 48. Hsu C.Y. Chang S.S. Lee E.S. Yip P.S. Geography of suicide in Hong Kong: Spatial patterning, and socioeconomic correlates and inequalities Soc. Sci. Med. 2015 130 190 203 10.1016/j.socscimed.2015.02.019 25706063 49. Yoshioka E. Hanley S.J. Sato Y. Saijo Y. Geography of suicide in Japan: Spatial patterning and rural–urban differences Soc. Psychiatry Psychiatr. Epidemiol. 2021 56 731 7466 10.1007/s00127-020-01978-7 33159535 50. Vaz E. Shaker R.R. Cusimano M.D. A geographical exploration of environmental and land use characteristics of suicide in the greater Toronto area Psychiatry Res. 2020 1 112790 10.1016/j.psychres.2020.112790 51. Pérez-Costillas L. Blasco-Fontecilla H. Benítez N. Comino R. Antón J.M. Ramos-Medina V. López A. Palomo J.L. Madrigal L. Alcalde J. Space–time suicide clustering in the community of Antequera (Spain) Rev. Psiquiatr. Salud Men. 2015 8 26 34 10.1016/j.rpsm.2014.01.007 24986472 52. Kassem A.M. Carter K.K. Johnson C.J. Hahn C.G. Peer Reviewed: Spatial Clustering of Suicide and Associated Community Characteristics, Idaho, 2010–2014 Prev. Chronic Dis. 2019 16 E37 10.5888/pcd16.180429 30925141 53. Takahashi K. Yokoyama T. Tango T. FleXScan v3. 1: Software for the Flexible Scan Statistic National Institute of Public Health Tokyo, Japan 2010 54. Lersch K.M. Exploring the geography of suicide threats and suicide attempts: An application of Risk Terrain Modeling Soc. Sci. Med. 2020 249 112860 10.1016/j.socscimed.2020.112860 32106000 55. Ceccato V. Uittenbogaard A. Suicides in commuting railway systems: The case of Stockholm county, Sweden J. Affect. Disord. 2016 198 206 2157 10.1016/j.jad.2016.02.051 27018939 56. Tidemalm D. Runeson B. Waern M. Frisell T. Carlström E. Lichtenstein P. Långström N. Familial clustering of suicide risk: A total population study of 11.4 million individuals Psychol. Med. 2011 41 2527 2534 10.1017/S0033291711000833 21733212 57. Dos Santos A.D. Guimarães L.M.L. De Carvalho Y.F. Viana L.D.C. Alves G.L. Lima A.C.R. Santos M.B. Góes M.A.D.O. De Araújo K.C.G.M. Spatial analysis and temporal trends of suicide mortality in Sergipe, Brazil, 2000–2015 Trends Psychiatry Psychother. 2018 40 269 276 10.1590/2237-6089-2017-0028 30156647 58. Bando D.H. Barrozo L.V. Volpe F.M. Geographical clusters and social risk factors for suicide in the city of São Paulo, 2006–2015: An ecologic study Int. J. Soc. Psychiatry 2020 66 460 468 10.1177/0020764020918618 32345084 59. Beringuel B.M. de Barros Canuto I.M. Silva A.P. do Bonfim C.V. Epidemiology and Spatiotemporal Clustering of Suicides in the State of Pernambuco (1999–2018) Trends Psychol. 2021 29 123 138 10.1007/s43076-020-00054-y 60. Guo Y. Chau P.P. Chang Q. Woo J. Wong M. Yip P.S. The geography of suicide in older adults in Hong Kong: An ecological study Int. J. Geriatr. Psychiatry 2020 35 99 112 10.1002/gps.5225 31663178 61. Santos E.G. Barbosa I.R. Severo A.K. Space-time analysis of mortality by suicide in the State of Rio Grande do Norte, Brazil, in the period from 2000 to 2015 Cien Saude Colet. 2020 25 633 643 10.1590/1413-81232020252.11042018 32022203 62. Reser J.P. Australian Aboriginal suicide deaths in custody: Cultural context and cluster evidence Aust. Psychol. 1989 24 325 342 10.1080/00050068908259573 63. Lazzarini T.A. Gonçalves C.C.M. Benites W.M. Silva LFd Tsuha D.H. Ko A.I. Rohrbaugh R. Andrews J.R. Croda J. Suicide in Brazilian indigenous communities: Clustering of cases in children and adolescents by household Rev. Saúde Públ. 2018 52 56 10.11606/S1518-8787.2018052000541 64. Austin A.E. van den Heuvel C. Byard R.W. Cluster hanging suicides in the young in South Australia J. Forensic Sci. 2011 56 1528 1530 10.1111/j.1556-4029.2011.01840.x 21790596 65. Mackenzie D.W. Lester D. Manson R. Yeh C. Do suicides from the Golden Gate Bridge cluster? Psychol. Rep. 2016 118 70 73 10.1177/0033294115625565 29693523 66. Torok M. Konings P. Batterham P.J. Christensen H. Spatial clustering of fatal, and non-fatal, suicide in new South Wales, Australia: Implications for evidence-based prevention BMC Psychiatry 2017 17 339 10.1186/s12888-017-1504-y 28985736 67. Gould M.S. Petrie K. Kleinman M.H. Wallenstein S. Clustering of attempted suicide: New Zealand national data Int. J. Epidemiol. 1994 23 1185 1189 10.1093/ije/23.6.1185 7721521 68. Leung M. Chow C.B. Ip P. Yip S. Pure spatial and space-time clusters of self-harm in Kwai Tsing 2004 to 2012 Spat Spatiotemporal Epidemiol. 2018 27 1 9 10.1016/j.sste.2018.07.002 30409371 69. Karami M. Yazdi–Ravandi S. Ghaleiha A. Olfatifar M. Comparison of the clusters and non-clusters areas of attempted suicide cases in Hamadan Province, western Iran: Findings from a pilot study (2016–2017) J. Res. Health Sci. 2018 18 e00425 10.5888/pcd16.18042935 30270217 70. Smeeton N. Wilkinson G. The identification of clustering in parasuicide Br. J. Psychiatry 1988 153 218 221 10.1192/bjp.153.2.218 3255435 71. Mittendorfer-Rutz E. Rasmussen F. Wasserman D. Familial clustering of suicidal behaviour and psychopathology in young suicide attempters Soc. Psychiatry Psychiatr. Epidemiol. 2008 43 28 36 10.1007/s00127-007-0266-0 17934681 72. Pisinger V.S. Hawton K. Tolstrup J.S. School-and class-level variation in self-harm, suicide ideation and suicide attempts in Danish high schools Scand. J. Public Health 2019 47 146 156 10.1177/1403494818799873 30222056 73. Goodman M.L. Puffer E.S. Keiser P.H. Gitari S. Suicide clusters among young Kenyan men Health Psychol. 2020 25 1004 1013 10.1177/1359105317743803 29192523 74. Queinec R. Benjamin C. Beitz C. Lagarde E. Encrenaz G. Suicide contagion in France: An epidemiologic study Inj. Prev. 2010 16 (Suppl. 1) A241 10.1136/ip.2010.029215.858 75. Jonas K. Modelling and suicide: A test of the Werther effect Br. J. Soc. Psychol. 1992 31 295 306 10.1111/j.2044-8309.1992.tb00974.x 1472984 76. Sinyor M. Williams M. Tran U.S. Schaffer A. Kurdyak P. Pirkis J. Niederkrotenthaler T. Suicides in Young People in Ontario Following the Release of “13 Reasons Why” Can. J. Psychiatry 2019 64 798 804 10.1177/0706743719870507 31434503 77. Niederkrotenthaler T. Stack S. Till B. Sinyor M. Pirkis J. Garcia D. Rockett I.R.H. Tran U.S. Association of increased youth suicides in the United States with the release of 13 Reasons Why JAMA Psychiatry 2019 76 933 940 10.1001/jamapsychiatry.2019.0922 31141094 78. Whitley R. Fink D.S. Santaella-Tenorio J. Keyes K.M. Suicide mortality in Canada after the death of Robin Williams, in the context of high-fidelity to suicide reporting guidelines in the Canadian media Can. J. Psychiatry 2019 64 805 812 10.1177/0706743719854073 31181955 79. Fink D.S. Santaella-Tenorio J. Keyes K.M. Increase in suicides the months after the death of Robin Williams in the US PLoS ONE 2018 13 e0191405 10.1371/journal.pone.0191405 29415016 80. Pirkis J. Currier D. Too L.S. Bryant M. Bartlett S. Sinyor M. Spittal M.J. Suicides in Australia following media reports of the death of Robin Williams Aust. N. Z. J. Psychiatry 2020 54 99 104 10.1177/0004867419888297 31749369 81. Sinyor M. Tran U.S. Garcia D. Till B. Voracek M. Niederkrotenthaler T. Suicide mortality in the United States following the suicides of Kate Spade and Anthony Bourdain. 2020 Aust. N. Z. J. Psychiatry 2021 55 613 619 10.1177/0004867420976844 33300363 82. Bridge J.A. Greenhouse J.B. Ruch D. Stevens J. Ackerman J. Sheftall A.H. Horowitz L.M. Kelleher K.J. Campo J.V. Association between the release of netflix’s 13 Reasons Why and suicide rates in the United States: An interrupted time series analysis J. Am. Acad. Child Adolesc. Psychiatry 2020 59 236 243 10.1016/j.jaac.2019.04.020 31042568 83. Romer D. Reanalysis of the Bridge et al. study of suicide following release of 13 Reasons Why PLoS ONE 2020 15 e0227545 10.1371/journal.pone.0227545 31945088 84. Jang S.A. Sung J.M. Park J.Y. Jeon W.T. Copycat suicide induced by entertainment celebrity suicides in South Korea Psychiatry Investig. 2016 13 74 10.4306/pi.2016.13.1.74 26766949 85. Chen Y.Y. Tsai P.C. Chen P.H. Fan C.C. Hung G.C.L. Cheng A.T. Effect of media reporting of the suicide of a singer in Taiwan: The case of Ivy Li Soc. Psychiatry Psychiatr. Epidemiol. 2010 45 363 369 10.1007/s00127-009-0075-8 19468661 86. Phillips D.P. Carstensen L.L. Clustering of teenage suicides after television news stories about suicide N. Engl. J. Med. 1986 315 685 689 10.1056/NEJM198609113151106 3748072 87. Ladwig K.H. Kunrath S. Lukaschek K. Baumert J. The railway suicide death of a famous German football player: Impact on the subsequent frequency of railway suicide acts in Germany J. Affect. Disord. 2012 136 194 198 10.1016/j.jad.2011.09.044 22036798 88. Cheng Q. Chen F. Yip P.S. The Foxconn suicides and their media prominence: Is the Werther Effect applicable in China? BMC Public Health 2011 11 841 10.1186/1471-2458-11-841 22044598 89. Lee S.Y. Media Coverage of Adolescent and Celebrity Suicides and Imitation Suicides among Adolescents J. Broadcast. Electron. Media 2019 63 130 143 10.1080/08838151.2019.1570200 90. Koburger N. Mergl R. Rummel-Kluge C. Ibelshäuser A. Meise U. Postuvan V. Roskar S. Székely A. Tóth M.D. Van der Feltz-Cornelis C. Celebrity suicide on the railway network: Can one case trigger international effects? J. Affect. Disord. 2015 85 38 46 10.1016/j.jad.2015.06.037 26143403 91. Etzersdorfer E. Sonneck G. Nagel-Kuess S. Newspaper reports and suicide N. Engl. J. Med. 1992 10.1056/NEJM199208133270720 92. Too L.S. Pirkis J. Milner A. Robinson J. Spittal M.J. Clusters of suicidal events among young people: Do clusters from one time period predict later clusters? Suicide Life Threat Behav. 2019 49 561 571 10.1111/sltb.12460 29578243 93. Lawson A.B. Kleinman K. Spatial and Syndromic Surveillance for Public Health John Wiley & Sons Hoboken, NJ, USA 2005 94. Maurya S.P. Ohri A. Mishra S. Open-Source GIS: A Review Proceedings of the National Conference on Open-Source GIS: Opportunities and Challenges Varanasi, India 9–10 October 2015 95. Nsubuga P. White M.E. Thacker S.B. Anderson M.A. Blount S.B. Broome C.V. Chiller T.M. Espitia V. Imtiaz R. Sosin D. Public Health Surveillance: A Tool for Targeting and Monitoring Interventions Disease Control Priorities in Developing Countries 2nd ed. Jamison D.T. Breman J.G. Measham A.R. Alleyne G. Claeson M. Evans D.B. Jha P. Mills A. Musgrove P. The International Bank for Reconstruction and Development/The World Bank Washington, DC, USA Oxford University Press New York, NY, USA 2006 Chapter 53 99 1015 -10 0-8213-6179-1 96. Cromley E.K. McLafferty S.L. GIS and Public Health Guilford Press New York, NY, USA 2011 97. Ishioka F. Kurihara K. Evaluation of Hotspot Detection Method Based on Echelon Structure Proceedings of the 59th ISI World Statistics Congress Hong Kong, China 25–30 August 2013 5366 98. Tango T. Takahashi K. A flexibly shaped spatial scan statistic for detecting clusters Int. J. Health Geogr. 2005 4 11 10.1186/1476-072X-4-11 15904524 99. Patil G.P. Taillie C. Upper-level set scan statistic for detecting arbitrarily shaped hotspots Environ. Ecol. Stat. 2004 11 183 197 10.1023/B:EEST.0000027208.48919.7e 100. Kurihara K. Hotspots Detection for Cellular Surface Data. Bull Int. Stat. Inst. 2003 54 115 116 101. Tango T. Takahashi K. A flexible spatial scan statistic with a restricted likelihood ratio for detecting disease clusters Stat. Med. 2012 31 4207 4218 10.1002/sim.5478 22807146 102. Myers W. Patil G.P. Joly K. Echelon approach to areas of concern in synoptic regional monitoring Environ. Ecol. Stat. 1997 4 131 152 10.1023/A:1018518327329 103. Kurihara K. Ishioka F. Kajinishi S. Spatial and temporal clustering based on the echelon scan technique and software analysis JJSD 2020 3 313 332 10.1007/s42081-020-00072-1 104. Terrell C. Predictions in Time Series Using Regression Models Scientific e-Resources New Delhi, India 2019 105. Värnik P. Sisask M. Värnik A. Laido Z. Meise U. Ibelshäuser A. Van Audenhove C. Reynders A. Kocalevent R.D. Kopp M. Suicide registration in eight European countries: A qualitative analysis of procedures and practices Forensic Sci. Int. 2010 202 86 92 10.1016/j.forsciint.2010.04.032 20483553 106. Baran A. Gerstner R. Ueda M. Gmitrowicz A. Implementing real-time data suicide surveillance systems Crisis 2021 42 321 327 10.1027/0227-5910/a000829 34528451 107. Kulldorff M. Prospective time periodic geographical disease surveillance using a scan statistic J. R. Stat. Soc. Ser. A 2001 164 61 72 10.1111/1467-985X.00186

PMC009xxxxxx/PMC9099649.txt

==== Front Molecules Molecules molecules Molecules 1420-3049 MDPI 10.3390/molecules27092836 molecules-27-02836 Article Optimization of an Ultra-Sonication Extraction Method for Major Compounds Found in Mondia whitei Using Design of Experiment Chokwe Ramakwala Christinah 1* Dube Simiso 1 https://orcid.org/0000-0001-7222-1505 Nindi Mathew Muzi 2* Morana Alessandra Academic Editor Squillaci Giuseppe Academic Editor 1 Department of Chemistry, The Science Campus, College of Science Engineering and Technology, University of South Africa, Corner Christiaan de Wet and Pioneer Avenue, Florida Park, Roodepoort 1709, South Africa; dubes@unisa.ac.za 2 Institute for Nanotechnology and Water Sustainability, College of Science Engineering and Technology, University of South Africa, Corner Christiaan de Wet and Pioneer Avenue, Florida Park, Roodepoort 1709, South Africa * Correspondence: 36340227@mylife.unisa.ac.za (R.C.C.); nindimm@unisa.ac.za (M.M.N.) 29 4 2022 5 2022 27 9 283614 3 2022 12 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Optimum extraction conditions are vital in quality control methods to enable accurate quantification of the compounds of interest. An ultra-sonication method was developed for the extraction of seven major compounds found in Mondia whitei. Extraction temperature, time, power, frequency, percentage of ethanol in water and solvent to sample ratio were screened to access their significance on the percentage recovery of the compounds of interest. These parameters were screened using Descriptive screening design. Extraction temperature, solvent to sample ratio and the interaction between temperature and percentage ethanol in water were found to have a significant effect on the response. These parameters were then optimized using central composite design. The optimum conditions were found to be 66.1% ethanol in water, 70 °C temperature and 3 mL: 5 mg solvent to sample ratio. This method was successfully applied in the development of a quality control method for the seven compounds in Mondia whitei samples. Mondia whitei extraction ultra-sonication descriptive screening design central composite design The University of South Africa Grow Your Own Timber The National Research Foundation of South Africa SFH150630122311 This research was funded by The University of South Africa Grow Your Own Timber and The National Research Foundation of South Africa, grant number SFH150630122311. ==== Body pmc1. Introduction Mondia whitei is a medicinal plant that is native to Africa. It is distributed across different parts of the continent and is therefore, known by various names such as Ogomo in Kenya, Limte in Cameroon, Mulondo in Uganda, and Umondi in South Africa, and in Zimbabwe, they call it Mungurawu [1]. It has been used for many years and still continue to be used traditionally to cure various ailments. The roots are the commonly used part of this plant. They are used to treat aches and pains, hypertension, stress, improve appetite, libido, fertility treatment, mental disorder, diabetes, asthma among other ailments [2,3,4,5]. Above all, the plant is popularly known and used as an aphrodisiac across the whole African continent [6]. The aphrodisiac activity of the roots of the plant has been proven scientifically by several researchers [6,7,8,9]. The popularity of Mondia whitei has necessitated the need to isolate and characterized compounds from the plant. This is so that the compounds responsible for the activities can be identified and to also enable quality control of the plant samples that are sold commercially. Most scientific work done on the plant so far was to corroborate the traditional medicinal use of the plant. A few researchers have reported on the isolation and characterization of compounds from the plant. Koorbanally et al., (2008), Mukonyi and Ndiege (2001) isolated 2-hydroxy-4-methoxybenzaldehyde and 3-hydroxy-4-methoxybenzaldehyde from the roots of the plant [10,11,12]. The authors reported that the compound 2-hydroxy-4-methoxybenzaldehyde was responsible for the taste modifying property of the plant. Wang J et al., (2010) reported its antimicrobial and antioxidant activities [13]. Patnam et al., (2004) isolated and identified 6-methoxy-7-hydroxycoumarin and 6-methoxy-7,8-dihydroxycoumarin from the roots [10]. These compounds were found to have antimicrobial activity by Yang et al., (2017) [14]. Ultra-sonic assisted extraction is a low-cost and efficient method whereby extraction can be carried out in short periods of time [15]. This makes it a suitable method for extraction of plant materials. Extraction yield and therefore recovery is known to be affected by factors such as the type of extracting solvent, temperature, ratio of solvent to sample and extraction time [16]. Therefore, it is of paramount importance that the extraction method is optimized for optimum extraction of the compounds of interest. Usually, a one-factor-at-a-time method is used to optimize the factors that are known to influence the response. In this method, one factor is optimized at a time while the other factors are kept constant. Moreover, this method does not consider the interaction between the factors. Therefore, design of experiment can be used as an alternative, since the method not only considers interactions between factors, but the number of experiments is fewer as compared to one-factor-at-a-time, especially when many factors are being investigated [17]. In this study, an extraction method was optimized for the extraction of 2-hydroxy-4-methoxybenzaldehyde (C1), 3-hydroxy-4-methoxybenzaldehyde (C2), 2,4-dihydroxy-6-methylbenzaldehyde (C3), 7-hydroxy-6-methoxycoumarin (C4), 7,8-dihydroxy-6-methoxycoumarin (C5), coumarin (C6) and phenantherene (C7) which have been identified in Mondia whitei. The optimum extraction conditions are not only important for maximum extraction of the compounds of interest but also to enable accurate quantification of the compounds in Mondia whitei products, thereby enabling quality control of Mondia whitei samples. In the design of the experiment, descriptive screening design has an advantage over other screening methods because it requires fewer experiments and the factors are accessed using three levels [18]. In this study, factors which were identified as being significant using descriptive screening design were further optimized using central composite design and response surface methodology. Furthermore, an HPLC-DAD method was used to separate and analyze the extracts. The optimum extraction method for these compounds has been used successfully for their extraction and subsequent accurate quantification in Mondia whitei samples and or products [19]. 2. Results and Discussion 2.1. Screening of the Extraction Factors Using Descriptive Screening Design Temperature, ratio of ethanol to water, ratio of solvent to sample, extraction time, sonication power and frequency are known to affect the percentage recovery of compounds from plants materials. Therefore, these factors were selected as the independent variables and were screened for their effect on the percentage recovery of the compounds of interest using ultra-sonication technique. The percentage recovery of the compounds was identified as the dependent variable. Descriptive screening design (DSD) has the advantage over other screening methods because less experiments are required for the same number of factors and main effects are not cofounded with two-factor interactions. Therefore, DSD is recommended if the number of independent variables is more than four [20]. The experiments generated using the Minitab software as well as the responses that were obtained experimentally are shown in Table 1, the experiments were randomized to avoid biases. An average percentage recovery of the compounds was chosen as the response because it was necessary to find optimum extraction conditions for simultaneous extracts of the compounds of interest. Stepwise selection was used to assess the significance of the factors on the response. The significant effects of the factors are shown in the pareto chart (Figure 1), temperature, the ratio of solvent to sample and the interaction between temperature and percentage ethanol in water had a significant effect on the percentage recovery of the compounds. Time, power, frequency and the other interactions were not significant as shown by their absences from the pareto chart. Even though percentage of ethanol in water does not have a significant effect on the response, it was included due to hierarchy, since its interaction with temperature is significant. Only the significant factors were chosen to be optimized. 2.2. Optimization of the Extraction Factors Using Central Composite Design (CCD) CCD was used for optimization of the significant factors identified by the screening method. Extraction time, power and frequency were kept constant at 20 min, 0.03 watts and high frequency respectively while optimizing the significant factors. The experiments generated by the Minitab software together with the response are given in Table 2. R2 and R2 adjusted for the model are given in Table 3, these were 90.10 and 81.20% respectively. These parameters indicate how well the model predict the response and descriptive ability of the model respectively [21]. The linear and second-order models were significant with p-values of 0.016 and 0.000, respectively (Table 3). The two-way interaction model was not significant shown by p > 0.05. The non-significant value of lack of fit with (p = 0.969) indicated that the model fits well with the experimental design and can be used to predict the response. Modelling of the response was done using second-order polynomial as shown in Section 3.4. The surface and contour plots were used to visualize the results. The surface plots are curved because the model has second order terms that are statistically significant (Figure 2). The full model that includes non-significant factors is given in Equation (1) below. Response = 72.07 − 0.02x + 7.76y+ 8.26z − 4.51x2 + 12.39y2 − 35.95z2 − 1.14xy − 1.34xz − 6.02yz(1) where x = %ethanol; y = extraction temperature, z = solvent: sample ratio. 2.2.1. Effects of Solvent: Sample Ratio This was found to be the most significant factor/variable with p = 0.013; the effect of this factor on the response was positive as shown by its positive regression coefficient. The surface plots (Figure 2a,b) show that the percentage recovery increases as the solvent: sample ratio increases up to 4 mL: 50 mg, this could be attributed to the mass transfer between the solid material and the solvent due to the difference in concentration gradient [22]. Above 4 mL: 50 mg the percentage recovery decreases with an increase in solvent: sample ratio. A similar trend was observed by Chen et al., (2020) in their investigation for the effects of solvent: sample ratio on the yield of phytochemicals from coffee leaves [16]. The contour plots show that the optimum solvent is between 2.7 mL and 3.8 mL, shown by the dark green regions in Figure 3b,c. 2.2.2. Effects of Temperature An increase in temperature is known to result in an increase in percentage recovery [21,22]. In this study, temperature had a significant effect on the response (p = 0.018 coefficient = 7.76). However, the effect of temperature on the response was the opposite up to 45 °C as shown by the surface plot thereafter it increases with an increase in temperature. The contour plots show that optimum extraction can be obtained with temperature from 68.55 °C, indicated by the dark green region (Figure 3a,c). 2.2.3. Effects of %Ethanol in Water Different ratios of ethanol: water have been reported for the extraction of the compounds from Mondia whitei. Therefore, it was necessary to optimize this factor. Percentage ethanol in water had no significant effect on the response in this study, as shown by p > 0.05 for this factor. This is also observed on the surface plots, where no change in response is observed when percentage ethanol increases. The contour plots show that the optimum percentage ethanol in water is between 46.6 and 91.5%. 2.3. Testing the Predicted Optimum Conditions The optimizer in the Minitab software was used to predict the optimum conditions for each independent factor (Figure 4). The setting for the independent factors was so that they should be maximized. The optimum conditions that were predicted are percentage ethanol (66.1%), temperature (70 °C) and solvent to sample ratio (3 mL: 50 mg) and the predicted desirability functions was 0.9928. These conditions were tested by extracting the compounds of interest under the predicted optimum conditions. The results that were obtained were used to calculate the desirability function. The practical and predicted desirability functions were compared and the percentage difference was calculated. The calculated desirability function was 0.9773, comparing that with the predicted desirability function, the percentage difference is 0.02% which shows that both the experimental and predicted values are in agreement. This method was applied successfully to extract the compounds of interest from Mondia whitei samples and products (Figure 5) in turn enabling their accurate quantification [19]. Figure 5 shows a chromatogram representing separation of the Mondia whitei roots powder which was spiked with the compounds of interest then extracted using the developed extraction method. The separation conditions are shown in the caption for Figure 5. 3. Materials and Methods 3.1. Chemicals and Material The solvents that were used in this work were of analytical grade with a purity of >95%, they were purchased from Sigma-Aldrich (Steinheim, Germany). The standards were also purchased from Sigma-Aldrich. The aqueous solutions were prepared using ultra-high-purity (UHP) water (18.2 mΩ) from a Milli-Q water purification system (Molsheim, France) and filtered using a 0.45 µm membrane filter Sigma-Aldrich (Steinheim, Germany). The plant material was obtained from Durban, Kwazulu-Natal in South Africa and was authenticated at the college of Agriculture and Environmental Sciences (University of South Africa, Florida Park, Roodepoort, South Africa). The roots were grinded using a kitchen blender after which the powder was separated from the fibers using 0.25 mm mesh Sieve. 3.2. Instrumentation Ultrasonic-assisted extraction was performed using an ultrasonic bath (ScientTech, Labotec, Midrand, South Africa). Separation of the compounds was performed using an Agilent HPLC 1260 system (Agilent Technologies, Waldbronn, Germany) which consisted of a binary high-pressure pump, autosampler, a thermostatted column compartment, a diode array detector and a fluorescence detector. Instrument control, data collection and processing were achieved using the ChemStation (version 1.9.0) software. The separation of the mixture was performed on an XTerra® MS C18 (150 mm × 4.6 mm, 3.5 µm) analytical column (Waters Corporation, Milford, MA, USA). The mobile phase used for the separation was 0.1% formic acid in water (A) and acetonitrile (B). The following gradient elution mode was used to separate the compounds: 0 min 25% (B), 1 min 35% (B), 2 min 45% (B), 3 min 55% (B), 5 min 100% (B). Injection volume was 5 µL; temperature was 25 °C; and flow rate was 1.3 mL min −1. The compounds were monitored at 254 and 331 nm. A typical chromatogram for the separation of the pure standards in the solvent is shown in Figure 6. 3.3. Preparation of the Samples for Extraction The stock solutions of the standards were prepared at a concentration of 500 mg L−1, by diluting 2.5 mg of each standard with ethanol in 5 mL volumetric flasks. Fifty milligrams of the Mondia whitei roots powder was spiked with 200 µL of each standard and extracted under the experimental conditions shown in Table 1. After extraction, the extracts were separated from the undissolved materials using a centrifuge (Eppendorf, Hamburg, Germany) at 3000 rpm for 6 min. The extracts were dried using a freeze dryer (BioBase, Shandong, China). The dried extracts were then reconstituted with 1 mL of ethanol in preparation for HPLC-DAD analyses. 3.4. Experimental Design Descriptive screening design was used to assess the significance of the following independent factors: Extraction temperature; ratio of ethanol: water, solvent: sample ratio, extraction time, ultrasound power and frequency on the percentage recovery of the seven compounds of interest. Percentage recovery of the compounds were calculated using Equation (2). % Recovery = Peak area found − peak area original/peak area spiked(2) where peak area found is the peak area of the compounds found in the spiked sample after extraction, peak area original is the peak area of the compound found in the unspiked sample after extraction, peak area spiked the peak area of the spiked compound in the pure solvent. The design consisted of six factors, three level, two centre points and fourteen experiments. Table 1 shows the minimum and maximum values for each factor. The experiments were randomized to avoid systemic error (Table 4). The Central Composite Design consisted of three factors, three levels, six centre points and six axial points with twenty experiments (Table 2). The experiments were run in triplicates. Analysis of variance was used to assess the effect of the factors on the response. Response surface methodology was used to visualize the results in the form of surface and contour plots. The response was analyzed using a second-order polynomial regression equation as follows:(3) Y=βo+∑i=1nβ1x1+∑i=1kβ11x12+∑1k−1β12x1x2. where Y is the response, βo; β1; β11 and β12 are the intercept, linear, quadratic and interaction regression coefficients, respectively. 3.5. Software Minitab Version 18 from Minitab Inc. (State College, PA, USA) was used for the creation of the experimental design, statistical and graphical analysis. 4. Conclusions A simple and fast extraction method was developed and optimized for seven major compounds found in Mondia whitei. Extraction temperature and the ratio of solvent to sample were found to have a statistically significant effect on the percentage recovery of the compounds of interest. The optimum conditions for extraction of these compounds were 66.1% ethanol in water, 70 °C temperature and 3 mL to 5 mg sample. The developed method can be used to extract the compounds of interest from Mondia whitei products to enable their quantification. Acknowledgments The Chemistry Department at the University of South Africa is acknowledged for making the facilities required for the research available. Author Contributions Supervision, M.M.N. and S.D.; Writing—original draft, R.C.C. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement The data presented in this study is available on request from the corresponding authors. Conflicts of Interest The authors declare no conflict of interest. Sample Availability Samples of the compounds used in this study, are available from the authors. Figure 1 Pareto chart showing the effects of the critical quality parameters on the response. Figure 2 Surface and contour plots showing percentage recovery as function of the effects temperature and solvent ratio (S:S) while keeping percentage ethanol constant (a); effects of percentage ethanol and solvent ratio (S:S) while keeping temperature constant (b); and effects of percentage ethanol and temperature while keeping solvent ratio constant (c). Figure 3 Surface and contour plots showing percentage recovery as function of the effects temperature and solvent ratio while keeping percentage ethanol constant (a); effects of percentage ethanol and solvent ratio while keeping temperature constant (b); and effects of percentage ethanol and temperature while keeping solvent ratio constant (c). Figure 4 The optimum conditions predicted using the Minitab v18 software (State College, PA, USA). Figure 5 Chromatogram representing the separation of the Mondia whitei roots powder spiked with the compounds of interest under optimised conditions by gradient elution mode; at 0 min 25% B, 2 min 40% B, 3 min 65% B, 4 min 100% B with a run time of 6.5 min. Injection volume of 5 µL, flow rate of 1.3 mL min−1 and temperature of 25 °C. Figure 6 Typical chromatogram of separation of the C 1–7 under optimised conditions by gradient elution mode; at 0 min 25% B, 2 min 40% B, 3 min 65% B, 4 min 100% B with a run time of 6.5 min. Injection volume of 5 µL, flow rate of 1.3 mL min−1 and temperature of 25 °C. molecules-27-02836-t001_Table 1 Table 1 Descriptive screening design experiments and the response. Run Order 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Temperature 70 30 70 50 50 50 30 30 30 70 70 30 70 50 Ethanol ratio 100 100 40 70 100 70 70 40 40 40 100 100 70 40 Solvent volume 3 5 4 3 5 3 1 3 5 1 1 1 5 5 Time 60 20 40 40 60 40 60 20 60 20 20 40 20 60 Power 0.01 0.03 0.05 0.03 0.05 0.03 0.05 0.05 0.01 0.05 0.05 0.01 0.01 0.01 Frequency Low Low Low High High High Low High Low Low Low High High High % Average recovery 44.58 53.53 72.91 46.50 42.03 65.98 55.92 45.22 31.67 73.72 66.41 68.34 44.88 58.85 molecules-27-02836-t002_Table 2 Table 2 CCD experimental conditions and response for optimization of the extraction method. Run Order % Ethanol Temperature Solvent: Sample Ratio % Average Recovery 1 40 50 3 68.62 2 100 70 1 49.99 3 100 30 5 47.67 4 100 30 1 27.05 5 40 30 5 50.17 6 70 50 1 22.52 7 40 70 1 51.70 8 70 50 5 80.85 9 40 30 1 19.58 10 100 70 5 51.15 11 70 50 3 78.24 12 70 70 3 92.86 13 70 50 3 78.99 14 70 50 3 82.18 15 40 70 5 53.58 16 70 50 3 76.57 17 70 30 3 77.17 18 100 50 3 67.63 19 70 50 3 57.54 20 70 50 3 56.67 molecules-27-02836-t003_Table 3 Table 3 ANOVA results for optimization of the extraction method using CCD. p-Value Coefficient Linear 0.016 Square 0.000 2-way interaction 0.303 Constant 0.000 72.07 %Ethanol 0.996 −0.002 Temperature 0.018 7.76 Solvent 0.013 8.26 %Ethanol*%Ethanol 0.410 4.51 Temperature*Temperature 0.040 12.39 Solvent*Solvent 0.000 −35.95 %Ethanol*Temperature 0.719 −1.14 %Ethanol*Solvent 0.673 −1.34 Temperature*Solvent 0.079 −6.02 Lack-of-fit 0.969 R2 90.10% R2(adj) 81.20% R2(pred) 76.65% molecules-27-02836-t004_Table 4 Table 4 The independent factors and their labels in uncoded form. Temperature Ethanol Ratio Solvent Volume Time Power Frequency Low 30 40 1 20 0.01 Low High 70 100 5 60 0.05 High Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. South African National Biodiversity Available online: http://pza.sanbi.org/mondia-whitei (accessed on 4 December 2021) 2. Oketch-Rabah H.A. Mondia whitei, a medicinal plant from Africa with aphrodisiac and antidepressant properties: A review J. Diet Suppl. 2012 9 272 284 10.3109/19390211.2012.726704 23039023 3. Aremu A.O. Cheesman L. Finnie J.F. Van Staden J. Mondia whitei (Apocynaceae): A review of its biological activities, conservation strategies and economic potential S. Afr. J. Bot. 2011 77 960 971 10.1016/j.sajb.2011.06.010 4. Balogun F.O. Tshabalala N.T. Ashafa A.O. Antidiabetic Medicinal Plants Used by the Basotho Tribe of Eastern Free State: A Review J. Diabetes Res. 2016 2016 4602820 10.1155/2016/4602820 27437404 5. Patrick-Iwuanyanwu K.C. Nkpaa K.W. Toxicity Effect of Sub-Chronic Oral Administration of Class Bitters®-A Polyherbal Formula on Serum Electrolytes and Hematological Indices in Male Wistar Albino Rats J. Xenobiot. 2015 5 5369 10.4081/xeno.2015.5369 30701040 6. Watcho P. Zelefack F. Ngouela S. Nguelefack T.B. Ngouela S. Telefo P.B. Kamtchouing P. Tsamo E. Kamanyi A. Effects of the Aqueous and Hexane Extracts of Mondia whitei o sexual behaviour and some fertility parameters of sexually inexperienced male rats Afr. J. Tradit. CAM 2007 4 37 46 10.4314/ajtcam.v4i1.31190 7. Gundidza A. Mmbengwa G.M. Magwa V.M. Ramalivhana M.L. Mukwevho N.J. Ndaradzi N.T. Samie W. Aphrodisiac properties of some Zimbabwean medicinal plants formulations Afr. J. Biotechnol. 2009 8 6402 6407 8. Watcho P. Donfack M.M. Zelefack F. Telesphore A. Nguelefack B. Kamtchouing W.S. Tsamo P. Kamanyi E. Effects of the hexane extract of Mondia whitei on the reproductive organs of male rats Afr. J. Tradit. CAM 2005 2 302 311 10.4314/ajtcam.v2i3.31129 9. Quasie O. Martey O.N.K. Nyarko A.K. Gbewonyo W.S.K. Okine L.K.N. Modulation of penile erection in rabbits by Mondia whitei: Possible mechanism of action Afr. J. Tradit. CAM 2010 7 241 252 10.4314/ajtcam.v7i3.54783 10. Patnam R. Kadali S.S. Koumaglo K.H. Roy R. A chlorinated coumarinolignan from the African medicinal plant, Mondia whitei Phytochemistry 2005 66 683 686 10.1016/j.phytochem.2004.11.012 15771890 11. Koorbanally N.A. Mulholland D.A. Crouch N.R. Isolation of Isovanillin from aromatic roots of the medicinal African Liane Mondia whitei J. Herbs Spices Med. Plants 2008 7 37 43 10.1300/J044v07n03_05 12. Mukonyi I. Ndiege K.W. 2-hydroxy-4-methoxybenzaldehyde: Aromatic taste modifying compound from Mondia whytei skeels Bull. Chem. Soc. Ethiop. 2001 15 137 141 13. Wang J. Zhao J. Gao H. Zhou L. Liu Z. Chen Y. Sui P. Antimicrobial and antioxidant activities of the root bark essential oil of Periploca sepium and its main component 2-hydroxy-4-methoxybenzaldehyde Molecules 2010 15 5807 5817 10.3390/molecules15085807 20736908 14. Yang J.Y. Park J.H. Lee M.J. Lee J.H. Lee H.S. Antimicrobial Effects of 7,8-Dihydroxy-6-Methoxycoumarin and 7-Hydroxy-6-Methoxycoumarin Analogues against Foodborne Pathogens and the Antimicrobial Mechanisms Associated with Membrane Permeability J. Food Prot. 2017 80 1784 1790 10.4315/0362-028X.JFP-17-050 15. Izadiyan P. Hemmateenejad B. Multi-response optimization of factors affecting ultrasonic assisted extraction from Iranian basil using central composite design Food Chem. 2016 190 864 870 10.1016/j.foodchem.2015.06.036 26213050 16. Chen X. Ding J. Ji D. He S. Ma H. Optimization of ultrasonic-assisted extraction conditions for bioactive components from coffee leaves using the Taguchi design and response surface methodology J. Food Sci. 2020 85 1742 1751 10.1111/1750-3841.15111 32449951 17. Wahid Z. Nadir N. Improvement of One Factor at a Time Through Design of Experiments 1 World Appl. Sci. J. 2013 21 56 61 18. Jones B. Nachtsheim C.J. Definitive screening designs with added two-level categorical factors J. Qual. Technol. 2013 45 121 129 10.1080/00224065.2013.11917921 19. Chokwe R.C. Dube S. Nindi M.M. Development of a Quantitative Method for Analysis of compounds found in Mondia whitei using HPLC-DAD Molecules 2020 25 4451 10.3390/pr9111864 20. Hocharoen L. Noppiboon S. Kitsubun P. Process Characterization by Definitive Screening Design Approach on DNA Vaccine Production Bioeng. Biotechnol. 2020 8 1164 10.3389/fbioe.2020.574809 21. Sulaiman I.S.C. Basri M. Masoumi H.R.F. Chee W.J. Efliza S. Effects of temperature, time, and solvent ratio on the extraction of phenolic compounds and the anti-radical activity of Clinacanthus nutans Lindau leaves by response surface methodology Chem. Cent. J. 2017 11 54 10.1186/s13065-017-0285-1 29086900 22. Belwal T. Pandey A. Bhatt I.D. Rawal R.S. Optimized microwave assisted extraction (MAE) of alkaloids and polyphenols from Berberis roots using multiple-component analysis Sci. Rep. 2020 10 917 10.1038/s41598-020-57585-8 31969583

PMC009xxxxxx/PMC9099650.txt

==== Front Sensors (Basel) Sensors (Basel) sensors Sensors (Basel, Switzerland) 1424-8220 MDPI 10.3390/s22093487 sensors-22-03487 Article Machine Learning-Based View Synthesis in Fourier Lightfield Microscopy https://orcid.org/0000-0002-6000-3655 Rostan Julen 1 Incardona Nicolo 2* https://orcid.org/0000-0001-5524-5302 Sanchez-Ortiga Emilio 23 https://orcid.org/0000-0002-1449-8976 Martinez-Corral Manuel 2 https://orcid.org/0000-0001-6984-5173 Latorre-Carmona Pedro 1 Bennamoun Mohammed Academic Editor Zhang Liang Academic Editor Feng Mingtao Academic Editor 1 Departamento de Ingenieria Informatica, Universidad de Burgos, E09006 Burgos, Spain; jrs1002@alu.ubu.es (J.R.); plcarmona@ubu.es (P.L.-C.) 2 3D Imaging and Display Laboratory, Department of Optics, University of Valencia, E46100 Burjassot, Spain; emilio.sanchez@uv.es (E.S.-O.); manuel.martinez@uv.es (M.M.-C.) 3 School of Science, Universidad Europea de Valencia, Passeig de l’Albereda, 7, E46010 Valencia, Spain * Correspondence: nicolo.incardona@uv.es 03 5 2022 5 2022 22 9 348731 3 2022 28 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Current interest in Fourier lightfield microscopy is increasing, due to its ability to acquire 3D images of thick dynamic samples. This technique is based on simultaneously capturing, in a single shot, and with a monocular setup, a number of orthographic perspective views of 3D microscopic samples. An essential feature of Fourier lightfield microscopy is that the number of acquired views is low, due to the trade-off relationship existing between the number of views and their corresponding lateral resolution. Therefore, it is important to have a tool for the generation of a high number of synthesized view images, without compromising their lateral resolution. In this context we investigate here the use of a neural radiance field view synthesis method, originally developed for its use with macroscopic scenes acquired with a moving (or an array of static) digital camera(s), for its application to the images acquired with a Fourier lightfield microscope. The results obtained and presented in this paper are analyzed in terms of lateral resolution and of continuous and realistic parallax. We show that, in terms of these requirements, the proposed technique works efficiently in the case of the epi-illumination microscopy mode. Fourier lightfield microscopy view synthesis neural radiance fields 3D microscopy Ministerio de Ciencia, Innovacion y Universidades (Spain) and European Regional Development FundRTI2018-099041-B-I00 Generalitat Valenciana (Spain)PROMETEO/2019/048 This research was funded by Grant RTI2018-099041-B-I00, which is co-founded by the Ministerio de Ciencia, Innovacion y Universidades (Spain), by the European Regional Development Fund, and by Generalitat Valenciana (Spain) under Grant PROMETEO/2019/048. ==== Body pmc1. Introduction The realistic generation of different views of an existing scene not only requires understading its three-dimensional (3D) geometry, but also modeling the complex and inherent viewpoint-dependent information resulting from a series of highly non-linear light propagation processes. In these situations, the 5D plenoptic function is a good strategy to model the radiance and direction of the light rays passing through each point in space [1,2,3]. The main difficulty in this approach, however, lies in adequately sampling the plenoptic function. As a recent alternative to this type of strategy, the novel view synthesis (NVS) framework aims to approximate the plenoptic function from only a sparse group of observations. One possible way of accomplishing this is through the use of a small set of images acquired from different viewpoints. This group of techniques is currently obtaining high-quality results, in particular when being used in a deep learning framework [4]. Their aim is to obtain new views that did not previously exist. Sometimes, other useful scene information might also be subsequently obtained—for instance, the depth maps or the 3D location of the acquisition cameras. Once the 3D representation is inferred, one might select any new (x,y,z) coordinate to generate a new view from it. Many of these methods allow renders and models of, for instance, people or monuments, therefore enhancing virtual reality (VR) capabilities [5,6]. The neural radiance field (NeRF) method [7] is based on the concept of neural rendering, which aims to substitute some of the rendering stages for a specific type of neural network surrogate. In this framework, once the 3D model is obtained, new views can be synthesized. NeRF aims at using the concept of high-quality real-time ray tracing [8]. The method presented in [7] may be able to represent a non-dynamic scene as a continuous 5D function that outputs the radiance emitted in each direction (θ,ϕ) at each point (x,y,z) in space, and a density parameter at each point which controls how much radiance is accumulated by a ray passing through the 3D (x,y,z) position. Here, θ and ϕ are the zenith and azimuth angles codifying the direction of a unit vector in 3D space. This method optimizes a multilayer perceptron (MLP) to represent this function by a regression procedure from a single 5D coordinate point (x,y,z,θ,ϕ) to a single volume density and view-dependent RGB color. The original NeRF version makes use of a software called COLMAP [9,10]. COLMAP is a general-purpose structure-from-motion (SfM) and multi-view stereo (MVS) framework (pipeline). It has a varied group of reconstruction methods implemented, allowing for the estimation of the camera system acquisition parameters. At present, the amount of methods based on (or inspired by) the original NeRF proposal have substantially increased. They have tried to tackle some of the drawbacks of the original theory. Some examples of these new methods include RegNeRF [11], where a series of regularization terms are included to help improve the scene reconstruction quality, InstantNeRF [12], which aimed to increase the convergence speed of the method, pixelNeRF [13], where the main problem to solve is related to the possibility to have one, two, or just a few input images, and D-NeRF [14], which aimed to reconstruct highly dynamic scenes (considering rigid and non-rigid motions), to cite only a few cases. More recently, Wang et al. [15] presented NeRF--, a more versatile version of the method proposed in [7], which can be applied without any prior knowledge of the acquisition system parameters. This was possible due to the use of a particular type of joint optimization strategy applied during the training process. Some further details about NeRF and NeRF-- will be given in Section 2. The aim of this paper is to show the utility of NeRF-- in Fourier lightfield microscopy (FLMic) [16,17,18,19,20,21], which is a promising technique for the rapid acquisition of 3D information of microscopic samples. In fact, this system is able to acquire angular and spatial information of the sample in a single shot. This makes FLMic specially suited for the study of dynamic processes in 3D microscopy, such as neural activity [22,23] or single-molecule localization [24]. The lightfield scattered or emitted by the sample can be computationally retrieved through different approaches. The 3D reconstruction of the sample can be computed based on back-propagation algorithms [25], deconvolution-based techniques [26,27], or even deep learning methodologies [28]. There are substantial differences between the conventional perspective images and the ones provided by FLMic. NeRF-- was developed and applied to images that show a conical perspective; that is, the magnification decreases proportionally to the distance. They also all have a positive disparity, which decreases with the distance and is zero for very far-away objects—ideally placed at the infinite. On the contrary, the elemental images captured with FLMic are orthographic, have no disparity (or parallax) at the object plane and, as a consequence, they may have a positive or negative disparity. In this context, we demonstrate the synthesis of elemental images in Fourier lightfield microscopy. This has great utility in the display of 3D microscopic images, since it allows the off-line observation of 3D samples from many different view points with smooth and realistic perspective changes. In this paper, we demonstrate the applicability of NeRF-- to the microscopic images captured with FLMic. We apply the method to different kinds of samples and with different configurations of the optical system. Then, we discuss the quality of the results, and their dependence on the sample type. 2. Materials and Methods The particular details of the acquisition and processing tools and methodologies are as follows. Section 2.1 is dedicated to describing the main properties of FLMic. Section 2.2 will give a brief summary about the particular NVS method applied on the images acquired by this microscope. 2.1. Fourier Lightfield Microscopy Lightfield microscopy is an active area of research whose origins can be traced back to the seminal work by Levoy et al. in [29]. FLMic is an upgrade of lightfield microscopy, based on the spatial multiplexing of the angular information of the sample, in such a way that the captured image (named here as the integral image) consists of several elemental images (EIs), each one of them obtained from a different perspective angle. This is achieved by inserting a microlens array (MLA) at the aperture stop (AS) of a telecentric microscope objective (MO). As shown in Figure 1, this telecentricity allows for the acquisition of orthographic EIs, whose lateral magnification does not depend on the object position. Another singular feature is that the object plane is captured with no parallax; i.e., all the EIs are identical for that plane. The number of EIs along one direction, NEI, is the result of dividing the AS diameter by the microlenses’ pitch (i.e., the distance between the microlenses’ centers). Plane objects in the space between the object plane and the MO are captured with positive parallax, while the rest are captured with negative parallax. Note that, in conventional perspective setups, the plane with no parallax is placed at the infinite. One of the main drawbacks of FLMic is that the angular information acquisition is achieved at the cost of reducing the lateral resolution of the EIs. This unavoidable effect occurs due to the spatial multiplexing of the aperture stop, which reduces the effective spatial bandwidth for each EI. In particular, the lateral resolution limit of captured EIs can be expressed as:(1) ρ=λNEI2NA, where λ is the wavelength of the light scattered (or emitted in the case of fluorescence) by the sample, NEI is the number of EIs along the corresponding direction, and NA is the numerical aperture of the host microscope objective. Equation (1) shows a loss in lateral resolution (with respect to a conventional microscope) by a factor NEI, and, therefore, a direct trade-off between the lateral and the angular resolutions of the system. Furthermore, the number of EIs is directly related to the 3D information that can be computationally extracted from the sample. For instance, in a set of refocused images (namely, z-stack) assessed with standard back-propagation algorithms, the thickness of the refocused region is inversely proportional to NEI. The same tendency occurs when applying deconvolution techniques to provide computational optical sectioning, as the thickness of the computed optical sections decreases (i.e., the optical sectioning capability is improved) when the number of EIs increases [25]. In a realistic implementation of FLMic, the microlenses are not placed directly at the aperture stop, but in a plane conjugated with it. The standard scheme of FLMic is shown in Figure 2. An afocal relay system (R1–R2) projects an image of the aperture stop onto the lens array. In addition, a field stop (FS) is used to avoid the vignetting and the overlapping between EIs. 2.2. Neural Radiance Field, with Simultaneous Inference of Calibration Parameters The number of EIs plays an important role during both the acquisition and the computational post-processing processes. We must also bear in mind that the 3D information obtained from a scene increases in quality with the increase in the number of EIs processed. That is the reason why NVS methods are attractive in applications that depend on the number of EIs processed. Given a set of images I={I1,…,IN} captured from N sparse viewpoints of a scene, with their associated (intrinsic and extrinsic) camera parameters Π={π1,…,πN}, the goal of NVS is to come up with a scene representation that enables the generation of realistic images from novel, unseen viewpoints. The neural radiance field method [7] represents an ideal tool to optimize the lateral resolution as well as the thickness of the reconstructed axial regions. Nevertheless, as previously stated, NeRF needs a previous calibration step (inference of the intrinsic and extrinsic parameters of the acquisition set-up used). Sometimes, these parameters are difficult or even impossible to obtain. Hence, we considered the use of NeRF--, since it is able to infer the camera parameters during a global optimization process. The creators of NeRF adopt a continuous function for constructing a volumetric representation of the scene from a sparse group of input views. In essence, it models the view-dependent appearance of the 3D space using a continuous function FΘ:(x,u)→(c,ρ), parameterized by a multi-layer perceptron (MLP). The function maps a 3D location x=(x,y,z), together with a viewing direction u=(ϑ,ψ), to a radiance color c=(r,g,b) and a density value, ρ. Rendering an image in a NeRF framework implies that the color at each pixel p=(px,py) on the image plane (I^i) is obtained by a so-called rendering function, which aggregates the radiance along a ray coming from the camera position oi and passing through a specific pixel (p) into a volume [3]:(2) I^i(p)=∫hnhfτ(h)ρ(r(h))c(r(h),u)dh where τ(h)=e−∫hnhρ(r(s))ds denotes the accumulated transmittance along the ray, i.e., the probability of the ray travelling from hn to h without hitting any other particle, and r(h)=o+hu denotes the camera ray that starts from camera origin o and passes through p, controlled by the camera parameter πi, with near and far bounds hn and hf. With this implicit scene representation FΘ(x,u), NeRF can be trained by minimizing the photometric error between the observed views and the synthesized ones, under known camera parameters: E=∑i=1N∥Ii−I^i∥22, where Θ*=argminΘ[E(I^|I,Π)]. I^ denotes the set of synthesized images {I^1,…,I^N}. The discretization approach corresponding to Equation (2), and other particular details related to the implementation strategy, can be found in [7]. In contrast to the work in [7], the authors in [15] show that the pre-processing step of estimating the camera parameters πi of the input images is unnecessary. Unlike the training setup of the original NeRF, the authors only assume a set of RGB images I as inputs, with unknown camera parameters, and they seek to jointly optimize the camera parameters and scene representation during the training. Mathematically, this can be written as: Θ*,Π*=argminΘ,Π[E(I^,Π*|I)]. Figure 3 shows a brief scheme of the NeRF-- method (further graphical and mathematical details can be found in [15]). On the other hand, Figure 4 shows an example of the view synthesis results that the NeRF-- method obtains. Starting from a few images (six, in this case), NeRF-- generates synthesized views corresponding to other camera positions. As we can see, the results look realistic, with no color or geometrical artifacts whatsoever. Input views in Figure 4(left) were obtained from the web page created by the authors of the NeRF method [30]. We should emphasize that the above method (and potential variants that might appear in the future) could be good options as post-processing algorithms for Fourier lightfield data, mainly because they are able to reliably encode the scene information that is being acquired, and do not need other, more human-involved algorithms, such as in [31]. 3. Results and Discussion A series of samples were prepared and subsequently acquired by an FLMic built in our laboratory in open configuration. In all the experiments, we used an MLA with pitch p = 1.0 mm and focal length fL = 6.44 mm, and a sensor with 2560×1920 square pixels with size δ = 2.2 μm. Figure 5 shows representative examples of them. The first sample (top left) is composed of cotton fibers stained with the fluorescent ink of a highlighter. The second sample (top right) is the condenser of an electronic circuit. The third (bottom) corresponds to a dried shark skin tissue. We will call these samples fiber, chip, and shark, respectively, from now on. Table 1 summarizes the main features of the FLMic setups used for the different acquisition sessions. The focal length of the second relay lens (fR2=100 mm) is not specified in the table because it is the same for all the samples. Looking at both Figure 5 and Table 1, we can correlate the appearance of each sample with the illumination technique used to observe it. The fiber sample was imaged in epi-fluorescence mode; that is, it was illuminated through the objective by a monochromatic beam (λ=480 nm). The same objective collected the light emitted by the fluorescent fibers, producing an integral image in the sensor plane after blocking out the illumination beam by means of a dichroic mirror. The chip sample was epi-illuminated in reflection mode with white light. The background is the printed circuit board (PCB) of the electronic chip: since this sample is not transparent, it was observed with reflected light. Finally, the shark sample shows a yellowish background because it was trans-illuminated with white light. Note that we have used different types of illumination techniques, aiming to test the performance of the algorithms for a diverse group of samples. The reader should also bear in mind that, in the integral images of fiber and chip, despite being NEI = 9, only five complete EIs can be seen in a row. This is because the exit pupil (the image of the AS through the relay system) was greater than the sensor we used. This fact has impacts on the lateral resolution of the captured EIs, but no impact on the performance of the view synthesis algorithm. Considering only the complete EIs (i.e., those that can be seen entirely), the configuration of the integral image for the fiber and chip samples is four–five–four: four EIs in the first row, five in the second row, and four in the third row. In the integral image of shark, we have NEI = 4. As this is an even number, in the central row (corresponding to the diameter of the exit pupil), we have three complete EIs, and two half EIs. Therefore, the integral image configuration is two–three–two. NeRF-- was applied, for each sample, using the acquired complete EIs as training images. The output is the synthesized images. Neural network convergence was analyzed based on the 2D inferred positions of the acquisition sensor. Those inferred positions accurately reproduce the hexagonal geometry of the MLA used for the image capture. The number of epochs considered was 500 for the fiber, and 1000 for the other two samples. Two hundred (200) new perspective images per scene were synthesized. To perform the tests, Google Colab was used, since it offers a greater computational power than our computers. The training time was 12 min for the fiber, 20 min for the chip, and 9 min for the shark. The test with the chip was repeated with a laptop with an Intel i7-6700HQ CPU and a Nvidia GTX 960M GPU. The training time in this case was 50 min. Figure 6 shows the following for the three selected samples: one of the captured EIs (left); one of the synthesized views (center). Comparison between the first and second columns shows us that the images synthesized by NeRF-- fully preserve the resolution in case of epi-illuminated sparse fluorescent samples, but suffer from a slight fall in resolution in the case of epi-illumination reflections in the brightfield samples. On the contrary, the resolution of the synthesized image in the case of the trans-illumination architecture is much poorer. The right-side column shows that the inferred capturing camera array reproduces the hexagonal structure of the physical MLA with good accuracy. Next, in Figure 7, we show a series of images extracted from the 200 images on the fiber specimen generated by NeRF--. A thorough observation of the synthesized images demonstrates the homogeneous resolution of all the images, and the small changes of parallax between neighbor ones. Take into account that, in fact, the shown images are not neighbors, since there are still eight images between any two of those shown. The maximum parallax is achieved between images 1 and 100. In order to gain a deeper insight into the utility of the proposed approach, it would be useful to generate some videos where the parallaxes of captured EIs, and the corresponding synthesized views, are compared. In this sense, we have prepared Video S1, which can be found in the Supplementary Material. In that video, we compare two movies. We show, on the left side, a movie whose frames are the captured EIs of the fibers sample, and on the right side, a movie with the 200 synthesized images. This is the best result achieved here. From the movie, we can confirm that all the synthesized images have been generated without any kind of resolution degradation. The perspective change is fairly smooth and continuous and, importantly, the occlusions are very well-resolved. In Video S2, also given in the Supplementary Material, we show the movies corresponding to the chip sample. In this case, the synthesized views suffer from a slight lost in resolution, but the perspective change is very continuous and realistic. Finally, in Video S3, we show the movies corresponding to the shark sample. Clearly, this is the worst case in terms of resolution. However, it can still be of utility, since the synthesized perspective views are very realistic, and the occlusions are well resolved. The reconstruction artifacts of Videos S2 and S3 might depend mostly on the nature of NeRF--. The estimation of the 3D model is difficult in images with low contrast and an absence of features, which are intrinsic characteristics of these microscopic images (mainly those of shark). In these conditions, the algorithm fails to predict a good 3D model of the sample, leading to artifacts in the reconstruction. This can be observed also in the original results presented by the authors of NeRF in [30]: in those regions of the scenes that present lack of features, defocus, or low contrast, the resolution of the synthesized images is substantially decreased. This effect seems to be more evident in images with a low number of pixels (455 × 455 pixels per EI). In addition, for the case of shark, the number of EIs is lower than that for the other two samples, which means that the model was trained with less images. This, along with the repetitive pattern and the typical background of brightfield images, might lead to a bad estimation of the 3D model, and hence to synthesized images with poorer resolutions and more artifacts. Finally, to assess the view synthesis quality, we have made a new computer experiment in which, for the case of the chip sample, we have trained NeRF-- with all the EIs but the central one. The central EI is, therefore, used to assess the view synthesis quality. Then, we compared the EI acquired with the central microlens and the corresponding synthesized view in terms of the peak signal-to-noise ratio (PSNR). We obtained a value of PSNR=34.82, which is similar to the one obtained when the test image is not omitted in the training, and indicates a good image synthesis quality. 4. Conclusions Fourier lightfield microscopy (FLMic) features a promising method for the fast capture of the plenoptic map of microscopic samples with high lateral resolution. The price paid is a low number of perspective images, which somehow compromises its utility in terms of display applications. This problem comes from the trade-off between the number of captured views and the effective numerical aperture (NA) value. Thus, a computational method that preserves the resolution but synthesizes new views is of great importance. In this sense, this paper presents, to the best of our knowledge, the first application of a neural radiance field-type method (NeRF--) for the synthesis of new views in FLMic. The application of the NeRF-- concept to FLMic is not trivial, due to some of its features, such as: (a) the telecentric nature of FLMic; (b) the object plane is acquired without any parallax, and the other planes show positive or negative parallax; (c) the particular type of illumination technique and the type of sample. In this sense, we found that, in the case of epi-illumination microscopy, the technique provides remarkable results, with only slight or no losses in lateral resolution, depending on the structural information of the images, but with continuous parallax and well-resolved occlusions. This can be of great utility for microscopists, since they can capture 3D dynamic samples in real time with the FLMic, and perform a post-processing application of NeRF-- for a realistic off-line observation of the sample from a continuous viewpoint. Future work will be focused on the development of a new mathematical framework that is able to take into account the particular details of the Fourier lightfield microscope used, in order to improve the quality of the reconstruction. Acknowledgments The authors acknowledge María del Carmen Fuentes-Albero, Daniel García Párraga, and Marta Muñoz Baquero from the Biology Department of Avanqua-Oceanogràfic S.L for providing the shark sample. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/s22093487/s1, Video S1: Animation showing the fiber sample from the inferred and synthesized views; Video S2: Animation showing the chip sample from the inferred and synthesized views; Video S3: Animation showing the shark sample from the inferred and synthesized views. Click here for additional data file. Author Contributions Conceptualization, M.M.-C., N.I. and P.L.-C.; methodology, all authors; software, J.R.; validation, N.I. and E.S.-O.; formal analysis, J.R., N.I. and M.M.-C.; investigation, J.R., N.I. and E.S.-O.; resources, N.I. and E.S.-O.; data curation, J.R. and N.I.; writing—original draft preparation, N.I., E.S.-O. and P.L.-C.; writing—review and editing, M.M.-C.; visualization, J.R., N.I. and M.M.-C.; supervision, M.M.-C. and P.L.-C.; project administration, M.M.-C. and P.L.-C.; funding acquisition, M.M.-C. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Data is available upon request to the authors on a reasonable basis. Conflicts of Interest The authors declare no conflict of interest. Abbreviations The following abbreviations are used in this manuscript: 3D Three-dimensional AS Aperture stop EI Elemental image FLMic Fourier lightfield microscopy IEP Inferred extrinsic parameters MLA Microlens array MLP Multilayer perceptron NeRF Neural radiance field NA Numerical aperture NVS Novel view synthesis PCB Printed circuit board R Relay SfM Structure-from-motion VR Virtual reality Figure 1 Basic scheme of the Fourier lightfield microscope. The MLA is placed at the aperture stop (AS) of the microscope objective (MO). Note that the object is assumed to be of a 3D nature. Thus, only one section of it (the one placed at the so-called object plane) is conjugated with the sensor. Figure 2 Fourier lightfield microscope optical scheme. The afocal relay R1–R2 projects an image of the aperture stop (AS) onto the lens array. Between them, a field stop (FS) is placed at the plane of the intermediate image, to avoid overlapping between the EIs formed by the lens array. The sensor is placed at the back focal plane of the lens array, so that it conjugated to the object plane. Figure 3 Summary of the NeRF-- framework. A NeRF model and the camera parameters of the input images are simultaneously optimized by minimizing the photometric reconstruction error (E=∑i=1N∥Ii−I^i∥22) between the input and the reconstructed images, and a pixel, p is rendered once these camera parameters are optimized (read more details in the text body, and in [15]). Figure 4 Image synthesis example. From a reduced number of input image views, NeRF-- renders a predefined number of high-quality new ones. Original input images can be downloaded from the corresponding repository in [30]. Figure 5 Integral image of the three types of scenes considered: fiber, chip, and shark. In the pictures, we have shaded the non-complete EIs. Figure 6 (Left column) Acquired elemental image; (central column) synthesized image; (right column) 2D inferred positions of the capturing cameras (blue full dots). Figure 7 Group of 12 synthesized images extracted from the series of 200 images provided by NeRF-- for the case of the fibers sample. The labels refer to the position of each image in the order created during the generation of these images. sensors-22-03487-t001_Table 1 Table 1 Acquisition data for the different samples. First column: the microscope objective information is shown in the classical magnification/numerical aperture format. Second column: R1 lens focal length. Third column: number of EIs. Fourth column: illumination technique. Sample MO f R1 NEI Illumination Type fiber 10×/0.45 200 mm 9 Epi-Fluorescence chip 10×/0.45 200 mm 9 Epi-Reflection shark 20×/0.40 180 mm 4 Brightfield Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Landy M. Movshon J.A. The Plenoptic Function and the Elements of Early Vision Comput. Model. Vis. Process. 1991 1 3 20 2. Levoy M. Hanrahan P. Light Field Rendering Proceedings of the 23rd Annual Conference on Computer Graphics and Interactive Techniques, SIGGRAPH’96 New Orleans, LA, USA 4–9 August 1996 Association for Computing Machinery New York, NY, USA 1996 31 42 10.1145/237170.237199 3. Gortler S.J. Grzeszczuk R. Szeliski R. Cohen M.F. The Lumigraph Proceedings of the 23rd Annual Conference on Computer Graphics and Interactive Techniques, SIGGRAPH’96 New Orleans, LA, USA 4–9 August 1996 Association for Computing Machinery New York, NY, USA 1996 43 54 10.1145/237170.237200 4. Wang H. Yan B. Sang X. Chen D. Wang P. Qi S. Ye X. Guo X. Dense view synthesis for three-dimensional light-field displays based on position-guiding convolutional neural network Opt. Lasers Eng. 2022 153 106992 10.1016/j.optlaseng.2022.106992 5. Moreau A. Piasco N. Tsishkou D. Stanciulescu B. Fortelle A.d.L. LENS: Localization enhanced by NeRF synthesis Proceedings of the 5th Conference on Robot Learning London, UK 8–11 November 2021 Volume 164 1347 1356 6. Martin-Brualla R. Radwan N. Sajjadi M.S.M. Barron J.T. Dosovitskiy A. Duckworth D. NeRF in the Wild: Neural Radiance Fields for Unconstrained Photo Collections Proceedings of the 2021 Computer Vision and Pattern Recognition conference (CVPR) Online 19–25 June 2021 7. Mildenhall B. Srinivasan P.P. Tancik M. Barron J.T. Ramamoorthi R. Ng R. NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis Commun. ACM 2021 65 99 106 10.1145/3503250 8. Christen M. Ray Tracing on GPU Master’s Thesis University of Applied Sciences Basel (FHBB) Basel, Switzerland 19 January 2005 9. Schönberger J. Zheng E. Pollefeys M. Frahm J.M. Pixelwise View Selection for Unstructured Multi-View Stereo Proceedings of the European Conference on Computer Vision (ECCV) Amsterdam, The Netherlands 11–14 October 2016 Volume 9907 10.1007/978-3-319-46487-9_31 10. Schönberger J.L. Frahm J.M. Structure-from-Motion Revisited Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Las Vegas, NV, USA 27–30 June 2016 4104 4113 10.1109/CVPR.2016.445 11. Niemeyer M. Barron J.T. Mildenhall B. Sajjadi M.S.M. Geiger A. Radwan N. RegNeRF: Regularizing Neural Radiance Fields for View Synthesis from Sparse Inputs arXiv 2021 2112.00724 12. Müller T. Evans A. Schied C. Keller A. Instant Neural Graphics Primitives with a Multiresolution Hash Encoding arXiv 2022 2201.05989 13. Yu A. Ye V. Tancik M. Kanazawa A. pixelNeRF: Neural Radiance Fields from One or Few Images Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Nashville, TN, USA 20–25 June 2021 4576 4585 10.1109/CVPR46437.2021.00455 14. Pumarola A. Corona E. Pons-Moll G. Moreno-Noguer F. D-NeRF: Neural Radiance Fields for Dynamic Scenes Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Nashville, TN, USA 20–25 June 2021 10313 10322 10.1109/CVPR46437.2021.01018 15. Wang Z. Wu S. Xie W. Chen M. Prisacariu V. NeRF--: Neural Radiance Fields without Known Camera Parameters arXiv 2021 2102.07064 16. Llavador A. Sola-Pikabea J. Saavedra G. Javidi B. Martínez-Corral M. Resolution improvements in integral microscopy with Fourier plane recording Opt. Express 2016 24 20792 20798 10.1364/OE.24.020792 27607682 17. Scrofani G. Sola-Pikabea J. Llavador A. Sanchez-Ortiga E. Barreiro J.C. Saavedra G. Garcia-Sucerquia J. Martínez-Corral M. FIMic: Design for ultimate 3D-integral microscopy of in-vivo biological samples Biomed. Opt. Express 2018 9 335 346 10.1364/BOE.9.000335 29359107 18. Guo C. Liu W. Hua X. Li H. Jia S. Fourier light-field microscopy Opt. Express 2019 27 25573 25594 10.1364/OE.27.025573 31510428 19. Scrofani G. Saavedra G. Martínez-Corral M. Sánchez-Ortiga E. Three-dimensional real-time darkfield imaging through Fourier lightfield microscopy Opt. Express 2020 28 30513 30519 10.1364/OE.404961 33115051 20. Hua X. Liu W. Jia S. High-resolution Fourier light-field microscopy for volumetric multi-color live-cell imaging Optica 2021 8 614 620 10.1364/OPTICA.419236 34327282 21. Galdon L. Yun H. Saavedra G. Garcia-Sucerquia J. Barreiro J.C. Martinez-Corral M. Sanchez-Ortiga E. Handheld and Cost-Effective Fourier Lightfield Microscope Sensors 2022 22 1459 10.3390/s22041459 35214359 22. Cong L. Wang Z. Chai Y. Hang W. Shang C. Yang W. Bai L. Du J. Wang K. Wen Q. Rapid whole brain imaging of neural activity in freely behaving larval zebrafish (Danio rerio) Elife 2017 6 e28158 10.7554/eLife.28158 28930070 23. Yoon Y.G. Wang Z. Pak N. Park D. Dai P. Kang J.S. Suk H.J. Symvoulidis P. Guner-Ataman B. Wang K. Sparse decomposition light-field microscopy for high speed imaging of neuronal activity Optica 2020 7 1457 1468 10.1364/OPTICA.392805 24. Sims R.R. Rehman S.A. Lenz M.O. Benaissa S.I. Bruggeman E. Clark A. Sanders E.W. Ponjavic A. Muresan L. Lee S.F. Single molecule light field microscopy Optica 2020 7 1065 1072 10.1364/OPTICA.397172 25. Sánchez-Ortiga E. Scrofani G. Saavedra G. Martinez-Corral M. Optical Sectioning Microscopy Through Single-Shot Lightfield Protocol IEEE Access 2020 8 14944 14952 10.1109/ACCESS.2020.2966323 26. Sánchez-Ortiga E. Llavador A. Saavedra G. García-Sucerquia J. Martínez-Corral M. Optical sectioning with a Wiener-like filter in Fourier integral imaging microscopy Appl. Phys. Lett. 2018 113 214101 10.1063/1.5049755 27. Stefanoiu A. Scrofani G. Saavedra G. Martínez-Corral M. Lasser T. What about computational super-resolution in fluorescence Fourier light field microscopy? Opt. Express 2020 28 16554 16568 10.1364/OE.391189 32549475 28. Vizcaino J.P. Wang Z. Symvoulidis P. Favaro P. Guner-Ataman B. Boyden E.S. Lasser T. Real-time light field 3D microscopy via sparsity-driven learned deconvolution Proceedings of the 2021 IEEE International Conference on Computational Photography (ICCP) Haifa, Israel 23–25 May 2021 1 11 29. Levoy M. Ng R. Adams A. Footer M. Horowitz M. Light Field Microscopy ACM Trans. Graph. 2006 25 924 934 10.1145/1141911.1141976 30. Representing Scenes as Neural Radiance Fields for View Synthesis Available online: https://www.matthewtancik.com/nerf (accessed on 16 March 2022) 31. Broxton M. Grosenick L. Yang S. Cohen N. Andalman A. Deisseroth K. Levoy M. Wave optics theory and 3-D deconvolution for the light field microscope Opt. Express 2013 21 25418 25439 10.1364/OE.21.025418 24150383

PMC009xxxxxx/PMC9099651.txt

==== Front J Clin Med J Clin Med jcm Journal of Clinical Medicine 2077-0383 MDPI 10.3390/jcm11092576 jcm-11-02576 Review Vertebral Body Tethering: Indications, Surgical Technique, and a Systematic Review of Published Results https://orcid.org/0000-0001-9114-2204 Raitio Arimatias 1 Syvänen Johanna 2 https://orcid.org/0000-0001-5200-3279 Helenius Ilkka 3* Nakajima Hideaki Academic Editor Trobisch Per Academic Editor 1 Department of Paediatric Surgery, Turku University Hospital, University of Turku, Kiinamyllynkatu 4-8, 20521 Turku, Finland; arimatias.raitio@fimnet.fi 2 Department of Paediatric Orthopaedics, Turku University Hospital, University of Turku, Kiinamyllynkatu 4-8, 20521 Turku, Finland; johanna.syvanen@tyks.fi 3 Department of Orthopaedics and Traumatology, Helsinki University Hospital, University of Helsinki, Topeliuksenkatu 5, 00260 Helsinki, Finland * Correspondence: ilkka.helenius@helsinki.fi 04 5 2022 5 2022 11 9 257622 3 2022 30 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Vertebral body tethering (VBT) represents a new surgical technique to correct idiopathic scoliosis using an anterior approach, spinal instrumentation with vertebral body screws, and a cable compressing the convexity of the curve. According to the Hueter-Volkmann principle, compression reduces and distraction increases growth on the growth plates. VBT was designed to modulate spinal growth of vertebral bodies and hence, the term ‘growth modulation’ has also been used. This review describes the indications and surgical technique of VBT. Further, a systematic review of published studies was conducted to critically evaluate the results and complications of this technique. In a total of 23 included studies on 843 patients, the preoperative main thoracic curve corrected from 49 to 23 degrees in a minimum 2 year follow-up. The complication rate of VBT was 18%. The results showed that 15% of VBT patients required reoperations for pulmonary or tether-related issues (10%) and less than 5% required conversion to spinal fusion. While the reported median-term results of VBT appear promising, long-term results of this technique are currently lacking. adolescent idiopathic scoliosis growth-friendly techniques surgery vertebral body tethering Helsinki University HospitalTurku University HospitalPaulo FoundationFinnish Paediatric Research FoundationFinska LäkaresällskapetNuvasiveMedtronic InternationalERP-2020-12238 Stryker (Clinical Research Institute HUCH)20631 University of HelsinkiThe authors declare research grants from Clinical Research Institute HUCH and Helsinki and Turku University Hospitals (Finnish State Funding). Raitio reports research grants from Paulo Foundation and Helenius reports institutional funding from the Finnish Paediatric Research Foundation, Finska Läkaresällskapet, Medtronic International, Stryker, and Nuvasive. This study was funded by Medtronic International (grant number ERP-2020-12238), Stryker (Clinical Research institute HUCH, project number 20631). Open access funding provided by University of Helsinki. ==== Body pmc1. Introduction Adolescent idiopathic scoliosis (AIS) is a three-dimensional deformity including a lateral deviation of the spine, reduced thoracic kyphosis, and rotation of the vertebral bodies. A curve of 45 degrees or higher is typically regarded as an indication to surgical treatment as these curves typically continue to progress even in skeletally mature patients [1]. Additionally, thoracic curves of over 50 degrees are associated with reduced lung volumes [2]. Three-dimensional correction of scoliosis and continued growth should be the aim of the treatment of spinal deformity on a growing child [3]. Posterior spinal fusion with pedicle screw instrumentation has been the traditional method to address these curves [4]. Normal lung development is dependent on the length of the thoracic spine and its final length is closely related to the lung volume obtained at skeletal maturity [5]. A recommended minimum length of the thoracic spine before posterior fusion is 22 cm [6,7,8]. Additional length obtained from correction of spinal deformity averages about 25 mm in normal AIS [9]. Spinal fusion provides sustainable long-term outcomes but is associated with reduced spinal mobility [10] and hence reduced functional outcomes as compared with the normal population [11]. On the other hand, it leads to an irreversible stage of permanent spinal fusion and straining of the remaining mobile segment due to reduced spinal mobility [12,13]. These disadvantages have led surgeons to investigate other methods to correct adolescent idiopathic scoliosis without spinal fusion. It is known that every human vertebral body between C3 and L5 has a growth plate (apophysis) on its upper and lower endplates. Both endochondral ossification (length) and appositional ossification (volume) lead to growth of the spine [14]. According to the Hueter-Volkman principle, distraction of the growth plate promotes and compression inhibits growth [15]. To control this growth, surgeons have attempted asymmetrical hemiepiphyseodesis to the spine, but this has remained unpredictable [14,16]. Asymmetrical growth plate inhibition with staples or unilateral plates has been used for decades in mechanical axis deviations (e.g., genu valgum or varum) of lower extremities in growing children [17]. A similar technique was applied by Betz and colleagues in the spine using stapling over the disc and growth plates. However, this was only successful in thoracic curves less than 35 degrees, which are typically treated with a brace on a growing child [18]. Additionally, movement of the spine often led to problematic loosening of the vertebral implants extending over the intervertebral disc [19]. Spinal tethering is the newest method to address scoliosis deformity correction three-dimensionally without fusion in preadolescent patients. It is based on the Hueter-Volkman principle and utilizes the patient’s own spinal growth to improve the initial correction rate after surgery. Currently, spinal tethering is mainly indicated only in children with suitable growth remaining (Sanders 2 to 5), while spinal arthrodesis can be performed whenever a minimum of 22 cm of the thoracic spine length has been achieved. Additionally, spinal tethering maintains spinal mobility and can be converted to anterior or posterior spinal fusion if necessary. A systematic review of the literature was conducted to analyze the published results and complications of VBT. Additionally, indications and technical considerations of this technique are described. 2. Methods 2.1. Identification and Selection of Studies A comprehensive search of the published literature in PubMed and EMBASE databases was performed based on PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [20] using ‘vertebral body tethering’ as a keyword. 2.2. Inclusion and Exclusion Criteria All articles reporting a minimum of one-year follow-up results of AVBT published up to February 28, 2022, were included. Non-English language papers, animal studies, and case reports (<3 patients) were excluded. 2.3. Data Extraction and Analysis Identified papers were independently reviewed by two authors (AR and JS) with final selection approved by the senior author (IH). The data on patient demographics, pre- and postoperative scoliosis curves, duration of surgery, intraoperative blood loss, length of follow-up, and complications were extracted from the original publications. 2.4. Statistical Analysis Analyses were performed using JMP Pro, version 16.1.0 for Windows (SAS Institute Inc., Cary, NC, USA). 3. Anterior Vertebral Body Tethering Lately, spinal tethering has become one of the options to treat AIS without spinal fusion. This method has been made possible with better understanding of spinal biomechanics, technical developments in minimally invasive techniques (endoscopic and mini-open), and improved instrument and device design [21,22,23]. The tethering system limits the progression of scoliotic deformity by mechanically restraining the remaining spinal growth internally [19]. Spinal tethering is carried out using an anterior thoracoscopic or mini-open thoracolumbar approach. One or two bicortical screws over a washer or a small plate are inserted to each vertebral body laterally. A polyethylene tetraphalate cable is used to connect these screws. Immediate correction of the deformity is obtained by compressing the cable and between the screws (typically 45–50% initial correction). Additional correction can be obtained via growth modulation of the vertebral bodies according to the Hueter-Volkmann principle (Figure 1). 3.1. Indications for Main Thoracic Curves The most well-documented indication for spinal tethering is a single major thoracic curve with non-structural lumbar and proximal thoracic curve (Lenke 1A or 1B curve in a preadolescent patient [19,23,24]. Recently, Krakow et al. [25] evaluated how many AIS patients would potentially be suitable candidates for VBT. In their study, approximately 25% of the patients fulfilled the growth parameters and curve characteristics (Lenke 1, 3, 5, or 6 curves, i.e., not including a structural upper thoracic curve) amenable to VBT. Therefore, the majority of scoliosis patients may still require posterior spinal fusion (PSF). Skeletal growth can be assessed using the hand radiograph and Sanders’ classification [26]. Patients with a relatively flexible right thoracic curve (40 to 60 degrees, bending to 30 degrees or below), rib hump of less than 20 degrees, and suitable amount of remaining growth (Sanders 3 to 4) are the ideal candidates for the tethering procedure [19,24]. 3.2. Timing of the Procedure Appropriate timing of VBT is of utmost importance. If carried out too early, the patient may undergo overcorrection (i.e., right-sided curve turns into left-sided curve). Additionally, the remaining growth modulation may not correct the curve enough and/or a tethering rupture may result if the procedure is performed too late. In cases with limited growth remaining, or no growth at all, immediate correction utilizing the mobility of the discs can be achieved intraoperatively. However, this correction may not be maintainable without substantial three-dimensional shape change of the vertebral bodies via growth [19]. Anterior shortening may help with the restoration of thoracic kyphosis, but according to the literature, this kyphosing effect seems to be minimal [27]. Alanay et al. [28] investigated the effects of skeletal maturity according to Sanders’ classification (hand radiograph) on postoperative growth modulation. They observed that growth modulation was unpredictable in Sanders 1 (prepubertal) resulting in up to 45 degrees and in Sanders 2 (start of puberty) resulting in up to 29 degrees of postoperative growth modulation. According to their findings, Sanders 3–5 were the most predictable in terms of growth modulation of VBT. In the study of Takahashi et al. [24], the average correction rate of thoracic segments was 1.8 per segment per year for the first 2 years. Significantly greater rates were observed for the Sanders stage two than the Sanders stage three cohort. In the same study, scoliosis correction correlated also with height velocity. Still, there is a lack of studies to determine the optimum timing of the procedure and the optimum amount of tension that should be placed. 3.3. Technical Considerations for Vertebral Body Tethering The procedure is carried out using a strict lateral decubitus position and single lung ventilation. Instrumentation is typically carried from end vertebra to end vertebra. The spine is accessed anteriorly with mini-open, open thoracotomy, or thoracoscopically [23]. To minimize the chest wall violation and associated deleterious effects on pulmonary function, most surgeons favor minimally invasive techniques [15]. Preoperative screw trajectory planning under C-arm fluoroscopy helps planning the portal placement. The right lung should be deflated during surgery. Parietal pleura is opened over the spine using a monopolar hook or ultrasonic sealing device such as Harmonic scalpel (Ethicon Endo-Surgery, Inc, Cincinnati, OH, USA). Segmental vessels are ligated or mobilized on the convex side. Fluoroscopic control is used to control the placement of staples and bicortical screws. A polyethylene tether is placed and tightened starting cranially. Tightening can be controlled using a force measurement. Typically, the apical segments are tightened into 300–400 Newtons and upper thoracic screws to maximum 150–200 N to prevent screw pull-out. A chest drain is typically placed and set into 10–20 cm H2O suction. Endoscopic vertebral body tethering involves a relatively long learning curve [29]. Reported operation time for AVBT ranges from 2.7 to 4.3 h with a mean of 3.8 h in this systematic review [30,31]. Intraoperative blood loss is typically minimal averaging at 180 mL, while the length of stay ranges from 3 to 5 days postoperatively [29]. Screw accuracy can be improved using CT-guided navigation, or as we have adopted, intraoperative imaging using intraoperative 3D evaluation of the screws and staples inserted before corrective maneuvers. Especially in the upper thoracic spine, the vertebral bodies are small with limited margins around the implants. If the mini-open technique with a small thoracotomy is used, the segmental vessels can be mobilized especially in the apical area, while with the thoracoscopic technique, all the segmental vessels need to be ligated. The mini-open technique also allows easier spinal manipulation in terms of derotation and tightening of the cord. On the other hand, the thoracoscopic technique might be associated with reduced postoperative pain, better cosmesis, and better pulmonary function due to less chest wall violation. However, there are no studies comparing these two approaches. It should be noted that in revision cases, spared segmental vessels might start to bleed profusely as the cable on top of them is firmly attached to them. 3.4. Indications for Thoracolumbar and Lumbar Curves Growth modulating techniques are not contraindicated in lumbar curves. However, caution needs to be taken as most techniques have been described for thoracic curves [32]. Thoracolumbar or lumbar idiopathic scoliosis (Lenke 5 type curve) may be an option for spinal tethering, as loss of spinal mobility in this area has an even greater impact on functional outcomes. However, there are few studies on this indication [33,34]. Approach includes a mini-open thoracoabdominal exposure with two incisions: one over the 10th rib and a second over the L3/4 disc. Lumbar vertebral bodies have larger diameters and the use of two screws and two cables is easier in this area. When two curves (thoracic and lumbar) are instrumented, T12 typically needs instrumentation from right (thoracic curve) and left (thoracolumbar) sides. Careful evaluation of the sagittal profile reduces the risk of flat back or decreased lordosis [32]. 4. Results Our literature search identified 163 publications after duplicates were excluded. Thirty-one papers met inclusion criteria and were selected for full text review (Table 1). After full text review of 31 articles, 23 papers met the eligibility criteria and were selected for review (Figure 2). A total of 843 patients (736, 87% women) with a mean age of 12.7 years underwent VBT and were followed-up for minimum of 2 years. 4.1. Curve Correction after AVBT in Thoracic Curves In the included studies, the mean preoperative main thoracic curve was 49 degrees, which corrected to 24 degrees in first postoperative imaging. VBT provided sustainable median-term results as the reported curves after a minimum of two-year follow-up averaged at 23 degrees. Kyphosis remained unchanged at 23 degrees. Samdani et al. [22] observed that the lumbar curves underwent spontaneous correction from 25 degrees to 7 degrees in two years. In addition, axial rotation measured by scoliometer improved from 12 degrees to 7 degrees at the latest follow-up in their cohort. Newton and coworkers recently published a follow-up study of 14 AVBT patients using biplanar radiographs (EOS) [50]. In their 3D models, seven patients (50%) showed progressive correction of scoliosis defined as ≥15 degrees scoliosis correction between postoperative and follow-up radiographs. Coronal vertebral wedging occurred at 0.11°/month in the progressive correction compared to 0.02°/month in the non-progressive group. Similarly, coronal disc wedging was more pronounced in the progressive than in the non-progressive group. They concluded that the symmetry of apical vertebrae and the height of the discs in immature patients with thoracic scoliosis could be restored. Progressive correction was dependent on the skeletal maturity. According to Takahashi et al. [24], twice as much correction occurred in the Sanders stage 2 compared to the Sanders stage 3 group. 4.2. Outcomes of Lumbar Curves and Double Curves Compared to thoracic scoliosis, there are limited studies on the correction of lumbar and double curves using VBT. A single lumbar tether seems to have a relatively high cord breakage up to 50% within two years [36,51]. Limited evidence suggests that using a double tether with double screws can reduce this risk to 16% during the first year [33]. Pehlivanoglu et al. [34] reported the outcomes of 13 patients (11.8 years at the time of surgery) undergoing both endoscopic tethering of the thoracic curves (mean preoperative curve of 48 degrees) and mini-open approaches for thoracolumbar and lumbar curves (mean curve 45 degrees). They observed an initial 64% correction of thoracic and 69% of lumbar curves with additional growth modulation resulting in 80% and 82% correction at 2 year follow-up, respectively. 4.3. Reported Complications The reported rate and nature of complications for AVBT appear acceptable. In the included studies, the complication rate was 18% with pulmonary (pneumothorax, pleural effusion) and instrumentation-related (tether breakage, overcorrection) being the most common. Reoperations related to tethering were required in 10% of cases. These included tether release(s) for overcorrection, replacing and extending tethers for breakage or curve progression, and chest tube insertions for pulmonary complications. The vast majority avoided spinal fusion, as only 4.7% of VBT patients required conversion to PSF after unsuccessful tethering. However, the published studies on outcomes and complications of AVBT are sparse. Hence, the reported rate of complications varies considerably between reports and the true complication rate remains to be established. Furthermore, long-term studies related to complication and reoperation rates are lacking. 4.4. Comparison between Spinal Fusion and Vertebral Body Tethering There was only one study comparing traditional fusion and AVBT. Newton et al. [44] compared the outcomes of AVBT and PSF using pedicle screw instrumentation at a mean of 3.5 years follow-up. The correction of major thoracic curves was significantly better in the PSF group (70%) as compared with AVBT (38%). There were nine revisions in the AVBT group including three conversions into PSF with three more pending. Twelve patients had a broken tether, but the majority (74%) of the patients in the AVBT cohort had avoided spinal fusion at the end of follow-up. Operative time was reported to be significantly shorter in AVBT than PSF while there was no difference in the length of postoperative stay [44]. Compared to AVBT, posterior spinal fusion is a permanent stage which cannot be reversed. The risk of revision after PSF remains low and is mainly related to deep surgical site infection, adding-on phenomenon, and rarely on pseudoarthrosis. The revision risk after VBT appears acceptable in the light of these comparisons given that PSF with pedicle screws is doable with small modifications and probably with similar outcomes than in primary surgery. 4.5. Spinal Mobility after AVBT Only a handful of studies have investigated spinal mobility after AVBT. Buyuk et al. [39] investigated the spinal mobility using flexion-extension and side bending radiographs in 32 children after thoracic VBT. These patients maintained both coronal (mean 7 degrees) and sagittal arc of motion (21 degrees) at one-year follow-up even though especially the coronal movement was significantly reduced from a preoperative value of 30 degrees. Another recent study demonstrated that AVBT in thoracolumbar curves yielded significantly superior lumbar range of motion and lumbar anterior and lateral flexibility compared to patients with spinal fusion. In addition, trunk flexor-extensor endurance and trunk motor strength were better in AVBT than PSF [34]. 4.6. Pulmonary Function after AVBT Baroncini et al. [52] evaluated the pulmonary function after mini-open VBT for AIS. Fifty-one patients completed pulmonary function testing including total lung capacity (TLC), forced vital capacity (FVC), and forced expiratory volume in one second (FEV1). There was a small reduction in FVC from 91% preoperatively to 86% at one-year follow-up, while TLC and FEV1 remained at the same level. They concluded that the mini-open approach does not result in a clinically significant reduction in pulmonary function. Further, Alanay et al. [53] and Samdani et al. [47] have reported significantly improved pulmonary function after VBT scoliosis correction. 4.7. Health-Related Quality of Life Newton et al. [44] and Qiu et al. [54] reported similar HRQoL total and domain scores between AVBT and PSF patients. On the other hand, HRQoL and patient satisfaction were also significantly better in tethered patients in the study of Pehlivanoglu et al. [34]. Further, Hegde et al. [30] reported significant improvement in SRS-22 scores from preoperation to 1 year after surgery. 4.8. Cost-Utility Analysis Only one study was found in the literature concerning costs of two different treatments [55]. It suggested that AVBT may be a cost-effective alternative to fusion. The results relied on HRQoL benefits over fusion patients. 5. Discussion The premise of spinal growth modulation is supported by the basic science and experimental studies, which have shown that asymmetric mechanical compression of vertebral body centers can slow the growth on the anterior and convex aspect of the spinal column [15,56,57,58,59,60]. Similarly, clinical experience and publications of the early outcomes have confirmed the reduction in scoliosis curves over time with growth [19] as also noted in our systematic review. As stated above, AVBT produces three-dimensional deformity correction during surgery which continues based on the Hueter-Volkman principle, producing asymmetrical growth to vertebras [61]. The same kind of technique can also be used from the posterior approach (costo-vertebral). However, the anterior technique has been proven to be more effective in all planes [62] in a finite element model. The anteriorly placed tether developed coronal correction, reduced axial rotation, and maintained kyphosis. In the same study, higher initial tensions produced overcorrection to the deformity. In another finite element model, tensioning of the cable 100 N vs. 200 N, and placing the screws on the lateral sides of the vertebral bodies (lateral, anterior, or triangulated) were important factors for ideal correction. That study demonstrated that a 200 N tightening and an anterior location provided better correction rates in all three planes [63]. Overall, the AVBT appears to be an effective technique in these models. 5.1. Advantages The main advantage of AVBT is allowing correction of the scoliotic deformity without reverting to spinal fusion, which could be avoided in the majority of patients according to this review. Initial correction is achieved with implants inserted thoracoscopically or through (mini-open) thoracotomy. Further correction is gained through the axial growth modulated by the inserted tether. Ideally, it has the potential to correct all three planes of deformity: compression of the apex in coronal plane via growth inhibition, correction of hypokyphosis by anteriorly placed screws and applied compression, and correction of axial rotation as these modulating forces are applied laterally. AVBT is also less invasive than PSF. Hence, a more rapid return to normal daily life and sports can be expected, especially if a minimally invasive technique is applied [64]. Further, thoracoscopic operation is likely to be associated with minimal respiratory issues as well as minimal blood loss [65]. Additionally, avoiding spinal fusion provides the advantage of preserving at least some extent of spinal column mobility. 5.2. Disadvantages The most common adverse events of AVBT consist of those related to thoracic surgery in general such as atelectasis and pneumo-, hemo-, and chylothorax, the need to convert endoscopic to open approach, and post-thoracotomy pain. Other adverse events include overcorrection, screw pull-out, and broken tether; all of which may require reoperation. Additionally, there are few published reports on the clinical outcomes of AVBT, and long-term outcomes remain elusive. Currently, only a small number of teams have published their data and the results remain to be substantiated by third parties. According to our systematic analysis, the rate of these adverse outcomes was 18%. Theoretical disadvantages include concerns on the long-term sustainability of the results. Currently, we are unable to predict the fate of the tethered intervertebral discs and the effects of AVBT on the development and growth of the spinal canal. Contrary to spinal fusion and brace treatment, we do not yet know the long-term outcomes of AVBT. Pekmezci et al. [66] observed a reduction in the enlargement of the spinal canal by growth on a porcine model of anterior fusion. Even though this was not demonstrated in a tethering model, there is a chance that patients may end up with stiff degenerated thoracic spines associated with spinal stenosis in these segments after AVBT. Recently, Hoernschemeyer et al. [67] reassuringly reported that AVBT did not produce degenerative changes in the intervertebral discs or facet joint during two-year follow-up. 5.3. Limitations There were no randomized controlled clinical trials or even prospective follow-up studies comparing the outcomes of AVBT and segmental pedicle screw instrumentation. Thus, we are currently lacking evidence-based recommendations on which to treat patients with instrumented spinal fusion and to use AVBT. Furthermore, long-term outcomes of AVBT are currently lacking. 6. Conclusions Treatment of deformities in growing children is complex. Until recently, the surgical treatment has led into spinal fusion surgery. As a relatively novel technique, vertebral body tethering allows correction of the scoliotic deformity while preserving motion especially in patients with moderate curvature. The majority of these patients seemed to avoid posterior spinal fusion with a major curve less than 35 degrees, when followed-up to skeletal maturity. However, some of these patients required replacement of a broken tether, and overall risk of revision surgery appears to be around 15%. Additionally, long-term studies are required to clarify curve characteristics, rate of complications, and their prevention. Author Contributions Conceptualization, all authors; methodology, all authors.; formal analysis, A.R. data curation, A.R. and J.S.; writing—original draft preparation, all authors; writing—review and editing, all authors; supervision, I.H. All authors have read and agreed to the published version of the manuscript. Conflicts of Interest The authors declare no conflict of interest. Figure 1 12 year-old girl, Sanders 2, Lenke 1 AN curve of 50 degrees (a). First erect postoperative radiograph (b), 1 year follow-up (c), 30 month follow-up demonstrating splaying of the screw heads at multiple levels and progression of the deformity (d), and after revision surgery (Sanders 6) for replacement of broken tether (e). Figure 2 PRISMA study selection flow diagram. jcm-11-02576-t001_Table 1 Table 1 Summary of eligible studies and their main findings. Values are given as mean (range) unless mentioned otherwise. Author/Setting/Year Number of Patients (% Women) Age (Years) Preoperative Curve Final Curve Length of Follow-Up (Years) Complications (%) Main Findings/Conclusion Abdullah [35]/Multi-center register study/2021 120 (84.2%) 12.6 (8.2–15.7) 51.2 (40–70) 27.5 (−5–52) 2 15.8 Higher than expected complication rate during learning curve. Alanay [28]/Single-center/2020 42 (95.2) 12.1 (SD 1.5) 47 (35–68) 17 (−6–28) 2.8 7.1 Curve behavior after VBT varied according to Sanders stage. Baker [36]/Single-center/2021 17 (70.6) 12.9 (SD 1.4) 45 (35–60) 20 (−40–25) 2 23.5 The majority of patients (53%) were successful despite four revisions and nine broken tethers. Baroncini [31]/2 centers/2021 86 (83.7) 13.2 (SD 2.4) 52.4 (SD 13.9) 26.6 (SD 12.7) 2 8.1 The majority of the patients had a physiologic sagittal profile after surgery. Bernard [37]/Single center/2022 20 (95.0) 13.8 (9–17) 56.5 (40–79) 19.4 (−17–56) 5.4 15 High success rate (95%) in helping children avoid fusion at five years post-surgery. Betz [38]/Single center/2019 71 (83.1) 14.5 N/A N/A 2 4.2 Results of showed clinical success in 93% of immature patients, 81% of maturing, and 86% of mature patients. Buyuk [39]/ Single center/2021 32 (93.8) 13 (11–15) 51 (42–70) 26 (7–43) 1 9.4 Particularly, sagittal plane motion was preserved postoperatively after anterior vertebral body tethering. Cebeci [40]/ Single center/2017 12 (100) 12.2 (11–13) 46 (35–59) 18 (6–26) 2 0 VBT resulted in a significant correction in both major and compensatory curves. Costanzo [41]/Single center/2022 23 (82.6) 12 (9–14) 56.5 (33–79) 37 (15–58) 2 8.7 Initial results were encouraging. Hegde [30]/Single center/2021 10 (100) 14.9 (12–17) 52 (42–80) 15.3 (3–28) 2 0 Preliminary experience was promising. Mackey [42]/Multicenter/2022 37 (97.3) 11.3 (IQR 10.9–11.8) 50 (IQR 43.5–58) 28 (IQR 21–35) 3 27 Satisfactory curve control and improved thoracic and spinal height. Miyanji [27]/Multicenter/2020 57 (94.7) 12.7 (8.2–16.7) 51 (31–81) 23 (−18–57) 3.4 28.1 Satisfactory curve correction and an acceptable complication rate in skeletally immature patients. Mladenov [43]/Single center/2021 20 (70.0) 13.4 (11.5–14.5) 46.5 (29–64) 23 (8–38) 1.6 5 Anticipated curve correction averaged 50%. Newton [44]/Single center/2020 23 (69.6) 12 (9–15) 53 (41–67) 33 (−5–62) 3.4 39.1 AVBT resulted in less deformity correction and more revision procedures than PSF, but resulted in the delay or prevention of PSF in the majority of patients. Pehlivanoglu [45]/Single center/2020 21 (71.4) 11.1 (9–14) 48.2 (IQR 44–52.1) 10.1 (IQR 7.7–11.2) 2.3 9.5 AVBT was a safe and effective option in skeletally immature patients with AIS. Rushton [46]/2 centers/2021 112 (92.9) 12.7 (8.2–16.7) 50.8 (31–81) 25.7 (−32–58) 3.1 22 Satisfactory deformity correction in majority of cases. Samdani [47]/Single center / 2021 57 (86.0) 12.4 (10.1–15.0) 40.4 (SD 6.8) 18.7 (SD 13.4) 4.6 12.3 Our current study suggested VBT as a viable option for skeletally immature children with scoliosis. Takahashi [24]/ Single center / 2021 23 (69.6) 12.2 (SD 1.6) 53 (SD 8) N/A 3.4 30.4 Correction occurred primarily within 2 to 3 years after surgery. Wong [48]/Single center/2019 5 (100) 12 (9–12) 40.1 (37.2–44.0) 25 (−12.4–58) 4 40 Of all patients, 60% avoided spinal fusion. Yucekul [49]/Single center/2021 28 (82.1) 12.2 (10–14) 46 (SD 7.7) 12 (SD 11.5) 3.2 28.6 Intermediate discs and facet joints were preserved after growth modulation with VBT surgery. IQR—interquartile range, N/A—not available, and SD—standard deviation. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Weinstein L.S. Zavala D.C. Ponseti I.V. Idiopathic scoliosis: Long-term follow-up and prognosis in untreated patients J. Bone Jt. Surg. Am. 1981 63 702 712 10.2106/00004623-198163050-00003 2. Newton O.P. Faro F.D. Gollogly S. Betz R.R. Lenke L.G. Lowe T.G. Results of preoperative pulmonary function testing of adolescents with idiopathic scoliosis. A study of six hundred and thirty-one patients J. Bone Jt. Surg. Am. 2005 87 1937 1946 10.2106/JBJS.D.02209 16140807 3. Akbarnia B.A. Management themes in early onset scoliosis J. Bone Jt. Surg. Am. 2007 89 42 54 10.2106/JBJS.F.01256 4. Suk I.S. Lee S.M. Chung E.R. Kim J.H. Kim S.S. Selective thoracic fusion with segmental pedicle screw fixation in the treatment of thoracic idiopathic scoliosis: More than 5-year follow-up Spine 2005 30 1602 1609 10.1097/01.brs.0000169452.50705.61 16025028 5. Karol A.L. Johnston C. Mladenov K. Schochet P. Walters P. Browne R.H. Pulmonary function following early thoracic fusion in non-neuromuscular scoliosis J. Bone Jt. Surg. Am. 2008 90 1272 1281 10.2106/JBJS.G.00184 18519321 6. Karol L.A. Early definitive spinal fusion in young children: What we have learned Clin. Orthop. Relat. Res. 2011 469 1323 1329 10.1007/s11999-010-1622-z 20957466 7. Sponseller D.P. Jain A. Newton P.O. Lonner B.S. Shah S.A. Shufflebarger H. Bastrom T.P. Marks M.C. Betz R.R. Posterior Spinal Fusion With Pedicle Screws in Patients With Idiopathic Scoliosis and Open Triradiate Cartilage: Does Deformity Progression Occur? J. Pediatr. Orthop. 2016 36 95 700 10.1097/BPO.0000000000000524 8. Oksanen H. Lastikka M. Helenius L. Pajulo O. Helenius I. Posterior Spinal Fusion Extended to Stable Vertebra Provides Similar Outcome in Juvenile Idiopathic Scoliosis Patients Compared with Adolescents with Fusion to the Touched Vertebra Scand. J. Surg. 2019 108 83 89 10.1177/1457496918798193 30207198 9. Spencer T.H. Gold M.E. Karlin L.I. Hedequist D.J. Hresko M.T. Gain in spinal height from surgical correction of idiopathic scoliosis J. Bone Jt. Surg. Am. 2014 96 59 65 10.2106/JBJS.L.01333 10. Helenius I. Remes V. Yrjonen T. Ylikoski M. Schlenzka D. Helenius M. Poussa M. Harrington and Cotrel-Dubousset instrumentation in adolescent idiopathic scoliosis. Long-term functional and radiographic outcomes J. Bone Jt. Surg. Am. 2003 85 2303 2309 10.2106/00004623-200312000-00006 11. Helenius L. Diarbakerli E. Grauers A. Lastikka M. Oksanen H. Pajulo O. Löyttyniemi E. Manner T. Gerdhem P. Helenius I. Back Pain and Quality of Life After Surgical Treatment for Adolescent Idiopathic Scoliosis at 5-Year Follow-up: Comparison with Healthy Controls and Patients with Untreated Idiopathic Scoliosis J. Bone Jt. Surg. Am. 2019 101 1460 1466 10.2106/JBJS.18.01370 31436653 12. Marks C.M. Bastrom T.P. Petcharaporn M. Shah S.A. Betz R.R. Samdani A. Lonner B. Miyanji F. Newton P.O. The Effect of Time and Fusion Length on Motion of the Unfused Lumbar Segments in Adolescent Idiopathic Scoliosis Spine Deform. 2015 3 549 553 10.1016/j.jspd.2015.03.007 27927557 13. Parsch D. Gaertner V. Brocai D.R. Carstens C. The effect of spinal fusion on the long-term outcome of idiopathic scoliosis. A case-control study J. Bone Jt. Surg. Br. 2001 83 1133 1136 10.1302/0301-620X.83B8.0831133 14. Roaf R. Vertebral growth and its mechanical control J. Bone Jt. Surg. Br. 1960 42 40 59 10.1302/0301-620X.42B1.40 13854527 15. Newton O.P. Upasani V.V. Farnsworth C.L. Oka R. Chambers R.C. Dwek J. Kim J.R. Perry A. Mahar A.T. Spinal growth modulation with use of a tether in an immature porcine model J. Bone Jt. Surg. Am. 2008 90 2695 2706 10.2106/JBJS.G.01424 16. Roaf R. The Treatment of Progressive Scoliosis by Unilateral Growth-Arrest J. Bone Jt. Surg. Br. 1963 45 637 651 10.1302/0301-620X.45B4.637 17. Bouchard M. Guided Growth: Novel Applications in the Hip, Knee, and Ankle J. Pediatr. Orthop. 2017 37 S32 S36 10.1097/BPO.0000000000001022 28799992 18. Betz R.R. Ranade A. Samdani A.F. Chafetz R. D’Andrea L.P. Gaughan J.P. Asghar J. Grewal H. Mulcahey M.J. Vertebral body stapling: A fusionless treatment option for a growing child with moderate idiopathic scoliosis Spine 2010 35 169 176 10.1097/BRS.0b013e3181c6dff5 20081512 19. Newton P.O. Spinal growth tethering: Indications and limits Ann. Transl. Med. 2020 8 27 10.21037/atm.2019.12.159 32055618 20. Moher D. Liberati A. Tetzlaff J. Altman D.G. Group P. Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement BMJ 2009 339 b2535 10.1136/bmj.b2535 19622551 21. Crawford H.C. 3rd Lenke L.G. Growth modulation by means of anterior tethering resulting in progressive correction of juvenile idiopathic scoliosis: A case report J. Bone Jt. Surg. Am. 2010 92 202 209 10.2106/JBJS.H.01728 20048114 22. Samdani F.A. Ames R.J. Kimball J.S. Pahys J.M. Grewal H. Pelletier G.J. Betz R.R. Anterior vertebral body tethering for idiopathic scoliosis: Two-year results Spine 2014 39 1688 1693 10.1097/BRS.0000000000000472 24921854 23. Newton O.P. Kluck D.G. Saito W. Yaszay B. Bartley C.E. Bastrom T.P. Anterior Spinal Growth Tethering for Skeletally Immature Patients with Scoliosis: A Retrospective Look Two to Four Years Postoperatively J. Bone Jt. Surg. Am. 2018 100 1691 1697 10.2106/JBJS.18.00287 30277999 24. Takahashi Y. Saito W. Yaszay B. Bartley C.E. Bastrom T.P. Newton P.O. Rate of Scoliosis Correction After Anterior Spinal Growth Tethering for Idiopathic Scoliosis J. Bone Jt. Surg. Am. 2021 103 1718 1723 10.2106/JBJS.20.02071 34288633 25. Krakow R.A. Magee L.C. Cahill P.J. Flynn J.M. Could have tethered: Predicting the proportion of scoliosis patients most appropriate for thoracic anterior spinal tethering Spine Deform. 2021 9 1005 1012 10.1007/s43390-021-00296-5 33534123 26. Sanders O.J. Khoury J.G. Kishan S. Browne R.H. Mooney J.F. 3rd Arnold K.D. McConnell S.J. Bauman J.A. Finegold D.N. Predicting scoliosis progression from skeletal maturity: A simplified classification during adolescence J. Bone Jt. Surg. Am. 2008 90 540 553 10.2106/JBJS.G.00004 27. Miyanji F. Pawelek J. Nasto L.A. Rushton P. Simmonds A. Parent S. Safety and efficacy of anterior vertebral body tethering in the treatment of idiopathic scoliosis Bone Jt. J. 2020 102 1703 1708 10.1302/0301-620X.102B12.BJJ-2020-0426.R1 28. Alanay A. Yucekul A. Abul K. Ergene G. Senay S. Ay B. Cebeci B.O. Dikmen P.Y. Zulemyan T. Yavuz Y. Thoracoscopic Vertebral Body Tethering for Adolescent Idiopathic Scoliosis: Follow-up Curve Behavior According to Sanders Skeletal Maturity Staging Spine 2020 45 E1483 E1492 10.1097/BRS.0000000000003643 32756290 29. Mathew S. Larson A.N. Potter D.D. Milbrandt T.A. Defining the learning curve in CT-guided navigated thoracoscopic vertebral body tethering Spine Deform. 2021 9 1581 1589 10.1007/s43390-021-00364-w 34003460 30. Hegde K.S. Venkatesan M. Akbari K.K. Badikillaya V.M. Efficacy of Anterior Vertebral Body Tethering in Skeletally Mature Children with Adolescent Idiopathic Scoliosis: A Preliminary Report Int. J. Spine Surg. 2021 15 995 1003 10.14444/8122 34551922 31. Baroncini A. Courvoisier A. Berjano P. Migliorini F. Eschweiler J. Kobbe P. Hildebrand F. Trobisch P.D. The effects of vertebral body tethering on sagittal parameters: Evaluations from a 2-years follow-up Eur. Spine J. 2022 31 1060 1066 10.1007/s00586-021-07076-9 34910244 32. Parent S. Shen J. Anterior Vertebral Body Growth-Modulation Tethering in Idiopathic Scoliosis: Surgical Technique J. Am. Acad. Orthop. Surg 2020 28 693 699 10.5435/JAAOS-D-19-00849 32618681 33. Trobisch D.P. Baroncini A. Preliminary outcomes after vertebral body tethering (VBT) for lumbar curves and subanalysis of a 1- versus 2-tether construct Eur. Spine J. 2021 30 3570 3576 10.1007/s00586-021-07009-6 34591175 34. Pehlivanoglu T. Oltulu I. Erdag Y. Korkmaz E. Sarioglu E. Ofluoglu E. Aydogan M. Double-sided vertebral body tethering of double adolescent idiopathic scoliosis curves: Radiographic outcomes of the first 13 patients with 2 years of follow-up Eur. Spine J. 2021 30 1896 1904 10.1007/s00586-021-06745-z 33611658 35. Abdullah A. Parent S. Miyanji F. Smit K. Murphy J. Skaggs D. Gupta P. Vitale M. Ouellet J. Saran N. Spine Study, Risk of early complication following anterior vertebral body tethering for idiopathic scoliosis Spine Deform. 2021 9 1419 1431 10.1007/s43390-021-00326-2 33837516 36. Baker E.C. Kiebzak G.M. Neal K.M. Anterior vertebral body tethering shows mixed results at 2-year follow-up Spine Deform. 2021 9 481 489 10.1007/s43390-020-00226-x 33113121 37. Bernard J. Bishop T. Herzog J. Haleem S. Lupu C. Ajayi B. Lui D.F. Dual modality of vertebral body tethering: Anterior scoliosis correction versus growth modulation with mean follow-up of five years Bone Jt. Open. 2022 3 123 129 10.1302/2633-1462.32.BJO-2021-0120.R1 35119295 38. Betz R. Bassett W.P. Cuddihy L. Cerrone J. Haas A. Antonacci M.D. Non-fusion anterior scoliosis correction (asc): Comparison of outcomes in skeletally immature vs. skeletally mature patients with adolescent idiopathic scoliosis Eur. Spine J. 2019 28 2856 10.1007/s00586-019-06103-0 39. Buyuk F.A. Milbrandt T.A. Mathew S.E. Larson A.N. Measurable Thoracic Motion Remains at 1 Year Following Anterior Vertebral Body Tethering, with Sagittal Motion Greater Than Coronal Motion J. Bone Jt. Surg. Am. 2021 103 2299 2305 10.2106/JBJS.20.01533 40. Cebeci O.B. Sogunmez N. Ergene G. Ay B. Yilgor C. Alanay A. Non-fusion growth modulation with anterior vertebral body tethering for adolescent idiopathic scoliosis: A promising minimal invasive alternative to traditional treatment Eur. Spine J. 2017 26 S291 10.1007/s00586-017-5224-2 41. Costanzo S. Pansini A. Colombo L. Caretti V. Popovic P. Lanfranchi G. Camporesi A. Pelizzo G. Video-Assisted Thoracoscopy for Vertebral Body Tethering of Juvenile and Adolescent Idiopathic Scoliosis: Tips and Tricks of Surgical Multidisciplinary Management Children 2022 9 74 10.3390/children9010074 35053697 42. Mackey C. Hanstein R. Lo Y. Vaughan M. Hilaire T.S. Luhmann S.J. Vitale M.G. Glotzbecker M.P. Samdani A. Parent S. Magnetically Controlled Growing Rods (MCGR) Versus Single Posterior Spinal Fusion (PSF) Versus Vertebral Body Tether (VBT) in Older Early Onset Scoliosis (EOS) Patients: How Do Early Outcomes Compare? Spine 2022 47 295 302 10.1097/BRS.0000000000004245 34610613 43. Mladenov K. Stucker R. Anterior Vertebral Body Tethering (VBT) for Idiopathic Scoliosis Eur. Spine J. 2021 30 3349 10.1007/s00586-021-07017-6 44. Newton O.P. Bartley C.E. Bastrom T.P. Kluck D.G. Saito W. Yaszay B. Anterior Spinal Growth Modulation in Skeletally Immature Patients with Idiopathic Scoliosis: A Comparison with Posterior Spinal Fusion at 2 to 5 Years Postoperatively J. Bone Jt. Surg. Am. 2020 102 769 777 10.2106/JBJS.19.01176 32379117 45. Pehlivanoglu T. Oltulu I. Ofluoglu E. Sarioglu E. Altun G. Korkmaz M. Yildirim K. Aydogan M. Thoracoscopic Vertebral Body Tethering for Adolescent Idiopathic Scoliosis: A Minimum of 2 Years’ Results of 21 Patients J. Pediatr. Orthop. 2020 40 575 580 10.1097/BPO.0000000000001590 32427800 46. Rushton P.P.R. Nasto L. Parent S. Turgeon I. Aldebeyan S. Miyanji F. Anterior Vertebral Body Tethering for Treatment of Idiopathic Scoliosis in the Skeletally Immature: Results of 112 Cases Spine 2021 46 1461 1467 10.1097/BRS.0000000000004061 34091563 47. Samdani F.A. Pahys J.M. Ames R.J. Grewal H. Pelletier G.J. Hwang S.W. Betz R.R. Prospective Follow-up Report on Anterior Vertebral Body Tethering for Idiopathic Scoliosis: Interim Results from an FDA IDE Study J. Bone Jt. Surg. Am. 2021 103 1611 1619 10.2106/JBJS.20.01503 48. Wong K.H. Ruiz J.N.M. Newton P.O. Liu K.P.G. Non-Fusion Surgical Correction of Thoracic Idiopathic Scoliosis Using a Novel, Braided Vertebral Body Tethering Device: Minimum Follow-up of 4 Years JB JS Open Access 2019 4 e0026 10.2106/JBJS.OA.19.00026 32043058 49. Yucekul A. Akpunarli B. Durbas A. Zulemyan T. Havlucu I. Ergene G. Senay S. Dikmen P.Y. Balci S.T. Karaarslan E. Does vertebral body tethering cause disc and facet joint degeneration? A preliminary MRI study with minimum two years follow-up Spine J. 2021 21 1793 1801 10.1016/j.spinee.2021.05.020 34033932 50. Newton O.P. Takahashi Y. Yang Y. Yaszay B. Bartley C.E. Bastrom T.P. Munar C. Anterior vertebral body tethering for thoracic idiopathic scoliosis leads to asymmetric growth of the periapical vertebrae Spine Deform. 2022 10 553 561 10.1007/s43390-021-00464-7 34982419 51. Yucekul A. Karaman I. Zulemyan T. Ergene G. Senay S. Balci S. Yavuz Y. Yilgor C. Alanay A. Thoracoscopic and Mini-open Lumbotomy Vertebral Body Tethering for Thoracolumbar/Lumbar Curves: Two to Three Years Follow-up Proceedings of the 28th International Meeting on Advanced Spine Techniques Virtual Meeting 23–25 April 2021 52. Baroncini A. Trobisch P. Blau C. Golias C. Kobbe P. Eschweiler J. Tingart M. Migliorini F. Analysis of the pulmonary function in patients undergoing vertebral body tethering for adolescent idiopathic scoliosis Eur. Spine J. 2022 31 1022 1027 10.1007/s00586-021-07029-2 34677678 53. Alanay A. Yucekul A. Abul K. Karaman I. Durbas A. Zulemyan T. Ergene G. Senay S. Balci S.T. Dikmen P.Y. Two to five years follow up results after thoracoscopic VBT: A single surgeon’s experience Spine J. 2021 21 S86 S87 10.1016/j.spinee.2021.05.201 54. Qiu C. Talwar D. Gordon J. Capraro A. Lott C. Cahill P.J. Patient-Reported Outcomes Are Equivalent in Patients Who Receive Vertebral Body Tethering Versus Posterior Spinal Fusion in Adolescent Idiopathic Scoliosis Orthopedics 2021 44 24 28 10.3928/01477447-20201119-02 33238017 55. Polly W.D. Larson A.N. Samdani A.F. Rawlinson W. Brechka H. Porteous A. Marsh W. Ditto R. Cost-Utility Analysis of Anterior Vertebral Body Tethering versus Spinal Fusion in Idiopathic Scoliosis from a US Integrated Healthcare Delivery System Perspective Clin. Outcomes Res. 2021 13 175 190 10.2147/CEOR.S289459 56. Newton O.P. Farnsworth C.L. Faro F.D. Mahar A.T. Odell T.R. Mohamad F. Breisch E. Fricka K. Upasani V.V. Amiel D. Spinal growth modulation with an anterolateral flexible tether in an immature bovine model: Disc health and motion preservation Spine 2008 33 724 733 10.1097/BRS.0b013e31816950a0 18379398 57. Newton O.P. Fricka K.B. Lee S.S. Farnsworth C.L. Cox T.G. Mahar A.T. Asymmetrical flexible tethering of spine growth in an immature bovine model Spine 2002 27 689 693 10.1097/00007632-200204010-00004 11923660 58. Braun T.J. Ogilvie J.W. Akyuz E. Brodke D.S. Bachus K.N. Fusionless scoliosis correction using a shape memory alloy staple in the anterior thoracic spine of the immature goat Spine 2004 29 1980 1989 10.1097/01.brs.0000138278.41431.72 15371698 59. Moal B. Schwab F. Demakakos J. Lafage R. Riviere P. Patel A. Lafage V. The impact of a corrective tether on a scoliosis porcine model: A detailed 3D analysis with a 20 weeks follow-up Eur. Spine J. 2013 22 1800 1809 10.1007/s00586-013-2743-3 23503934 60. Newton O.P. Faro F.D. Farnsworth C.L. Shapiro G.S. Mohamad F. Parent S. Fricka K. Multilevel spinal growth modulation with an anterolateral flexible tether in an immature bovine model Spine 2005 30 2608 2613 10.1097/01.brs.0000188267.66847.bf 16319746 61. Jain V. Lykissas M. Trobisch P. Wall E.J. Newton P.O. Sturm P.F. Cahill P.J. Bylski-Austrow D.I. Surgical aspects of spinal growth modulation in scoliosis correction Instr. Course Lect. 2014 63 335 344 24720319 62. Aubin E.C. Clin J. Rawlinson J. Biomechanical simulations of costo-vertebral and anterior vertebral body tethers for the fusionless treatment of pediatric scoliosis J. Orthop. Res. 2018 36 254 264 10.1002/jor.23648 28685857 63. Cobetto N. Parent S. Aubin C.E. 3D correction over 2years with anterior vertebral body growth modulation: A finite element analysis of screw positioning, cable tensioning and postoperative functional activities Clin. Biomech. 2018 51 26 33 10.1016/j.clinbiomech.2017.11.007 29169117 64. Baroncini A. Trobisch P.D. Berrer A. Kobbe P. Tingart M. Eschweiler J. da Paz S. Migliorini F. Return to sport and daily life activities after vertebral body tethering for AIS: Analysis of the sport activity questionnaire Eur. Spine J. 2021 30 1998 2006 10.1007/s00586-021-06768-6 33638722 65. Samdani F.A. Ames R.J. Kimball J.S. Pahys J.M. Grewal H. Pelletier G.J. Betz R.R. Anterior vertebral body tethering for immature adolescent idiopathic scoliosis: One-year results on the first 32 patients Eur. Spine J. 2015 24 1533 1539 10.1007/s00586-014-3706-z 25510515 66. Pekmezci M. Yilmaz G. Daglioglu K. Gulsen M. Alanay A. Acaroglu E. Yazici M. The effect of anterior spinal fusion on spinal canal development in an immature porcine model Spine 2009 34 E501 E506 10.1097/BRS.0b013e3181a958fa 19564754 67. Hoernschemeyer G.D. Boeyer M.E. Tweedy N.M. Worley J.R. Crim J.R. A preliminary assessment of intervertebral disc health and pathoanatomy changes observed two years following anterior vertebral body tethering Eur. Spine J. 2021 30 3442 3449 10.1007/s00586-021-06972-4 34494139

PMC009xxxxxx/PMC9099652.txt

==== Front Polymers (Basel) Polymers (Basel) polymers Polymers 2073-4360 MDPI 10.3390/polym14091687 polymers-14-01687 Article Synthesis of a New Chelating Iminophosphorane Derivative (Phosphazene) for U(VI) Recovery Atia Bahig M. 1† https://orcid.org/0000-0001-8619-3154 Sakr Ahmed K. 1*† Gado Mohamed A. 1 El-Gendy Hassan S. 1 Abdelazeem Nagwa M. 2 El-Sheikh Enass M. 1 https://orcid.org/0000-0001-6619-642X Hanfi Mohamed Y. 13 https://orcid.org/0000-0003-3040-8878 Sayyed M. I. 45 Al-Otaibi Jamelah S. 6 https://orcid.org/0000-0002-9297-9804 Cheira Mohamed F. 1* Kammakakam Irshad Academic Editor Khodakarami Mostafa Academic Editor 1 Nuclear Materials Authority, P.O. Box 530, El Maadi, Cairo, Egypt; dr_bahig.atia@yahoo.com (B.M.A.); mag.nma@yahoo.com (M.A.G.); elgendy_nma@yahoo.com (H.S.E.-G.); elsba3y@hotmail.com (E.M.E.-S.); mokhamed.khanfi@urfu.ru (M.Y.H.) 2 National Research Center (NRC), 33 El-Buhoth Street, Dokki, Cairo 12622, Egypt; nagwamorad@yahoo.com 3 Institute of Physics and Technology, Ural Federal University, St. Mira, 19, 620002 Yekaterinburg, Russia 4 Department of Physics, Faculty of Science, Isra University, Amman 11622, Jordan; dr.mabualssayed@gmail.com 5 Department of Nuclear Medicine Research, Institute for Research and Medical Consultations (IRMC), Imam Abdulrahman Bin Faisal University (IAU), P.O. Box 1982, Dammam 31441, Saudi Arabia 6 Department of Chemistry, College of Science, Princess Nourah bint Abdulrahman University, P.O. Box 84428, Riyadh 11671, Saudi Arabia; jsalotabi@pnu.edu.sa * Correspondence: akhchemist@gmail.com (A.K.S.); mf.farid2008@yahoo.com (M.F.C.) † These authors contributed equally to this work. 21 4 2022 5 2022 14 9 168726 3 2022 15 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). A new synthetic chelating N–hydroxy–N–trioctyl iminophosphorane (HTIP) was prepared through the reaction of trioctylphosphine oxide (TOPO) with N–hydroxylamine hydrochloride in the presence of a Lewis acid (AlCl3). Specifications for the HTIP chelating ligand were successfully determined using many analytical techniques, 13C–NMR, 1H–NMR, FTIR, EDX, and GC–MS analyses, which assured a reasonable synthesis of the HTIP ligand. The ability of HTIP to retain U(VI) ions was investigated. The optimum experimental factors, pH value, experimental time, initial U(VI) ion concentration, HTIP dosage, ambient temperature, and eluents, were attained with solvent extraction techniques. The utmost retention capacity of HTIP/CHCl3 was 247.5 mg/g; it was achieved at pH = 3.0, 25 °C, with 30 min of shaking and 0.99 × 10−3 mol/L. From the stoichiometric calculations, approximately 1.5 hydrogen atoms are released during the extraction at pH 3.0, and 4.0 moles of HTIP ligand were responsible for chelation of one mole of uranyl ions. According to kinetic studies, the pseudo–first order model accurately predicted the kinetics of U(VI) extraction by HTIP ligand with a retention power of 245.47 mg/g. The thermodynamic parameters ΔS°, ΔH°, and ΔG° were also calculated; the extraction process was predicted as an exothermic, spontaneous, and advantageous extraction at low temperatures. As the temperature increased, the value of ∆G° increased. The elution of uranium ions from the loaded HTIP/CHCl3 was achieved using 2.0 mol of H2SO4 with a 99.0% efficiency rate. Finally, the extended variables were used to obtain a uranium concentrate (Na2U2O7, Y.C) with a uranium grade of 69.93% and purity of 93.24%. Uranium N–Hydroxy–N–trioctyl–iminophosphorane (HTIP) solvent extraction G. Gattar granite leach liquor ==== Body pmc1. Introduction The arrangement of the reactive center in a ligand has been the basis for many advances in inorganic as well as organometallic chemistry. Accordingly, significant time and effort have gone into designing and synthesizing ancillary ligands, besides studying the properties, reactivities, and geometries of the resulting complexes. In some cases, these efforts have resulted in new applications in stoichiometric chemistry, catalysis, and materials science [1,2,3]. A large number of structural studies for systems with phosphinimide ligands have been published in this area. These chelating and anionic ligands (R3PN) are easily obtained from neutral phosphinimine precursors, prepared utilizing the simple and long–established Staudinger reaction [4,5,6]. Iminophosphorane compounds, also known as phosphoranimines, phosphinimines, or phosphazenes, were detected in 1919; they are organic composites with the general structure R3P=NR. Iminophosphoranes contain nitrogen alongside phosphorus atoms, which coordinate to transition metals using the nitrogen atom’s lone pair of electrons. Multidentate ligands are produced by adding extra donor sites into the iminophosphorane ligand, and they are gaining popularity in both coordination chemistry and catalysis [7,8,9]. Iminophosphoranes are resonance hybrids of the two recognized forms A and B, and have a highly polarized P=N bond (Scheme 1). They can also integrate transition metals through the sp2–hybridized nitrogen atom’s lone pair, resulting in stable complexes (C in Scheme 1) [10]. Iminophosphoranes, which are primarily s–donor ligands with only minimal p–acceptor characteristics, have a limited inherent coordinating capability since they can be easily substituted by other ligands. Polydentate mixed ligands, such as those found in the generic metal complex structures shown in Scheme 2 (D–G), can stabilize a wider range of metal ions than their monodentate R3P = NR counterparts due to the addition of further donor sites to the iminophosphoranes. The reactivity of the metal center can be modified as needed by selecting and adjusting these new donor sites, as well as the appropriate linkers. It is regarded as a critical component in achieving high levels of activity and selectivity [7]. Several procedures for the production of iminophosphoranes are recognized; however, the most dominant and extensively used are the Staudinger as well as Kirsanov reaction (Scheme 3). A phosphine (PR3) is used as a starting material in both of them. The Staudinger reaction involves direct oxidation with an organic azide, whereas the Kirsanov reaction involves initial bromination and consecutive reaction of the resulting phosphine dibromide compound with a primary amine in the presence of a base [11,12,13]. There are various methods for the synthesis of phosphazenes, such as the free radical process, thermal ring opening polymerization, substitution polymerization and cationic or anionic living polymerization, and solution polymerization method [14,15,16,17]. Phosphazenes have highly tunable chemical as well as physical properties that count on the substituents attached to the phosphorus atom. Therefore, they are employed in an enormous range of fields, such as rechargeable batteries [18], liquid crystals [19], membranes [20] and lubricants [21], anticancer agents [22], flame retardants [23], antibacterial reagents [24], biological materials [25], synthetic bones [26], photosensitive substrates [27], thermally stable macromolecules [28], a high limiting oxygen index (LOI) [29], coordination metal supports and reagents for organometallic chemistry [30], silicon-containing compounds [31], and catalysts [32]. Uranium is a chemical element with the symbol U, which belongs to the actinides. Due to its strategic significance in the energy sector, uranium became the most valuable heavy metal [32,33,34,35]. In nature, the Earth’s crust contains an average of 4.0 mg/L uranium [36]; however, it deposits in different types of rocks. It is a hexavalent state in the secondary minerals, such as kasolite, uranophane, autunite, etc., while it has a tetravalent oxidation state in the form of primary minerals, e.g., coffinite, pitchblende, uraninite, etc. Firstly, uranium is leached from its ore in an acidic or alkaline process, and extracted from the resulting leach liquor using an ion exchanger or a solvent extraction technique [37,38,39,40,41,42,43,44,45,46,47]. Solvent extraction (SX) is a widely used technique for the recovery and separation of base metals and strategic metals in hydrometallurgy. Solvent extraction efficacy is subject to several factors including the type of separation apparatus (pulsed columns, mixer–settler, etc.), the type of feed solution, the applied flowsheets, the chemical composition of the organic solvent, the flow rates of both aqueous and organic phases in the extractions along with elution steps, etc. In the SX method, the extractant plays an essential role. It should have tremendous solubility in the organic solvent. Nevertheless, it must have an extremely low aqueous phase solubility. The solvent ought to be nonvolatile, industrially serviceable, nontoxic, nonflammable, and affordable for the element extraction process [48,49]. Previous studies have used several extractants for uranium extraction by solvent extraction. The derivatives of organophosphorus compounds have been employed since 1900 in uranium extraction processes. Compounds such as Cyanex 302, Cyanex 272, TBP, DEHPA, D2EHPA/TOPO mixtures, Primene JM–T/Alamine–336 mixture [50], DDPA, HDPA, CMPO, DNPPA, and DBBP, which are used to extract uranium from different matrices, were studied [51,52,53,54,55,56]. Several ligands with high selectivity are employed for uptake of uranium ions, comprising organic crown ethers [57], calixarenes [58], and Schiff bases. The long-chain amines have proven to be outstanding extractants for a large number of anionic metal complexes applied widely in uranium removal, such as Adogen–383, TOA, and Aliquate–336 [59,60,61]. Amides are deemed one of the outstanding as well as vital organic functional groups in pharmaceuticals, agrochemicals, polymers, and naturally occurring molecules. In addition, carboxamides have considerable value in coordination, medicinal, and organic chemistry. They can be obtained via the amidation process where a condensation reaction occurs between carboxylic acid and amine. One-pot synthesis of the chelating carboxamides using different catalysts were proposed. For instance, pyridine–2,6–dicarboxylic acid bis–(3–hydroxy phenyl) amide (Pydca) along with tetra–kis (2–ethylhexyl) pyridine–2,6–dicarboxamide (EHPyCA) were successfully synthesized and utilized for removal of thorium as well as uranium from leach liquors of ore samples in Egypt [62,63,64]. In this study, a new synthetic N–hydroxy–N–trioctyl iminophosphorane (HTIP) chelating ligand was synthesized using an effective alternative technique compared to the traditional Staudinger and Kirsanov methods and employed for uranium extraction from acidic solution. Both the removal and elution factors were optimized. Furthermore, the study dealt with the equilibrium, kinetic, and thermodynamic characteristics of uranium extraction from G. Gattar leach fluid, North Eastern Desert of Egypt. 2. Materials and Methods 2.1. Apparatus The acidity and alkalinity of solutions were detected using a digital pH meter (VSTAR10 series, Thermo Scientific™, Waltham, MA, USA) with an error of ±0.1. An analytical balance (AUW220D Series, Shimadzu, Kyoto, Japan) with standard deviation of 0.05 mg was utilized to measure all samples. A Vibromatic-384 shaker was employed to mix the contents in separating funnels. The crystal structure of materials was examined using X-ray diffraction (XRD) technique (D8 Discover Family, Bruker, Billerica, MA, USA). Quantitative analysis of U(VI) was executed with a double beam spectrometer (T80 UV/Vis, PG Instrument, Leicestershire, UK) using the arsenazo (III) indicator and 650 nm wavelength against an appropriate standard solution [65]. Furthermore, oxidometric titration against ammonium metavanadate and sodium diphenyl amine sulfonate as an indicator using automatic titrator [66,67] (SCOTT Instrument, GmbH, DE, Sialkot, Pakistan) was also employed to confirm the concentration of U(VI) ions. An ICP-OES spectrometer (OPTIMA 5300 DV, PerkinElmer, Waltham, MA, USA) was applied to specify the concentration of uranium and metal ions of G. Gattar leachate. A Reichert Thermovar was used to determine the melting point. The elemental analysis of uranium concentrate product and HTIP ligand were recorded using EDX (JSM–7900F, Jeol, Tokyo, Japan). An FTIR spectrophotometer (IRPrestige–21, Shimadzu, Kyoto, Japan) was employed to record the IR spectra using KBr disc. The 1H and 13C–NMR spectra were obtained at 500 MHz using an NMR spectrometer (Bruker Avance TM 500, Bruker, Billerica, MA, USA). The coupling constant (J) was measured in Hertz (Hz), while the chemical shift (δ) was measured in ppm. A mass spectrometer (Finnigan SSQ 7000 spectrometer, Thermo Finnigan, San Jose, CA, USA) was used for the molecular formula. The laboratories of National Research Center (NRC), Cairo, Egypt performed the FTIR, GC–MS, 1H, and 13C-NMR analyses. 2.2. Reagents All of the reagents were made with analytical grade chemicals. HCl, H2SO4, NaOH, and HNO3 were purchased from POCH S.A., Gliwice, Poland. Trioctylphosphine oxide (TOPO) and N–hydroxylamine hydrochloride were obtained from Thermo Fisher Scientific–Acros Organics Inc., Geel, Belgium. Ammonium metavanadate, sodium nitrite, FeSO4.7H2O, AlCl3, and urea were supplied by Scharlau Chemie. S.A., Barcelona, Spain. Uranyl acetate dihydrate, arsenazo III, and sodium diphenyl amine sulfonate were obtained from Merck, Darmstadt, Germany. Furthermore, methanol, DMF, and ethyl acetate were purchased from Fluka, Gillingham, UK. All reactions were performed utilizing flame-dried glassware. Thin paper chromatography (PC) was used to observe the reaction’s development. Ethanol plus ethyl acetate (50:50 v/v) was adopted as an eluent. A UV lamp was used to visualize spots on the PC plates (250 nm). 2.3. Synthesis of N–Hydroxy–N–Trioctyl Iminophosphorane (HTIP) Chelating Ligand Two primary procedures were used to produce N–Hydroxy–N–trioctyl iminophosphorane (HTIP). The first stage in neutralization was to mix 0.1 mole of NaOH (5.0 g, over the stoichiometric quantity) with 0.1 mole of NH2OH.HCl (7.0 g) in 50.0 mL of DMF as diluent. For 2.0 h, the mixture was refluxed at 50 °C. The vital goal of the neutralization phase was to make NH2OH more nucleophilic. The second swelling process started with 0.1 mole trioctylphosphine oxide (TOPO, 38.6 g) and 0.1 mole (13.3 g) AlCl3 hard Lewis acid in 50.0 mL DMF in a condenser for 2.0 h at 50 °C. Finally, the two additions were added to each other and allowed to condense for 6.0 h at 100 °C. The reaction was monitored using paper chromatography (PC) sheets and a solvent mixture of ethanol plus ethyl acetate 50:50 v/v. A UV lamp was used to detect the spots. The resulting HTIP appeared as a crystalline white to pale yellow solid with a density of ≈0.943 g/cm3. After completion of the reaction, the product was obtained by washing it several times with deionized water to remove any leftover DMF and AlCl3. The residue was washed, and the recrystallization procedure was performed with an ethanol/DMF mixture. 2.4. Preparation of U(VI) Standard Stock Solution A 1000 mg/L (4.2 × 10−3 mol/L) U(VI) standard stock solution was prepared by dissolving 1.872 g of UO2(CH3COOH)2.2H2O in deionized water that had been acidified with 5.0 mL concentrated HNO3 to avoid hydrolysis in a 1000 mL volumetric flask. In addition, numerous standard stock solutions of 1000 mg/L of probable different ions during U(VI) extraction by HTIP/CHCl3 chelating ligand were generated by dissolving proper amounts of their salts in 1000 mL deionized water. 2.5. Extraction and Stripping Procedures The pH value, shaking time, initial uranium(VI) conc., HTIP conc., temperature, and different ions were all tuned to optimize U(VI) ion extraction from synthetic solution by HTIP/CHCl3. In these experiments, 25.0 mL of a 100 mg/L (4.2 × 10−4 mol/L) synthetic U(VI) ions solution was mechanically shaken for a predefined period of time with 25.0 mL of varied concentrations of HTIP/CHCl3. Both the extraction distribution ratio D and stripping ratio D′ were calculated using the following Equation (1) [68]:(1) D=CoCa;D′=CaCo where Co and Ca (mg/L) correspond to the concentration of U(VI) in organic and aqueous phase, respectively. Moreover, the distribution coefficient (Kd) and extraction percentage (E%) were calculated using Equations (2) and (3), respectively:(2) Kd=Ci−CeCe×VaVo (3) E%=100DD+(VaVo) where Ci and Ce (mg/L) symbolise for the initial and equilibrium concentration of U(VI) ions, respectively. Vo and Va (mL) represent organic and aqueous phase volumes, respectively. Nonetheless, the stripping procedures were performed by shaking different volumes of the loaded organic solvent with the eluent (2.0 mol H2SO4) for 10 min at ambient temperature. After equilibration, the two layers were entirely separated, and the U(VI) ion concentration was measured. The stripping percentage (S%) can be expressed using the next Equation (4):(4) S%=100D′D′+(VoVa) 2.6. Production of G. Gattar Granite Leach Solution Uranium–rich ore sample was collected from G. Gattar granite, North Eastern Desert, Egypt. Percolation leaching technique was applied; H2SO4 was used as a leachant. The leaching parameters were optimized at 75.0 g/L of H2SO4, −100 mesh particle size for 4.0 h leaching time, and 1:1 S/L phase ratio at room temperature. ICP-OES and colorimetric analysis were used to detect the chemical composition of the G. Gattar ore sample and its leachate. 3. Results and Discussion 3.1. Characterization of N–Hydroxy–N–Trioctyl Iminophosphorane (HTIP) Chelating Ligand The synthesis procedures for N–Hydroxy–N–trioctyl iminophosphorane (HTIP) chelating ligand and the suggested mechanism of the reaction are illustrated in Scheme 4. A very important clarification should be mentioned concerning the role of the Lewis acid (AlCl3) in the fabrication of HTIP ligand. The hard Lewis acid was characterized by a great charge density with vacant orbitals, which could attract electrons from the oxygen phosphine group. This operation facilitated the breaking of –P=O bonds. After that, the nucleophilic attack of NH2OH upon the phosphine group could take place easily, as the phosphorous atom acted as an electrophile. This method is considered as an effective alternative technique to the Staudinger and Kirsanov methods. The obtained yield was ≈40.0 g (≈71.0%) with a melting point equal to 135–140 °C. Numerous functional groups in the synthesized HTIP were predicted using important observations in Fourier transformation infrared spectroscopy (FTIR) [69]. Figure 1a displays the FTIR spectra of the HTIP and its complex with U(VI) ions. The HTIP had a strong peak, centered at roughly 3409.48 cm−1, which was linked to the OH group’s stretching mode. After chelation with U(VI) ions, the stretching vibration related to the OH group disappeared, indicating that the OH group took part in the chelation. The features located at 2850–2918, 701.38, 819.83, 1145.18 and 1464.72 cm−1 were more likely caused by the CH aliphatic, (CH2)n aliphatic, –P=N, –P–N, and N–O bonds, respectively. The frequency of –P=N and –P–N stretching vibrations in the HTIP–U(VI) complex were shifted to a lower frequency as compared with the free ligand (776.3 and 1118.41 cm−1), indicating that there was an appreciable chelation between HTIP and U(VI). The peak in the HTIP–U(VI) complex spectrum observed at 925 cm−1 corresponded to coordinated U=O bond [40,70]. Gas chromatography–mass spectrometry (GC–MS) is a powerful and effective tool for predicting chemical formulas and purity; the more stable fragment, [m/z]+, is an influential and strong tool. The synthetic ligand’s molecular weight was represented by the molecular ion peak with a value of 401.63. Some important fragmentation patterns related to the synthesized HTIP were observed, including [P=N–OH]˙ with a molecular weight of 61.98, [P=N–O]˙ with a molecular weight of 60.98, [CH3(CH2)7]˙ with a molecular weight of 113.25, and a fragment with a 384.66 molecular weight, which denoted the formation of the [(CH3(CH2)7)3P=N]˙ moiety. The results of the entire investigation indicated that the HTIP ligand could be synthesized successfully. Figure 1b demonstrates the GC–MS spectrum of the HTIP chelating ligand. 1H–NMR analysis with a 500.15 MHZ energy and CDCl3 as a diluent is a useful and efficient technology that provides important information about protons in the produced substance and aids in structure estimates. The primary δ (ppm) assignments were 7.259, 0.815–0.842, and 1.222–1.637 ppm, which corresponded to the protons of OH, methyl, and methylene, respectively. The assignments of the OH proton (δ = 7.259 ppm) were more deshielded than the assignments of the –CH2 and –CH3 protons, whereas –CH2 (δ = 1.222–1.637 ppm) was more deshielded than –CH3 (δ = 0.815–0.842 ppm). Figure 2a illustrates the 1H–NMR characterization of the HTIP chelating ligand. 13C–NMR analysis with a 125.76 MHZ energy and CDCl3 as a diluent is a useful method for determining the number of carbon atoms in an HTIP ligand. The major δ (ppm), which is connected to alkyl carbon, occurred at 14.126–31.848 ppm (Figure 2b). The –CH3 carbon appeared at 14.126 ppm, which was more protected than the other –CH2 carbons that appeared in chemical shift ranges of 21.728–31.848 ppm. A characteristic EDX analysis was performed to illustrate the elements composing HTIP chelating ligand after and before uranyl ion chelation. A significant peak from 0–0.5 keV represented carbon, oxygen and nitrogen atoms, while the peak at 1.65 keV represented the phosphorous atom. The appearance of uranium gave an obvious sign for the chelation of uranyl ions by HTIP chelating ligand. The identification of HTIP and HTIP–U(VI) complex by EDX spectrum is presented in Figure 3. 3.2. Extraction Procedures 3.2.1. The Influence of pH The role of pH was very important in the retention of U(VI) on the ligand’s active sites because it impacted the form of uranium in aqueous media as well as the features of the active sites of the HTIP ligand. Figure 4 shows the species of uranyl ions in the HYDRA–MEDUSA program at different pH vales. There are many U(VI) species in solution, according to earlier research [71,72]. Uranium can be found in cationic, neutral, or anionic species. Until pH 5.0, uranium is primarily present in the cationic forms UO22+, UO2(OH)+, and (UO2)2(OH)22+, whilst UO2SO4, UO3.2H2O, UO2(OH)3–, and UO2(OH)42– are neutral and anionic species that are identified till pH 12.0. The retention of U(VI) on HTIP was examined using 25.0 mL of a 4.2 × 10−4 mol/L U(VI) solution (100 mg/L) and 25.0 mL of 0.99 × 10−4 mol/L HTIP/CHCl3 (10 mg of HTIP in 25.0 mL CHCl3) at pH ranges from 0.25–6.0 and room temperature for 30 min. Figure 5a depicts the acquired data, which reveal that the retention capacity was enhanced from pH 0.25 (qe = 12.5 mg/g) to pH 3.0 (qe = 247.5 mg/g) and remained constant up to pH 6.0. It was noticed that in highly acidic medium, pH 0.25–1.0, there was a small variation in HTIP retention due to the high conc. of hydrogen ions, which may compete with uranyl ions and protonated HTIP ligand. The ultimate retention of U(VI) was detected at pH 3.0 (qe = 247.5 mg/g) because UO22+, UO2(OH)+, and (UO2)2(OH)22+ species are predominant in this range. Therefore, pH 3.0 has been proposed as the best pH value for U(VI) ion retention on HTIP/CHCl3, with qe = 247.5 mg/g retention capacity (99.0%). A graph of logD versus pH shows a straight line with a slope of 1.5 and an intersection of 2.7157 in linear regression analysis (slope analysis), as displayed in Figure 5b. The value of slope represents the amount of hydrogen ions set free in the aqueous medium during the formation of the HTIP–U(VI) complex, indicating that about 1.5 moles of hydrogen ions were released during the extraction procedure. Moreover, at pH value 3.0 (logβ = 2.7157), the stability constant (β) of the HTIP–U(VI) complex was computed and found to be 519.63; it showed that the HTIP ligand had a high affinity for uranyl ions. Lastly, based on the preceding analysis, the suggested complex structure and chelation mechanism are presented in Scheme 5. It is clear that the first mechanism mainly depends on the pH value. Competition occurs in highly acidic media between both hydrogen and uranyl ions, causing the tendency of equilibrium to shift towards the left, but in slightly acidic, neutral and alkaline medium, may cause hydrogen ion withdrawal with increased chelation effect and uptake capacity of the HTIP ligand. The second mechanism depends on pH value and the affinity of the HTIP ligand for uranyl ions. The third mechanism is a mixture between the two later mechanisms. 3.2.2. The Influence of Equilibration Time Contact time is one of the highly essential factors on the financial side. The effect of equilibrium time on U(VI) retention was examined using 9.96 × 10−4 mol/L (10 mg/25.0 mL) HTIP/CHCl3 and a 25.0 mL aqueous uranium(VI) ion solution with a concentration of 0.42 ×10−3 mol/L at pH 3.0. U(VI) ion retention rose with rising equilibrium time and reached an ultimate value (247.5 mg/g, 99.0%) after 30 min, which remained roughly constant for the next 120 min, as shown in Figure 6. As a result, 30 min was deemed sufficient for achieving equilibrium in subsequent testing, and it was applied in all successive investigations. Kinetic studies were used to express the rate of uranyl ion entrapment using HTIP/CHCl3, providing crucial information for the design and modeling of the extraction process. Both pseudo first order (PFO) and second order (PSO) kinetic models were employed to identify the mechanism of U(VI) ion entrapment using HTIP/CHCl3 and the rate constant of the process. The PFO model is presented by the equation below [73]:(5) log(qe−qt)=logqe−(k1t2.303) where qe and qt (mg/g) correspond to the quantity of U(VI) ions entrapped per unit mass at equilibrium and time t, respectively, and k1 (min−1) denotes the rate constant. Figure 7a displays a straight line plot of log(qe − qt) against t; the values of both k1 and qe were calculated from the slope and intercept, respectively. The calculated value of qe was 245.47 mg/g, and of k1, 0.1648 min−1, with R2 = 0.9586. It is obvious that the calculated value of qe was extremely close to the practical retention capacity of 247.5 mg/g (Table 1). Nonetheless, the PSO kinetic model was calculated using the equation below [74]:(6) tqt=1k2qe2+tqe where k2 (g/mg.min) is in agreement with the rate constant. Figure 7b demonstrates the straight line of graphing t/qt vs. t; it has a slope of 1/qe and an intercept of 1/k2qe2. The PSO model was found to be ineffective in explaining the practical data. Table 1 illustrates the calculated value of qe as 277.47 mg/g, which was somewhat greater than the experimental retention capacity, whereas the value of k2 was 0.001 min−1 with R2 equal to 0.9991. Therefore, the PFO kinetic model was more reliable in characterizing the extraction process of U(VI) ions using HTIP/CHCl3, as it was appropriate for the experimental data. 3.2.3. The Impact of Initial U(VI) Ion Concentration The impact of the initial concentration of U(VI) ions on extraction efficiency is crucial to investigate since it helps us to anticipate retention power. A diagram of retention power against different initial U(VI) concentrations is shown in Figure 8. Two different stages can be noted. In the first stage, the retention power of HTIP ligand increases conspicuously from 24.75 to 247.5 mg/g with the rising initial concentration of U(VI) ions because the number of active points on the HTIP ligand exceeds the number of uranyl ions in solution. On the other hand, the retention capacity remains constant in the second stage, when the concentration of uranyl ions increases from 100 to 300 mg/L U(VI), because uranyl ions have entirely interacted with the HTIP active sites. The number of active HTIP sites is less than that of uranyl ions. At an initial U(VI) concentration of 100 mg/L, the maximum value of U(VI) ion retention on HTIP/CHCl3 is 247.5 mg/g (99.0%). 3.2.4. The Influence HTIP Dose The HTIP amount is important for better uranyl ion sequestration because it impacts the equilibrium of the system. The effect of HTIP doses varying from 0.0025 to 0.1 g/25.0 mL CHCl3 on U(VI) retention efficiency was investigated. Figure 9a shows that augmenting the HTIP quantity from 0.0025 to 0.01 g/25.0 mL CHCl3 improved U(VI) ion retention efficiency, followed by retention diminishing from 0.025 g to 0.1 g/25.0 mL CHCl3 because the number of HTIP incorporation sites surpassed the number of uranyl ions. The retention capacity of 0.01 g/25.0 mL HTIP/CHCl3 was 247.5 mg/g with a 99.0% removal efficiency. As a consequence, the best concentration for consequent extraction studies was 0.01 g/25.0 mL HTIP/CHCl3. It is obvious that the retention efficiency of HTIP is higher than that of other materials reported in earlier studies, as shown in Table 2. The regression analysis was studied to guarantee the configuration of the HTIP–U(VI) complex formed. Figure 9b illustrates a straight line with a slope of 4.876 and R2 = 0.964 when plotting logD versus logHTIP. The linear regression analysis (slope analysis) can elaborate the stoichiometry mechanism between HTIP chelating ligand and U(VI) ions, which suggests that 1.0 mole of U(VI) ions is chelated by 4.0 moles of HTIP. 3.2.5. U(VI) Ion Distribution Isotherm (McCabe–Thiele Isotherm) Figure 10a demonstrates the McCabe–Thiele diagram of the extraction distribution isotherm of U(VI) ions by HTIP in a system consisting of U(VI) concentration 0.42 × 10−4 mol/L, HTIP/CHCl3 conc. 0.99 × 10−3 mol/L (10 mg/25.0 mL), 3.0 pH value, and A/O phase ratio 1:1 for 30 min. It was observed that two theoretical extraction phases are required to extract nearly all of the U(VI) ions. Furthermore, the elution of U(VI) ions from the HTIP–U(VI) complex was conducted using 2.0 M of H2SO4. The HTIP/CHCl3 organic phase had a concentration of U(VI) ions of 95.0 mg/L. Figure 10b depicts the McCabe–Thiele diagram of the U(VI) ion stripping distribution isotherm; to release almost all of the entrapped U(VI) ions from the HTIP–U(VI) complex, four theoretical stripping stages are required with a 2:1 A/O phase ratio. 3.2.6. The Influence Temperature The effectiveness of temperature on the removal of U(VI) ions using HTIP/CHCl3 was investigated by mixing 25.0 mL of 0.99 × 10−4 mol/L (10 mg/25.0 mL) HTIP/CHCl3 and 25.0 mL of uranium(VI) ion solution with a concentration of 4.2 × 10−4 mol/L (100 mg/L) for 30 min at pH 3.0 and agitation temperature varying from 25–65 °C. The uptake efficiency of HTIP/CHCl3 for U(VI) ions dropped from 247.5 mg/g to 200 mg/g when the temperature was elevated from 25 to 65 °C, (Figure 11a). The pattern suggests that the process of extracting U(VI) ions through HTIP is exothermic. Thermodynamic analysis was also carried out to characterize the extraction mechanism using the Gibbs free energy formula [82,83,84]:(7) ∆G°=∆H°−T∆S° (8) logKd=∆S°2.303R−∆H°2.303RT where ∆G° (kJ/mol) points to Gibbs free energy, ∆H° (kJ/mol) is attributed to enthalpy change, ∆S° (J/mol.K) is defined as entropy change, T (K) stands for the temperature, and R (8.314 J/mol.K) corresponds to the universal gas constant. Figure 11b displays a logKd versus 1/T graph with R2 = 0.9974; ∆H° and ∆S° were generated by the slope and intercept, respectively. The fact that ∆G° is negative implies that the retention of U(VI) on HTIP is thermodynamically spontaneous and attainable (Table 3). In addition, the rise in ∆G° values as temperature rises, from −8.646 kJ/mol at 298 K to −3.926 kJ/mol at 338 K, suggests that the retention of uranium at low temperature is desirable. Moreover, the negative value of ∆H° indicates that U(VI) ion retention on HTIP is an exothermic process, suggesting that heat is produced during the extraction. Finally, the negative value of ∆S° affirms that the capture of U(VI) ions on HTIP is more practicable and less disorganized. The Arrhenius equation was employed to predict the activation energy (Ea) of entrapped U(VI) ions at various temperatures using the slope of the straight line in Figure 11b. The Arrhenius equation was estimated by the subsequent equation [85]:(9) logKd=−2.303EaRT+logA where Ea (kJ/mol) corresponds to the activation energy of extraction and A correlates with the pre-exponential factor, which is independent of temperature. According to calculations, the extraction of U(VI) ions on HTIP ligand requires an activation energy of −8.261 kJ/mol; this implies that the extraction procedure occurs spontaneously and exothermically at room temperature, with no need for activation energy. 3.2.7. The Influence of Co-Ions The investigated co-ions were identified as co-ions accompanied by U(VI) during the leaching process. The impact of co-existing ions was examined separately under optimal extraction conditions by introducing each one into 25.0 mL of a 4.2 × 10−4 mol/L (100 mg/L) U(VI) ion solution. Each ion was identified as an interfering ion when the extraction efficiency differed by more than ±5.0%. Thus, the co-ion concentration that produced a ±5.0% error in uranyl extraction efficiency was used to set the tolerance limit. Table 4 shows that none of the studied co-ions had a negative impact on the retention of U(VI) ions. The findings emphasize the selective extraction of U(VI) on HTIP chelating ligand, which could be used to extract U(VI) ions from leachates of ores in the presence of other ions. 3.3. Stripping and Precipitation Three types of acids with different concentrations ranging from 0.025–2.0 M were utilized as stripping agents for stripping of U(VI) ions from the HTIP–U(VI) complex; the experiments were carried out using 10.0 mL acid volume for 25.0 mL of HTIP–U(VI) complex, shaking well for 10 min at 25 °C. According to the data listed in Table 5, the U(VI) stripping efficiency dropped at low acidic concentrations, but increased when the acid concentration was higher. It was notable that 99.0% stripping efficiency could be achieved with 10.0 mL of 0.5 M HCl, 0.5 M HNO3, or 2.0 M H2SO4. Subsequently, the eluted solution was precipitated with 30.0% NaOH solution until pH 7.0–8.0 was attained, where uranium precipitated as sodium diuranate (Na2U2O7), yielding a final uranium concentrate (yellow cake). The precipitate was left to settle for 24.0 h and filtered. Lastly, the precipitate was dried for 3.0 h at 110 °C in an electrical oven to obtain the uranium concentrate as a final product. 3.4. Case Study: U(VI) Recovery from G. Gattar Ore Sample by HTIP Chelating Ligand The previous data suggest that the HTIP chelating ligand can extract uranium from leachates of geological ores. Accordingly, this hypothesis was investigated using G. Gattar granite leach liquor [85,86]. Uranium (1340 mg/kg) was leached from a G. Gattar ore sample using the percolation leaching technique. The leachate contained approximately 0.45 g/L (450 mg/L) of uranium ions in the presence of a variety of metal ion impurities (Table 6 and Table 7). The recovery experiment was carried out by mixing 1.0 L of HTIP/CHCl3 (0.99 × 10−3 mol/L) with 1.0 L of leach liquor under the previously defined optimum conditions (pH 3.0, 1:1 A/O phase ratio, 30 min, and 25 °C). According to the study, the extraction efficiency of U(VI) reached 99.0%. In addition, it was affirmed that the released U(VI) ions from the HTIP–U(VI) complex could be easily eluted in 10 min using 2.0 M H2SO4. After the elution process, 30.0% NaOH was used to precipitate U(VI) ions as sodium diuranate precipitate Na2U2O7 by adjusting the pH to 7.0–8.0. The uranium concentrate was characterized using XRD and EDX, in addition to ICP-OES analysis techniques in order to determine the U(VI) content alongside other associated metal ions. The results are given in Figure 12, as well as Table 8. According to the analyses, the uranium(VI) content in the uranium concentrate product “yellow cake’’ was 69.93%, with a purity of 93.24%. Figure 13 demonstrates a flow chart of the recovery of U(VI) ions from G. Gattar granite ore mineralization using HTIP chelating ligand. 4. Conclusions A novel synthetic chelating agent, N–hydroxy–N–trioctyl–iminophosphorane (HTIP), was synthesized using an effective substitutional technique compared to the conventional methods and was utilized to uptake U(VI) ions from leach liquor of G. Gattar, North Eastern Desert, Egypt. Characterization was performed successfully using numerous analytical techniques, including 1H–NMR, 13C–NMR, FTIR, EDX, and GC–MS analyses. The uranium sequestration procedures were optimized by mixing 25.0 mL of U(VI) ion solution containing 0.42 × 10−4 mol/L with 0.99 × 10−3 mol/L HTIP/CHCl3 at pH 3.0, 1:1 A/O phase ratio, at 25 °C for 30 min. The utmost retention capacity of HTIP/CHCl3 was 247.5 mg/g. From the stoichiometric calculations, approximately 1.5 hydrogen atoms were released during the extraction at pH 3.0, and 4.0 moles of HTIP ligand were responsible for chelation of 1 mole of uranyl ions. The kinetic modeling data were well-suited to the pseudo first-order model. Furthermore, thermodynamic study demonstrated a negative ΔS° value, indicating that the uptake process is less disordered. Moreover, the rise in ΔG° value pointed to the spontaneousness and possibility of removing U(VI) ions at low temperatures. The elution of uranium loaded on HTIP was achieved using 2.0 M of H2SO4 as uranyl sulfate with 99.0% efficiency. Lastly, uranium concentrate (Na2U2O7, Y.C) with a purity of 93.24% was obtained by adding 30.0% NaOH to the elution and adjusting the pH to 7.0–8.0 with continuous stirring for 2 h. Acknowledgments The authors express their gratitude for the support from Princess Nourah bint Abdulrahman University Researchers Supporting Project number (PNURSP2022R13), Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia. Author Contributions Conceptualization, B.M.A. and A.K.S.; methodology, B.M.A. and A.K.S.; software, B.M.A., A.K.S. and M.F.C.; validation, M.Y.H. and M.I.S.; formal analysis, B.M.A. and A.K.S.; investigation, A.K.S.; resources, B.M.A. and A.K.S.; data curation, M.A.G., H.S.E.-G. and N.M.A.; writing—original draft preparation, B.M.A.; writing—review and editing, A.K.S.; visualization, B.M.A. and A.K.S.; supervision, E.M.E.-S. and M.F.C.; project administration, A.K.S., E.M.E.-S. and M.F.C.; funding acquisition, J.S.A.-O. All authors have read and agreed to the published version of the manuscript. Funding The authors express their gratitude for the support from Princess Nourah bint Abdulrahman University Researchers Supporting Project number (PNURSP2022R13), Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figures, Schemes and Tables polymers-14-01687-sch001_Scheme 1 Scheme 1 The recognized structures of an iminophosphorane and its coordination to a metal. polymers-14-01687-sch002_Scheme 2 Scheme 2 Common structures of complexes of metal with polydentate iminophosphorane–based ligands. polymers-14-01687-sch003_Scheme 3 Scheme 3 Principal techniques of iminophosphorane synthesis. polymers-14-01687-sch004_Scheme 4 Scheme 4 Synthesis of N–Hydroxy–N–trioctyl iminophosphorane (HTIP). Figure 1 (a) FTIR spectra of HTIP and HTIP–U(VI) complex; (b) GS–MS spectrum of HTIP chelating ligand. Figure 2 (a) 1H–NMR spectrum; (b) 13C–NMR spectrum of HTIP chelating ligand. Figure 3 EDX spectra of (a) HTIP; (b) HTIP–U(VI) complex. Figure 4 The HYDRA–MEDUSA diagram of uranyl ions at different pH values. Figure 5 (a) The influence of pH on the retention of uranium by HTIP chelating ligand; (b) the regression analysis of U(VI) ion extraction by HTIP at various pH values. (U(VI) conc.: 0.42 × 10−4 mol/L, HTIP/CHCl3 conc.: 0.99 × 10−3 mol/L (10 mg/25 mL), A/O 1:1, T: 25 °C, time: 30 min). polymers-14-01687-sch005_Scheme 5 Scheme 5 Proposed mechanisms of UO22+ ion extraction by HTIP chelating ligand. Figure 6 The influence of time on the retention of uranium by HTIP chelating ligand. (U(VI) conc.: 0.42 × 10−4 mol/L, HTIP/CHCl3 conc.: 0.99 × 10−3 mol/L (10 mg/25 mL), pH 3.0, A/O 1:1, T: 25 °C). Figure 7 (a) PFO Kinetic model; (b) PSO kinetic model of U(VI) ion extraction by HTIP chelating ligand. (U(VI) conc.: 0.42 × 10−4 mol/L, HTIP/CHCl3 conc.: 0.99 × 10−3 mol/L (10 mg/25 mL), pH 3.0, A/O 1:1, T: 25 °C). Figure 8 The influence of initial U(VI) ion concentration on the retention of uranium by HTIP chelating ligand. (pH 3.0, HTIP/CHCl3 conc.: 0.99 × 10−3 mol/L (10 mg/25 mL), time: 30 min, A/O 1:1, T: 25 °C). Figure 9 (a) The influence of HTIP dose on the retention of uranium; (b) the regression analysis of U(VI) ion extraction by HTIP at different concentrations. (U(VI) conc.: 0.42 × 10−4 mol/L, pH 3.0, Time: 30 min, A/O 1:1, T: 25 °C). Figure 10 McCabe–Thiele diagram of (a) U(VI) ion extraction by HTIP chelating ligand; (b) U(VI) ion stripping isotherm from HTIP–U(VI) complex. (U(VI) conc.: 0.42 × 10−4 mol/L, HTIP/CHCl3 conc.: 0.99 × 10−3 mol/L (10 mg/25 mL), pH 3.0, time: 30 min, T: 25 °C). Figure 11 (a) The influence of temperature on the retention of uranium by HTIP chelating ligand; (b) thermodynamic study of U(VI) ion extraction by HTIP chelating ligand (U(VI) conc.: 0.42 × 10−4 mol/L, HTIP/CHCl3 conc.: 0.99 × 10−3 mol/L (10 mg/25 mL), pH 3.0, A/O 1:1, time: 30 min). Figure 12 (a) XRD spectrum; (b) EDX spectrum of Na2U2O7 product from G. Gattar granite ore sample. Figure 13 Flow chart illustrating the production of Na2U2O7 from G. Gattar granite ore sample using HTIP chelating ligand. polymers-14-01687-t001_Table 1 Table 1 The kinetic constants of U(VI) ion extraction by HTIP chelating ligand. Extraction Capacity qe, mg/g PFO PSO qe k 1 R 2 qe k 2 R 2 247.5 mg/g 245.47 0.1648 0.9586 277.47 0.001 0.9991 polymers-14-01687-t002_Table 2 Table 2 Comparison of retention efficiency of different materials. Materials Retention Efficiency, mg.g−1 Reference Poly (aminophosphonic)–functionalized poly (glycidyl methacrylate)–magnetic nanocomposite 262.5 [75] Nitrogen-enriched carbon–nitrogen polymer, C3N5 207.0 [76] Cyanex 923 loaded polymer beads 54.5 [77] 2–Hydroxy–4–aminotriazine–anchored activated carbon 135.0 [78] D2EHPA–TOPO/SiO2–P 48.0 [79] D2EHPA–TOPO@MCM–41 2.88 [80] Eggplant (Solanum melongena) leaves 110.97 [81] HTIP 247.5 This study polymers-14-01687-t003_Table 3 Table 3 Thermodynamic analysis of U(VI) ion extraction by HTIP chelating ligand. ∆S°, J/mol.K ∆H°, kJ/mol ∆G°, kJ/mol −118 × 10−6 −43.81 298 K 308 K 318 K 328 K 338 K −8.645 −7.466 −6.286 −5.106 −3.926 polymers-14-01687-t004_Table 4 Table 4 The influence of co-ions on U(VI) ion extraction by HTIP chelating ligand. Co-Ions Tolerance Limit *, mg/L E% Co-Ions Tolerance Limit *, mg/L E% Si4+ 3000 99.0 Ba2+ 1000 95.0 K+ 3000 99.0 Ni2+ 550 95.0 Na+ 3000 99.0 Rb+ 650 95.0 Al3+ 3000 99.0 Pb2+ 1000 95.0 Ca2+ 1500 95.0 Mo6+ 850 95.0 Mg2+ 1500 95.0 Zr4+ 950 95.0 Fe3+ 750 95.0 Sr2+ 700 95.0 P5+ 3000 99.0 V5+ 1000 95.0 Ti4+ 850 95.0 Zn2+ 900 95.0 * Tolerance limit: the conc. of co-ion that produces an error does not exceed ± 5.0%. polymers-14-01687-t005_Table 5 Table 5 Effect of stripping agent conc. on U(VI) ion elution from HTIP–U(VI) complex. Acid Conc., (M) Stripping Efficiency, (%) HNO3 HCl H2SO4 0.025 63.51% 58.6% 57.3% 0.05 85.7% 69.8% 68.2% 0.1 95.8% 93.5% 75.4% 0.5 99.0% 99.0% 89.3% 1.0 99.0% 99.0% 96.2% 2.0 99.0% 99.0% 99.0% polymers-14-01687-t006_Table 6 Table 6 Chemical composition of G. Gattar granite ore sample. Major Oxides Wt., % Elements Conc., mg/kg SiO2 48.60 U6+ 1340 Al2O3 8.16 Ni2+ 34 FeO 0.14 Rb2+ 404 Fe2O3 2.59 Sr2+ 38 TiO2 0.15 Y3+ 64 MnO 0.07 Cu2+ 65 K2O 3.38 Zn2+ 200 CaO 6.73 Ba2+ 432 Na2O 2.35 Zr4+ 170 P2O5 0.02 Nb5+ 25 MgO 0.60 Pb2+ 36 Mo6+ 228 polymers-14-01687-t007_Table 7 Table 7 The chemical analysis of G. Gattar granite leach liquor. Metal ions Si4+ Al3+ Fe3+ Mo6+ V5+ P5+ Na+ K+ Zr4+ Ti4+ U6+ Conc., g/L 2.33 3.11 1.31 0.06 0.11 0.07 2.11 2.87 0.05 0.03 0.45 polymers-14-01687-t008_Table 8 Table 8 ICP-OES analysis of Na2U2O7 product from G. Gattar granite ore sample. Elements Content, (%) Elements Content, (%) U6+ 69.93 Cr6+ 0.0031 Na+ 4.041 Cd2+ 0.0007 Al3+ 0.0158 Si4+ 0.0091 Fe3+ 0.0207 K+ 0.0185 Mg2+ 0.0391 Zr4+ 0.0035 Ca2+ 0.0521 Pb5+ 0.0028 Co2+ 0.0053 Zn2+ 0.0037 Cu2+ 0.0023 V5+ 0.0035 Ni2+ 0.0031 Mn2+ 0.0031 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Dehnicke K. Krieger M. Massa W. Phosphoraneiminato complexes of transition metals Coord. Chem. Rev. 1999 182 19 65 10.1016/S0010-8545(98)00191-X 2. Dehnicke K. Weller F. Phosphorane iminato complexes of main group elements Coord. Chem. Rev. 1997 158 103 169 10.1016/S0010-8545(97)90055-2 3. Staudinger H. Meyer J. Über neue organische Phosphorverbindungen III. Phosphinmethylenderivate und Phosphinimine Helv. Chim. Acta 1919 2 635 646 10.1002/hlca.19190020164 4. Stephan D.W. Sterically Demanding Phosphinimides: Ligands for Unique Main Group and Transition Metal Chemistry Adv. Organomet. Chem. 2006 54 267 291 10.1016/S0065-3055(05)54006-1 5. Gröb T. Geiseler G. Harms K. Greiner A. Dehnicke K. Phosphaniminato-Komplexe des Zirconiums: Die Kristallstrukturen von [ZrCl3(NPPh3)(HNPPh3)2] und [ZrCl2(NPPh3)2(HNPPh3)2] Z. Für Anorg. Und Allg. Chem. 2002 628 217 221 10.1002/1521-3749(200201)628:1<217::AID-ZAAC217>3.0.CO;2-G 6. Gololobov Y.G. Kasukhin L.F. Recent advances in the staudinger reaction Tetrahedron 1992 48 1353 1406 10.1016/S0040-4020(01)92229-X 7. García-Álvarez J. García-Garrido S.E. Cadierno V. Iminophosphorane–phosphines: Versatile ligands for homogeneous catalysis J. Organomet. Chem. 2014 751 792 808 10.1016/j.jorganchem.2013.07.009 8. Allcock H.R. Chemistry and Applications of Polyphosphazenes Wiley Interscience Hoboken, NJ, USA 2003 1 725 9. Gleria M. De Jaeger R. Applicative Aspects of Poly (Organophosphazenes) UK ed. Nova Science Publishers Inc. Hauppauge, NY, USA 2004 1 308 10. Steiner A. Zacchini S. Richards P.I. From neutral iminophosphoranes to multianionic phosphazenates. The coordination chemistry of imino–aza-P(V) ligands Coord. Chem. Rev. 2002 227 193 216 10.1016/S0010-8545(02)00012-7 11. Gololobov Y.G. Zhmurova I.N. Kasukhin L.F. Sixty years of staudinger reaction Tetrahedron 1981 37 437 472 10.1016/S0040-4020(01)92417-2 12. van Berkel S.S. van Eldijk M.B. van Hest J.C.M. Staudinger Ligation as a Method for Bioconjugation Angew. Chem. Int. Ed. 2011 50 8806 8827 10.1002/anie.201008102 13. Cates L.A. Ferguson N.M. Phosphorus-Nitrogen Compounds II: Some p-Toluidine Derivatives J. Pharm. Sci. 1964 53 973 975 10.1002/jps.2600530835 14241823 14. Amin A.M. Wang L. Wang J. Yu H. Huo J. Gao J. Xiao A. Recent Research Progress in the Synthesis of Polyphosphazene and Their Applications Des. Monomers Polym. 2009 12 357 375 10.1163/138577209X12486896623373 15. Rothemund S. Teasdale I. Preparation of polyphosphazenes: A tutorial review Chem. Soc. Rev. 2016 45 5200 5215 10.1039/C6CS00340K 27314867 16. Strasser P. Teasdale I. Main-Chain Phosphorus-Containing Polymers for Therapeutic Applications Molecules 2020 25 1716 10.3390/molecules25071716 32276516 17. Xu G.-X. Lu Q. Yu B.-T. Wen L. Inorganic polymer phosphazene disulfide as cathode material for rechargeable lithium batteries Solid State Ion. 2006 177 305 309 10.1016/j.ssi.2005.10.029 18. Singh A. Krogman N.R. Sethuraman S. Nair L.S. Sturgeon J.L. Brown P.W. Laurencin C.T. Allcock H.R. Effect of side group chemistry on the properties of biodegradable L-alanine cosubstituted polyphosphazenes Biomacromolecules 2006 7 914 918 10.1021/bm050752r 16529431 19. Brandt K. Kruszyński R. Bartczak T.J. Porwolik-Czomperlik I. AIDS-related lymphoma screen results and molecular structure determination of a new crown ether bearing aziridinylcyclophosphazene, potentially capable of ion-regulated DNA cleavage action Inorg. Chim. Acta 2001 322 138 144 10.1016/S0020-1693(01)00557-6 20. Keller M.A. Saba C.S. Oxidative Stability and Degradation Mechanism of a Cyclotriphosphazene Lubricant Anal. Chem. 1996 68 3489 3492 10.1021/ac960632x 21619281 21. Davarcı D. Beşli S. Demirbas E. Synthesis of a series of triple-bridged cyclotriphosphazene hexa-alkoxy derivatives and investigation of their structural and mesomorphic properties Liq. Cryst. 2013 40 624 631 10.1080/02678292.2013.773093 22. Cil E. Tanyildizi M.A. Ozen F. Boybay M. Arslan M. Gorgulu A.O. Synthesis, characterization, and biological-pharmacological evaluation of new phosphazenes bearing dioxybiphenyl and schiff base groups Arch. Der Pharm. 2012 345 476 485 10.1002/ardp.201100412 23. Allcock H.R. Kwon S. Covalent linkage of proteins to surface-modified poly(organophosphazenes): Immobilization of glucose-6-phosphate dehydrogenase and trypsin Macromolecules 1986 19 1502 1508 10.1021/ma00160a002 24. Sun J. Yu Z. Wang X. Wu D. Synthesis and Performance of Cyclomatrix Polyphosphazene Derived from Trispiro-Cyclotriphosphazene as a Halogen-Free Nonflammable Material ACS Sustain. Chem. Eng. 2014 2 231 238 10.1021/sc400283d 25. Greish Y.E. Bender J.D. Lakshmi S. Brown P.W. Allcock H.R. Laurencin C.T. Low temperature formation of hydroxyapatite-poly(alkyl oxybenzoate)phosphazene composites for biomedical applications Biomaterials 2005 26 1 9 10.1016/j.biomaterials.2004.02.016 15193876 26. Bortolus P. Gleria M. Photochemistry and photophysics of poly(organophosphazenes) and related compounds: A review. I. Monomolecular processes J. Inorg. Organomet. Polym. 1994 4 1 29 10.1007/BF00684025 27. Critchley J.P. Knight G.J. Wright W.W. Heat-Resistant Polymers Springer Berlin/Heidelberg, Germany 1983 1 403 10.1007/978-1-4899-0396-9 28. Potin P. De Jaeger R. Polyphosphazenes: Synthesis, structures, properties, applications Eur. Polym. J. 1991 27 341 348 10.1016/0014-3057(91)90185-Q 29. Chandrasekhar V. Thomas K.R.J. Coordination and organometallic chemistry of cyclophosphazenes and polyphosphazenes Appl. Organomet. Chem. 1993 7 1 31 10.1002/aoc.590070102 30. Prange R. Allcock H.R. Telechelic Syntheses of the First Phosphazene Siloxane Block Copolymers Macromolecules 1999 32 6390 6392 10.1021/ma991136o 31. Gleria M. De Jaeger R. Aspects of Phosphazene Research J. Inorg. Organomet. Polym. 2001 11 1 45 10.1023/A:1013276518701 32. Gado M. Atia B. Morcy A. The role of graphene oxide anchored 1-amino-2-naphthol-4-sulphonic acid on the adsorption of uranyl ions from aqueous solution: Kinetic and thermodynamic features Int. J. Environ. Anal. Chem. 2019 99 996 1015 10.1080/03067319.2019.1617283 33. Sakr A.K. Mohamed S.A. Mira H.I. Cheira M.F. Successive leaching of uranium and rare earth elements from El Sela mineralization J. Sci. Eng. Res. 2018 5 95 111 34. Xue Y. Cao M. Gao J. Gui Y. Chen J. Liu P. Ma F. Yan Y. Qiu M. Electroadsorption of uranium on amidoxime modified graphite felt Sep. Purif. Technol. 2021 255 117753 10.1016/j.seppur.2020.117753 35. Wang Y. Wang J. Ding Z. Wang W. Song J. Li P. Liang J. Fan Q. Light Promotes the Immobilization of U(VI) by Ferrihydrite Molecules 2022 27 1859 10.3390/molecules27061859 35335223 36. Hammond C.R. The Elements, in Handbook of Chemistry and Physics 81st ed. CRC Press Boca Raton, FL, USA 2000 37. Cheira M.F. El-Didamony A.M. Mahmoud K.F. Atia B.M. Equilibrium and Kinetic Characteristics of Uranium Recovery by the Strong Base Ambersep 920U Cl Resin IOSR J. Appl. Chem. 2014 7 32 40 10.9790/5736-07533240 38. Kumar J.R. Kim J.-S. Lee J.-Y. Yoon H.-S. A Brief Review on Solvent Extraction of Uranium from Acidic Solutions Sep. Purif. Rev. 2011 40 77 125 10.1080/15422119.2010.549760 39. Cote G. Hydrometallurgy of strategic metals Solvent Extr. Ion Exch. 2000 18 703 727 10.1080/07366290008934704 40. Cheira M.F. Mira H.I. Sakr A.K. Mohamed S.A. Adsorption of U(VI) from acid solution on a low-cost sorbent: Equilibrium, kinetic, and thermodynamic assessments Nucl. Sci. Tech. 2019 30 156 10.1007/s41365-019-0674-3 41. Hassanin M.A. Negm S.H. Youssef M.A. Sakr A.K. Mira H.I. Mohammaden T.F. Al-Otaibi J.S. Hanfi M.Y. Sayyed M.I. Cheira M.F. Sustainable Remedy Waste to Generate SiO2 Functionalized on Graphene Oxide for Removal of U(VI) Ions Sustainability 2022 14 2699 10.3390/su14052699 42. Hassanin M.A. El-Gendy H.S. Cheira M.F. Atia B.M. Uranium ions extraction from the waste solution by thiosemicarbazide anchored cellulose acetate Int. J. Environ. Anal. Chem. 2021 101 351 369 10.1080/03067319.2019.1667984 43. Yu H. Dai Y. Zhou L. Ouyang J. Tang X. Liu Z. Adesina A.A. Selective biosorption of U(VI) from aqueous solution by ion-imprinted honeycomb-like chitosan/kaolin clay composite foams Int. J. Biol. Macromol. 2022 206 409 421 10.1016/j.ijbiomac.2022.02.168 35245572 44. Wang Y. Chen B. Xiong T. Zhang Y. Zhu W. Immobilization of U(VI) in wastewater using coal fly ash aerogel (CFAA) as a low-cost adsorbent Process Saf. Environ. Prot. 2022 160 900 909 10.1016/j.psep.2022.03.006 45. Chen W. Feng J. Liu S. Zhang J. Cai Y. Lv Z. Fang M. Tan X. A green and economical MgO/biochar composite for the removal of U(VI) from aqueous solutions Chem. Eng. Res. Des. 2022 180 391 401 10.1016/j.cherd.2022.02.031 46. Liu Y. Tang X. Zhou L. Liu Z. Ouyang J. Dai Y. Le Z. Adesina A.A. Nanofabricated chitosan/graphene oxide electrodes for enhancing electrosorptive removal of U(VI) from aqueous solution Sep. Purif. Technol. 2022 290 120827 10.1016/j.seppur.2022.120827 47. Wang P. Tan K. Li Y. Xiao W. Liu Z. Tan W. Xu Y. The adsorption of U(VI) by albite during acid in-situ leaching mining of uranium J. Radioanal. Nucl. Chem. 2022 1 9 10.1007/s10967-022-08254-9 48. Collet S. Chagnes A. Courtaud B. Thiry J. Cote G. Solvent extraction of uranium from acidic sulfate media by Alamine® 336: Computer simulation and optimization of the flow-sheets J. Chem. Technol. Biotechnol. 2009 84 1331 1337 10.1002/jctb.2184 49. Eskandari Nasab M. Solvent extraction separation of uranium(VI) and thorium(IV) with neutral organophosphorus and amine ligands Fuel 2014 116 595 600 10.1016/j.fuel.2013.08.043 50. Amaral J.C.B.S. Morais C.A. Thorium and uranium extraction from rare earth elements in monazite sulfuric acid liquor through solvent extraction Miner. Eng. 2010 23 498 503 10.1016/j.mineng.2010.01.003 51. Sole K.C. Cole P.M. Feather A.M. Kotze M.H. Solvent Extraction and Ion Exchange Applications in Africa’s Resurging Uranium Industry: A Review Solvent Extr. Ion Exch. 2011 29 868 899 10.1080/07366299.2011.581101 52. Mondal S. Singh D.K. Anitha M. Sharma J.N. Hubli R.C. Singh H. New synergistic solvent mixture of DNPPA and bidentate octyl (phenyl) CMPO for enhanced extraction of uranium (VI) from phosphoric acid medium Hydrometallurgy 2014 147–148 95 102 10.1016/j.hydromet.2014.04.023 53. Quinn J.E. Wilkins D. Soldenhoff K.H. Solvent extraction of uranium from saline leach liquors using DEHPA/Alamine 336 mixed reagent Hydrometallurgy 2013 134–135 74 79 10.1016/j.hydromet.2013.01.014 54. Beltrami D. Chagnes A. Haddad M. Laureano H. Mokhtari H. Courtaud B. Jugé S. Cote G. Solvent extraction studies of uranium(VI) from phosphoric acid: Role of synergistic reagents in mixture with bis(2-ethylhexyl) phosphoric acid Hydrometallurgy 2014 144–145 207 214 10.1016/j.hydromet.2014.02.010 55. Zhu Z. Pranolo Y. Cheng C.Y. Uranium recovery from strong acidic solutions by solvent extraction with Cyanex 923 and a modifier Miner. Eng. 2016 89 77 83 10.1016/j.mineng.2016.01.016 56. Hassaballa H. Steed J.W. Junk P.C. Elsegood M.R.J. Formation of Lanthanide and Actinide Oxonium Ion Complexes with Crown Ethers from a Liquid Clathrate Medium Inorg. Chem. 1998 37 4666 4671 10.1021/ic9802597 11670619 57. Salmon L. Thuéry P. Ephritikhine M. Uranium(iv) complexes of calix[n]arenes (n = 4, 6 and 8) Chem. Commun. 2006 8 856 858 10.1039/b516438a 58. Sopo H. Goljahanpoor K. Sillanpää R. Aminoalkylbis(phenolate) [O,N,O] donor ligands for uranyl(VI) ion coordination: Syntheses, structures, and extraction studies Polyhedron 2007 26 3397 3408 10.1016/j.poly.2007.03.032 59. El-Nadi Y.A. Daoud J.A. Aly H.F. Modified leaching and extraction of uranium from hydrous oxide cake of Egyptian monazite Int. J. Miner. Processing 2005 76 101 110 10.1016/j.minpro.2004.12.005 60. Zhu Z. Pranolo Y. Cheng C.Y. Uranium Solvent Extraction and Separation From Vanadium in Alkaline Solutions Sep. Sci. Technol. 2013 48 1402 1408 10.1080/01496395.2012.738277 61. Reiller P.E. Mariet C. Luminescence of uranium(VI) after liquid-liquid extraction from HCl by Aliquat® 336 in n-dodecane:1-decanol by time-resolved laser-induced luminescence spectroscopy Radiochim. Acta 2020 108 527 533 10.1515/ract-2019-3177 62. Atia B.M. Khawassek Y.M. Hussein G.M. Gado M.A. El-Sheify M.A. Cheira M.F. One-pot synthesis of pyridine dicarboxamide derivative and its application for uranium separation from acidic medium J. Environ. Chem. Eng. 2021 9 105726 10.1016/j.jece.2021.105726 63. Atia B.M. Gado M.A. Cheira M.F. El-Gendy H.S. Yousef M.A. Hashem M.D. Direct synthesis of a chelating carboxamide derivative and its application for thorium extraction from Abu Rusheid ore sample, South Eastern Desert, Egypt Int. J. Environ. Anal. Chem. 2021 101 1 24 10.1080/03067319.2021.1924161 64. Spjuth L. Liljenzin J.O. Skålberg M. Hudson M.J. Chan G.Y.S. Drew M.G.B. Feaviour M. Iveson P.B. Madic C. Extraction of Actinides and Lanthanides from Nitric Acid Solution by Malonamides Radiochim. Acta 1997 78 39 46 10.1524/ract.1997.78.special-issue.39 65. Abd El-Rahem K.A. Taher M.A. Cheira M.F. Peculiarities of u (vi) adsorption from acidic solution using chitin-derived chitosan as a low-priced bio-adsorbent Al-Azhar Bull. Sci. 2021 32 57 70 10.21608/absb.2021.79764.1121 66. Mathew K. Bürger S. Vogt S. Mason P. Morales-Arteaga M. Narayanan U. Uranium assay determination using Davies and Gray titration: An overview and implementation of GUM for uncertainty evaluation J. Radioanal. Nucl. Chem. 2009 282 939 944 10.1007/s10967-009-0186-4 67. Cheira M.F. Solvent extraction of uranium and vanadium from carbonate leach solutions of ferruginous siltstone using cetylpyridinium carbonate in kerosene Chem. Pap. 2020 74 2247 2266 10.1007/s11696-020-01073-w 68. Hecht M. Anaya S.S. Hagenbach A. Abram U. Rhenium(V) and Technetium(V) Complexes with Phosphoraneimine and Phosphoraneiminato Ligands Inorg. Chem. 2005 44 3172 3180 10.1021/ic048210q 15847424 69. Li L. Sakr A.K. Schlöder T. Klein S. Beckers H. Kitsaras M.-P. Snelling H.V. Young N.A. Andrae D. Riedel S. Searching for Monomeric Nickel Tetrafluoride: Unravelling Infrared Matrix Isolation Spectra of Higher Nickel Fluorides Angew. Chem. Int. Ed. 2021 60 6391 6394 10.1002/anie.202015501 70. Dal Bo F. Aksenov S.M. Hatert F. Burns P.C. Synthesis, IR spectroscopy and crystal structure of [(UO2)2{Be(H2O)2(PO4)2}]·(H2O), the first compound with a trimer beryllophosphate anion Z. Für Krist. Cryst. Mater. 2018 233 391 398 10.1515/zkri-2017-2113 71. Nekhunguni P.M. Tavengwa N.T. Tutu H. Sorption of uranium(VI) onto hydrous ferric oxide-modified zeolite: Assessment of the effect of pH, contact time, temperature, selected cations and anions on sorbent interactions J. Environ. Manag. 2017 204 571 582 10.1016/j.jenvman.2017.09.034 72. Müller K. Brendler V. Foerstendorf H. Aqueous Uranium(VI) Hydrolysis Species Characterized by Attenuated Total Reflection Fourier-Transform Infrared Spectroscopy Inorg. Chem. 2008 47 10127 10134 10.1021/ic8005103 18831578 73. Abdel Geleel M. Atwa S.T. Sakr A.K. Removal of Cr (III) from aqueous waste using Spent Activated Clay J. Am. Sci. 2013 9 256 262 10.7537/marsjas090213.35 74. Allam E.M. Lashen T.A. Abou El-Enein S.A. Hassanin M.A. Sakr A.K. Hanfi M.Y. Sayyed M.I. Al-Otaibi J.S. Cheira M.F. Cetylpyridinium Bromide/Polyvinyl Chloride for Substantially Efficient Capture of Rare Earth Elements from Chloride Solution Polymers 2022 14 954 10.3390/polym14050954 35267777 75. Galhoum A.A. Eisa W.H. El-Tantawy El-Sayed I. Tolba A.A. Shalaby Z.M. Mohamady S.I. Muhammad S.S. Hussien S.S. Akashi T. Guibal E. A new route for manufacturing poly(aminophosphonic)-functionalized poly(glycidyl methacrylate)-magnetic nanocomposite-Application to uranium sorption from ore leachate Environ. Pollut. 2020 264 114797 10.1016/j.envpol.2020.114797 32559874 76. Wang Z. Li W. Wu L. Wang Z. Cao Y. Cheng J. Chen G. Zhao Q. Jiang M. Chen Z. Nitrogen-rich carbon nitrogen polymers for enhancing the sorption of uranyl Chin. Chem. Lett. 2022 10.1016/j.cclet.2022.03.097 77. Gujar R.B. Lakshmi D.S. Figoli A. Mohapatra P.K. Polymeric beads containing Cyanex 923 for actinide uptake from nitric acid medium: Studies with uranium and plutonium J. Chromatogr. A 2013 1305 48 54 10.1016/j.chroma.2013.07.032 23891375 78. Mayyas M. Al-Harahsheh M. Wei X.-Y. Solid phase extractive preconcentration of Uranium from Jordanian phosphoric acid using 2-hydroxy-4-aminotriazine-anchored activated carbon Hydrometallurgy 2014 149 41 49 10.1016/j.hydromet.2014.07.005 79. Khayambashi A. Wang X. Wei Y. Solid phase extraction of uranium (VI) from phosphoric acid medium using macroporous silica-based D2EHPA-TOPO impregnated polymeric adsorbent Hydrometallurgy 2016 164 90 96 10.1016/j.hydromet.2016.05.013 80. Abdel-Magied A.F. Solid phase extraction of uranium from phosphoric acid: Kinetic and thermodynamic study Radiochim. Acta 2017 105 813 820 10.1515/ract-2017-0001 81. Mahmoud M.A. Sorption of U(VI) ions from aqueous solution by eggplant leaves: Isotherm, kinetics and thermodynamics studies Prog. Nucl. Energy 2021 138 103829 10.1016/j.pnucene.2021.103829 82. Allam E.M. Lashen T.A. Abou El-Enein S.A. Hassanin M.A. Sakr A.K. Cheira M.F. Almuqrin A. Hanfi M.Y. Sayyed M.I. Rare Earth Group Separation after Extraction Using Sodium Diethyldithiocarbamate/Polyvinyl Chloride from Lamprophyre Dykes Leachate Materials 2022 15 1211 10.3390/ma15031211 35161155 83. Sakr A.K. Cheira M.F. Hassanin M.A. Mira H.I. Mohamed S.A. Khandaker M.U. Osman H. Eed E.M. Sayyed M.I. Hanfi M.Y. Adsorption of Yttrium Ions on 3-Amino-5-Hydroxypyrazole Impregnated Bleaching Clay, a Novel Sorbent Material Appl. Sci. 2021 11 10320 10.3390/app112110320 84. Mahmoud N.S. Atwa S.T. Sakr A.K. Abdel Geleel M. Kinetic and thermodynamic study of the adsorption of Ni (II) using Spent Activated clay Mineral N. Y. Sci. J. 2012 5 62 68 85. Zidan I.H. Cheira M.F. Bakry A.R. Atia B.M. Potentiality of uranium recovery from G. Gattar leach liquor using Duolite ES-467 chelating resin: Kinetic, thermodynamic and isotherm features Int. J. Environ. Anal. Chem. 2020 100 1 23 10.1080/03067319.2020.1748613 86. Atia B.M. 1,5-Diphenylthiocarbazone-Modified Amberlite XAD-7: A Chelating Collector for Upgrading Uranium Concentrate Produced from G. Gattar Pilot Plant, North Eastern Desert, Egypt Radiochemistry 2019 61 728 732 10.1134/S1066362219060158

PMC009xxxxxx/PMC9099653.txt

==== Front Int J Environ Res Public Health Int J Environ Res Public Health ijerph International Journal of Environmental Research and Public Health 1661-7827 1660-4601 MDPI 10.3390/ijerph19095136 ijerph-19-05136 Article More Income, Less Pollution? How Income Expectation Affects Pesticide Application Su Xiaoshan 1 Shi Jingyi 1 https://orcid.org/0000-0002-3783-6472 Wang Tianxi 2 Shen Qinghui 1 Niu Wentao 1* Xu Zhenzhen 3 Xu Erqi Academic Editor Hua Xiaobo Academic Editor 1 School of Management, Zhengzhou University, Zhengzhou 450001, China; suxiaoshan@zzu.edu.cn (X.S.); shi__jingyi@163.com (J.S.); shenqh1022@163.com (Q.S.) 2 Business School, University of Edinburgh, 29 Buccleuch Place, Edinburgh EH8 9JS, UK; s1819714@ed.ac.uk 3 School of Architecture and Built Environment, Deakin University, Geelong 3219, Australia; xuzhenz@deakin.edu.au * Correspondence: wentaoniu@zzu.edu.cn 23 4 2022 5 2022 19 9 513601 3 2022 21 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Farmers are still the foundation of China’s current “small, scattered, and weak” agricultural production pattern. As such, increasing guidance for reduction response behavior is central to reducing agricultural pesticide use. Following this pesticide reduction logic, four of the most widely promoted pesticide reduction technologies, including light trapping, biopesticide application, healthy crop growth, and insect-proof net technologies, were selected, and a theoretical analysis framework of farmers’ willingness to adopt these technologies was constructed based on the theories of value perception and planned behavior. An ordered logistic regression model is used to explore key factors behind current pesticide reduction technology perceptions, technology response willingness, and behavioral decisions of farmers in China, with survey data from 516 farmers in Henan Province. The results show that among the four pesticide reduction technologies, healthy crop growth technology is the most-appealing one for farmers, followed by insect-proof net technology and biopesticide application technology. The least-appealing one for farmers is the light trapping technology. Farmers’ perceived degree of income improvement from technology adoption is the main determinant of their willingness, which is positively significant at a 1% confidence level in all four models. In addition, farmers’ willingness to respond to technologies is also significantly influenced by farmers’ perception of technical operational ability, perception of risk from adopting technology, government-related subsidies, government technical training guidance, trust in government promotion of technology, and perception of the government’s role in improving the external environment for adopting technology. pesticide reduction farmer response behavior value perception theory of planned behavior ==== Body pmc1. Introduction Overreliance on chemical pesticides, together with long-term excessive and inefficient use of pesticides in small-scale decentralized agriculture, has led to a series of problems such as non-point source pollution, environmental damage, declines in agricultural product safety, and risks to human and animal safety [1]. Therefore, more attention should be paid to protecting consumers and field workers (farmers) from the risks to their health and harmful effects of using pesticides [2] and ingesting them with food and drinking water [3]. The Chinese government is concerned with the excessive and inefficient application of pesticides and has introduced a series of policies to guide and reduce pesticide use in agriculture. In February 2021, the “Guiding Opinions of the State Council on Accelerating the Establishment and Improvement of a Green, Low-Carbon and Circular Development Economic System” and the “Central Document No. 1” (the first policy statement released by central authorities each year) were released. Both documents proposed to “accelerate green development of agriculture” and emphasized this paradigm shift as a priority. Since the Ministry of Agriculture issued the “Action Plan for Zero Growth in Pesticide Use by 2020” in 2015, the demand for pesticides in China has shrunk steadily. In 2019, the commodity quantity of pesticides used in China was 1.456 million tons, down 3.2% compared with that in 2018. In 2020, the pesticide utilization rate of China’s three major staple foods reached 40.6% (Ministry of Agriculture and Rural Affairs, 2021), with an increase of 4% over 2015. However, the pesticide application rate in China is not only much higher than the world average but also far exceeds the optimal level in the economic sense [4]. Therefore, there is still ample opportunity to reduce the quantity of formulated product applied as well as the increase application efficiency. Thus, it is necessary to explore the path of pesticide reduction in depth. In literature, researchers have focused on pesticide reduction from four main perspectives. The first one is the farmer’s perspective, which mainly includes the characteristics of resource endowment and psychological perceptions. Farmers’ decision-making behavior around reducing pesticide use is affected by the gender of the household head, age [5], education level, family planting size [6], degree of part-time employment [7], health status [8], number of family laborers, source of family income [9], identity characteristics [10], etc. In addition, farmers’ expected cost savings and income improvement [11], ecological environment perceptions [12], risk preferences [1], public image [13], learning and operating ability perceptions, and sense of moral responsibility [14] can also significantly affect their decision-making behaviors on reducing pesticides. Among these factors, farmers are more inclined to focus on whether reducing pesticide use can optimize their private economic goals by saving costs and increasing income [11]. Moreover, farmers’ understanding of the relationship between pesticide use and their own health can also influence their decisions to reduce pesticides. Empirical studies have found that farmers’ health awareness is one of the most important determinants influencing the intensity of pesticide application [3,15,16]. Farmers will experience varying degrees of pesticide exposure during pesticide application, which in severe cases can even lead to acute poisoning. About 7% of the agricultural population in the world is affected by pesticide poisoning [17,18]. From 2007 to 2019, only Chongming District, Shanghai, China, reported as many as 1182 cases of pesticide poisoning [19]. The increasing exposure to pesticides has led to an increasing number of farmers perceiving a greater health risk in the process of pesticide application reduction. The second perspective is based on the organization of the agricultural business. Pesticide reduction is influenced by the characteristics of agricultural business [20]. More moderate-scale agriculture operations [21], more embedded cooperative organizations [22], more service of scale operations, and a better division of labor transactions [23] can effectively contribute to pesticide reduction. The third perspective focuses on the external institutional environment. Agricultural pesticide reduction has a positive economic externality [23], which calls for government intervention to encourage farmers to reduce pesticide use. Government intervention mainly includes policy advocacy, project support [24], government subsidies, a system of proper rewards and punishments [25], institutional trust [26], social capital [27], and social norms [28]. The fourth perspective is the consumer perspective. Consumers’ purchasing preferences can significantly affect agricultural production behavior. Therefore, through green product marketing, health information communication, and other methods [29], consumers can have a stronger passion for green agricultural products, and pesticide use can be reduced [30]. Although pesticide application involves multiple stakeholders, farmers are still the foundation of the “small, scattered, and weak” production pattern of China’s agriculture. Thus, raising farmers’ pesticide reduction awareness is an important logical prerequisite for reducing pesticides in agriculture. Following this logic, many scholars have studied pesticide reduction decision-making behavior from the farmer’s perspective. Wu et al. [31] find that farm size is closely related to farmers’ pesticide use. Their study shows that a 1% increase in farm size induces a 0.5% decrease in pesticide use per hectare. Cheng et al. [21] examine the impact of networks embedded on farmers adopting green farming technologies and find that the role of the family farming area is more substantial than that of education level or annual family income in green technology adoption. Guo et al. [32] study farmers’ pesticide reduction behavior from a mutual perspective of social learning and social network and find that farmers’ production experience and technical knowledge based on their social networks can motivate pesticide reduction behavior. Wang et al. [13] argue that the effect of adopting green agricultural technologies is heterogeneous in terms of risk perceptions and public image, and thus the government should implement various guidance measures accordingly. Zhao et al. [14] study the impact of differences in socio-economic status among farmers on pesticide application behavior based on their sense of moral responsibility and conclude that some social norms such as rules of conduct, regulations, customs, and value standards can encourage farmers to reduce pesticide use. Xiang et al. [27] find that the accumulation of farmers’ social capital can promote their willingness to adopt fertilizers and technologies for pesticide reduction. Among different dimensions of social capital, social trust plays a major role. In addition, given the obvious lack of incentives for pesticide reduction in China [33], Yang et al. [34] compare the effects of two types of incentives, “green technology training” and “green agricultural subsidies”, on farmers’ biopesticide application behavior. Li et al. [35] argue that promoting green production behavior depend more on value orientation, disciplinary supervision, and internalization of transmission, as well as guiding and incentive regulations. They argue that restraint regulation is prone to the “relative institutional failure” phenomenon, suggesting the necessity to build an interactive regulatory system that integrates formal and informal institutions. Yang et al. [30] argue that agricultural socialization services can effectively boost farmers’ participation in reducing agricultural pesticide use, and Qin et al. [36] argue that market agents such as cooperatives, contractors, and purchasers can also effectively curb excessive pesticide use among farmers. Compared with traditional high-toxicity and high-residue chemical pesticide application technology, pesticide reduction technology is a technology to reduce and control pests with higher efficiency. It uses physical, biological, ecological, and other pest prevention and control methods to replace traditional chemical pesticides. It aims to reduce the amount of pesticide application and improve the efficiency of pest control [37]. It is characterized by low toxicity, low residue, high efficiency, and low dosage requirements. Common physical prevention and control technology mainly includes light trapping technology and insect-proof net technology, which takes advantage of the characteristics of insects (e.g., phototaxis) to trap and kill crop pests. Biological prevention and control technology mainly includes biopesticide application technology and natural enemy preying, which controls insects and bacteria by using insects and bacteria themselves. Ecological prevention and control technology mainly includes improving water and fertilizer management and promoting farmland ecological engineering, intercropping and other biodiversity control methods of healthy crop growth technology, and artificially enhancing crop resistance to pests and diseases (as shown in Figure 1). Pesticide reduction technologies not only help reduce the health risks to farmers but also help improve the ecological environment and ensure the safety of agricultural products, and promote high-quality agricultural development [38]. The above studies on farmers’ pesticide reduction behavior provide a rich theoretical foundation and empirical basis for this paper. In this study, the four most widely used pesticide reduction technologies, including light trapping, biopesticide application, healthy crop growth, and insect-proof net technologies, are selected, and a theoretical framework of farmers’ willingness to adopt pesticide reduction technologies is constructed based on the theories of value perception and planned behavior. This study also takes into account the current status of the adoption of pesticide reduction technologies by farmers in China. This paper uses survey data from 516 farmers in Henan Province to compare and analyze the commonalities and differences in farmers’ willingness and behavior in response to the four existing pesticide reduction technologies. It also reveals the key factors behind farmers’ perceptions of pesticide reduction technology, their willingness to adopt the technologies, and farmer behavioral decisions in China. This paper contributes to a deeper understanding of farmers’ willingness to respond to pesticide reduction technologies in a psychological sense. It also provides empirical implications for the government to promote pesticide reduction and also to maintain efficient production in agriculture. 2. Theoretical Analysis Based on the Theories of Value Perception and Planned Behavior Research on perceived value dates back to Porter, who argues that perceived value is the difference between a decision-maker’s perceived benefits and perceived costs [39]. With this initial claim, scholars define perceived value from different perspectives: the overall utility based on comparing gains and losses [40], the ratio of perceived benefits to perceived costs [41], and the trade-off between benefits and costs in the whole process of multiple transaction behaviors [42]. Nowadays, the view of ‘balance between perceived benefits and efforts’ is widely accepted [43]. Since then, some researchers have started to study farmers’ production behavior based on a value perception perspective. To explore the differences and influencing factors of farmers’ behaviors and willingness, they analyze the costs and benefits behind farmers’ behavioral decisions mainly based on perceived benefits and perceived risks [44,45], or construct farmers’ value systems from different value dimensions such as economic dimension, ecological dimension, emotional dimension [46,47], monetary dimension, social dimension, conditional dimension, and perceived dimension [48]. Numerous studies have shown that farmers’ value perceptions significantly affect their willingness to adopt technologies [49,50], and generally, the stronger the value perception, the stronger the willingness [51]. Among them, expected benefits and integrated value perceptions have positive effects [52], while cost and benefit risk perceptions have a negative effect [53]. The theory of planned behavior, proposed by Ajzen [54], suggests that behavioral attitudes, subjective norms, and perceived behavioral control can jointly affect an individual’s intentions, which then affect their behavioral outcomes. Behavioral attitudes arise from the people’s expected outcome. Subjective norms are pressures and constraints from other individuals or organizational groups that make people perform or not perform a behavior. Subjective norms include ‘legal norms’ and ‘descriptive norms’ [55]. Perceived behavioral control is an individual’s prior self-perception of the ease or difficulty of performing a behavior based on past experience and future expectations. It is influenced by the individual’s perception of their own resources, skills, opportunities, and other factors [56]. In general, the more positive the behavioral attitude of the subject is, the greater the subjective normative constraints are, and the stronger the perception of behavioral control and the intention to perform a certain behavior is. The theory of planned behavior has been widely used in research on farmers’ production behavior decisions because of its explanatory power in human’s general decision-making behavior [54]. Based on the theory of planned behavior, Xie et al. [57] explore the intrinsic attribution of the heterogeneity in farmers’ willingness to adopt ecological farming, and Hu et al. [58] studied farmers’ heterogeneous willingness to adopt rice and shrimp co-cropping models. Some researchers further expand the research framework of this theory, however, because they argue that the theory of planned behavior does not explain the deviation between the empirical results and the actual behavior of farmers [59]. In addition to behavioral attitudes, subjective norms, and perceived behavioral control factors, Shi et al. [60] introduce economic rationale and environmental values to study the factors influencing farmers’ willingness to adopt green production. Shi and Yu [61] introduce risk expectations and perceptions of citizenship to study the mechanisms of farmers’ homestead withdrawal and analyze the moderating factors behind farmers’ behavioral decisions. Zhang et al. [49] analyze the factors influencing farmers’ adoption of straw return technology and its relation with external variables such as perceived value and awareness of environmental responsibility based on an extended Technology Acceptance Model (TAM). In integrating the theories of value perception and planned behavior, this paper constructs a theoretical, analytical framework for farmers’ willingness to adopt pesticide reduction technologies based on three key factors: behavioral attitudes, subjective norms, and perceived behavioral control, as described below. Farmers’ behavioral attitudes, subjective norms, and perceived behavioral control jointly influence their willingness to adopt pesticide reduction technologies. Behavioral attitudes are farmers’ expectations and evaluations of the outputs of adopting pesticide reduction technology, which is formed based on farmers’ value perception of technology adoption, risk perception, and perception of technical operational ability, and may also be related to farmers’ personal characteristics [28]. Three aspects of farmers’ value perception of technology adoption are reflected in their personal characteristics: economic values, ecological values, and social values. Subjective norms are a collection of various external constraints farmers face when deciding to adopt pesticide reduction technology. The effect of subjective norms can be understood as the interactive influence of people around and the policy environment in which farmers live. Policy environment factors include government subsidies, publicly available government technology information, government-related technical training guidance, and the role government plays in improving the external environment for adopting technology. Perceived behavioral control is a farmer’s psychological perception of the difficulties in the practical application of pesticide reduction technologies. This psychological perception is also related to the farmer’s personal characteristics, production conditions, relevant experience, and perception of technical operational ability. The interactive relations among all these factors are shown in Figure 2. 3. Data Sources and Model Construction for Farmers’ Responses to Pesticide Reduction Technologies 3.1. Data Sources The data for this study are mainly collected from a field survey of farmers conducted by the research team in Kaifeng, Henan Province (as shown in Figure 3) from July to September 2020, with questionnaires (as shown in Supplementary Materials) and interviews by the trained researchers. Six villages in the administrative area of Kaifeng were randomly selected for field research. The survey focuses on farmers’ willingness to adopt pesticide reduction and pest control technologies in their agricultural production process. The contents of the survey mainly include farmers’ basic characteristics, production conditions, perceptions and psychology around pesticide reduction technology, relevant experience of technology adoption, the current status of technology response, and policy perceptions. A total of 516 valid observations were obtained from this survey. 3.2. Basic Characteristics of Samples Table 1 shows that most of the farmers in the sample are middle-aged and elderly people with a relatively low level of education. The proportions of male and female farmers interviewed are similar, accounting for 54.8% and 45.2% of the total sample, respectively. The farmers’ age varies from 20 to 78 years old, and those who are above 45 years old account for 73.4% of the total sample. The education level of farmers is generally low; 80.6% of them only have a primary school (43.6%) and junior/middle school education. Moreover, most farmers have been engaged in agricultural production for a long time, and 76.7% of farmers have more than 25 years of experience in agricultural production. The overall family planting area of the sample farmers is of small or medium size. A total of 62.6% of farmers have a planting area that is between 6 and 15 mu (1 mu = 0.067 hectares). The planting area for 17.6% of the farmers is less than 5 mu. Farmers whose planting area is more than 35 mu account for only 3.9%. In total, agricultural income is the major source of family income for 76.7% of the sample farmers. A total of 36.2% of farmers claim that pesticide expenditure is a substantial component of total family agricultural expenditure. It is not common for the sample farmers to have either part-time employment or membership in cooperatives. More specifically, 36.2% of farmers have part-time employment, and 42.1% of farmers are members of professional cooperatives. A total of 97.1% of the sample farmers have ever encountered technical problems in the process of agricultural production. The above descriptive statistics are in line with the current situation of agricultural production in underdeveloped regions of China. 3.3. Model Construction The question on willingness to adopt technology in the questionnaire has three options: “unwilling”, “doesn’t matter”, and “willing”. This feature makes the answer an ordered multi-classification variable. Therefore, this study chooses an ordered logistic regression model for quantitative analysis. The model formulations are as follows:(1) In(p(y≤j|x)1−p(y≤j|x))=μj−(α+∑i=1kβixi) (2) p(y≤j|x)=eμj−(α+∑i=1kβixi)1+eμj−(α+∑i=1kβixi)  where y denotes the dependent variable, representing farmers’ willingness to adopt pesticide reduction technologies; xi denotes the explanatory variable, representing the ith factor influencing farmers’ adoption of pesticide reduction technologies; μ denotes the threshold or critical value; α denotes the intercept; and βi denotes the corresponding parameter to be estimated for xi, representing the degree and direction of the influence of each explanatory variable on the dependent variable. 3.4. Variable Selection 3.4.1. Dependent Variables The pesticide reduction technologies involved in this study mainly focus on four types of pesticide reduction: light trapping technology, biopesticide application technology, healthy crop growth technology, and insect-proof net technology. Four ordered logistic regression models are constructed to study the factors affecting farmers’ willingness to adopt pesticide reduction technologies. The dependent variable is farmers’ willingness to adopt each technology. There are three options for measuring farmers’ willingness: “unwilling”, “doesn’t matter”, and “willing”, which correspond to values 1, 2, and 3, respectively. 3.4.2. Explanatory Variables “Farmers” are the interested subjects that are expected to respond to the promotion of pesticide reduction technologies in this study. Indicator designs related to farmers have been widely used and proven effective. Li et al. [35], Guo et al. [62], and Gao et al. [63] describe the individual characteristics of farmers in terms of gender, age, education level, years of agricultural production, part-time employment and farmer status; Tian et al. [6], Li et al. [22], and Li et al. [64] describe the family characteristics of farmers in terms of family planting size, income structure, and cost expenditure. Huang et al. [25], Yan et al. [38], and Zhang et al. [65] describe farmers’ profit perceptions in terms of income improvement, ecological environment improvement, and agricultural product safety; Huang et al. [11], Wang et al. [13], and Su et al. [66] describe farmers’ risk perceptions in terms of operational risk, market risk, and perception of operational ability. Cheng et al. [21], Gai et al. [24], Xiang et al. [27], He et al. [67], and He et al. [68] describe the policy environment in terms of government subsidies, technical information publicity, technology training, interpersonal trust, and government credibility. Based on previous literature, this study selects 21 indicators on farmer characteristics, production conditions, perceived value and risk of technology adoption, ability perception, peer influence, and policy environment, as well as relevant experience to construct a model including the factors influencing farmers’ response to pesticide reduction technologies. These indicators can be classified into eight categories. A conceptual description of these indicators is found in Table 2. First, the farmer characteristics category mainly includes two factors: personal features and production conditions. Among them, four indicators are selected for personal features: the gender of the household head, age, education level, and years of production experience. Three indicators are selected for factors of production conditions: the size of the family planting land, the income structure, and the proportion of pesticide expenditure in total family agricultural expenditure. Age, gender, and education level of the household head are expected to affect farmers’ technology response behavior, and the expansion of family planting land is expected to increase the risk of technology adoption by farmers. Farmers with larger agricultural income tend to be more cautious in technology adoption, and a higher proportion of pesticide expenditure in total family agricultural expenditure predicts a higher likelihood of farmers adopting pesticide reduction technologies. Second, farmers’ psychological perception consists of three categories of technology adoption: value perception, risk perception, and ability perception. Five indicators related to technology adoption are selected: farmers’ perception of the degree of income improvement, perception of improved ecological environment, perception of product safety, farmers’ attitudes towards risk, and perception of technical operational ability. Pesticide reduction technologies have substantial value on ecological, economic, and social, and farmers’ motivation to pursue profit and food and ecological safety will promote their technology response behavior. Meanwhile, pesticide reduction technologies are also characterized by uncertain effects, complex technical operations, and high market operation risks. This cognitive conflict brings a significant impact on farmers’ technology response behavior: stronger risk perception generally leads to lower technology adoption [13]. Third, subjective norms include peer influence and relevant policy environment factors. The policy environment is measured by five indicators: farmer satisfaction with government subsidies, satisfaction with technical information publicized by the government, satisfaction with government technical training guidance, trust in government promotion of technology, and perception of the government’s role in improving the external environment for technology adoption. Fourth, the category of relevant experience includes three aspects that can affect farmers’ technology response behavior: cooperative organizations, part-time employment, and relevant experience of farmers. Therefore, three indicators, participation in professional cooperatives, part-time employment, and the frequency of technical problems encountered in industrial operations, are selected as other variables. 4. Analysis of Farmers’ Response Behavior to Pesticide Reduction Technologies 4.1. Farmers’ perceptions of Pesticide Reduction Technologies Based on the statistical analysis of the sample farmers’ psychological perceptions of pesticide reduction technologies (Table 3), the farmers generally agree that the four pesticide reduction technologies have a higher safety level in terms of ecology and product quality. Among them, farmers strongly emphasize the ecological safety of the healthy crop growth technology, with 84.9% of farmers believing it is relatively safe, followed by light trapping technology. However, 11.2% of farmers are skeptical about the biopesticide application technology, considering it is relatively unsafe ecologically. However, based on the benign nature of new biopesticides, such should be the most desirable green option at present [69], which indicates that the relevant government departments need to further strengthen technical publicity and continuously improve farmers’ perceptions of this technology. The vast majority of farmers believe that adopting pesticide reduction technologies have helped to increase agricultural income. The scale of the impact, however, is considered to be relatively negligible. The vast majority of farmers considered that technology adoption has a “relatively large” or “very large” impact on income growth. A total of 26.5% of the sample farmers think the impact of healthy crop growth technology is substantial, followed by the biopesticide application technology (10.9%). However, only a few of them believed that adopting the biopesticide application technology would have a “very large” impact on agricultural income growth. Due to the wide variety of biopesticides’ poor stability and complex application process [70], it is necessary to scientifically select proper biopesticide products and suitable equipment [71]. At the same time, it is also essential to accurately grasp the time, dose, and time interval of application; otherwise, it will be less effective and even induce additional cost [72]. Therefore, farmers generally consider that biopesticide application technologies are difficult to operate [73] and require training on the application technologies of different biopesticide varieties. In the sample of this paper, similar observations were obtained. The biopesticide application technology is said to be the most difficult one, with 40.3% and 26.6% of farmers saying it is “relatively difficult” and “very difficult”, respectively. The healthy crop growth technology is considered to be the easiest one to implement. The survey finds that farmers are risk-sensitive, and the majority of farmers believe that there are risks in adopting pesticide reduction technologies. Among them, most farmers believe that the risks of the biopesticide application technology are “relatively large” or “very large”, accounting for 51.4% and 6.6% of the sample, respectively. By contrast, the adoption of insect-proof net technology is considered to be the least risky among the four pesticide reduction technologies. 4.2. Farmers’ Willingness and Behavior in Response to Pesticide Reduction Technologies In order to clarify farmers’ attitudes to pesticide reduction technologies, this study further surveys and interviews the sample farmers with additional questions focusing on four aspects: “whether they have heard”, “whether they are concerned”, “ whether they need”, and “whether they are willing” (Table 4). The results show that: (1) 65.5% of farmers have heard of these four pesticide reduction technologies. The most well-known ones are insect-proof net technology and biopesticide application technology, which reflects the efforts to promote them by relevant government departments. (2) Farmers show the greatest interest in the healthy crop growth technology (e.g., soil test and formula fertilization, crop rotation and intercropping, deep loosening and tilling of the soil, etc.). The proportion of concern and need reached 80.6% and 74.2%, respectively, followed by insect-proof net technology. (3) Light trapping technology is the least attractive option for farmers, and the percentage of those unwilling to adopt the technology is also the highest (79.8%). One possible reason for this is that they are skeptical about the insecticidal effect. Moreover, the insecticidal equipment needs to be set up in the field, which can induce many accompanied problems in practice. (4) Among the respondents, the number of farmers who are willing to adopt the insect-proof net technology is the largest, accounting for 73.6% of the total sample. By contrast, the number of farmers who are unwilling to adopt it is also relatively substantial (19.8%). Although the insect-proof net technology is relatively easy to apply, farmers also express greater difficulties in selecting suitable specifications for insect-proof nets, choosing the follow-up of supporting measures for covering and cultivation of insect-proof nets, and selecting suitable varieties of agricultural products. (5) In total, 58.3% of the interviewed farmers claim they are reluctant to adopt the biopesticide application technology, which is the second-highest proportion among the four technologies after light trapping. The reasons for this might be the high cost of biopesticide application, its poor quick-acting properties, narrow insecticidal and bactericidal spectrum [12], and relatively complex application procedures. These factors make farmers think that the technology is less cost-effective and reduce their willingness to adopt it. 5. Factors Influencing Farmers’ Responses to Pesticide Reduction Technologies 5.1. Correlation Analysis of Independent Variables and Willingness to Utilize Pesticide Reduction Technologies SPSS20.0 is used to analyze correlations in the survey data, and Kendall’s Tau-b coefficients of the independent variables and farmers’ willingness to respond to light trapping technology, biopesticide application technology, healthy crop growth technology, and insect-proof net technology are obtained separately. For brevity, only the results of the analysis related to the willingness to respond to the light trapping technology are presented in Table 5. The results show that farmers’ age, participation in professional cooperatives, years of working in agricultural production, perception of income improvement, peer influence, satisfaction with government-related subsidies, satisfaction with technical information publicized by the government, trust in government promotion of technology, satisfaction with government technical training, and perception of the government’s role in improving the external environment for technology adoption are all significantly and positively correlated with farmers’ willingness to implement pesticide reduction and pest control technologies. In contrast, gender, part-time employment, perception of operational ability, and risk perception are negatively correlated with farmer willingness. The results may indicate that farmers can be prompted to respond to pesticide reduction technologies by improving the government-related subsidy system, enhancing government-related technical information publicity, and training guidance. These measures can lower the risk of technology adoption and optimize the external environment for technology adoption. 5.2. Analysis of Influencing Factors of Farmers’ Responses to Pesticide Reduction Technologies The ordered logistic regression models of farmers’ responses to the four technologies are estimated using the stepwise regression analysis method in Stata 12.0. The results are shown in Table 6, Table 7, Table 8 and Table 9. Combined with the goodness-of-fit test index of each model, the model chi-square statistic is significant at the 1% level, which indicates that the models are powerful in predicting the dependent variable and the effects of the explanatory variables are strong. First, the influence of farmer characteristics on their willingness to respond to pesticide reduction technologies is discussed in terms of the following aspects. Among the results of the four ordered logistic regression models, only age and farmers’ years of working in agricultural production significantly affect farmers’ willingness to respond to the insect-proof net technology. The effect is positively significant at a 1% confidence level. This result indicates that older farmers who have been in agricultural production longer are more willing to respond to the insect-proof net technology. The government has been promoting the insect-proof net technology for several years, and the technology is relatively easy to apply, so older farmers who are more experienced in agricultural production have better knowledge and richer experience with it. Therefore, they can better solve the problems of suitable insect-proof nets selection, soil disinfection, wind, and flood prevention, etc., which can more effectively improve the effectiveness of insect-proof nets. Second, the impact of farmers’ psychological perception on their willingness to respond to pesticide reduction technologies is as follows. Farmers’ perception of the degree of income improvement from technology adoption is positively significant at a 1% confidence level in all four regression models, indicating that the more the technology response contributes to agricultural income improvement, the more willing farmers are to adopt the technology. Farmers’ perception of technical operational ability significantly influenced their willingness to respond to healthy crop growth and insect-proof net technologies, and the effect is negatively significant for both technologies at the 1% confident level. This indicates that the harder the technology is perceived to operate, the less willing farmers are to adopt the technology. Farmers’ perception of technology adoption risk significantly affects their willingness to respond to light trapping and insect-proof net technologies. The effect is significant at a 1% level for both technologies. The biopesticide application passes the significance test at the 5% level, both in a negative direction, indicating that the greater the perceived risk of technology adoption is, the less willing farmers are to respond to the technologies. Agricultural production has many uncertainties. Farmers adopting new agricultural technologies are exposed to not only natural and social risks but also market and technological risks. Therefore, reducing risks and maximizing returns is the fundamental motivation for farmers’ behavioral decisions. In promoting pesticide reduction and pest control technologies, on the one hand, it is necessary to vigorously advertise the efficient, environment-friendly, safe, and harmless features of these technologies to enhance farmers’ confidence in technology adoption through word-of-mouth publicity through demonstration households and also enhance their perception of technical operational ability through field demonstrations. On the other hand, the adoption of some technologies will improve product quality by sacrificing yield to some extent. This trade-off requires the government to optimize planting structures, which refer to the combination and optimization of different varieties of crops guided by the local government through better regulations. It also requires the government to actively seek markets for products to ensure farmers’ income. In this way, farmers will be more willing to respond to pesticide reduction technologies. Third, the influence of subjective norms on farmers’ willingness to respond to pesticide reduction technologies also deserves researchers’ attention. Among the environmental policy factors, farmers’ satisfaction with government subsidies is positively significant at a 1% confidence level in all four regression models. Farmers’ satisfaction with government technical training guidance significantly affects their willingness to respond to light trapping, biopesticide application, and healthy crop growth technologies, with a positive significance level of 1%. The estimated effect of farmers’ trust in government promotion of technology is positively significant at the 1% level in the response model for biopesticide application and insect-proof net technologies and is positively significant at the 5% level in the response model for light trapping and healthy crop growth technologies. Farmers’ perception of the government’s role in improving the environment for technology adoption is positively significant at the 1% level in the biopesticide application technology response model. These results show that a higher level of farmers’ satisfaction with the government-related subsidy system and the effectiveness of government technical training, a higher level of farmers’ trust towards government-promoted technologies, and a more crucial role of the government in improving the external environment for technology adoption, can make farmers more willing to respond to new technologies. As “rational economic individuals”, farmers make their decisions by comparing costs and benefits. In general, the lower the cost, the higher the expected benefits. After technology adoption, they can attain more compensation from the government to hedge the risk of technology adoption. The government can improve the environment for technology adoption with more detailed and clearer technical training guidance and improve the corresponding infrastructure. Meanwhile, farmers will be more convinced that the technologies can improve productivity if the government is more credible. Consequently, they will have stronger motivation to adopt pesticide reduction technologies. Fourth, the impact of relevant experience on farmers’ willingness to respond to pesticide reduction technologies is as follows. The frequency of the technical problems encountered is negatively significant at the 5% level in the model of willingness to respond to light trapping technology, indicating that frequent technical problems in practice hinder farmers’ willingness to respond to light trapping technology. Although the light trapping technology has been used for decades, the function of insecticidal lamps has not seen substantial improvements for many years [74]. Moreover, many technical problems arise in farmers’ long-term practice [75], which then affects farmers’ willingness to respond to the technology. 6. Conclusions 6.1. Main Conclusions Based on the survey of 516 farmers in Henan Province, this paper constructs four ordered logistic regression models for light trapping, biopesticide application, healthy crop growth, and insect-proof net technologies to explore farmers’ willingness and determinants of their response to pesticide reduction technologies. Three main conclusions are drawn below. First, the interviewed farmers have a high perception of the ecological safety and product safety of all four pesticide reduction technologies, and most farmers are familiar with the four technologies to some degree. The four technologies can be ranked from high to low in terms of farmers’ concern and need: healthy crop growth technology, insect-proof net technology, biopesticide application technology, and light trapping technology. Farmers are overall passive in adopting new technologies, and most farmers believe the costs are high while the benefits are uncertain. Large-scale demonstrations which can help farmers learn more about the technologies are necessary to increase farmers’ willingness to adopt them. Second, most farmers believe that adopting pesticide reduction technologies has slightly increased agricultural income, but the impact is relatively small. Among them, 16.9% and 6.6% of farmers believe adopting healthy crop growth technology has a “relatively large” and “very large” impact on income growth. Very few farmers believe adopting biopesticide application technology will have a “very large” impact on agricultural income growth. Moreover, farmers generally believe that biopesticide application technology is the riskiest and hardest technology to apply. The least-appealing technology farmers are unwilling to adopt is the light trapping technology, followed by the biopesticide application technology. Third, the effects of the farmers’ characteristics, their technology perceptions and psychology, the relevant policy environment, and the relevant experience on farmers’ response to different technologies are heterogeneous. Overall, farmers’ willingness to adopt technologies is mainly affected by their perception of many factors, including income improvement, their technical operational ability, their estimated risks, government-related subsidies, government technical training guidance, trust in government promotion of the technology, and the government’s role in improving the external environment for technology adoption. Most of these factors have positive effects on farmers’ willingness to adopt technologies, while higher risks can inhibit their willingness. 6.2. Future Implications and Recommendations Typically, farmers make decisions in response to technologies by considering whether the technology can reduce productive inputs and labor intensity, save working time, and increase productivity and income. However, with the increasing demand for sustainable environmental practice and product quality, farmers’ technological needs are no longer limited to simply increasing productivity and income. They also begin to emphasize the quality of the products and the sustainability of the ecological environment. Farmers have an intrinsic motivation to adopt new technologies only when the perceived value of their adoption exceeds the risk of loss. Therefore, encouraging farmers to positively respond to pesticide reduction technologies can start from two key points: first, strengthen technology publicity and demonstrations to guide farmers to form a more positive value perception of technology adoption, thus stimulating their intrinsic motivation to adopt pesticide reduction technologies; second, use technical subsidies, and incentive mechanisms to encourage farmers to more positively respond to pesticide reduction technologies. First, expand the breadth and depth of technical publicity and training programs to raise farmers’ awareness of reducing pesticide use. Farmers’ perceptions of pesticide reduction technologies, especially whether technology is promising in improving productivity and quality and whether it contributes to high product prices and best-selling products are the key concerns for farmers’ behavioral decisions. Information publicity and training guidance on pesticide reduction technologies are important ways of improving farmers’ perception of pesticide reduction technologies and promoting their technology response behavior [76]. On the one hand, it is necessary to advertise pesticide reduction technologies through multiple channels such as radio, television, the Internet, MMS/texting, technical training courses, etc. These measures can facilitate the top-down publicity and promote the role of government agricultural technology extension departments. They can also highlight the benefits of adopting pesticide reduction technologies, making farmers fundamentally aware of the importance of pesticide reduction technologies to ecological protection and their own health. The prevalence of such information can enhance their confidence in technology adoption [77]. At the same time, the ecological education of farmers should be continuously strengthened to guide their choice of pesticide reduction behavior from the dimension of ecological protection and personal health. On the other hand, through government regulation, efforts should be made to optimize planting structures and actively seek broader markets for products. The incentive mechanisms such as government subsidies should also be improved to ensure farmers’ income, strengthen the direct perception of the results and effectiveness of pesticide reduction technology responses, enhance willingness to respond to technology, and finally promote technology adoption behavior. Second, promote the innovation of pesticide reduction technologies to balance technology supply and demand and achieve cost savings and income enhancement. A scientific and technological innovation mechanism targeting the needs of farmers should be established and continuously improved. It should highlight the transformation from researchers’ technical innovation to economic outputs and also take into account the public and social nature of pesticide reduction technologies, and facilitate government investment in technological innovation research and development. In this way, the mechanism should be able to contribute to the diffusion and adoption of pesticide reduction technologies through scientific and technological innovations [78]. The technical needs of farmers should be carefully clarified, and the help and guidance provided to the farmer trainees should precisely match their demands and mitigate their difficulties. The channels of education and training should also be further broadened, and the role of private training institutions, professional associations, and leading enterprises should be used to promote high-quality training programs. At the same time, training contents and methods should be further improved to enhance farmers’ technical mastery and application abilities. After training, additional efforts must be made to improve land cultivation, finance, insurance, and other related supporting systems to promote the adoption of new pesticide reduction technologies. Third, effectively improve farmers’ satisfaction with technical training and promote farmers’ active participation in technical training. Due to the outflow of young and strong laborers from rural areas to urban areas, most of the individuals actually engaged in agricultural production are older and limited in relevant knowledge. Therefore, more on-site field guidance and instruction should be developed. By making better use of agricultural leisure time and appropriately extending the training time, the government can hire agricultural experts from inside and outside the province to carry out field guidance to improve the applicability and practicality of the technologies and help farmers to operate them proficiently [79]. Farmers who have received such training should also be publicized more vigorously to enhance their demonstration effect and to stimulate other farmers’ intrinsic motivation to participate in agricultural green education and pesticide reduction production skills training. Fourth, improve the government subsidy system, strengthen institutional trust, and build a favorable environment for technology adoption. Government subsidies play an important role in promoting the adoption of pesticide reduction technologies as an explicit incentive. Farmers’ trust in the government also significantly influences their willingness to adopt pesticide reduction technologies. Further improving the government subsidy system, strengthening government trust-building, and reducing transaction costs and institutional costs in technology promotion can effectively incentivize farmers to adopt pesticide reduction technologies. The importance of pesticide reduction technology subsidies should be highlighted as equivalent to that of other mainstream subsidies such as “direct subsidies for grain cultivation”, “subsidies for good seeds”, “subsidies for the purchase of agricultural machinery”, and “comprehensive subsidies for agricultural materials” so that subsidies can become a regular form of incentive for farmers to adopt pesticide reduction technologies [76]. Relevant policies, laws, and regulations should be strictly implemented to support and benefit farmers. This can help to ensure that compensation or subsidies for pesticide reduction technologies are in place and maintain the credibility of the government. As long as farmers are convinced that they will benefit from adopting the technologies and the compensation or subsidies can hedge the additional costs and risks stemming from using the technologies, they will be more willing to try these technologies [80]. The risk compensation system should be designed to suit those high-income, well-educated, and large-scale planting farmers and also coordinate with differentiated subsidy policies. The transparency of policy implementation must be ensured to establish a foundation of mutual trust and communication between the government and farmers. Such transparency can also mitigate farmers’ worries when participating in environmental protection and governance. The whole system should also focus on solving the problems and obstacles encountered by farmers in adopting pesticide reduction technologies. Fifth, it is vital to strengthen the training of new professional farmers and solidify the demonstration effect. Stakeholders, including family farms and large professional households of new professional farmers with moderate-scale operations, as well as intensive and specialized production organization structure, provide the favorable potential for applying pesticide reduction technologies. Strengthening the guidance and promotion of pesticide reduction technologies for new professional farmers can become a “field classroom” for other small and scattered farmers, allowing many small farmers to learn operational skills of new pesticide reduction technologies in the field. At the same time, risk-averse small farmers will become increasingly aware of the economic and social benefits of pesticide reduction technologies, which promotes their adoption of pesticide reduction technologies. In order to strengthen the training of new professional farmers and guide the intensive and specialized operation in agriculture, it is first necessary to clarify farmers’ contracting rights on land, establish a long-term mechanism for land transfer, secure farmers’ land tenure, and create an environment for moderate scale operations; second, a strict approval system should be established, and organizations that are qualified for approval should be allocated differentiated financial support and subsidies according to their operations so that these large demonstration households can truly benefit from the whole process and better exert their demonstration effect [14,81]. Acknowledgments The authors wish to thank the anonymous reviewers and editors for their helpful reviews and critical comments. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ijerph19095136/s1. Farmers’ Green Farming Technology Adoption Questionnaire. Click here for additional data file. Author Contributions X.S. and W.N. conceptualized the study and analyzed the results; X.S. ran model simulations and wrote the first draft; X.S. provided basic data; X.S. provided support in modelling. J.S., T.W. and Z.X. reviewed, edited, and improved the manuscript. X.S., J.S. and Q.S. wrote the revised version. All authors have read and agreed to the published version of the manuscript. Funding This research was financially supported by the Soft Science Project of Henan Province in China (Grant No. 192400410054), the Philosophy and Social Science planning project of Henan Province in China (Grant No. 2018BJJ052), the Foundation for Key Scientific Research of Henan Province in China (Grant No. 20A790026), the Foundation for University Key Teacher of Henan Province in China (Grant No. 2018GGJS014), the Philosophy and Social Science planning project of Henan Province in China (Grant No. 2021BJJ093), the Think Tank Project for Philosophy and Social Sciences of Henan Province in China (Grant No. 2021ZKYJ10), and the National Social Science Foundation of China (Grant No. 20BJL073). Institutional Review Board Statement Not applicable. Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The data presented in this study are mainly from a field survey of farmers conducted by the research team in Kaifeng, Henan Province from July to September 2020, which was conducted by questionnaires and interviews by the researchers after the training. Based on the principle of random sampling, six villages in the administrative area of Kaifeng were randomly selected for field research in this study. A total of 516 valid questionnaires were retained for this survey. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Related images to pesticide reduction technology. (Source: Photo by the author.). Note: (a) Light trapping technology (solar wind type insect trap for vegetable fields). (b) Healthy crop growth technology based on scientific water and fertilizer management (water and fertilizer sprinkler irrigation). (c) Insect-proof net technology (insect-proof net for vegetable greenhouses). (d) Drones apply biopesticides in tobacco field (Validamycin + Vivo-Bacillus cereus Mixture). Figure 2 Theoretical framework of farmers’ response behavior to adopting pesticide reduction technology. Figure 3 Topographic location of Kaifeng, Henan Province, China. ijerph-19-05136-t001_Table 1 Table 1 Basic characteristics of sample farmers. Features Classification Frequency Percentage % Gender Male = 1 283 54.8 Female = 0 233 45.2 Age ≤25 = 1 15 2.9 26~35 = 2 41 7.9 36~45 = 3 81 15.7 46~55 = 4 219 42.4 ≥56 = 5 160 31.0 Education level Primary school = 1 225 43.6 Junior/middle school = 2 191 37.0 Technical secondary school and high school = 3 84 16.3 College and above = 4 16 3.1 Part-time employment Working part-time = 1 187 36.2 Not working part-time = 0 329 63.8 Years of agricultural production ≤5 = 1 10 1.9 6~10 = 2 40 7.8 11~25 = 3 70 13.6 26~39 = 4 289 56.0 ≥40 = 5 107 20.7 Technology problems Encountered = 1 501 97.1 Not encountered = 0 15 2.9 Planting size (mu) ≤5 = 1 91 17.6 6~15 = 2 323 62.6 16~25 = 3 45 8.7 26~35 = 4 37 7.2 ≥36 = 5 20 3.9 Income structure Agricultural income dominated = 1; 396 76.7 Non-farm income dominated = 0 119 23.1 Professional cooperatives Joined = 1 217 42.1 Did not join = 0 299 57.9 Proportion of pesticide Expenditure in total Family agricultural expenditure Very small = 1 24 4.6 Relatively small = 2 105 20.3 Neutral = 3 185 36.0 Relatively large = 4 187 36.2 Very large = 5 15 2.9 ijerph-19-05136-t002_Table 2 Table 2 Model variable definitions and assignments. Categories Variable Names and Assignment Definitions Farmers’ characteristics X1 Gender: Male = 1; Female = 0 X2 Age: ≤20 = 1; 21~30 = 2; 31~40 = 3; 41~50 = 4; ≥50 = 5 X3 Education level: Primary schools = 1; Junior/middle School =2; Technical secondary school and high school = 3; College and above = 4 X4 Years of agricultural production: ≤5 = 1; 6~10 = 2; 11~20 = 3; 21~29 = 4; ≥30 = 5 X5 Planting size: ≤5 = 1; 6~15 = 2; 16~25 = 3; 25~35 = 4; ≥35 = 5 X6 Income structure: Agricultural income dominated = 1; Non-farm income dominated = 0 X7 Proportion of pesticide expenditure in total family agricultural expenditure: Very small = 1; Relatively small = 2; Neutral =3; Relatively large = 4; Very large = 5 Value perception X8 Perception of degree of income improvement from technology adoption: Very small = 1; Relatively small = 2; Neutral = 3; Relatively large = 4; Very large = 5 X9 Perception of technology adoption in improving ecological environment: Completely unecological = 1; Relatively unecological = 2; Neutral = 3; Relatively ecological = 4; Completely ecological = 5 X10 Perception of product safety through technology adoption: Very unsafe = 1; Relatively unsafe = 2; Neutral =3; Relatively safe = 4; Completely safe = 5 Risk perception X11 Risk attitude towards technology adoption: Relatively small = 1; Neutral = 2; Relatively large = 3 Ability perception X12 Perception of easiness in technical operation: Very easy = 1; Relatively easy= 2; Neutral = 3; Relatively difficult = 4; Very difficult = 5 Peer influence X13 Peer influence on your own technology adoption: Very small = 1; Relatively small = 2; Neutral = 3; Relatively large = 4; Very large = 5 Policy environment X14 Satisfaction with government subsidies: Completely dissatisfied = 1; Relatively dissatisfied = 2; Neutral = 3; Relatively satisfied = 4; Completely satisfied = 5 X15 Satisfaction with government technical information publicity: Completely dissatisfied = 1; Relatively dissatisfied = 2; Neutral =3; Relatively satisfied = 4; Completely satisfied = 5 X16 Satisfaction with government technical training guidance: Completely dissatisfied = 1; Relatively dissatisfied = 2; Neutral =3; Relatively satisfied = 4; Completely satisfied = 5 X17 Trust in government promotion of technology: Completely distrust= 1; Relatively distrust = 2; Neutral = 3; Relatively trust = 4; Completely trust = 5 X18 Government’s role in improving the external environment for technology adoption: Very unimportant = 1; Relatively unimportant = 2; Neutral = 3; Relatively important = 4; Very important = 5 Relevant experience X19 Part-time employment: Working part-time = 1; Not working part-time = 0 X20 Membership in professional cooperatives: Joined = 1; Did not join = 0 X21 Frequency of technical problems encountered in industrial operations: Encountered = 1; Not encountered = 0 Willingness to adopt Y1 Willingness to adopt light trapping technology: Unwilling = 1; Doesn’t matter = 2; Willing = 3 Y2 Willingness to adopt biopesticide application technology: Unwilling = 1; Doesn’t matter = 2; Willing = 3 Y3 Willingness to adopt healthy crop growth technology: Unwilling = 1; Doesn’t matter = 2; Willing = 3 Y4 Willingness to adopt insect-proof net technology: Unwillingness= 1; Doesn’t matter = 2; Willing = 3 ijerph-19-05136-t003_Table 3 Table 3 Farmers’ psychological perceptions of pesticide reduction technologies. Variables Classification Light Trapping Technology Biological Pesticide Application Technology Healthy Crop Growth Technology Insect-Proof Net Technology Frequency % Frequency % Frequency % Frequency % Perception of improved agricultural income Very small = 1; 57 11.0 29 5.6 48 9.3 43 8.3 Relatively small = 2; 167 32.4 232 45 221 42.8 196 38.0 Neutral = 3; 241 46.7 199 38.6 126 24.4 228 44.2 Relatively large = 4; 37 7.2 56 10.9 87 16.9 46 8.9 Very large = 5 14 2.7 0 0 34 6.6 3 0.6 Perception of improved ecological environment Completely unecological = 1; 0 0 0 0 0 0 0 0 Relatively unecological = 2; 18 3.5 58 11.2 2 0.4 15 2.9 Neutral = 3; 49 9.5 25 4.8 45 8.7 102 19.8 Relatively ecological = 4; 423 82.0 398 77.1 438 84.9 378 73.3 Completely ecological = 5 26 5.0 35 6.8 31 6.0 21 4.1 Perception of product safety Very unsafe = 1; 0 0 0 0 0 0 0 0 Relatively unsafe = 2; 0 0 6 1.1 0 0 0 0 Neutral = 3; 17 3.3 17 3.3 28 5.4 17 3.3 Relatively safe = 4; 478 92.6 472 91.5 465 90.1 479 92.8 Completely safe = 5 21 4.1 21 4.1 23 4.5 20 3.9 Perception of easiness in technical operation Very easy = 1; 0 0 0 0 0 0 0 0 Relatively easy = 2; 175 33.9 61 11.8 181 35.1 130 25.2 Neutral = 3; 155 30.0 110 21.3 193 37.4 154 29.8 Relatively difficult = 4; 130 25.2 208 40.3 79 15.3 172 33.3 Very difficult = 5 56 10.9 137 26.6 63 12.2 60 11.6 Perception of technology adoption risk Very small = 1; 33 6.4 11 2.1 23 4.5 17 3.3 Relatively small = 2; 67 13.0 71 13.7 34 6.6 19 3.7 Neutral = 3; 126 24.4 135 26.2 187 36.2 226 43.8 Relatively large = 4; 261 50.6 265 51.4 243 47.1 231 44.7 Very large = 5 29 5.6 34 6.6 29 5.6 23 4.5 ijerph-19-05136-t004_Table 4 Table 4 Statistics of farmers’ responsive attitudes to pesticide reduction technologies (number/proportion). Variable Name Light Trapping Technology Biopesticide Application Technology Healthy Crop Growth Technology Insect-Proof Net Technology Heard 338/65.5% 445/87.2% 447/86.6% 487/94.4% Concerned 167/32.4% 247/47.8% 416/80.6% 378/73.2% Needed 137/26.6% 118/22.65% 387//74.2% 224/43.4% Willing to adopt 58/11.2% 81/15.7% 274/53.1% 380/73.6% Doesn’t matter to adopt 46/8.9% 134/25.9% 104/20.1% 34/6.6% Unwilling to adopt 412/79.8% 301/58.3% 138/26.7% 102/19.8% ijerph-19-05136-t005_Table 5 Table 5 Correlation analysis of independent variables and farmers’ willingness to respond to light trapping technology. Independent Variable Name Maximum Value Minimal Value Mean Value Variance Kendall’s Tau-b Correlation Coefficient Significance (Bilateral p-Value) Gender 1.00 0.00 0.548 0.498 −0.113 *** 0.008 Age 5.00 1.00 3.907 1.020 0.124 *** 0.002 Education level 4.00 1.00 1.789 0.824 0.096 ** 0.018 Part-time employment 1.00 0.00 0.357 0.479 −0.063 0.114 Professional cooperatives 1.00 0.00 0.614 0.491 0.266 *** 0.000 Years of agricultural production 5.00 1.00 3.859 0.899 0.152 0.231 Planting size 5.00 1.00 2.171 0.933 0.044 ** 0.028 Income Structure 2.00 1.00 0.614 0.491 0.073 * 0.090 Proportion of pesticide expenditure 5.00 1.00 2.171 0.934 0.044 0.280 Technology problems 1.00 0.00 0.971 0.028 0.026 0.537 Perception of improved income 5.00 1.00 2.378 0.878 0.632 *** 0.000 Perception of improved environment 5.00 1.00 3.767 0.763 0.012 0.767 Perception of product safety 5.00 1.00 4.014 0.282 0.008 0.841 Peer influence 5.00 1.00 4.174 0.743 0.121 *** 0.004 Operational ability 5.00 1.00 3.364 0.943 −0.047 *** 0.000 Risk perception 3.00 1.00 3.664 2.473 −0.362 *** 0.000 Government subsidies 4.00 1.00 2.324 0.932 0.750 *** 0.000 Information publicity 5.00 1.00 2.244 0.549 0.165 *** 0.000 Technical guidance 5.00 1.00 2.804 0.946 0.750 *** 0.000 Government trust 5.00 1.00 2.248 0.727 0.151 *** 0.000 Government role 5.00 1.00 4.021 0.821 0.109 *** 0.000 Note: ***, **, and * denote significance at the 1%, 5%, and 10% level respectively. ijerph-19-05136-t006_Table 6 Table 6 Estimation result of farmers’ willingness to adopt light trapping technology. Variable Name Coefficient Standard Error Sig Perception of improved income 2.086 *** 0.294 0.000 Perception of technology adoption risk −0.687 *** 0.185 0.000 Government subsidies 2.149 *** 0.324 0.000 Technical guidance 1.642 *** 0.288 0.000 Trust in government promotion of technology 0.696 ** 0.303 0.022 Frequency of technical problems encountered −2.519 ** 1.069 0.018 Critical value 1 17.135 3.897 Critical value 2 8.977 3.761 Note: Log likelihood = 249.491, LR chi2 = 173.567, Prob > chi2 = 0.000, Pseudo R2 = 0.729. ***, ** denote significance at the 1%, 5% level respectively. ijerph-19-05136-t007_Table 7 Table 7 Estimation result of farmers’ willingness to adopt biopesticide application technology. Variable Name Coefficient Standard Error Sig Perception of improved income 2.197 *** 0.838 0.000 Perception of technology adoption risk −1.644 ** 0.719 0.022 Government subsidies 4.019 *** 0.957 0.000 Technical guidance 5.716 *** 1.189 0.000 Trust in government promotion of technology 1.997 *** 0.740 0.007 Perception of government in improving the environment for technology adoption 4.533 *** 1.111 0.000 Critical value 1 46.394 13.882 Critical value 2 27.936 12.415 Note: Log likelihood = 508.719, LR chi2 = 152.694, Prob > chi2 = 0.000, Pseudo R2 = 0.256. ***, ** denote significance at the 1%, 5% level respectively. ijerph-19-05136-t008_Table 8 Table 8 Estimation result of farmers’ willingness to adopt healthy crop growth technology. Variable Name Coefficient Standard Error Sig Perception of improved income 8.414 *** 1.466 0.000 Perception of technical operational ability −2.455 *** 0.686 0.000 Government subsidies 2.509 *** 0.758 0.001 Technical guidance 5.496 *** 1.127 0.000 Trust in government promotion oftechnology 1.850 ** 0.786 0.019 Critical value 1 7.773 1.394 Critical value 2 5.596 1.192 Note: Log likelihood = 624.094, LR chi2 = 414.969, Prob > chi2 = 0.000, Pseudo R2 = 0.553. ***, ** denote significance at the 1%, 5% level respectively. ijerph-19-05136-t009_Table 9 Table 9 Estimation result of farmers’ willingness to adopt insect-proof net technology. Variable Name Coefficient Standard Error Sig Age 1.001 *** 0.288 0.001 Years of agricultural production 0.622 *** 0.238 0.009 Perception of improved income 3.737 *** 0.406 0.000 Perception of technical operational ability −0.741 *** 0.167 0.000 Perception of technology adoption risk −2.294 *** 0.230 0.000 Government subsidies 1.752 *** 0.312 0.000 Trust in government promotion of technology 0.986 *** 0.223 0.000 Critical value 1 3.602 3.289 Critical value 2 9.115 3.331 Note: Log likelihood = 624.094, LR chi2 = 414.969, Prob > chi2 = 0.000, Pseudo R2 = 0.553. *** denote significance at the 1% level respectively. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Gao Y. Zhang X. Lu J. Wu L. Yin S. Adoption behavior of green control techniques by family farms in China: Evidence from 676 family farms in Huang-huai-hai Plain Crop Prot. 2017 99 76 84 10.1016/j.cropro.2017.05.012 2. Tessema R.A. Nagy K. Ádám B. Pesticide Use, Perceived Health Risks and Management in Ethiopia and in Hungary: A Comparative Analysis Int. J. Environ. Res. Public Health 2021 18 10431 10.3390/ijerph181910431 34639728 3. Han H. Cai S. Research on the health cost of pesticide application and its influencing factors—Based on the research data of farmers in the main grain producing areas J. Chin. Agric. Univ. 2011 16 163 170 4. Cai R. Wang Z. Qian L. Does joining cooperatives promote family farms to choose environmentally friendly production methods? —Taking fertilizer and pesticide application reduction as an example Chin. Rural Surv. 2019 1 51 65 5. Xue B. Zheng S. A study on farmers’ production technology selection behavior under the threshold of agricultural product quality and safety: An example of kiwifruit farmers in Shaanxi Province J. Northwest A F Univ. 2019 19 104 110 6. Tian Y. Zhang J. He K. Analysis of farm households’ agricultural low-carbon production behavior and its influencing factors —Taking fertilizer application and pesticide use as examples Chin. Rural Surv. 2015 4 61 70 7. Xia Q. Li D. Zhou H. Study on the impact of farming part-time on agricultural surface source pollution Chin. Popul., Resour. Environ. 2018 28 131 138 8. Li R. Chen Y. Research on farmland conservation and utilization behavior of farming households in typical black soil areas in Northeast China—An empirical analysis based on a survey of farming households in Suihua City, Heilongjiang Province J. Agrotech. Econ. 2017 11 80 91 9. Cai R. Cai S. An empirical analysis of conservation tillage technology adoption and its impact on crop yields—Based on survey data from rice farmers in Anhui Province Resour. Sci. 2012 34 1705 1711 10. Zhao P. Yan B. Liu T. Socio-economic status differences and the inverted U-shaped relationship between green control technology diffusion in agriculture: The mediating effect of social learning J. Arid Land Resour. Environ. 2021 35 18 25 11. Huang Y. Luo X. Tang L. The cost-saving and income-generating effects of green control technologies Chin. Popul. Resour. Environ. 2020 30 174 184 12. Guo L. Wang Y. Why do farmers say one thing but do another in the application of biopesticides? J. Huazhong Agric. Univ. 2018 169 71 80 13. Wang X. Zhang J. He K. He P. The influence of risk perception and public image appeal on farmers’ adoption of green farming technologies J. Chin. Agric. Univ. 2020 25 213 226 14. Zhao Q. Xia X. How social norms influence farmers’ pesticide application reduction—An analysis of the moderating effect based on the mediating effect of moral responsibility and socio-economic status differences J. Agrotech. Econ. 2020 10 61 73 15. Wang H. Xu X. Micro-behavior and agricultural product safety: An analysis of farmers’ production and consumers’ consumption J. Nanjing Agric. Univ. 2004 4 23 28 16. Galt R.E. Regulatory risk and farmers’ caution with pesticides in Costa Rica Trans. Inst. Br. Geogr. 2007 32 377 394 10.1111/j.1475-5661.2007.00256.x 17. Pimentel D. Acquay H. Biltonen M. Rice P. Silva M. Nelson J. Lipner V. Giordano S. Horowitz A. D’Amore M. Environmental and economic costs of pesticide use Bioscience 1992 42 750 760 10.2307/1311994 18. Eskenazi B. Bradman A. Castorina R. Exposures of children to organophosphate pesticides and their potential adverse health effects Environ. Health Perspect. 1999 107 409 419 10.1289/ehp.99107s3409 10346990 19. Ma W. Epidemiological analysis of pesticide poisoning in Chongming District, Shanghai, 2007–2019 Occup. Health 2022 38 484 487 20. Liu S. Shen X. Zhu S. Research on the green production behavior of farmers under the evolution of agricultural industrialized management organization system Rural Econ. 2020 11 37 44 21. Cheng L. Zhang J. He K. Analysis of the influence of network embedding and risk perception on farmers’ green farming technology adoption behavior—Based on survey data from 615 farmers in Hubei Province Resour. Environ. Yangtze Basin 2019 28 1736 1746 22. Li C. Zhou H. Organizational embedding and pesticide reduction among farmers: An analysis based on rice farmers in Jiangsu Province Res. Agric. Mod. 2021 42 694 702 23. Zhang L. Luo B. Agricultural reduction and its path options: Evidence from green energy companies Rural Econ. 2019 10 9 21 24. Gai H. Yan T. He K. Zhang J. Research on farmers’ conservation tillage technology adoption behavior from the perspective of social embeddedness—Based on data from 668 farmer surveys in three provinces of Ji, Wan and E Resour. Environ. Yangtze Basin 2019 28 2141 2153 25. Huang Y. Luo X. Li R. Zhang J. Farmers’ perceptions, external environment and willingness to produce green agriculture: Based on research data from 632 farmers in Hubei Province Resour. Environ. Yangtze Basin 2018 27 680 687 26. Li F. Zhang J. He K. Substitution and complementarity: Informal and formal institutions in farmers’ green production J. Huazhong Univ. Sci. Technol. 2019 33 51 60 27. Xiang C. Ji N. The influence of social capital on farmers’ willingness to adopt fertilizer and pesticide reduction technologies-mediated by learning ability and moderated by ecological cognition J. Chin. Agric. Univ. 2021 26 150 163 28. Li X. Zhang L. An empirical analysis of farmers’ willingness to adopt cleaner production technologies and the influencing factors Res. Agric. Mod. 2019 40 299 307 29. Wu B. A review of green consumption research Bus. Manag. J. 2014 36 178 189 30. Yang G. Zhang L. Yue M. Zhang J. Can socialized agricultural services promote reduced agricultural production-an empirical analysis based on micro-survey data of rice farmers in the Jianghan Plain World Agric. 2020 5 85 95 31. Wu Y. Xi X. Tang X. Luo D. Gu B. Lam S.K. Chen D. Policy distortions, farm size, and the overuse of agricultural chemicals in China Proc. Natl. Acad. Sci. USA 2018 115 7010 7015 10.1073/pnas.1806645115 29915067 32. Guo Q. Li S. Nan L. Social learning, social networks and pesticide reduction-empirical evidence from micro data of farmers J. Arid Land Resour. Environ. 2020 34 39 45 33. Zhang L. Smallholder differentiation, behavioral differences, and agricultural reduction Iss. Agric. Econ. 2020 6 131 142 34. Yang Y. He Y. Yan G. The impact of different incentives on farmers’ green production behaviors: An example of biopesticide application World Agric. 2021 4 53 64 35. Li F. Zhang J. He K. Influence of informal institutions and environmental regulations on farmers’ green production behavior: Based on 1105 farmers’ survey data in Hubei Resour. Sci. 2019 41 1227 1239 36. Qin S. Lv X. Can the participation of market players reduce pesticide overuse by rice farmers? J. Huazhong Agric. Univ. 2020 4 61–70, 176–177 37. Wei W. Practice and prospect of pesticide use reduction in China Agric. Technol. 2021 41 46 49 38. Yan A. Luo S. Huang Y. Influence of socialized services on farmers’ pesticide reduction behavior J. Arid Land Resour. Environ. 2021 35 91 97 39. Porter M.E. Competitive Advantage The Free Press New York, NY, USA 1985 40. Zeithaml V.A. Consumer Perceptions of Price, Quality and Value: A Means-End Model and Synthesis of Evidence J. Mark. 1988 52 2 22 10.1177/002224298805200302 41. Monroe G.C. Benbasat I. Development of An Instrument to Measure the Perceptions of Adopting an Information Technology Innovation Inf. Syst. Res. 1991 2 192 222 42. Gronroos C. Value-Driven Relational Marketing: From Products to Resources and Competencies J. Mark. Manag. 1997 13 407 419 10.1080/0267257X.1997.9964482 43. Petrick J.F. Development of A Multi-Dimensional Scale for Measuring the Perceived Value of a Service J. Leisure Res. 2002 34 119 134 10.1080/00222216.2002.11949965 44. Wu X. Zhang J. Feng J. Factors influencing farmers’ green farming technology perception and its hierarchical structure decomposition—Based on Probit-ISM model J. Huazhong Agric. Univ. 2017 5 36 45 45. Shen M. Gan C. Chen Y. Mei Y. Research on factors influencing farmers’ willingness to transfer farmland out based on DTPB theory—An example from Wuhan city circle Res. Agric. Mod. 2019 40 441 449 46. Ren L. Lin Z. Li N. A study on residents’ willingness to pay for ecological watershed governance Res. Financ. Educ. 2018 31 3 11 47. Wu J. Zan M. Wang Z. Influence of perceived value on farmers’ willingness to participate in arable land quality conservation: The case of Shaanxi Province Chin. Land Sci. 2020 34 66 74 48. Biswas A. Roy M. Leveraging factors for Sustained green consumption behavior based on consumption value perceptions: Testing the structural model J. Clean. Prod. 2015 95 332 340 10.1016/j.jclepro.2015.02.042 49. Zhang J. Yan T. Jiang X. Value perceptions, environmental responsibility awareness and straw resource utilization by farmers—A multi-group analysis based on an extended technology acceptance model Chin. J. Agric. Resour. Reg. Plan. 2021 42 99 107 50. Zhao H. Hu W. The influence of farmers’ perceptions on their willingness to participate in agricultural waste resource utilization under environmental regulations Chin. J. Eco-Agric. 2021 29 600 612 51. Cao H. Zhao K. Factors influencing farmers’ intention to reduce fertilizer application and decomposition of its effect—An empirical analysis based on VBN-TPB J. Huazhong Agric. Univ. 2018 6 29 38 52. Yang F. Zheng X. The influence of ecological compensation methods on farmers’ green production behavior from the perspective of value perception Chin. Popul., Resour. Environ. 2021 31 164 171 53. Mo J. Yu Z. Influence of value perception on farmers’ propensity to adopt technology and its conditional response—Empirical evidence based on structural equation modeling Chin. J. Agric. Resour. Reg. Plan. 2020 41 238 245 54. Ajzen I. The Theory of Planned Behavior Organ. Behav. Hum. Dec. Processes 1991 50 179 211 10.1016/0749-5978(91)90020-T 55. Hagger M. Chatzisarantis N. First -and higher-order models of attitudes, normative influence, and perceived behavioral control in the theory of planned behavior Br. Psychol. Soc. 2005 44 513 536 10.1348/014466604X16219 56. Greaves M. Zibarras L.D. Stride C. Using the theory of planned behavior to explore environmental behavioral intentions in the workplace J. Environ. Psychol. 2013 34 109 120 10.1016/j.jenvp.2013.02.003 57. Xie X. Chen M. Analysis of farmers’ willingness to adopt ecological farming and its heterogeneity—An empirical study based on TPB framework Resour. Environ. Yangtze Basin 2019 28 1185 1196 58. Hu N. Wang Y. Chen Q. Zhu L. Factors influencing farmers’ willingness to adopt rice and shrimp crop models and their heterogeneity Chin. J Eco-Agric. 2021 29 1752 1761 59. Ruan W. Yu H. Song X. A study on the behavioral intention of international tourists under the threat of haze based on the theory of extended planned behavior—An example of international tourists in Beijing J. Arid Land Resour. Environ. 2019 33 195 201 60. Shi Z. Cui M. Zhang H. A study on farmers’ willingness to produce green based on extended theory of planned behavior J. Arid Land Resour. Environ. 2020 34 40 48 61. Shi P. Yu J. The influence of risk expectation, citizenship perception and farmers’ cognition on the withdrawal of homestead bases of farmers relocated to alleviate poverty Resour. Sci. 2021 43 1387 1402 62. Guo Q. Li H. Li S. Influence of social norms on adoption behavior of fertilizer reduction measures by farm households J. Northwest A F Univ. 2019 19 112 120 63. Gao J. Shi Q. The influence of farmers’ productive characteristics on pesticide application: Mechanisms and evidence Chin. Rural Econ. 2019 11 83 99 64. Li F. He K. Zhang J. The influence of value perception on the willingness of large-scale pig farmers to participate in agricultural carbon trading and the expected carbon price: An example of a household biogas CCER project Res. Agric. Mod. 2021 42 889 899 65. Zhang T. Yan T. He K. Zhang J. Contrary of farmers’willingness of straw utilization to the behavior: Based on the MOA model J. Arid Land Resour. Environ. 2019 33 30 35 66. Su S. Zhou X. Zhou Y. Can organic fertilizer substitution increase farmers’ income?—A case study of vegetable growers in Shandong J. Arid Land Resour. Environ. 2022 36 24 31 67. He L. Wang Y. Subsidies on organic fertilizer, perception of the effect of organic fertilizer use and farmers’use of organic fertilizer—A survey based on some pilot and non-pilot counties of organic fertilizer subsidies in Shaanxi Province J. Arid Land Resour. Environ. 2019 33 85 91 68. He P. Zhang J. He K. Zeng Y. Cognitive social capital and farmers’ participation behavior in environmental governance—An example of straw resource utilization approach Chin. J. Agric. Resour. Reg. Plan. 2019 40 187 194 69. Huang Y. Luo X. Eating and selling at the same time: An analysis of differences in biopesticide application behavior of rice farmers Chin. Rural Econ. 2018 7 63 78 70. Guo L. Zhao J. Research on farmers’ willingness to apply biopesticides from the perspective of cognitive conflict—Empirical evidence based on 639 rice farmers in Jiangsu J. Nanjing Agric. Univ. 2017 17 123 133 71. Wang J. Wang Q. Zhao Q. Current status and countermeasures for the development of professional pest control in China Chin. Plant Protect. 2014 34 76 80 72. Meng Y. The promotion and application of biopesticide technology in agricultural cultivation in the new era Seed Sci. Technol. 2021 1 55 56 73. Du S. Luo X. Huang Y. Tang L. Yu W. Risk perception, agricultural social services and rice farmers’ biopesticide technology adoption behavior Resour. Environ. Yangtze Basin 2021 30 1768 1779 74. Liu Y. Application status and development trend of light pest control technology J. Chin. Agric. Mech. 2012 4 34 36 10.3901/JME.2012.03.034 75. Sang W. Huang Q. Wang X. Guo S. Lei C. Development, achievements and prospects of phototropism of insects and light trapping technology in China Chin. J. Appl. Entomol. 2019 56 907 916 76. Li W. Zhang W. Wang L. Jin Y. Analysis of the main influencing factors of pesticide use in Shanxi Province and discussion of countermeasures to reduce the amount Chin. Plant Protect. 2018 38 81 84 77. Guo Z. Xie Y. Xu R. Wang J. Xie H. Zhang Z. Practice and reflection on the promotion of pesticide reduction in Hubei Province Chin. Plant Protect. 2020 40 76 79 78. Yang P. Wang K. Li J. Li W. Yin J. Helping agricultural green development with pesticide reduction and pest control Plant Protect. 2018 44 95 100 79. Wang X. Li Z. Tan X. Lei Q. Shao C. Jiao M. Wu Q. Measures and effectiveness of zero growth action of pesticide use in Guizhou Province Chin. Plant Protect. 2021 41 98 102 80. Jin X. Zhang Q. Yao Z. Zhu W. Survey and analysis of pesticide use in Tongxiang City Chin. Plant Protect. 2020 40 105 114 81. Xiong Y. Li X. Zhong Y. A study on farmers’ application behavior based on reduction targets—Micro data from grain farmers in seven provinces Chin. J. Eco-Agric. 2021 29 1262 1273

PMC009xxxxxx/PMC9099654.txt

==== Front Materials (Basel) Materials (Basel) materials Materials 1996-1944 MDPI 10.3390/ma15093268 materials-15-03268 Article Effect of Chlorides Content on the Structure and Properties of Porous Glass Ceramics Obtained from Siliceous Rock https://orcid.org/0000-0002-8080-9808 Rodin Alexander * https://orcid.org/0000-0002-2560-0948 Ermakov Anatoly Erofeeva Irina Erofeev Vladimir Zanotto Edgar Dutra Academic Editor Faculty of Architecture and Construction Engineering, National Research Mordovia State University, 430005 Saransk, Russia; anatoly.ermakov97@mail.ru (A.E.); ira.erofeeva.90@mail.ru (I.E.); yerofeevvt@mail.ru (V.E.) * Correspondence: al_rodin@mail.ru 02 5 2022 5 2022 15 9 326831 3 2022 29 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Porous glass-ceramic materials are used in the construction engineering and repair of various objects. The article investigates the method for obtaining porous glass ceramics from siliceous rock with a high calcite content. To obtain samples with an even fine porous structure, a small amount (≤0.386%) of chloride (NaCl, KCl, MgCl2·6H2O, CaCl2) was added to the charge mixture. At the first stage, mechanochemical activation of raw materials was carried out. Siliceous rock, Na2CO3 and additives (chlorides) were grinded together in a planetary ball mill. The resulting charge was annealed at a temperature of 850 °C. The influence of the type and amount of chloride on the properties of the charge mixture and glass ceramics has been defined by thermal analysis (TA), X-ray diffraction (XRD), scanning electron microscopy (SEM), etc. The chlorides in the charge mixture decreased the calcite’s decarbonization temperature and had an effect on the macro- and microstructure of the material. As a result, samples of glass ceramics with an even finely porous structure in the form of blocks were obtained. The samples consist of quartz, wollastonite, devitrite, anorthoclase and an amorphous phase. On average, 89–90% of the resulting material consists of with small pores. The apparent density of the samples is in the range of 245–267 kg/m3. Bending and compressive strength reaches 1.75 MPa and 3.8 MPa, respectively. The minimum thermal conductivity of the modified samples is 0.065 W/(m∙°C). The limiting operating temperature is 860 °C, and the minimum thermal shock resistance is 170 °C. The material has a high chemical stability. They can be used as thermal insulation for some types of industrial and civil facilities. glass ceramic foaming construction material thermal insulation siliceous rocks chlorides compressive strength thermal conductivity thermal analysis Russian Science Foundation21-79-10422 This research was funded by Russian Science Foundation, grant number 21-79-10422, https://rscf.ru/project/21-79-10422/. ==== Body pmc1. Introduction Porous glass-ceramic materials have many unique properties. They are poor heat conductors, have relatively high strength and chemical stability, are non-inflammable, can be operated at high temperatures, etc. [1,2]. They find a wide utility as thermal insulation in the construction and repair of civil and industrial facilities [3,4]. Raw materials such as glass waste [5,6], slag from metallurgical industries [7], fly ash [1,2,8,9], siliceous rocks [10,11,12], etc., are utilized to obtain porous glass ceramics. The porous structure of the material is obtained in different ways [1,12,13]. One of the most employed is powder foaming. Under this method, a foaming component is added to the charge mixture, which foams the material at a certain temperature. The types of foaming components are diverse: carbonates [1,2,14], carbon mixed with MnxOy [4,15], Na2CO3 [16], etc. Porous glass ceramics from siliceous rocks are mainly obtained by the alkaline activation of components [10,11,12]. Diatomite, zeolite-containing tripoli, opoka is mixed with a NaOH aqueous solution of high concentration. The resulting mixture is granulated and annealed. The material comes out in the form of granules. We have obtained porous glass ceramics from siliceous rocks by powder foaming without the use of foaming components. The charge mixture is foamed owing to the zeolite minerals contained in the rock. The method allows to obtain samples in the form of blocks [17]. The main requirement is the joint mechanochemical activation of siliceous (zeolite-containing) rocks and Na2CO3. Siliceous rocks have different chemical and mineralogical composition [11,18]. The increased calcite content in the rock makes the macrostructure of glass ceramics uneven. This is due to the superimposition of the effects of the beginning of charge mixture’s melting and decarbonization of calcite. There is little molten in the charge mixture while already a strong gas formation [19]. A partial decrease of the calcite’s decarbonization temperature was achieved by changing the activation mode of the charge mixture: by increasing the activation time, the rotation speed of the mill cups, etc. [18]. We have managed to obtain a fine-pored macrostructure of glass ceramics from siliceous rocks with calcite by introducing amorphous SiO2 into the charge mixture as an additive. With the introduction of this additive, the amount of calcite decreases, and the remaining CaCO3 decarbonizes at a lower temperature [20]. A similar effect was observed when we had an increase in the amount of Na2CO3 in the charge mixture [21]. However, following the increase in the amount of amorphous SiO2 and Na2CO3 in the charge mixture composition, the cost of porous glass ceramics increased and some properties of the material deteriorated (the limiting operating temperature fell, chemical resistance decreased, etc.) The scholarly literature has some evidence on the effect of certain types of additives on the temperature and rate of carbonate minerals’ decomposition during heating. This is known in relation to the decrease in the decomposition temperature of magnesite, dolomite and calcite with different Na salts in the composition [22,23]. Joint annealing of MgCO3 and a small amount of LiNO3, NaNO3, and KNO3 sharply accelerated decarbonization [24,25,26]. Joint annealing of brucite, dolomite, serpentite and a small amount of chlorides (NaCl, KCl, etc.), also decreased their decomposition temperature [23,27]. The effect of nitrates, chlorides and some other salts on the decomposition temperature of carbonate minerals is described differently in the literature. Some authors claim that when heated, salts melt, and MgO (CaO) dissolve in the molten and are decarbonized faster [28]. Other researchers claim that molten salts of alkali act primarily as a diffusion medium for CO2. The molten prevents the formation of a one-component carbonate impermeable to carbonate ions which facilitates the intensive formation of CO2 [29]. So far, we have not been able to find any evidence about the use of chlorides or nitrates as additives that accelerate the decomposition of calcite when obtaining porous glass ceramics from siliceous rocks. The research objective was to reveal the effect of chlorides (NaCl, KCl, MgCl2∙6H2O, CaCl2) added to mechanochemically activated charge mixture composition (siliceous rock + Na2CO3) on the structure and properties of porous glass-ceramic materials. Tasks:To employ thermal analysis (TA) and X-ray diffraction (XRD) to define the effect of additives on phase transformations in the charge mixture during heating and the phase composition of annealed porous glass ceramics; To reveal the effect of additives in the charge mixture composition on the macro- and microstructure of porous glass ceramics; To determine the effect of the content of NaCl, KCl, MgCl2·6H2O, and CaCl2 in the charge mixture composition on the physico-mechanical and thermophysical properties, as well as chemical stability of porous glass ceramics samples. 2. Materials and Methods 2.1. Materials Porous glass-ceramic samples were fabricated by mixing siliceous rock, sodium carbonate and additives. Siliceous rock of the following chemical composition was used: SiO2—67.86%; CaO—7.74%; Al2O3—7.61%; Fe2O3—1.99%; K2O—1.56%; MgO—1.07%; TiO2—0.34%; Na2O—0.17%; P2O5—0.15%; SO3—0.06%; SrO—0.06%; BaO—0.02%; ZrO2—0.01%; V2O5—0.01%; MnO—0.01%; Cr2O3—0.01%; LOI—11.32%. The rock’s mineralogical composition: cristobalite—20.5%; heulandite—20.4%; quartz—15.5%; calcite—10.5%; muscovite—13.1%; amorphous phase—20.0%. The siliceous rock was dried to a moisture content of <1%. Sodium carbonate (Na2CO3) was used to reduce the melting point and foaming of the charge mixture. The purity was ≥99%. To obtain an even macrostructure of pores in glass ceramics, chlorides and their crystallohydrates were used as additives: NaCl, KCl, MgCl2·6H2O, CaCl2. The purity was ≥97%. 2.2. Compositions and Fabrication Technology of Samples Samples of porous glass-ceramic materials were fabricated using the following technology:Mechanochemical activation of raw materials. Siliceous rock, sodium carbonate and additives were being grinded in a planetary ball mill Retsch PM 400 for 35 min (overload inside the crushing cylinder—20G). The required concentration of components was determined by us on the basis of preliminary tests, as well as by the results of previous works [17]. The charge mixture compositions are presented in Table 1. Annealing. The charge mixture obtained after mechanochemical activation was put into metal molds with a size of 120 mm × 120 mm × 260 mm and annealed in a muffle furnace (the molds were previously coated with clay). Annealing schedule: heating to a temperature of 670 °C at a rate of 4.5 °C/min, holding at a temperature of 670 °C for 1 h, heating to a temperature of 850 °C at a rate of 4.5 °C/min, holding at a temperature of 850° C for 30 min; cooling to room temperature inside the furnace. Preparation of samples for testing. The molds with the resulting material were removed from the furnace and dismantled. The resulting porous glass ceramics were sawn into samples of the required sizes and tested. 2.3. Analytical Techniques 2.3.1. X-ray Diffraction (XRD) The X-ray patterns were identified by using an Empyrean PANalytical device (PANalytical, Almelo, The Netherlands) with a PIXcel3D semiconductor detector. The experiment was carried out on grinded glass ceramic samples (fraction < 90 µm). The detector operated in linear scanning mode. Diffraction patterns were identified in CuKa emission in the scan range 2Θ = 4–80°. The step size was 0.0067°/min, the counting time was 150 s. The phase composition of the samples was determined by the Hanawalt method. We used the ICDD PDF-2 database. 2.3.2. Thermal Analysis (TA) Differential thermal analysis (DTA) and differential thermogravimetry (DTG) of the charge mixture was made using a TGA/DSC1 device (Mettler-Toledo, Greifensee, Switzerland). The charge mixture weighing 20 ± 0.1 mg was put into an alumina crucible with a volume of 150 mcl and compacted. The crucible with the sample was placed in the apparatus and heated from 30 to 850 °C at a rate of 10 °C/min. Images were processed by STARe software (Version SW 10.00, Mettler-Toledo, Greifensee, Switzerland). 2.3.3. Scanning Electron Microscopy (SEM) SEM micrographs of porous glass-ceramic samples were obtained using the Quanta 200 I 3d apparatus (FEI Company, Hillsboro, OR, USA). Sample scanning mode: pressure—60 Pa, accelerating voltage—30 kV, operating distance—15 mm. 2.3.4. Apparent Density and Porosity At the first stage, the true density of glass-ceramic materials (ρ0, g/cm3) was found. Then, the cube-shaped samples with 50 ± 5 mm on edge were dried, weighed (m0, g) and measured (V, cm3). The samples were put in a cylindrical tank for vacuuming. Water was poured into the tank in such a way that its level in the tank was 20 mm higher than that of the samples. Preliminarily, the water density (ρw, g/cm3) was determined. Air was pumped out of the tank to a residual pressure equal to 2000 Pa. Testing samples were taken out of the tank after weight stabilization but not less than after 2 h exposure. The water-saturated samples (m1, g) were weighed indoors. Calculations were done according to the Equations (1)–(4): Apparent density (ρ, g/cm3):(1) ρ=m0V, Total porosity (Pt, %):(2) Pt=ρ0−ρρ·100, Open porosity (Po, %):(3) Po=m1−m0V·ρw·100, Closed porosity (Pc, %):(4) Pc=Pt−Po. The arithmetic mean of the test results of three samples of each composition was taken as the final result. 2.3.5. Bending and Compressive Strength To define the bending strength, dry samples in the form of rectangular prisms with 120 mm × 30 mm × 30 mm on edges were used. The sample was placed horizontally on two cylindrical supports with a distance of 100 ± 1 mm between them. A cylindrical rod was installed on top along the entire width of the sample at an equal distance from the supports. The diameter of the supports and the rod was 6 ± 0.1 mm. The load on the sample was applied by the rod at a speed of 5 mm/min. The maximum destructive force was taken to be the value at which the sample cracked. The bending strength was calculated according to the standard formula. The arithmetic mean of the test results of three samples of each composition was considered as the final result. To find out the compressive strength, dry cube-shaped samples of porous glass ceramics with 90 ± 5 mm on edge were tested. The maximum destructive force was identified when the sample was fractured (cracks appeared) or deformed in the surface layers by 10 % of the initial height. The compressive strength of the samples was calculated as the relation of maximum destructive force against the sample’s cross-section area. The arithmetic mean of the test results of five samples of each composition was considered as the final result. 2.3.6. Thermal Conductivity The thermal conductivity of samples was determined by the probe method using the mobile thermal conductivity meter (MIT–1, LLC “Scientific and production enterprise INTERPRIBOR”, Chelyabinsk, Russian Federation). The testing was done on dry cube-shaped samples with a 90 ± 5 mm on edge. A 6 mm diameter hole with a depth of 50 to 60 mm was drilled in the center of the sample’s side plane. A probe was put into the hole to make readings. The arithmetic mean of the test results of five samples of each composition was taken as the final result. 2.3.7. Thermal Shock Resistance To identify the values of the thermal shock resistance, cube-shaped samples 50 ± 5 mm on edge were used. The samples were kept in a thermostat at a temperature (TT) 110 °C ≥ 2 h and then quickly (<10 s) immersed in a water tank. The water temperature (Tw) was 20 ± 2 °C. There, they were kept for 65 ± 5 s. The appearance of cracks on the surface of samples was monitored. The experiment was repeated raising the thermostat temperature by 10 °C until cracks appeared on all samples. The thermal shock resistance of each sample was calculated by the Equation (5):(5) ΔT=TT−Tw−10. The arithmetic mean of the test results of four samples of each composition was taken as the final result. 2.3.8. Limiting Operating Temperature To identify the values of the limiting operating temperature, samples in the form of rectangular prisms 90 mm × 40 mm × 40 mm on edges were used. They were installed vertically in a muffle furnace and tested in the following way: heating to 50 °C less than the set temperature—10 °C/min, heating to the set temperature—2 °C/min, exposure at the set temperature—2 h. During the experiment, the change in the size of the samples was monitored. The limiting operating temperature of the material was determined against the highest test temperature at which the sample sizes changed <1% of the initial values. The experiment was repeated raising the set temperature by 10 °C until the sample sizes changed >1% of the initial values. The arithmetic mean of the test results of three samples of each composition was taken as the final result. 2.3.9. Chemical Stability To define the values of the chemical resistance of glass-ceramic materials, we tested samples grinded to a fraction of 0.315–0.630 mm. During the experiment, the change in the samples’ weight was controlled after boiling for 3 h in a chemical medium (distilled water, an aqueous solution of 6 N HCl, a mixture of equal volumes of 1 N Na2CO3 and NaOH solutions). The samples were dried and put into a test jar in an amount of 5 ± 0.0005 g. Then, we dispensed 100 ± 0.5 cm3 of reagent into the test jar, connected a backflow condenser to it and boiled the solution. The liquid was drained after boiling, and the sample was washed 5 times with distilled water. The washed sample was drained through a funnel with a paper ash-free filter and placed in a quartz crucible. The crucible with the sample was calcined for 1 h in a muffle furnace at a temperature of 800 ± 10 °C and cooled to 150 °C. The crucible with the sample was cooled to room temperature in a desiccator with CaCl2. The samples were weighed, and their weight loss was measured. The arithmetic mean of the test results of two samples of each composition was taken as the final result. 3. Results and Discussion 3.1. Charge Mixture’s TA The effect of chlorides on phase transformations in the charge mixture during heating has been established by thermal analysis. The results of differential thermal analysis (DTA) and differential thermogravimetry (DTG) of the charge mixture are presented in Figure 1. According to Figure 1, the following processes occur in the charge mixture from siliceous rock when heated. The first intensive endothermic effect (Figure 1a) and a significant weight loss of the sample (Figure 1b) were detected at ≈120 °C. According to scholarly literature, sodium hydrosilicates release water in the siliceous charge mixture in this temperature range [30]. When the charge mixture is heated above 300 °C, it triggers the reaction of formation of sodium silicates [11,30]. The reaction is accompanied by an endothermic effect and a significant loss of samples’ weight. The DTG curve of C1 (Figure 1b) shows that the sample is strongly losing its weight until the temperature of ≈550 °C. The next endothermic effect and weight loss of all samples until the temperature reaches ≈650 °C is attributed to the superimposition of the decomposition effects of select minerals of siliceous rock: calcite, muscovite and heulandite. This conclusion is confirmed by the findings in [11,17,30]. Many authors believe that in this temperature range, in the siliceous (zeolite-containing) charge mixture, there is a sealing of surface hydroxyl groups (Si–O–H) in micropores which form water vapor and foam it during the melting of the charge mixture [11,30]. Insignificant endothermic effect and the loss of sample’s weight of C1 with the peak at the temperature ≈680 °C are linked with decarbonization of non-reacting CaCO3. The onset of the endothermic melting effect of the charge mixture occurs at ≈680 °C. The temperature at the beginning of the crystallization of the charge mixture sample C1 (the beginning of the exothermic effect) is ≈710 °C. Crystallization is characterized by an intensive exothermic effect with the peak at ≈760 °C. According to Figure 1, chlorides (NaCl, KCl, MgCl2·6H2O and CaCl2) affect the charge mixture’s phase transformations at >300 °C. By increasing the amount of these chlorides to 0.368% of the charge mixture weight, the intensive loss of the samples’ weight stops at ≈500 °C. This is 50 °C less in comparison with the sample without the additive. With the introduction of chlorides into the charge mixture, the intensity of the endothermic effect and the loss of weight attributable to the decomposition of calcite, muscovite and heulandite decreases. The endothermic effect with a peak at ≈680 °C shifts to the region of lower temperatures (≈665 °C) and is almost imperceptible. According to Figure 1a, the type and amount of chlorides used in the work does not produce an effect on the beginning of charge mixture melting. The increase in the amount of chlorides in the charge mixture to 0.368% shifts the beginning of crystallization by ≈10–20 °C to the region of higher temperatures. The exothermic effects of crystallization are also shifted to the region of higher temperatures and are less intensive compared to samples without additives. The results of thermal analysis of the charge mixture samples confirmed the positive effect of chlorides on reducing the decomposition temperature of carbonate minerals. Additionally, the effect of chlorides on the crystallization of the charge mixture was revealed. 3.2. Grass Ceramics Samples’ XRD Figure 2 shows the X-ray patterns of samples of annealed glass ceramics. Samples of the control composition (C1) and those modified with chlorides (NaCl, KCl, MgCl2·6H2O and CaCl2) (samples C4, C7, C10 and C13) were tested. For visual clarity, the X-ray patterns are presented in the scan range of 2Θ = 5–45°. Based on the results of the XRD of the samples (Figure 2), the qualitative phase composition of glass-ceramic materials has been identified. All samples of porous glass ceramics from siliceous rocks consist of a crystalline and amorphous phase. The presence of an amorphous phase in the samples is evidenced on all X-ray patterns by a non-monotonic change in the background (halo) in the scan range from 17 to 37° (2θ). The change of the amorphous halo depending on the type and amount of chlorides in the composition of the material has not been detected. The crystal phase of the control sample (C1) and chloride-modified samples (C4, C7, C10 and C13) consists of quartz [SiO2, ICDD: 01-075-8320], wollastonite [CaSiO3, ICDD: 01-076-0186], sodium calcium silicate (devitrite) [Na2Ca3Si6O16, ICDD: 00-023-0671] and anorthoclase [(Na0.85K0.14) (AlSi3O8), ICDD: 01-075-1634]. When modifying the charge mixture with chlorides (NaCl, KCl, MgCl2·6H2O and CaCl2 in an amount up to 0.386% of the total weight), the phase composition of glass ceramics changes slightly. The results of the phase composition of glass-ceramic materials from siliceous rocks are consistent with the scholarly literature findings [17]. 3.3. Porous Glass Ceramics Macrostructure To visually confirm the effect of chlorides in the charge mixture composition on the change in the macrostructure of porous glass-ceramic samples, the surface of the latter was scanned. Figure 3 shows scans of the surface of porous glass ceramics samples of the control composition (C1) and samples with NaCl (C2–C4). According to Figure 3, the sample of the control composition has an uneven macrostructure over the entire surface area. The scan of C1 shows hollows with a diameter of up to 7 mm and channels with a length of ≈10 mm in one cross-sectional plane. Samples of glass-ceramic materials from a charge mixture with NaCl in the amount of 0.092–0.386% have a uniform fine-porous structure (C2–C4). The pore size decreases from ≈ 1 mm to ≈ 0.5 mm against the increase in NaCl from 0.092% to 0.386% in charge mixture composition. The dependence of the macrostructure of porous glass-ceramic samples from the charge mixture with KCl, MgCl2·6H2O, CaCl2 is similar to samples from the charge mixture with NaCl. 3.4. SEM Micrographs of Samples The effect of NaCl in the charge mixture composition on the microstructure of porous glass ceramics samples is shown in Figure 4. The results were obtained by the SEM. A control sample (C1) and samples from a charge mixture with 0.092% and 0.386% of NaCl (C2 and C4) were tested. According to Figure 4, most of the pores in the control sample (C1, without additives) have the spherical shape. The pores have different diameters (up to 1 mm or more). The surface of the pore walls is not smooth, but rough. The SEM micrograph of C1 shows small pores in the walls connecting large pores with a diameter of up to 0.3 mm. There are many closed and open microscopic pores <50 µm in the pore walls. The SEM micrographs of C2 display pores with a diameter of ≤1 mm. Some pores are connected to each other by wide channels. The pore walls consist of closed micropores. At a higher magnification, microscopic holes with a diameter of ≤20 µm are visible in the pore walls. Sample C2 was obtained from the charge mixture containing NaCl in an amount of 0.092%. The increase in the amount of NaCl in the charge mixture to 0.386% (SEM micrograph C4) causes the appearance of many pores in glass ceramics with a diameter of <0.5 mm. Some pores are linked to each other by holes with a diameter of <0.3 mm. At a higher magnification, microscopic holes are not visible in the pores’ walls. The maximum micropore diameter is less than 20 µm. Following the completed analysis (Figure 4), it can be stated that small quantities of chlorides in the charge mixture composition produce a significant effect on the microstructure of porous glass ceramics from siliceous rocks. The increase in the amount of NaCl in the charge mixture to 0.386% decreases the pore size and diameter of small pores in the walls connecting large pores. It is known from the scholarly literature that open pores in glass ceramics are a consequence of intensive crystallization of samples [4]. The results of the DTA (Figure 1a) confirm this. The beginning of crystallization in samples with chlorides is shifted to higher temperatures by 10–20 °C, and its intensity is much less. Additives KCl, MgCl2·6H2O and CaCl2 in the charge mixture produce a similar effect on the microstructure of porous glass ceramics as NaCl does. 3.5. Apparent Density and Porosity of Samples Figure 5 shows dependence graphs of the apparent density and porosity of glass-ceramic samples on the type and amount of additives (NaCl, KCl, MgCl2·6H2O and CaCl2) in the charge mixture composition. The effect of chlorides in the charge mixture composition on the apparent density of porous glass-ceramic samples is as follows (Figure 5). The apparent density increased from ≈220 kg/m3 (C1) to 245–250 kg/m3 when NaCl, KCl, MgCl2·6H2O and CaCl2 were introduced into the charge mixture in the amount of 0.092% (C2, C5, C8 and C11). The increase of all these chlorides to 0.386% in the charge mixture composition increased the apparent density of the samples almost linearly to 257–267 kg/m3 (depending on the type of additive). A similar influence of all chlorides used in the work on the apparent density of glass ceramics has been found. According to the findings (Figure 5b), the total porosity of glass-ceramic samples decreases slightly against the increase of chlorides content in the charge mixture composition. In control samples, this indicator was 91.3%. The samples from the charge mixture containing 0.386% CaCl2 have the lowest value of total porosity (≈89%). Regardless of the type of additive, 66–69% of the volume of all samples is occupied by open pores. The effect of all chlorides used in the work on the open porosity of glass ceramics is similar. A small amount of additive in the charge mixture (0.092%) decreased the number of open pores in the samples. According to SEM findings (Figure 4), chlorides in small amounts in the charge mixture composition caused the decrease in the diameter of small pores in the walls connecting large pores in the material. When the amount of additives was increased to 0.386% (of the charge mixture’s weight), the open porosity of the samples increased but did not exceed the value of the control composition (Figure 5). The effect may be attributed to the appearance of a system of channels linking adjacent pores inside the sample. The SEM micrograph of C4 (Figure 4) displays that some pores are linked to each other with holes up to 0.3 mm in diameter. 3.6. Strength of Samples The influence of NaCl, KCl, MgCl2·6H2O and CaCl2 admixed to charge mixture on the strength values of porous glass ceramics samples from siliceous rocks is shown in Figure 6. Following the completed studies (Figure 6a), the bending strength of porous glass-ceramic samples is linearly related to their apparent density. When we increase the content of NaCl, KCl, MgCl2 6H2O and CaCl2 in the charge mixture composition to 0.386%, the apparent density of the samples increased and, as a consequence, their bending strength did as well. This indicator also depends on the type of additive. For example, samples (C2) from a charge mixture containing 0.092% of NaCl have an average bending strength of 1.2 MPa. In samples (C11) from the charge with the same amount of CaCl2, this indicator is 25% higher (1.5 MPa). The apparent densities of samples C2 and C11 are almost equal (≈245 kg/m3). Increasing the amount of NaCl and CaCl2 in the charge mixture to 0.386% leads to the increase in bending strength of porous glass ceramic samples (C4 and C13) to 1.35 and 1.75 MPa, respectively. C13 samples are almost 30% stronger than C4 samples with a slight difference in apparent density. The bending strength of samples from the charge mixture containing equal amount of chloride increased with respect to the type of additive in the following sequence NaCl→KCl→MgCl2·6H2O→CaCl2. The effect can be attributed to the change in the phase composition of the samples. According to the X-ray patterns (Figure 2), the sample of glass ceramics from the NaCl charge mixture has the smallest amount of wollastonite, and from the CaCl2 charge mixture, the largest. The bending strength of glass-ceramic materials containing different amounts of wollastonite was investigated in [31]. We also have to draw attention to the different patterns of the crystallization process in samples from charge mixture containing different chlorides (Figure 1). In samples from the charge mixture with NaCl, crystallization takes place intensively and at a lower temperature. Crystallization in samples with CaCl2 is less intensive with a maximum at a higher temperature. According to the research findings (Figure 6b), the compressive strength of porous glass-ceramic samples from the charge mixture with chlorides also increases linearly against the increase in their apparent density. The highest value belongs to samples C10 and C13 (≈3.8 MPa). The average density of the samples is in the range of 256–267 kg/m3. The samples of the control composition (C1) have the lowest compressive strength ≈2.2 MPa and apparent density ≈220 kg/m3. Following the findings for the bending strength, the same relationship between the increase in compressive strength of samples from the charge mixture with an equal amount of chloride and the type of additive (NaCl→KCl→MgCl2·6H2O→CaCl2) remains. The introduction of chloride additives into the siliceous charge mixture with sodium carbonate allowed to obtain porous glass-ceramic materials which having equal apparent density, surpass foam glass and glass ceramics from industrial waste in strength values [1,2,10,11,12]. 3.7. Thermal Conductivity of Samples The results of thermal conductivity of porous glass-ceramic samples are shown in Figure 7. The results are shown in relation to the apparent density of the samples. Based on the results of the completed analysis (Figure 7), a linear dependence of the samples’ thermal conductivity on their apparent density has been revealed. The dependence is valid for glass ceramic samples with an apparent density from 210 to 270 kg/m3. It can be written down in the following Equation (6):(6) λ=17.1·10−5·ρ+0.023, where: λ—thermal conductivity (W/(m∙°C)), ρ—apparent density of a dry material (kg/m3). Correlation coefficient (R2) ≈ 0.92. According to experimental findings (Figure 7), the control samples (C1) have the lowest thermal conductivity equal to an average of 0.06 W/(m∙°C). The apparent density of the samples is 220 kg/m3. According to Figure 3, the macrostructure of the sample C1 is uneven. The lowest thermal conductivity of porous glass ceramics with a uniform fine-porous structure is in samples from the charge mixture containing 0.092% NaCl, KCl, MgCl2·6H2O and CaCl2. It is equal on average to 0.065–0.066 W/(m∙°C) with an apparent sample density from 245 to 249 kg/m3. Samples from the charge mixture with the maximum amount of chlorides (0.386%) have the highest value of thermal conductivity which is ≈0.068 W/(m∙°C). The apparent density of these samples is from 262 to 267 kg/m3. The results obtained correlate with the findings of previous studies [3]. 3.8. Thermal Shock Resistance In order to expand the scope of application of the developed materials, their thermal shock resistance has been defined. Porous glass ceramics with high values of this parameter can be used as thermal insulation of industrial equipment [3,32]. The results of experimental testing are presented in Figure 8. After analyzing the values of Figure 8, it was found that the thermal shock resistance of porous glass-ceramic samples decreased against the increase in the amount of chlorides in the charge mixture (NaCl, KCl, MgCl2∙6H2O and CaCl2) to 0.184% or more. The lowest values of thermal shock resistance on average equal to 170 °C in samples C3, C4, C6, C7, C10, C12 and C13. The highest thermal shock resistance of the control samples is ≈190 °C. It was not possible to find the influence of the phase composition of the material on the samples’ thermal shock resistance. According to Figure 2, the effect of chlorides in the charge mixture on the phase composition of glass ceramics is insignificant. The decrease in the samples’ thermal shock resistance as response to the increase in the amount of chloride in the charge mixture can probably be attributed to the ordering of the structure of the material. According to Figure 3 and Figure 4, sample C1 consists of large macropores up to 7 mm in diameter and small pores in the walls connecting large pores up to 0.3 mm. Increasing the content of additive in samples decreases the pore diameter (<1 mm), and the holes are practically not visible on SEM micrographs. Probably, the initial temperature retained longer inside the material with a uniform fine-pored structure. A large temperature variation was formed between the surface of the sample and its middle part. As a result, the thermal shock resistance of the samples fell. The findings are consistent with the results of other researchers. The relationship between thermal shock resistance and porosity of glass ceramics is described in detail in the scholarly literature [4]. The obtained values of thermal shock resistance of porous glass-ceramic materials are almost identical to those obtained from industrial waste [32]. The developed materials can be recommended as thermal insulation for some types of industrial equipment [3,32]. 3.9. Limiting Operating Temperature One of the main criteria for the use of porous material in the thermal insulation of industrial equipment is the limiting temperature of its operation. The effect of the type and amount of additive (NaCl, KCl, MgCl2·6H2O and CaCl2) in the charge mixture composition on the limiting operating temperature of glass-ceramic samples from siliceous rocks is shown in Figure 9. The parameter value was determined by measuring the size of samples after holding them for 2 h at a given temperature. Based on the completed experimental studies, it was found that the addition of NaCl, KCl, MgCl2·6H2O and CaCl2 to the charge mixture in the amount of up to 0.386% (of the charge mixture weight) had almost no effect on the limiting operating temperature of the samples (Figure 9). The developed porous glass-ceramic materials can be operated at temperatures under 860 °C. The residual sizes of the samples after holding them for 2 h at this temperature were more than 99% of their initial values. As mentioned above, chlorides were introduced into the siliceous charge mixture with a high calcite content to obtain a uniform fine-pored structure of glass ceramics. A similar effect can be achieved by increasing the amount of amorphous SiO2 (for example, diatomite) and Na2CO3 [21]. However, the limiting operating temperature of such glass-ceramic materials rarely exceeds 800 °C [20]. Moreover, in terms of the limiting operating temperature, porous glass-ceramic materials from siliceous rocks (modified with chlorides) are significantly superior to foam glass, as well as porous glass-ceramics from siliceous rocks obtained by alkaline charge mixture activation [10,11,12,19,30]. The developed materials can be used as thermal insulation for some types of industrial equipment. 3.10. Chemical Stability The effect of the type and amount of chlorides in the charge mixture composition on the chemical stability of porous glass-ceramic samples is shown in Table 2. The criterion for estimating this parameter was the weight loss of grinded samples (fraction 0.315–0.63 mm) after boiling them for 3 h in various chemical media. According to the findings (Table 2), the developed glass-ceramic materials from siliceous rocks have high chemical stability in water. This indicator decreases slightly in samples from the charge mixture with chloride. After boiling for 3 h, the samples from the charge mixture with the maximum amount of additive (NaCl, KCl, MgCl2·6H2O and CaCl2) 0.386% lost about 1% in weight on average. The result obtained makes it possible to recommend the developed materials for use in wet conditions. The effect of chlorides on the chemical stability of glass ceramic samples in an aqueous solution of HCl (6 N) was found. The stability increased slightly or was almost equal to the value of the control composition when we increased the chlorides content in the charge mixture to 0.092% and 0.184% (Table 2). After the increase in the amount of the additive to 0.386%, the chemical stability of the samples decreased. Weight loss of samples containing NaCl (C4) and KCl (C7) in the charge mixture was >5%. It is known from the scholarly literature that different minerals have different resistance to the action of acids [33]. According to Figure 2, glass ceramics samples C4 and C7 contain the largest amount of devitrite. This probably led to the decrease in the chemical stability of the samples in an aqueous solution of HCl (6N). The stability of glass ceramics from siliceous rocks to the alkaline solutions exposure (Na2CO3(1 N) + NaOH(1 N)) decreased slightly due to the increase of NaCl, KCl, MgCl2·6H2O and CaCl2 (Table 2) in the charge mixture composition. With the maximum additive content in the charge mixture (0.386%), the weight loss of glass ceramic samples (C4, C7, C10 and C13) after boiling in an alkaline solution for 3 h increased by 1–1.5% on average. The effect may also be attributed to a slight change in the phase composition of glass ceramics caused by the type and amount of additive in the charge mixture (Figure 2). Following the results of the completed studies, high chemical stability of glass-ceramic samples from siliceous rocks has been revealed. The values obtained are higher or equal to the values of analogues [34,35]. Even if we consider a slight decrease in the chemical stability of select samples from the charge mixture with chlorides. The developed materials can be used as thermal insulation for some types of pipelines, industrial installations, etc. 4. Conclusions Porous glass-ceramic materials were obtained from siliceous rock with a high calcite content. Mechanochemical activation of the rock, sodium carbonate and an additive was carried out in a planetary ball mill. Chlorides (NaCl, KCl, MgCl2·6H2O and CaCl2) were used as additives to obtain glass-ceramic samples with an even fine-pored structure. After activation, the charge mixture was annealed at a temperature of 850 °C. The influence of the type and quantity of additives on the properties of the charge mixture and glass ceramics has been identified by thermal (TA) and X-ray diffraction (XRD) analysis, scanning electron microscopy (SEM), etc. Main conclusions:A small amount of chloride produces a significant effect on the phase transformations in the charge mixture obtained from siliceous rock. The increase in the amount of additive in the charge mixture to 0.368% has accelerated the formation of sodium silicates and decreased the calcite’s decarbonization temperature. The peak of the exothermic effect shifted to higher temperatures by ≤40 °C. Samples of glass ceramics with a uniform fine-porous structure from siliceous rock with a calcite content (10.5%) have been obtained by modifying the charge mixture with chlorides (NaCl, KCl, MgCl2·6H2O and CaCl2) in the amount of 0.096 to 0.368%. The pore diameter in the material decreased to ≈0.5 mm after the increase in the amount of additive. Glass ceramic samples consist of quartz, wollastonite, devitrite, anorthoclase and amorphous phase. The additives used (chlorides) have a negligible effect on the phase composition of the samples. The developed porous glass ceramic has an apparent density of 245–267 kg/m3; bending and compressive strength up to 1.75 MPa and 3.8 MPa, respectively; thermal conductivity 0.065–0.068 W/(m∙°C); thermal shock resistance 170–180 °C; limiting operating temperature up to 860 °C; and high chemical stability. Comparing some indicators, the obtained materials are superior to foam glass and other analogues. They can be used as thermal insulation for some types of industrial and civil facilities. Author Contributions Conceptualization: A.R.; methodology: A.R. and V.E.; software: A.R. and I.E.; validation: A.R. and A.E.; formal analysis: A.E.; investigation: A.R. and A.E.; resources: A.R.; data curation: A.R. and V.E.; writing—original draft preparation: A.R.; writing—review and editing: A.R. and V.E.; visualization: A.R. and A.E.; supervision: A.R. and V.E.; project administration: A.R.; funding acquisition: A.R. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement All data are freely available. Conflicts of Interest The authors declare no conflict of interest. Figure 1 DTA (a) and DTG (b) curves of charge mixture samples. Figure 2 X-ray patterns of glass ceramics samples. Figure 3 Scan of the surface of porous glass-ceramic samples. Figure 4 SEM micrographs of samples. Figure 5 Apparent density (a) and porosity (b) of samples. Figure 6 Bending (a) and compressive (b) strength of samples. Figure 7 Thermal conductivity of samples. Figure 8 Thermal shock resistance of samples. White square – without chlorides; Gray squares – with chlorides. Figure 9 Residual size of samples after exposure to the set temperature for 2 h (samples from charge mixture containing: (a) no additive and with NaCl; (b) with KCl; (c) with MgCl2·6H2O; (d) with CaCl2; C1–C13—numbers of compositions). materials-15-03268-t001_Table 1 Table 1 Charge mixture’s compositions. No. of Composition Charge Mixture’s Compositions, % Siliceous Rock Na2CO3 NaCl KCl MgCl2·6H2O CaCl2 C1 81.6 18.4 – – – – C2 18.308 0.092 – – – C3 18.216 0.184 – – – C4 18.032 0.368 – – – C5 18.308 – 0.092 – – C6 18.216 – 0.184 – – C7 18.032 – 0.368 – – C8 18.308 – – 0.092 – C9 18.216 – – 0.184 – C10 18.032 – – 0.368 – C11 18.308 – – – 0.092 C12 18.216 – – – 0.184 C13 18.032 – – – 0.368 materials-15-03268-t002_Table 2 Table 2 Change in the samples’ weight after boiling in chemical media for 3 h. Composition No. Change in the Samples’ Weight after Boiling in Chemical Media for 3 h, % * H2O 6 N HCl Solution 1 N Na2CO3 Solution + 1 N NaOH Solution (1:1) C1 0.34 4.34 7.34 C2 0.65 3.61 7.67 C3 0.76 4.48 7.97 C4 1.05 5.04 8.07 C5 0.89 4.29 8.19 C6 0.98 4.50 8.36 C7 1.21 5.34 8.81 C8 1.10 4.10 7.55 C9 1.07 4.29 8.05 C10 1.05 4.72 8.57 C11 1.22 3.86 7.86 C12 0.46 3.81 8.10 C13 0.91 4.49 8.45 * The differences in the test results of the samples of each composition did not exceed 5% of the average value. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Fernandes H.R. Tulyaganov D.U. Ferreira J.M.F. Preparation and characterization of foams from sheet glass and fly ash using carbonates as foaming agents Ceram. Int. 2009 35 229 235 10.1016/j.ceramint.2007.10.019 2. Zhu M. Ji R. Li Z. Wang H. Liu L. Zhang Z. Preparation of glass ceramic foams for thermal insulation applications from coal fly ash and waste glass Constr. Build. Mater. 2016 112 398 405 10.1016/j.conbuildmat.2016.02.183 3. Guo H. Ye F. Li W. Song X. Xie G. Preparation and characterization of foamed microporous mullite ceramics based on kyanite Ceram. Int. 2015 41 14645 14651 10.1016/j.ceramint.2015.07.186 4. König J. Lopez-Gil A. Cimavilla-Roman P. Rodriguez-Perez M.A. Petersen R.R. Østergaard M.B. Iversen N. Yue Y. Spreitzer M. Synthesis and properties of open- and closed-porous foamed glass with a low density Constr. Build. Mater. 2020 247 118574 10.1016/j.conbuildmat.2020.118574 5. Hisham N.A.N. Zaid M.H.M. Aziz S.H.A. Muhammad F.D. Comparison of foam glass-ceramics with different composition derived from ark clamshell (ACS) and soda lime silica (SLS) glass bottles sintered at various temperatures Materials 2021 14 570 10.3390/ma14030570 33530370 6. Kyaw Oo D’Amore G. Caniato M. Travan A. Turco G. Marsich L. Ferluga A. Schmid C. Innovative thermal and acoustic insulation foam from recycled waste glass powder J. Clean. Prod. 2017 165 1306 1315 10.1016/j.jclepro.2017.07.214 7. Cao J. Lu J. Jiang L. Wang Z. Sinterability, microstructure and compressive strength of porous glass-ceramics from metallurgical silicon slag and waste glass Ceram. Int. 2016 42 10079 10084 10.1016/j.ceramint.2016.03.113 8. Romero A.R. Toniolo N. Boccaccini A.R. Bernardo E. Glass-ceramic foams from ‘weak alkali activation’ and gel-casting of waste glass/fly ash mixtures Materials 2019 12 588 10.3390/ma12040588 9. Zeng L. Sun H. Peng T. Zheng W. Preparation of porous glass-ceramics from coal fly ash and asbestos tailings by high-temperature pore-forming Waste Manag. 2020 106 184 192 10.1016/j.wasman.2020.03.008 32229415 10. Ivanov K.S. Optimization of the structure and properties of foam-glass ceramics Mag. Civ. Eng. 2019 89 52 60 11. Kazantseva L.K. Rashchenko S.V. Optimization of porous heat-insulating ceramics manufacturing from zeolitic rocks Ceram. Int. 2016 42 19250 19256 10.1016/j.ceramint.2016.09.091 12. Yatsenko E.A. Goltsman B.M. Klimova L.V. Yatsenko L.A. Peculiarities of foam glass synthesis from natural silica-containing raw materials J. Therm. Anal. Calorim. 2020 142 119 127 10.1007/s10973-020-10015-3 13. Huo W. Yan S. Wu J.-M. Liu J. Chen Y. Qu Y. Tang X. Yang J. A novel fabrication method for glass foams with small pore size and controllable pore structure J. Am. Ceram. Soc. 2017 100 5502 5511 10.1111/jace.15089 14. Ding L. Ning W. Wang Q. Shi D. Luo L. Preparation and characterization of glass-ceramic foams from blast furnace slag and waste glass Mater. Lett. 2015 141 327 329 10.1016/j.matlet.2014.11.122 15. König J. Petersen R.R. Iversen N. Yue Y. Application of foaming agent–oxidizing agent couples to foamed-glass formation J. Non-Cryst. Solids. 2021 553 120469 10.1016/j.jnoncrysol.2020.120469 16. Xi C. Zheng F. Xu J. Yang W. Peng Y. Li Y. Li P. Zhen Q. Bashir S. Liu J.L. Preparation of glass-ceramic foams using extracted titanium tailing and glass waste as raw materials Constr. Build. Mater. 2018 190 896 909 10.1016/j.conbuildmat.2018.09.170 17. Erofeev V. Rodin A. Bochkin V. Ermakov A. The formation mechanism of the porous structure of glass ceramics from siliceous rock Mag. Civ. Eng. 2020 100 A6 18. Kruszewski Ł. Palchik V. Vapnik Y. Nowak K. Banasik K. Galuskina I. Mineralogical, geochemical, and rock mechanic characteristics of zeolite-bearing rocks of the hatrurim basin, Israel Minerals 2021 11 1062 10.3390/min11101062 19. Kazantseva L.K. Lygina T.Z. Rashchenko S.V. Tsyplakov D.S. Preparation of Sound-Insulating Lightweight Ceramics from Aluminosilicate Rocks with High CaCO3 Content J. Am. Ceram. Soc. 2015 98 2047 2051 10.1111/jace.13581 20. Erofeev V.T. Rodin A.I. Bochkin V.S. Ermakov A.A. Properties of porous glass ceramics based on siliceous rocks Mag. Civ. Eng. 2021 102 10202 21. Erofeev V.T. Rodin A.I. Kravchuk A.S. Kaznacheev S.V. Zaharova E.A. Biostable silicic rock-based glass ceramic foams Mag. Civ. Eng. 2018 84 48 56 22. Dunstan M.T. Donat F. Bork A.H. Grey C.P. Müller C.R. CO2 Capture at Medium to High Temperature Using Solid Oxide-Based Sorbents: Fundamental Aspects, Mechanistic Insights, and Recent Advances Chem. Rev. 2021 121 12681 12745 10.1021/acs.chemrev.1c00100 34351127 23. Wieczorek-ciurowa K. Paulik J. Paulik F. Influence of foreign materials upon the thermal decomposition of dolomite, calcite and magnesite part I. Influence of sodium chloride Thermochim. Acta 1980 38 157 164 10.1016/0040-6031(80)87056-0 24. Dal Pozzo A. Armutlulu A. Rekhtina M. Abdala P.M. Müller C.R. CO2 Uptake and Cyclic Stability of MgO-Based CO2 Sorbents Promoted with Alkali Metal Nitrates and Their Eutectic Mixtures ACS Appl. Energy Mater. 2019 2 1295 1307 10.1021/acsaem.8b01852 25. Qiao Y. Wang J. Zhang Y. Gao W. Harada T. Huang L. Hatton T.A. Wang Q. Alkali Nitrates Molten Salt Modified Commercial MgO for Intermediate-Temperature CO2 Capture: Optimization of the Li/Na/K Ratio Ind. Eng. Chem. Res. 2017 56 1509 1517 10.1021/acs.iecr.6b04793 26. Rekhtina M. Dal Pozzo A. Stoian D. Armutlulu A. Donat F. Blanco M.V. Wang Z.J. Willinger M.G. Fedorov A. Abdala P.M. Effect of molten sodium nitrate on the decomposition pathways of hydrated magnesium hydroxycarbonate to magnesium oxide probed by in situ total scattering Nanoscale 2020 12 16462 16473 10.1039/D0NR01760D 32478776 27. Chernykh T.N. Orlov A.A. Kramar L.Y. Trofimov B.Y. Perminov A.V. Temperature reduction during brucite-based magnesia cement production Mag. Civ. Eng. 2013 38 29 35 10.5862/MCE.38.4 28. Zhang K. Li X.S. Li W.Z. Rohatgi A. Duan Y. Singh P. Li L. King D.L. Phase Transfer-Catalyzed Fast CO2 Absorption by MgO-Based Absorbents with High Cycling Capacity Adv. Mater. Interfaces 2014 1 1400030 10.1002/admi.201400030 29. Harada T. Simeon F. Hamad E.Z. Hatton T.A. Alkali Metal Nitrate-Promoted High-Capacity MgO Adsorbents for Regenerable CO2 Capture at Moderate Temperatures Chem. Mater. 2015 27 1943 1949 10.1021/cm503295g 30. Kazantseva L.K. Rashchenko S.V. Chemical processes during energy-saving preparation of lightweight ceramics J. Am. Ceram. Soc. 2014 97 1743 1749 10.1111/jace.12980 31. Almasri K.A. Sidek H.A.A. Matori K.A. Zaid M.H.M. Effect of sintering temperature on physical, structural and optical properties of wollastonite based glass-ceramic derived from waste soda lime silica glasses Results Phys. 2017 7 2242 2247 10.1016/j.rinp.2017.04.022 32. Costa F.P.D. Morais C.R.D.S. Pinto H.C. Rodrigues A.M. Microstructure and physico-mechanical properties of Al2O3-doped sustainable glass-ceramic foams Mater. Chem. Phys. 2020 256 123612 10.1016/j.matchemphys.2020.123612 33. Türk T. Üçerler Z. Burat F. Bulut G. Kangal M.O. Extraction of potassium from feldspar by roasting with CaCl2 obtained from the acidic leaching of wollastonite-calcite ore Minerals 2021 11 1369 10.3390/min11121369 34. Abdel-Hameed S.A.M. El-kheshen A.A. Thermal and chemical properties of diopside-wollastonite glass-ceramics in the SiO2–CaO–MgO system from raw materials Ceram. Int. 2013 29 265 269 10.1016/S0272-8842(02)00114-1 35. Montoya-Quesada E. Villaquirán-Caicedo M.A. Mejía de Gutiérrez R. Muñoz-Saldaña J. Effect of ZnO content on the physical, mechanical and chemical properties of glass-ceramics in the CaO–SiO2–Al2O3 system Ceram. Int. 2020 46 4322 4328 10.1016/j.ceramint.2019.10.154

PMC009xxxxxx/PMC9099655.txt

==== Front Nutrients Nutrients nutrients Nutrients 2072-6643 MDPI 10.3390/nu14091787 nutrients-14-01787 Review Dietary Supplements for Weight Management: A Narrative Review of Safety and Metabolic Health Benefits Mah Eunice 1* https://orcid.org/0000-0002-6273-9017 Chen Oliver 2 Liska DeAnn J. 3 https://orcid.org/0000-0003-3871-8635 Blumberg Jeffrey B. 2 Arjmandi Bahram H. Academic Editor 1 Biofortis Research, Addison, IL 60101, USA 2 Friedman School of Nutrition Science and Policy, Tufts University, Boston, MA 02111, USA; nutrategic@gmail.com (O.C.); jeffrey.blumberg@tufts.edu (J.B.B.) 3 Consultant, Ridgefield, WA 98642, USA; deannliska@gmail.com * Correspondence: eunice.mah@mxns.com 24 4 2022 5 2022 14 9 178727 3 2022 19 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Dietary supplements for weight management include myriad ingredients with thermogenic, lipotropic, satiety, and other metabolic effects. Recently, the safety of this product category has been questioned. In this review, we summarize the safety evidence as well as relevant clinical findings on weight management and metabolic effects of six representative dietary supplement ingredients: caffeine, green tea extract (GTE), green coffee bean extract (GCBE), choline, glucomannan, and capsaicinoids and capsinoids. Of these, caffeine, GTE (specifically epigallocatechin gallate [EGCG]), and choline have recommended intake limits, which appear not to be exceeded when used according to manufacturers’ instructions. Serious adverse events from supplements with these ingredients are rare and typically involve unusually high intakes. As with any dietary component, the potential for gastrointestinal intolerance, as well as possible interactions with concomitant medications/supplements exist, and the health status of the consumer should be considered when consuming these components. Most of the ingredients reviewed also improved markers of metabolic health, such as glucose, lipids, and blood pressure, although the data are limited for some. In summary, weight management supplements containing caffeine, GTE, GCBE, choline, glucomannan, and capsaicinoids and capsinoids are generally safe when taken as directed and demonstrate metabolic health benefits for overweight and obese people. cardiometabolic diabetes energy body weight BMI liver blood pressure ==== Body pmc1. Introduction Obesity is a global epidemic that is associated with a higher risk of a multiplicity of devastating diseases, including type 2 diabetes mellitus (T2DM) and cardiovascular disease (CVD), among others [1]. The etiology of obesity is complicated, multifactorial, and involves the dysregulation of the body’s energy balance system, mediated by a complex interplay between neural, hormonal, and metabolic pathways, sedentary lifestyle, genetic predisposition, and excess calorie consumption [1,2]. Modest weight loss is considered beneficial to people with overweight/obesity due to its impact on disease risk reduction [3,4,5]. However, successful weight management, which includes not only weight loss but also weight loss maintenance (i.e., limiting weight regain), is challenging [5]. Calorie restriction and increased physical activity are the cornerstones of traditional weight management programs, but often do not lead to significant, sustained weight loss alone [6,7]. Dietary supplements, particularly those with thermogenic, lipotropic, or satiety properties, are used by many consumers to support dietary and lifestyle programs for weight management. Caffeine and green tea extract (GTE) are commonly used supplements with purported thermogenic properties. Choline and glucomannan represent supplements with expected lipotropic and satiety effects, respectively. Green coffee bean extract (GCBE) and capsaicinoids and capsinoids are newer to the market and becoming popular for weight management support. In the U.S., dietary supplements are regulated by the Food and Drug Administration (FDA); however, despite the existence of federal regulatory oversight, the safety of dietary supplement products targeted for weight management support has been questioned [8]. The FDA defines dietary supplements as products taken orally that contain dietary ingredient(s) intended to supplement the diet, but not intended to treat, diagnose, mitigate, prevent, or cure diseases [9]. Specifically, these products may make structure and/or function claims associated with a disclaimer stating the claims have not been reviewed by the FDA, although manufacturers are required to maintain supporting evidence [9]. Dietary supplements are also required by the FDA to be manufactured under Good Manufacturing Practices, with mandatory reporting of serious adverse events and FDA notification prior to marketing a new dietary ingredient (NDI), and manufacturers are prohibited from marketing supplements with unsafe ingredients [10]. Given the continued high interest in strategies for successful weight management through diet and lifestyle approaches, as well as recent questions on the safety of products in the dietary supplement weight management category, this review has identified six commonly used components that represent the breadth of the category. The safety evidence for each ingredient is reviewed, and the status of the clinical evidence on the weight management and relevant clinical health benefits are summarized. Given the large body of research for some of these ingredients, this review includes the highest quality evidence, focusing primarily on reports from authoritative sources and published systematic reviews and meta-analyses. 2. Evidence Standards Over the past several decades, much work has been directed on basing health and nutrition policy and recommendations on scientific evidence over expert opinions. Reports from authoritative bodies and government agencies that provide transparency on the scientific evidence review process are among the highest form of evidence for regulatory and policy applications. Within the biomedical sciences, systematic reviews and meta-analyses following established standards, such as PRISMA and MOOSE, and those published in peer-reviewed journals are critical resources [11,12]. The quality of the research studies included in these reviews is an important factor as well, and randomized clinical trials are the preferred evidence for establishing cause-and-effect relationships [13,14]. Other evidence, such as observational studies, can help define questions for further investigation, but because these studies are not randomized and often confounded, they may not be considered conclusive [13]. In addition, it is important that the science is assessed not only for quality but also for relevance to the key questions. For example, key aspects that authoritative bodies consider when selecting and assessing whether the evidence for decision-making is relevant generally follow the population, intervention, comparator, and outcome (PICO) approach [13,15]. In particular, the intervention tested and comparator control should be relevant and practical. Specifically, the form of delivery (supplement vs. food) and intake level (e.g., testing in the range of a typical product) are important to consider when translating science for decision-making [15]. For this review, authoritative sources were searched for evidence, including the U.S. FDA, the European Food Safety Authority (EFSA), Health Canada, the National Academies of Sciences and Medicine (NASEM, formerly the Institute of Medicine (IOM)), the National Institutes of Health’s Office of Dietary Supplements (ODS), and the most recent U.S. Dietary Guidelines for Americans (DGA) 2020–2025. Discussion of individual clinical trial evidence is included when more recent studies were available and/or no authoritative sources were found. In addition, to provide context around the recommended amounts of the ingredients in products provided to consumers, the Mintel Global New Products Database [16], which monitors worldwide product innovation and new product activity in consumer-packaged goods markets, with data going back to 1996, was searched for products with weight management claims (e.g., slimming, weight and muscle gain) having one or more of the ingredients of interest. 3. Caffeine Caffeine (1,3,7-trimethylxanthine) is one of the most frequently consumed dietary bioactive substances across the globe, is known as a central nervous system stimulant, and is also proposed to increase thermogenesis and fat oxidation [17,18]. Most caffeine intake is from beverage sources, including brewed coffee (56–100 mg caffeine/100 mL), instant coffee and tea (20–73 mg caffeine/100 mL), and cola soft drinks (9–19 mg caffeine/100 mL) [18]. Caffeine is generally recognized as safe (GRAS) for use in cola-type beverages by the FDA [19]. As a supplement, the ODS identifies sources of added caffeine as guarana, kola nut, yerba maté, and other herbs [17]. A search of the Mintel GNPD found 579 caffeine-containing supplements with weight management claims whereby 164 listed the caffeine content. Of these, 93% recommended intake levels of ≤400 mg/day, with the remaining 7% recommending 400 to 420 mg caffeine/day. Various studies on caffeine intake have been conducted, generally addressing beverage sources. In a summary of data from 18 nationally representative sources, the average caffeine consumption was found to be 23.7 mg/day, 36.6 mg/day, and 83.2 mg/day for infants, children, and adolescents, respectively, compared to 122.1–225.5 mg/day for adults [20]. These authors reported energy drinks contributed little to total caffeine intake across all groups. Specific to the U.S., examination of the 2001–2010 NHANES data showed an average daily intake of 186 mg (90th percentile, 436 mg/day) by adults from foods and beverages, including energy drinks [21]. In another study using data from the 2013–2016 NHANES, those aged 1–80 years who consume caffeine (>1 mg/d) had a mean intake of 195 mg/day, with 14% of those 30–80 years old and only 4.1% of those 13–29 years old having intakes >400 mg caffeine/day [22]. Overall, the consumption of caffeine has remained relatively stable over the past >15 years, with coffee consumers, regardless of age, being the most likely to have higher intakes. Consumption of caffeine by children is generally reported below 2.5 mg/kg/day, with the U.S. having the highest intake (average of 14 mg/day or 0.82 mg/kg/day) [20]. 3.1. Caffeine Safety The safety of caffeine has been the focus of many federal agencies and scientific and non-governmental organizations over the past several decades. In 2003, Health Canada conducted a comprehensive review concluding that an intake of ≤400 mg caffeine/day was not associated with adverse effects in healthy adults [18]. Additionally, Health Canada concluded that the consumption of ≤300 mg/day for pregnant or lactating women as well as those planning to become pregnant, and 2.5 mg/kg/day for children is not associated with adverse effects. Similarly, the EFSA has also indicated that ≤400 mg caffeine/day does not lead to safety concerns for non-pregnant adults, but identified a limit of 200 mg/day for pregnant women [23]. In 2017, an updated extensive systematic review also found consumption of ≤400 mg/day in healthy adults is not associated with overt, adverse cardiovascular, behavioral, reproductive, and developmental effects, acute effects, or bone status [24]. Additionally, the review found that the consumption of ≤300 mg caffeine/day in healthy pregnant women is generally not associated with adverse reproductive and developmental effects [24]. Limited new data were identified for children (3–12 years old) and adolescents (12–19 years old), and the authors indicated that “there is no evidence to suggest a need for a change from the recommendation of 2.5 mg/kg caffeine/day” [24]. The DGA 2020–2025 addressed caffeine intake and, noting that the FDA has identified ≤400 mg/day of caffeine as the amount not generally associated with dangerous, negative effects, and indicated that caffeine can be consumed ≤400 mg/day [19]. The DGA also reviewed the evidence of caffeine consumption by pregnant and lactating women. Small amounts of caffeine can pass from the mother to the infant through breast milk, but a review of the evidence indicated that consumption of ≤300 mg/day (equivalent to about 2–3 cups of coffee) by the mother does not adversely affect the infant [19]. No safe limits have been established for children aged 2 years or younger. When used according to the manufacturer’s instructions, most weight management supplements provide caffeine within the recommended amount of ≤400 mg caffeine/day. However, concern has been raised over the consumption of caffeine at higher intake levels. Acute intake of caffeine > 500 mg/day potentially results in various untoward events such as headache, jitteriness, agitation, anxiety, dizziness, and tinnitus [17,24]. In a review of safety data, the FDA has noted that caffeine metabolism is slowed after consumption of >500 mg, with adverse effects, such as tachycardia, ventricular arrhythmia, and seizures, at consumptions > 1200 mg [25]. Responses are variable and likely depend on individual sensitivity to caffeine, the existence of co-morbidities, and intake of concomitant medications or supplements [24]. Fatality due to caffeine poisoning is rare, and caffeine’s lethal dose is unclear. Reviews by Health Canada, ILSI, and FDA have concluded that there is a potential for death following acute exposures of ~10 g of caffeine for adults and adolescents [18,24,25]. However, due primarily to uncertainty in the estimates of exposure and the high risk of bias (e.g., use of case reports), there is low confidence in this evidence base. The FDA has recognized that pure or highly concentrated powdered and liquid caffeine has the potential to deliver higher amounts and thus has moved to restrict the sale of these highly concentrated items [25]. Consuming 10 g of caffeine from supplements would require ~25–26 servings of the most highly concentrated supplements, and not surprisingly, most case reports of caffeine-induced fatalities have been associated with suicide or abuse [25,26]. Very few reports of fatal caffeine poisoning in children are published [26]. 3.2. Caffeine and Weight Management Caffeine has been used as an approach in weight management because of its ability to stimulate both noradrenaline and dopamine secretions, which, in turn, may decrease BW and body fat (BF), as well as increase thermogenesis in brown adipose tissue via an unknown mechanism(s) [27]. A meta-analysis of 13 studies providing 60–4000 mg caffeine/day for 4–36 weeks showed that caffeine intake led to a reduction in BW, BF, and body mass index (BMI), and that this effect is dose-dependent [28]. However, all but three of the 13 included studies provided caffeine with other substances with potential weight loss properties. Of the three with only caffeine, one study in normal weight and one in overweight/obese subjects reported modest weight loss, whereas another in overweight/obese subjects did not. Taken together, there may be a modest effect of caffeine on weight management, but more clinical evidence with caffeine alone is needed for confirmation. 3.3. Caffeine and Metabolic Health Observational evidence suggests that caffeine may be protective against T2DM [29], [30], which has led to further investigations in this area, but results from clinical studies have been mixed. A meta-analysis of seven randomized clinical trials with primarily male normal weight participants consuming caffeine from 3–6 mg/kg [31], and a trial with healthy men and women [32] found significantly decreased insulin sensitivity indexes compared with a placebo, suggesting caffeine might possess a hyperglycemic activity by acutely impairing insulin action in adipose and muscle tissue. The hyperglycemic impact of caffeine may also extend to adolescents [33]; however, more research is needed on the long-term effect of caffeine consumption on glycemic status [29]. 3.4. Caffeine Summary Caffeine has been extensively researched and various reviews on caffeine safety, including from regulatory and authoritative sources, indicate that consumption of ≤400 mg/day for adults is safe with minimal risk of adverse events. The majority of supplements providing caffeine are within the recommended amount of ≤400 mg caffeine/day. Recommendations also indicate that pregnant and lactating women can safely consume ≤300 mg caffeine/day. Current evidence supports a beneficial effect of acute caffeine consumption on cognition, particularly on attention [34], as well as a benefit on exercise performance [35]. The evidence on the effect of caffeine on weight management is mixed, with some data suggesting caffeine might have benefits in supporting healthy blood glucose levels. Overall, as noted by the U.S. FDA “When formulated and marketed appropriately, caffeine can be an ingredient in a dietary supplement that does not present a significant or unreasonable risk of illness or injury” [25]. 4. Green Tea Extract (GTE) Green tea (Camellia sinensis) is one of the most popular beverages globally. Due to the association of its consumption with a wide range of health benefits, supplements formulated with GTE are of interest for a wide range of health benefits including as an antioxidant [36,37]. Extracted from green tea, GTE powder delivers polyphenols, which include flavonoids (45–90%) and caffeine (0.4–10%) at concentrations that vary for different extracts [38]. In all GTE powders, the major flavonoids are catechins, with the dominant four being epicatechin, epigallocatechin, epicatechin-3-gallate, and epigallocatechin-3-gallate (EGCG) [38]. A review of the Mintel GNPD indicates products with GTE or EGCG, when consumed under the manufacturers’ recommended servings ranged from 100–1200 mg/day and 27–350 mg/day, respectively. Some of these products also deliver 20–400 mg/day of caffeine. In most populations, a major source of dietary catechins is green tea infusions, which deliver an average of 0.7 mg EGCG/g of brewed green tea (~178 mg/8.5 fl oz) [39]. EFSA assessed the amount of EGCG delivered in green tea infusions and found a wide range, from 2.3 to 203 mg/100 g (~3.3 fl oz) infusion [39]. In the U.S., adults consume an average of 219 mg/day of flavonoids (~34% catechins), with tea drinkers consuming 610 mg/day flavonoids (data from NHANES 2011–2016) [40]. The EFSA panel has estimated an average EGCG intake from green tea of 90–300 mg/day, with the largest tea users consuming up to 856 mg EGCG/day [39]. EFSA also found average intakes of ECGC of 11.66–87.96 mg/day among children (3–9 years old) and 68.16–232.03 mg/day among adolescents (10–17 years old) based on food consumption surveys from various countries in Europe [39]. 4.1. Green Tea Extract (GTE) Safety The consumption of green tea infusions has traditionally been considered safe, even at levels of >5 cups/day. However, concerns have been raised regarding a link between GTE and liver function; therefore, extensive safety reviews have been conducted. In a 2017 review, Health Canada concluded that there might be a link between the use of GTE and the risk of hepatotoxicity and thus requires a cautionary statement on GTE products and limits their sale to adults [41]. EFSA published a review in 2017 and concluded that there is evidence from interventional clinical studies that intake of doses ≥800 mg EGCG/day induced a significant increase in serum transaminases (a marker of liver injury) in treated subjects compared to control [39]. Hu et al. conducted a comprehensive systematic review on the safety of various green tea preparations and proposed a safe intake limit of 338 mg EGCG/day for healthy adults (in a fed or fasted condition) delivered in solid dosage form based on no-observed-adverse-effect level (NOAEL) of 676 mg EGCG/day established from clinical evidence [42]. Yates et al. proposed a tolerable upper intake level (UL) of 300 mg/day of purified EGCG, based on human data in healthy adults in a fed state, and an acceptable daily intake (ADI) of 4.6 mg/kg/day, derived from animal toxicity data [43]. Dekant et al. also proposed a UL of 300 mg EGCG/day for food supplements based on clinical evidence indicating that liver effects were not observed after intakes ≤600 mg EGCG/person/day and animal toxicity studies [44]. In 2020, the USP Dietary Supplement Information Expert Committee reviewed case reports, animal data, and human clinical data and concluded the risk of hepatotoxicity due to GTE can be serious, but variable, stating that GTE rarely leads to severe hepatotoxicity in humans and its manifestation is dependent on the concentration of catechins, the bolus dose, fasting condition as well as genetic susceptibility, idiosyncrasy, and/or underlying liver health [45]. Clinical studies with beverages fortified with GTE at intakes up to 498.6 mg EGCG/person/day and durations of up to 1 year (median 12 weeks) have been published and do not show evidence of hepatotoxicity [44]. Of 15 clinical studies on GTE-containing beverages (100–704 mg EGCG/day; 1–24 weeks) reporting adverse events identified by Hu et al. [42], only five reported GI adverse events (i.e., abdominal pain and discomfort, diarrhea, and dyspepsia/indigestion) following consumption of green tea beverage and none reported liver-related adverse events. Additionally, safety recommendations for GTE are only for adults due to the limited data on children and adolescents. Limited data are available on children. In their systematic review, Hu et al. only found one study in children that assessed adverse events [42]. A clinical trial in obese Japanese children (6–16 years old; BMI > 28 kg/m2) consuming 576 mg catechins (30 mg EGCG)/day for 24 weeks reported no safety concerns [46]. 4.2. Green Tea Extract (GTE) and Weight Management The weight management potential of green tea has been postulated to be associated with multiple mechanisms of action, e.g., appetite control, inhibition of adipogenesis, increase in energy expenditure, and modulation of substrate oxidation [47,48,49]. The efficacy of green tea or GTE has been extensively researched, and for the most part, the results are favorable. A meta-analysis was published with 11 studies in T2DM patients, in which 10 used GTE and one used brewed green tea at amounts ranging from 400–10,000 mg/day and durations from 8–12 weeks [50]. Significant effects were reported (pooled weighted mean difference compared to control) for BW (−0.40 kg; 95% CI, −0.64 to −0.16; p = 0.001), BMI (−0.05; 95% CI, −0.10 to −0.00; p = 0.046), and BF (−0.56%; 95% CI, −0.73 to +0.38; p < 0.001), but not WC [50]. The results of another meta-analysis of 22 studies, which included many of the same studies of T2DM subjects, as well as additional results from studies in subjects described as overweight/obese, menopausal, and other conditions, also showed that GTE significantly decreased BW, BMI, and waist circumference (WC) [51]. Since caffeine in green tea or GTE may contribute to weight loss, a meta-analysis was conducted on 15 studies comparing green tea/GTE from 583–714 mg/day with caffeine-matched controls. The combination of green tea/GTE and caffeine led to an additional decrease in BW, BMI, and WC compared to caffeine alone over a median of 12 weeks [52]. Green tea/GTE interventions do not appear to affect obesity hormones, leptin, and adiponectin, probably due to the modest reduction in BF [53,54]. Moreover, a meta-analysis of green tea or GTE supplementation of four clinical studies with subjects having non-alcoholic fatty liver (NAFLD) reported a decrease in BMI of −2.08 kg/m2 (95% CI, −2.81 to −1.36) [55]. Thus, the current evidence supports the addition of GTE to lifestyle regimens for weight management. 4.3. Green Tea Extract (GTE) and Metabolic Health GTE constituents can be beneficial for cardiometabolic health. The reduction in fasting blood glucose (FBG) following green tea/GTE has been supported by several meta-analyses, although the effective dose and supplementation duration are unclear [56,57,58,59,60,61]. A meta-analysis of 22 studies showed that green tea catechins with or without caffeine for >12 weeks significantly reduced FBG but did not affect insulin, HbA1c, and homeostasis model assessment-estimated insulin resistance (HOMA-IR) independent of dose [56]. Another meta-analysis of 17 studies showed that green tea/GTE significantly reduced FBG and HbA1c, especially at ≥457 mg/day, regardless of supplementation duration [57]. A more recent meta-analysis with 27 studies showed that green tea/GTE with catechin doses >500 mg/day for ≤12 weeks significantly lowered FBG [60]. However, the benefits on FBG appear not to extend to people at increased risk for diabetes or with the disease based on the results of three other meta-analysis studies [58,59,61]. These meta-analyses included subjects who were overweight/obese, normo-weight, and at risk due to other metabolic factors. Mechanisms of action of green tea catechins on glucose regulation include reducing intestinal carbohydrate digestion and absorption, inhibiting hepatic gluconeogenesis, and sensitizing insulin action [60]. Meta-analyses have shown that green tea/GTE reduces total cholesterol (TC) and LDL cholesterol (LDL-C) but not HDL cholesterol (HDL-C) in both normo-weight and overweight/obese subjects [62,63,64,65]. The green tea/GTE amount included in these studies ranged from 80–3000 mg catechins/day, and the reduction in TC and LDL-C appears to be independent of dose or duration. In contrast, one meta-analysis showed that GTE improved blood triglycerides (TG) but not blood cholesterol in patients with T2DM when provided at ≥800 mg/day for ≥8 weeks [66]. Additionally, the positive effect on TC, LDL-C, and TG was noted in the meta-analysis with patients with NAFLD, as well as decreases in the enzymes ALT and AST, which are indicative of liver health [55]. Obesity is a risk factor for NAFLD, and these findings suggest green tea/GTE could be an option for NAFLD patients. The beneficial effect of green tea/GTE on blood pressure (BP) is supported by several meta-analyses [67]. A recent meta-analysis with 24 studies providing 208–1344 mg catechins/day showed that green tea/GTE reduced systolic BP (SBP) and diastolic BP (DBP) independent of caffeine content [68]. A meta-analysis of 13 studies with diverse participant demographics, including normo-weight and overweight/obese, showed decreases in SBP and DBP following green tea catechin supplementation, whereby a catechin amount of <500 mg/day was more efficacious. Additionally, the greatest reduction in both measures was noted in participants with a baseline SBP > 130 mmHg and BMI < 30 kg/m2 and in studies administering GTE [69]. However, an earlier meta-analysis showed green tea/GTE remained effective in improving SBP and DBP in overweight and obese people but to a smaller extent [70]. The decrease in BP may be due to a vasodilation effect following an increase in nitric oxide. Thus, the overall evidence supports the benefits of green tea/GTE on BP in the general population and overweight/obese individuals at increased risk for CVD. 4.4. Green Tea Extract (GTE) Summary Current evidence suggests that GTE may be beneficial for weight management, glucose regulation, and reducing TC and LDL-C as well as BP in people who are overweight or obese. Undesirable effects of GTE consumption reported in clinical trials are largely GI-related tolerance issues; however, hepatotoxicity concerns have led to a proposed UL of 300 mg/day of purified EGCG. For the most part, products marketed for weight management provide EGCG at an amount at or below the proposed EGCG UL for adults when used according to the manufacturers’ instructions. 5. Green Coffee Bean Extract (GCBE) Green coffee bean extracts (GCBE) are produced from water or alcohol extraction of green coffee beans [71], and together with yerba mate (Ilex paraguariensis), are considered to be the richest source of chlorogenic acid (CGA) in nature, containing about 6–12% w/w total CGA [72]. Like other coffee products, GCBE may or may not contain caffeine. In addition, although GCBE is a rich source of CGA, roasted coffee beans, and thus coffee beverages, are not because CGA is markedly degraded during the roasting process [71,73]. CGA has diverse bioactions, including antioxidant, anti-inflammatory, glucose and lipid metabolism regulatory, and cardiovascular protective activities, which are considered the principal mechanisms underlying GCBE effects [72]. The daily intake of CGA among coffee drinkers has been estimated to be 100–300 mg/day for medium drinkers and ≤600 mg/day for heavy drinkers [74]. Per capita consumption of CGA has been estimated to be 120–594 mg/day [74]. A search of the Mintel GNPD revealed many weight management supplements containing GCBE ranging between 50–3996 mg/day. Of those that also provided CGA content, these ranged from 100–1170 mg/day. Several GCBE-containing products also include added caffeine ranging from 40–420 mg/day, whereas those that provided caffeine as a percentage of GCBE had caffeine concentrations ranging from 8–20 mg/day. 5.1. Green Coffee Bean Extract (GCBE) Safety No authoritative reports were identified specifically for GCBE; however, humans have a high metabolic capacity for CGA with no evidence of saturation of metabolic pathways for doses ≤2 g/day [75]. Animal studies suggest CGA is of low toxicity with an LD50 of 100 mg/kg (7 g for a 70 kg adult) [76,77]. Of the studies included in four recent meta-analyses on GCBE and cardiometabolic outcomes [78,79,80,81], only seven assessed adverse events, and of these, two reported some undesirable side effects during the GCBE or CGA intervention. One study reported stomach irritation and dizziness following 800 mg GCBE (372 mg CGA)/day for 8 weeks [81], and the other reported nausea and headache following 200 mg CGA/day for 6 weeks [82]. Both used decaffeinated supplements. Those that reported no undesirable side effects provided GCBE at 100–1000 mg/day or CGA at 54–500 mg/day for 4–24 weeks; all provided decaffeinated supplements, except for one study that provided CGA with or without 200 mg caffeine [83]. Observations from these studies suggest that CGA side effects are individual-dependent. Considering the low occurrence of undesirable side effects reported by clinical studies testing GCBE or CGA along with the low toxicity suggested by animal studies, it is unlikely that the consumption of GCBE providing CGA ≤ 1 g (and possibly higher) would result in undesirable side effects. 5.2. Green Coffee Bean Extract (GCBE) and Weight Management Caffeine and CGA in GCBE can stimulate fatty acid oxidation and inhibit lipogenesis, which, in turn, decreases fat accumulation in the body [73]. A meta-analysis of 15 randomized clinical trials showed that GCBE in amounts ranging from 90–6000 mg/day containing 30–1200 mg/day CGA for 1–12 weeks significantly reduced BW, BMI, and WC compared to the control group, but did not affect BF and waist-to-hip ratio [84]. Additionally, the analysis did not find a dose–response relationship between CGA dosage and anthropometric measures. GCBE appears effective to support weight management but clinical studies with interventions > 12 weeks are warranted to substantiate the efficacy for weight loss and maintenance. 5.3. Green Coffee Bean Extract (GCBE) and Metabolic Health Several meta-analyses have been performed to evaluate the effect of GCBE products on cardiometabolic risk factors, such as impaired glucose regulation, dyslipidemia, and hypertension. Two meta-analyses of randomized studies showed that GCBE or CGA between 200–2000 mg/day for 2–24 weeks significantly reduced FBG and insulin, and this was not influenced by dose level [78,79]. However, the results of another meta-analysis of 10 studies found a significant decrease in FBG but not in insulin, and that the decrease in FBG was only significant for GCBE ≥400 mg/day [85]. Regarding lipid profile, a meta-analysis of 13 randomized studies showed that GCBE or CGA between 200 and 2000 mg/day for 2–24 weeks significantly reduced fasting blood TC but not TG, LDL-C, or HDL-C [78]. Finally, the beneficial effect of GCBE on SBP and DBP was demonstrated in a meta-analysis of nine studies in people with hypertension, whereby the effect size in SBP was larger with GCBE doses ≥400 mg/day [86]. In that meta-analysis, the hypertensive subjects were diverse and included overweight/obese and those with metabolic syndrome. These results appear most supportive of GCBE in blood glucose while the effect of GCBE on BP and lipids is promising but requires further investigation. 5.4. Green Coffee Bean Extract (GCBE) Summary Considering the low occurrence of undesirable side effects reported in clinical studies providing GCBE or CGA along with the low toxicity suggested by animal studies, it is unlikely that consumption of GCBE or CGA through weight management supplements is harmful. Limited clinical evidence supports a potential beneficial effect of GCBE for short-term (<12 weeks) weight loss and management of blood glucose and BP but not that of lipids [87], and CGA exhibits anti-diabetic and anti-lipidemic properties [73]. 6. Choline Choline is an essential water-soluble micronutrient that exerts diverse functions in cellular maintenance and growth throughout one’s lifetime. Specifically, choline is involved in metabolic pathways for the synthesis of acetylcholine, betaine, phospholipids, and trimethylamine, which, in turn, play roles in lipid transport, membrane synthesis, neurotransmission, and one-carbon metabolism [88,89]. Choline is particularly important for fetal brain development, and low choline status or deficiency is associated with negative health outcomes. The NASEM Dietary Reference Intake Report on choline has defined adequate intakes (AI) for the U.S. population that range between 425–550 mg/day for adults, 375–550 mg/day for adolescents aged 14–18 years, 250 mg/day for children age 4–8 years, 200 mg/day for young children (1–3 years old), and 125–150 mg/day for infants up to 1 year old [90]. Similarly, the EFSA set the recommended AI for choline at 400–550 mg/day for all adults and adolescents 14 years old and above, 375 for adolescents 9–13 years old, and 150–250 mg/day for infants and young children [91]. Average dietary choline intake in several countries has been reported to be below recommendations for most populations and age/sex groups, ranging from 230–468 mg/day for adults, and mostly from animal-based food sources [89]. For example, in the U.S., the average choline intake of men and women is 421 mg/day and 279 mg/day, respectively [88]. The DGA 2020–2025 has addressed choline nutriture and noted that choline intake needs to be higher during pregnancy and lactation for replenishment of maternal stores and support for fetal development, particularly of the brain and spinal cord [19]. Given that most women do not meet recommended intakes of choline during pregnancy and lactation, particularly women following a vegetarian or vegan dietary pattern, and that most prenatal vitamins do not contain choline, the DGA also states that supplementation may be necessary [19]. In dietary supplements, choline is delivered in different forms, including phosphatidylcholine, CDP-choline (citicoline), L-alpha-glycerylphosphorylcholine (alpha-GPC), and choline salts (e.g., choline bitartrate, choline chloride). A search of the Mintel GNPD database revealed some supplements with weight management claims that contain choline, whereby a majority were in the form of choline bitartrate, with a handful as citrate salts, phosphatidylcholine, and alpha-GPC. Most of these products appear to include choline as a vitamin for nutrition purposes, but some claim that the added choline helps with fat metabolism and increases energy. Among the products that listed the amount of choline per serving, most provide <300 mg/day, with the highest providing <993 mg/day. Meanwhile, most top-selling choline supplements are marketed for cognition and provide between 200–500 mg choline/day, with the highest being 1000 mg/day [92]. Thus, compared to other types of choline supplements, those marketed for weight management are not more likely to substantially contribute to excess dietary choline. 6.1. Choline Safety The 1998 Dietary Reference Intake report established upper levels of choline based on their review of safety data and the identification of the LOAEL as 3.5 g/day in adults, 3 g/day for adolescents 14–18 years old, 2 g/day for children aged 9–13 years, and 1 g/day for children aged 1–8 years [90]. The daily oral administration of high choline levels, such as 10 g choline chloride (7.5 g choline), has produced a slight hypotensive effect in patients with Alzheimer’s disease, and mild reports of transient Parkinsonian signs (bradykinesia, tremor, and rigidity) were observed at 12.7 g/day of choline chloride in people with tardive dyskinesia [90]. Excess choline intake (e.g., 10–16 g/day choline) has also been associated with fishy body odor, vomiting, salivation, sweating, and GI effects in patients with tardive dyskinesia and cerebellar ataxia, as well as generation of the metabolite trimethylamine-N-oxide [90,91]. However, as noted by EFSA, “these are indirect adverse effects of choline, depending on a ‘high’ dietary amount and a specific gut microbiome” [91]. Given that choline is a required micronutrient, and most people consume lower than the recommended amount, a main focus in the literature has been on inadequate intakes, not excess consumption. 6.2. Choline and Weight Management Clinical evidence on choline and weight loss is scarce, with no authoritative reports or systematic reviews specifically on choline and weight management identified. Based on its known lipotropic activity, some clinical studies have explored its effect on fat deposition and weight loss. In clinical studies, one report with 22 young female athletes undergoing normal athletic training found 1-week supplementation of 2 g choline/day led to a larger rapid weight loss as compared to the placebo [93]. Reduced BF accounted for the majority of the weight loss, mainly due to the effect of choline on fat metabolism. However, another study testing the effect of 6-week supplementation of 500 or 2000 mg/day citicoline (CDP-choline, a dietetic source of choline) in 16 overweight adults did not affect BW despite a lower appetite rating at the high dose compared to the baseline [94]. An observational study of 3054 adults in Newfoundland, Canada showed that dietary choline intake was associated inversely with BW, BMI, WC, BF, and positively with lean mass in both women and men [95], but an association was not found in an observational study with 788 6-year-old Iranian children [96]. 6.3. Choline and Metabolic Health Some evidence is available for the impact of choline on obesity-related conditions. For example, choline plays a key role in fat metabolism in the liver via phosphatidylcholine, a product of choline metabolism, which is essential for the assembly/secretion of very low-density lipoproteins [89,97]. Mechanisms including abnormal phospholipid synthesis, defects in lipoprotein secretion, and oxidative damage can exacerbate liver damage and contribute to liver disorders such as NAFLD. A link between choline deficiency and hepatic lipid disposition has been recognized for over 50 years [98]. The significance of choline in NAFLD is manifested in the occurrence of hepatic steatosis with choline deficiency and the reversal of lipid infiltration with choline replacement in patients receiving total parenteral nutrition [99]. Additionally, decreased choline intake was found to be significantly associated with an increased incidence of hepatic fibrosis in postmenopausal women with NAFLD [100]. Moreover, higher dietary choline intake was associated with a 28% lower risk of NAFLD only in normal-weight Chinese women but not in overweight/obese women or men [101]. Although clinical evidence on the efficacy of choline supplementation in protecting against the NAFLD development and progression is lacking [102,103], the current evidence suggests the potential of choline for nutritional support and management of NAFLD. Choline serves as a dietary precursor for the gut microbial-generated trimethylamine (TMA), which is oxidized by the hepatic enzyme flavin-containing monooxygenase 3 (FMO3) to form trimethylamine-N-oxide (TMAO) [104,105]. High levels of TMAO have been linked to detrimental vascular and inflammatory effects in animal and cell culture studies. In humans, increased levels of TMAO have been associated with greater CVD risk, although it remains unclear whether this relationship is a non-pathogenic indirect marker, or if TMAO has a direct role in disease progression [104]. Further, many of the choline-rich foods that may contribute to TMAO are considered cardioprotective, and the contribution of their consumption to clinically relevant levels of TMAO is unknown [104,105]. Therefore, the role of TMAO, and choline as a precursor, remains unclear. Choline supplementation has been explored in randomized, double-blind, placebo-controlled trials on lipid profiles with no effect seen with 400 mg/day choline in T2DM subjects [106] or with 1 g choline/day in 42 healthy postmenopausal women [107]. The supplementation of betaine, which is a metabolite of choline produced in the human body, at 1.5, 3, and 6 g/day for 6 weeks, significantly increased TC, LDL-C, and TG compared to placebo [108]. In the same study, 2-week supplementation with 2.6 g/day of choline (provided as phosphatidylcholine) increased serum TG by (8%) but did not alter TC, LDL-C, or HDL-C. Taken together, although choline is involved in fat metabolism, its effect following supplementation on lipid profile in humans is equivocal. 6.4. Choline Summary Choline is an essential micronutrient with lipotropic activity that helps catalyze fat metabolism and prevent fat disposition in the liver. The clinical evidence on the efficacy of choline supplementation on weight management is limited and inconsistent. Given that choline is a required micronutrient, and most people consume less than the recommended intake, the main focus in the literature has been on inadequate intakes, not excess consumption. 7. Glucomannan Glucomannan is a high molecular-weight polysaccharide mainly composed of d-mannose and d-glucose linked by β-1,4 glycosidic bonds with side chains [109]. Although it can be present in other plants, such as lily and orchid, glucomannan for human consumption is commonly derived from the tuber or root of Amorphophallus konjac or elephant yam [110]. Glucomannan is a non-digestible carbohydrate that has been approved for labeling as a dietary fiber by the FDA based on its ability to attenuate cholesterol levels [111]. As a fiber, glucomannan is not digested in the upper GI and thus is available for fermentation by gut microbiota. Like other fibers, glucomannan has been promoted for effects on weight management and associated benefits, such as glucose and cholesterol-lowering, laxative, prebiotic, and anti-inflammatory activities. It is a hygroscopic fiber, however, which puts it in a subcategory of fibers, such as guar gum, which are viscous and able to form a large volume of mucilage after absorbing water in the upper GI tract. The mucilage can then affect satiety as well as nutrient digestion, absorption, and metabolism in the GI tract [109]. Overall, dietary fiber has been identified as a nutrient of public health concern in all age/sex groups of the U.S. population, with more than 90% of females and 97% of males not meeting recommended levels of 21–38 g/day for most adults and 19–38 g/day for children aged 1–18 years [19,112]. A search of the Mintel GNPD indicated the recommended intake of glucomannan for weight management products containing glucomannan ranged from 10–3330 mg/day. 7.1. Glucomannan Safety Following their review of animal and human studies of konjac food supplements, the EFSA concluded there was no safety concern at <3 g/day konjac intake for the general population and noted that abdominal discomfort, including diarrhea or constipation, may occur after this daily dose in adults [113]. This advice is similar to that for other fibers, which can cause untoward GI side effects, particularly if large amounts are consumed without a stepped-up dosing regimen to allow for acclimation. Usually, these GI symptoms, if they occur, are mild to moderate and transient and not a safety issue. For example, a meta-analysis of randomized weight loss clinical trials in adults found GI-related symptoms in six of eight studies that assessed adverse events when konjac glucomannan was provided at 1.5–10 g for 3–12 weeks [114]. Another meta-analysis of randomized studies on the effect of 3 g for 12 weeks and 3.99 g for 8 weeks konjac glucomannan supplementation on BW also reported GI-related symptoms in two of the six studies that assessed adverse events [115]. Similarly, GI-related adverse events were also reported by some studies in children; however, in two studies in obese children, no adverse events were reported with 2–3 g/day glucomannan in capsules [116,117]. Another study in children (6–17 years old) provided 3 g/day noted that both the active and placebo (maltodextrin) groups experienced a similar number of adverse events that were mostly GI related [118]. Glucomannan aids in weight loss due to its hygroscopic property, whereby it swells rapidly in the stomach to produce a feeling of satiety and fullness. Water-soluble gums, such as konjac glucomannan, psyllium, and guar gum, may cause esophageal obstructions if the dry powder is consumed with inadequate water, thus allowing expansion to occur in the esophagus. The risk of esophageal obstructions for glucomannan products is dependent on the amount of glucomannan and how the products are manufactured, as well as individual consumer characteristics (e.g., ability to swallow, esophageal anatomy, compliance to consumption instructions, etc.) [119]. The FDA has reviewed this category of ingredient related to safety and requires that over-the-counter products in a dry or incompletely hydrated form must carry a warning indicating the product should be taken with adequate water and should not be consumed by individuals who have difficulty swallowing [120]. Specific to dietary supplements, the ODS has stated, “Significant safety concerns reported for tablet forms, which might cause esophageal obstructions, but few safety concerns with up to 15.1 g/day of other forms for several weeks” [17]. 7.2. Glucomannan and Weight Management In 2010, the EFSA approved a reduction for BW claim for glucomannan at a daily dose of ≥3 g consumed over three eating occasions throughout the day and in the context of an energy-restricted diet in overweight adults [121]. However, two meta-analyses published after the EFSA approval reported inconsistent results. A 2014 meta-analysis of eight human studies providing 1–3 g glucomannan/day reported a non-significant reduction of 0.5 kg [114], and a 2021 meta-analysis of eight studies reported a 1.27 kg weight reduction [122]. Thus, the data to date are inconsistent. The role of the microbiota in healthy weight management is of high interest and in vitro studies have shown konjac fiber may beneficially alter the human gut microbiome [123]; however, human data on overweight/obese subjects are scarce. In a randomized, double-blind, controlled study, konjac flour, which provides primarily konjac glucomannan, resulted in a reduction in BMI and fat mass and significantly increased α-diversity of the gut microbiome in 69 obese subjects over 5 weeks [124]. Maintaining lean muscle mass is an important issue in weight management regimes. Supplementation of 3 g/day of konjac glucomannan with 300 mg calcium carbonate for 60 days decreased BW by 1.2 kg and BF by 1.1 kg while maintaining lean muscle mass in compliant overweight participants compared to the placebo (300 mg of calcium carbonate alone) [125]. More studies on glucomannan in weight management are needed to confirm these findings and determine optimum clinical application conditions. In addition, understanding the impact of changes in physiochemical properties of glucomannan (e.g., high vs. low polymerization) is also an important area for future research. 7.3. Glucomannan and Metabolic Health Glucomannan has been investigated for its effects on many obesity-related conditions [126]. The FDA has reviewed data on konjac glucomannan and its effect on blood lipids for determining a beneficial physiological effect and thus its ability to be labeled as a fiber [111]. Their review indicated nine studies from which conclusions could be drawn had been published on this effect, with six of these studies showing a reduction in TC and/or LDL-C [111]. The evidence included children and adults with hypercholesterolemia, with amounts of glucomannan between 2–13 g/day delivered in a capsule or a food product, over 3–8 weeks [111]. Due to the delay in gastric emptying and interference of nutrient accessibility for digestion and absorption, glucomannan is anticipated to also impact blood glucose [110]. The results of another meta-analysis with glucomannan amounts ranging from 1.2–15.1 g/day showed significant decreases in FBG, TC, LDL-C, non-HDL-C, and TG in a diverse population including obese, hyperlipidemic, and diabetic adults [127]. Subgroup analysis did not show similar results in pediatric subjects. A more recent meta-analysis of 12 studies similarly reported that Konjac glucomannan significantly lowered LDL-C, and non-HDL-C, but not apolipoprotein B [128]. Recent clinical studies have also suggested that glucomannan may influence carbohydrate digestion and absorption when it is co-consumed with high-carbohydrate foods [129,130]. 7.4. Glucomannan Summary Glucomannan has an EFSA-approved claim for weight loss; however, the effect on weight loss is modest and more investigations are warranted to confirm the consistency of this effect. Existing clinical evidence supports beneficial effects on blood glucose, lipid profile, and GI function. Like other fibers, consumption of glucomannan is expected to cause transient undesirable GI side effects in some people, particularly if a large amount is consumed without allowing for acclimation to the increase in fiber intake. Consumers of glucomannan supplements are urged to follow the manufacturer’s consumption instructions, consume adequate liquid with glucomannan products, and be aware of any deficits in swallowing or esophageal anatomy that could increase their risk for esophageal obstruction. 8. Capsaicinoids and Capsinoids Capsaicinoids are the active ingredients that give chili peppers their characteristic pungent flavor [131]. Capsaicinoids are a family of alkaloids that include capsaicin, dihydrocapsaicin, nordihydrocapsaicin, homocapsaicin, and homodihydrocapsaicin, with capsaicin being the dominant capsaicinoid. Capsaicin has been traditionally used for muscular pain, headaches, and GI protection and to improve circulation [132]. Understanding the actions of capsaicin led to the discovery of its binding intracellularly to Transient Receptor Potential Vanilloid subfamily member 1 (TRPV1) that is expressed predominantly in sensory neurons, and whose activation can regulate pain sensation [133]. More recently, capsaicinoids have been investigated for impact on several lifestyle health outcomes, including weight management and obesity-related conditions [131]. A similar but independent group of compounds named capsinoids has also been investigated in several clinical trials on weight management [134]. Unlike capsaicinoids, capsinoids are non-spicy compounds and include capsiate (4-hydroxy-3-methoxybenzyl [E]-8-methyl-6-nonenoate), dihydrocapsiate, and nordihydrocapsiate. The consumption of capsaicinoids across different countries ranges from 1–30 mg/day. Average consumption in Korea is ~3.25 mg/day, and consumption in Mexico averages 24.5–32 mg/day, though intake up to 250 mg/day has been reported [135,136]. A search of the Mintel GNPD database indicated weight management supplements recommend <10 mg capsaicin/day or <10 mg dihydrocapsiate/serving. 8.1. Capsaicinoids and Capsinoids Safety Few authoritative and government reports related to capsaicinoids safety have been published. The ODS has stated that few safety concerns have been reported for ≤33 mg/day for 4 weeks or 4 mg/day for 12 weeks of capsaicin and other capsaicinoids [17]. Capsaicinoids have been evaluated for food uses as well. The capsinoid dihydrocapsiate is GRAS at 1 and 3 mg/serving, and this conclusion has been submitted to the FDA with no objection from the agency [137]. EFSA has also concluded that dihydrocapsiate is safe for use as a food ingredient or food uses [137]. A NDI 75-day premarket notification for dihydrocapsiate marketed as a dietary ingredient (i.e., supplement) with a maximum daily intake of 15 mg is on file with the FDA [138]. Related to supplements, EFSA evaluated phenylcapsaicin, a chemically synthesized analog of capsaicin for use in food supplements and foods for special medical purposes at a maximum of 2.5 mg/day in the general population above 11 years old [139]. In contrast, a meta-analysis [140] on the thermogenic properties of capsaicinoids reported that intervention amounts of capsaicin supplements were generally low at 0.2–7 mg, with one study providing a single acute intervention of 150 mg/day [141,142] and another providing 135 mg/day for 3 months [143]. The study providing 150 mg/day [141,142] did not assess adverse effects, whereas the study providing 135 mg/day [143] noted that 10 of the 42 participants who were randomized to the capsaicin group “complained” about the capsules, but no further details on the complaints were provided. Some safety data are available from clinical studies; however, many meta-analyses combine data of foods that contain capsaicinoids with that of supplement delivery of the concentrates or isolated components. For example, a meta-analysis of metabolic syndrome clinical studies providing capsaicinoids ranging from 2–9 mg/day for 4–12 weeks combined food and supplement sources noted no serious adverse events and no events leading to withdrawal were reported in the studies [144]. Ten of the 12 studies included in this analysis used a supplement form ranging from fermented red pepper paste pills, red pepper capsules, capsinoid capsules, and dihydrocapsiate capsules. Eight of the 10 studies using supplements reported adverse events, with six indicating no adverse events during the study [144]. The other two studies indicated some tolerance-related adverse events of leg cramps (with 3 or 9 mg/day dihydrocapsiate over 4 weeks), dyspepsia, bowel irregularities, and skin rash (with 9 mg capsinoids capsules over 12 weeks) [144]. A tolerability study of a proprietary product comprising of a blend of capsaicinoids obtained from Capsicum annuum has been reported as well and showed no change in tolerability at ≤500 mg/day (10 mg/day of capsaicinoids) when provided for 1 week [145]. A controversial area of safety has been the proposed carcinogenic properties of hot peppers. Early studies assessed the food and not capsaicinoid or capsinoid extracts but attributed putative effects to capsaicin. This has complicated the available data. For example, a 2014 meta-analysis reported that low intake of capsaicin showed protection, whereas medium-high intakes showed susceptibility to gastric cancer, but this only searched for studies that assessed chili pepper food and only included case–control design studies [146]. Similarly, more recent meta-analyses have shown inconsistent relationships but have evaluated the whole food and not supplements [147,148]. Interestingly, a nonlinear association of gastric cancer risk with capsaicin intake was shown with potential protective effects at intakes of chili delivering 0–30 mg capsaicin/day, no clear relationship at 30–90 mg/day, and risk at >90 mg/day [148]. These all represented observational studies, and therefore, no causal relationship can be established from these data. Moreover, supplement compositions are variable and most intervention studies on health effects are conducted at amounts much lower than those associated as a putative risk factor for gastric cancer. Information on gastric cancer risk needs to be determined for the supplement forms of these ingredients, and more data are needed to understand if a protective effect exists at lower levels. 8.2. Capsaicinoids and Capsinoids and Weight Management The consumption of capsaicinoid-containing foods has been associated with a lower incidence of obesity, leading to an interest in the role of capsaicinoids themselves [149]. The EFSA evaluated health claims for capsaicin and maintenance of weight loss and increased carbohydrate oxidation in 2011 and found limited evidence [150]; however, since then, several new studies have been published. A meta-analysis reported a marginal decrease in BW and BMI when assessing results from four and five studies, respectively [144]. The effect of capsaicinoids has also been explored for helping promote a negative energy balance through increased energy expenditure (EE) and increasing fat oxidation [140]. For example, a meta-analysis of nine studies with different durations and amounts showed that capsaicinoids significantly increased fat oxidation and EE (+58.56 kcal/day) [151]. Additionally, Ludy et al. [140] conducted a meta-analysis to examine whether there was a dose-dependent effect of capsaicin on EE and fat oxidation, and found that a positive effect on EE was only noted for red pepper or capsaicin delivering 135–150 mg/day but not for amounts < 35 mg. This increased EE can be a consequence of enhanced brown adipose density and activity, which is involved in thermogenesis [152,153]. Capsaicinoids may also affect food intake, and a meta-analysis of eight studies showed that capsaicinoid ingestion with a minimum dose of 2 mg before a meal significantly decreased ad libitum calorie (−74.0 kcal) during the meal, suggesting an appetite suppression effect [154]. A novel study with an intraduodenal capsaicin infusion of 1.5 mg reported significantly increased satiety, likely related to GI stress but not to satiety hormones, such as PYY or GLP-1 [155]. Nevertheless, these observations are not supported by the results of another meta-analysis that found a non-significant reduction in BW of −0.19 kg [144], although a limitation of this study is the heterogeneity of the interventions, which include capsaicin supplements and hot pepper foods. 8.3. Capsaicinoids and Capsinoids and Metabolic Health Preclinical evidence shows that capsaicinoids can modulate glucose homeostasis and lipid metabolism through the TRPV1-dependent pathway and others [156]. Clinical trials have been conducted on FBG and have shown a null effect [157,158]. However, these null results could be ascribed to study participants having normal glucose regulation. No significant changes in lipid profile were reported in overweight adults after 12 weeks of capsaicinoid supplementation at 2 or 4 mg/day [157]. However, HDL-C increased and TG decreased following 4 mg capsaicin/day for 3 months in adults with low HDL-C [157]. Null results were reported for SBP and DBP in two meta-analyses of studies in normotensive participants [159,160]. Taken together, the evidence of capsaicinoids and capsaicinoid-containing foods on glucose regulation and lipid profile is equivocal, so more studies are warranted, particularly in populations with overweight and obesity. 8.4. Capsaicinoid and Capsainoid Summary Capsaicinoids have been shown to increase EE, although subsequent effects on weight loss are modest. Similarly, the evidence of capsaicinoids and capsainoids on glucose regulation and lipid profile is inconsistent and limited. Weight-loss supplements typically provide <10 mg capsaicin/day. Although safety data are confounded by data on high levels of acute hot pepper foods in some, few safety concerns have been reported in clinical trials with capsaicinoid/capsinoid supplements at ≤33 mg/day for 4 weeks and lower amounts for longer periods. Based on these observations, the ODS has stated few safety concerns have been reported for 33 mg/day for 4 weeks or 12 mg/day for 12 weeks for capsaicin and other capsaicinoids. 9. Conclusions Eighty percent of Americans take dietary supplements as part of a healthy lifestyle, including weight management products [161]. We reviewed six ingredients that are representative of the broad range of dietary supplements used for weight management—caffeine, GTE, GCBE, choline, glucomannan, and capsaicinoids and capsainoids—for safety as well as evidence related to weight management and metabolic health. Published authoritative reports and comprehensive systematic reviews were available for assessing the safety of caffeine, GTE, choline, and glucomannan, with information on the main bioactive of GCBE extract (i.e., CGA) also well studied. Further, caffeine, GTE (specifically EGCG), and choline have recommended intake limits. Most dietary supplement products marketed for weight management that include these ingredients have levels within the amounts identified as generally safe in these reports. Serious events are rare and often involve intentionally high intakes; however, as with any dietary component, these ingredients have the potential for intolerance effects, such as GI symptoms, and inter-individual variations, concomitant medications/supplements, and health status may also play a role in these sensitivities. Calorie restriction, improvements in dietary quality, and physical activity are key components in weight management programs, and successful approaches are multifactorial. Due to this complexity, weight management studies that assess dietary supplements in the context of other modalities can lead to challenges in discerning the effect of a specific dietary ingredient. In our review of authoritative reports and human clinical research, most of the ingredients appear to improve some aspects of metabolic health in people with overweight and obesity, such as supporting healthy blood glucose, lipids, and BP. Of these components, GTE has the most plentiful and consistent clinical evidence for weight management benefits. In summary, supplements providing caffeine, GTE, GCBE, choline, glucomannan, and/or capsaicinoids and capsainoids may provide some benefit for weight management and related measures of metabolic health and are generally safe when used in accordance with the manufacturers’ directions. Author Contributions Conceptualization, O.C. and E.M.; writing—original draft preparation, O.C. and E.M.; writing—revisions, E.M. and D.J.L.; writing—review and editing, E.M., O.C., D.J.L. and J.B.B.; funding acquisition, O.C. All authors have read and agreed to the published version of the manuscript. Funding This research was funded by the Council of Responsible Nutrition, grant number C222. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. E.M. is an employee of Biofortis Research. O.C. was employed by Biofortis Research during the conduct of this study, and D.J.L. was employed by Biofortis Research as a consultant. J.B.B. serves on scientific advisory boards for Advocare International, Bragg Live Foods, and SmartyPants Vitamins. The funders had no role in interpreting the literature and writing the manuscript. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Larsson S.C. Burgess S. Causal role of high body mass index in multiple chronic diseases: A systematic review and meta-analysis of Mendelian randomization studies BMC Med. 2021 19 320 10.1186/s12916-021-02188-x 34906131 2. Kumar R.B. Srivastava G. Reid T.J. Aronne L.J. Understanding the pathophysiologic pathways that underlie obesity and options for treatment Expert Rev. Endocrinol. Metab. 2021 16 321 338 10.1080/17446651.2021.1991310 34904501 3. Klein S. Burke L.E. Bray G.A. Blair S. Allison D.B. Pi-Sunyer X. Hong Y. Eckel R.H. Clinical Implications of Obesity With Specific Focus on Cardiovascular Disease Circulation 2004 110 2952 2967 10.1161/01.CIR.0000145546.97738.1E 15509809 4. Williamson D.A. Bray G.A. Ryan D.H. Is 5% weight loss a satisfactory criterion to define clinically significant weight loss? Obes. Silver Spring 2015 23 2319 2320 10.1002/oby.21358 26523739 5. Heymsfield S.B. Aronne L.J. Eneli I. Kumar R.B. Michalsky M. Walker E. Wolfe B.M. Woolford S.J. Yanovski S. Clinical Perspectives on Obesity Treatment: Challenges, Gaps, and Promising Opportunities. NAM Perspectives. Discussion Paper 2018 Available online: https://nam.edu/ (accessed on 8 February 2022) 6. Tobias D.K. Hall K.D. Eliminate or reformulate ultra-processed foods? Biological mechanisms matter Cell Metab. 2021 33 2314 2315 10.1016/j.cmet.2021.10.005 34699743 7. Ludwig D.S. Aronne L.J. Astrup A. de Cabo R. Cantley L.C. Friedman M.I. Heymsfield S.B. Johnson J.D. King J.C. Krauss R.M. The carbohydrate-insulin model: A physiological perspective on the obesity pandemic Am. J. Clin. Nutr. 2021 114 1873 1885 10.1093/ajcn/nqab270 34515299 8. Geller A.I. Shehab N. Weidle N.J. Lovegrove M.C. Wolpert B.J. Timbo B.B. Mozersky R.P. Budnitz D.S. Emergency Department Visits for Adverse Events Related to Dietary Supplements N. Engl. J. Med. 2015 373 1531 1540 10.1056/NEJMsa1504267 26465986 9. FDA Questions and Answers on Dietary Supplements. US Food and Drug Administration Available online: https://www.fda.gov/food/information-consumers-using-dietary-supplements/questions-and-answers-dietary-supplements (accessed on 15 December 2021) 10. Dwyer J.T. Coates P.M. Smith M.J. Dietary Supplements: Regulatory Challenges and Research Resources Nutrients 2018 10 41 10.3390/nu10010041 29300341 11. Page M.J. Moher D. Bossuyt P.M. Boutron I. Hoffmann T.C. Mulrow C.D. Shamseer L. Tetzlaff J.M. Akl E.A. Brennan S.E. PRISMA 2020 explanation and elaboration: Updated guidance and exemplars for reporting systematic reviews BMJ 2021 372 n160 10.1136/bmj.n160 33781993 12. Stroup D.F. Berlin J.A. Morton S.C. Olkin I. Williamson G.D. Rennie D. Moher D. Becker B.J. Sipe T.A. Thacker S.B. Meta-analysis of observational studies in epidemiology: A proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group JAMA 2000 283 2008 2012 10.1001/jama.283.15.2008 10789670 13. FDA Guidance for Industry: Evidence-Based Review System for the Scientific Evaluation of Health Claims Available online: https://www.fda.gov/regulatory-information/search-fda-guidance-documents/guidance-industry-evidence-based-review-system-scientific-evaluation-health-claims (accessed on 8 February 2022) 14. Blumberg J. Heaney R.P. Huncharek M. Scholl T. Stampfer M. Vieth R. Weaver C.M. Zeisel S.H. Evidence-based criteria in the nutritional context Nutr. Rev. 2010 68 478 484 10.1111/j.1753-4887.2010.00307.x 20646225 15. Dwyer J.T. Rubin K.H. Fritsche K.L. Psota T.L. Liska D.J. Harris W.S. Montain S.J. Lyle B.J. Creating the Future of Evidence-Based Nutrition Recommendations: Case Studies from Lipid Research Adv. Nutr. 2016 7 747 755 10.3945/an.115.010926 27422509 16. Mintel Global New Products Database (GNPD) Available online: https://www.mintel.com/global-new-products-database (accessed on 1 December 2021) 17. NIH Office of Dietary Supplements Dietary Supplements for Weight Loss Fact Sheet for Health Professionals National Institutes of Health Office of Dietary Supplements Bethesda, MD, USA 2021 Available online: https://ods.od.nih.gov/factsheets/WeightLoss-HealthProfessional/#h5 (accessed on 8 February 2022) 18. Nawrot P. Jordan S. Eastwood J. Rotstein J. Hugenholtz A. Feeley M. Effects of caffeine on human health Food Addit. Contam. 2003 20 1 30 10.1080/0265203021000007840 12519715 19. U.S. Department of Agriculture and U.S. Department of Health and Human Services Dietary Guidelines for Americans, 2020–2025 9th ed. December 2020 Available online: DietaryGuidelines.gov (accessed on 8 February 2022) 20. Verster J.C. Koenig J. Caffeine intake and its sources: A review of national representative studies Crit. Rev. Food Sci. Nutr. 2018 58 1250 1259 10.1080/10408398.2016.1247252 28605236 21. Fulgoni V.L. III Keast D.R. Lieberman H.R. Trends in intake and sources of caffeine in the diets of US adults: 2001–2010 Am. J. Clin. Nutr. 2015 101 1081 1087 10.3945/ajcn.113.080077 25832334 22. Benson S.M. Unice K.M. Glynn M.E. Hourly and daily intake patterns among U.S. caffeinated beverage consumers based on the National Health and Nutrition Examination Survey (NHANES, 2013–2016) Food Chem. Toxicol. 2019 125 271 278 10.1016/j.fct.2018.12.053 30602125 23. EFSA NDA (EFSA Panel on Dietetic Products, Nutrition and Allergies) Scientific Opinion on the Safety of Caffeine EFSA J. 2015 13 4102 4222 10.2903/j.efsa.2015.4102 24. Wikoff D. Welsh B.T. Henderson R. Brorby G.P. Britt J. Myers E. Goldberger J. Lieberman H.R. O’Brien C. Peck J. Systematic review of the potential adverse effects of caffeine consumption in healthy adults, pregnant women, adolescents, and children Food Chem. Toxicol. 2017 109 585 648 10.1016/j.fct.2017.04.002 28438661 25. FDA Guidance for Industry: Highly Concentrated Caffeine in Dietary Supplements Available online: https://www.fda.gov/regulatory-information/search-fda-guidance-documents/guidance-industry-highly-concentrated-caffeine-dietary-supplements (accessed on 4 December 2021) 26. Cappelletti S. Piacentino D. Fineschi V. Frati P. Cipolloni L. Aromatario M. Caffeine-Related Deaths: Manner of Deaths and Categories at Risk Nutrients 2018 10 611 10.3390/nu10050611 27. Van Schaik L. Kettle C. Green R. Irving H.R. Rathner J.A. Effects of Caffeine on Brown Adipose Tissue Thermogenesis and Metabolic Homeostasis: A Review Front. Neurosci. 2021 15 621356 10.3389/fnins.2021.621356 33613184 28. Tabrizi R. Saneei P. Lankarani K.B. Akbari M. Kolahdooz F. Esmaillzadeh A. Nadi-Ravandi S. Mazoochi M. Asemi Z. The effects of caffeine intake on weight loss: A systematic review and dos-response meta-analysis of randomized controlled trials Crit. Rev. Food Sci. Nutr. 2019 59 2688 2696 10.1080/10408398.2018.1507996 30335479 29. Neves J.S. Leitão L. Magriço R. Bigotte Vieira M. Viegas Dias C. Oliveira A. Carvalho D. Claggett B. Caffeine Consumption and Mortality in Diabetes: An Analysis of NHANES 1999–2010 Front. Endocrinol. Lausanne 2018 9 547 10.3389/fendo.2018.00547 30294299 30. Grosso G. Godos J. Galvano F. Giovannucci E.L. Coffee, Caffeine, and Health Outcomes: An Umbrella Review Annu. Rev. Nutr. 2017 37 131 156 10.1146/annurev-nutr-071816-064941 28826374 31. Shi X. Xue W. Liang S. Zhao J. Zhang X. Acute caffeine ingestion reduces insulin sensitivity in healthy subjects: A systematic review and meta-analysis Nutr. J. 2016 15 103 10.1186/s12937-016-0220-7 28031026 32. Beaudoin M.S. Allen B. Mazzetti G. Sullivan P.J. Graham T.E. Caffeine ingestion impairs insulin sensitivity in a dose-dependent manner in both men and women Appl. Physiol. Nutr. Metab. 2013 38 140 147 10.1139/apnm-2012-0201 23438224 33. Shearer J. Reimer R.A. Hittel D.S. Gault M.A. Vogel H.J. Klein M.S. Caffeine-Containing Energy Shots Cause Acute Impaired Glucoregulation in Adolescents Nutrients 2020 12 3850 10.3390/nu12123850 34. Irwin C. Khalesi S. Desbrow B. McCartney D. Effects of acute caffeine consumption following sleep loss on cognitive, physical, occupational and driving performance: A systematic review and meta-analysis Neurosci. Biobehav. Rev. 2020 108 877 888 10.1016/j.neubiorev.2019.12.008 31837359 35. Grgic J. Grgic I. Pickering C. Schoenfeld B.J. Bishop D.J. Pedisic Z. Wake up and smell the coffee: Caffeine supplementation and exercise performance-an umbrella review of 21 published meta-analyses Br. J. Sports Med. 2020 54 681 688 10.1136/bjsports-2018-100278 30926628 36. Pastore R.L. Fratellone P. Potential health benefits of green tea (Camellia sinensis): A narrative review Explore 2006 2 531 539 10.1016/j.explore.2006.08.008 17113495 37. Rasaei N. Asbaghi O. Samadi M. Setayesh L. Bagheri R. Gholami F. Soveid N. Casazza K. Wong A. Suzuki K. Effect of Green Tea Supplementation on Antioxidant Status in Adults: A Systematic Review and Meta-Analysis of Randomized Clinical Trials Antioxidants 2021 10 1731 10.3390/antiox10111731 34829602 38. Chacko S.M. Thambi P.T. Kuttan R. Nishigaki I. Beneficial effects of green tea: A literature review Chin. Med. 2010 5 13 10.1186/1749-8546-5-13 20370896 39. EFSA Panel on Food Additives Nutrient Sources added to Food (ANS) Younes M. Aggett P. Aguilar F. Crebelli R. Dusemund B. Filipic M. Frutos M.J. Galtier P. Gott D. Scientific Opinion on the Safety of Green Tea Catechins EFSA J. 2018 16 e05239 10.2903/j.efsa.2018.5239 32625874 40. Vieux F. Maillot M. Rehm C.D. Drewnowski A. Flavonoid Intakes in the US Diet are Linked to Higher Socioeconomic Status and to Tea Consumption: Analyses of NHANES 2011–2016 Data J. Nutr. 2020 150 2147 2155 10.1093/jn/nxaa145 32470977 41. Health Canada Summary Safety Review—Green Tea Extract-Containing Natural Health Products—Assessing the Potential Risk of Liver Injury (Hepatotoxicity) Available online: https://www.canada.ca/en/health-canada/services/drugs-health-products/medeffect-canada/safety-reviews/green-tea-extract-containing-natural-health-products-assessing-potential-risk-liver-injury.html (accessed on 6 December 2021) 42. Hu J. Webster D. Cao J. Shao A. The safety of green tea and green tea extract consumption in adults—Results of a systematic review Regul. Toxicol. Pharm. 2018 95 412 433 10.1016/j.yrtph.2018.03.019 29580974 43. Yates A.A. Erdman J.W. Jr. Shao A. Dolan L.C. Griffiths J.C. Bioactive nutrients—Time for tolerable upper intake levels to address safety Regul. Toxicol. Pharm. 2017 84 94 101 10.1016/j.yrtph.2017.01.002 44. Dekant W. Fujii K. Shibata E. Morita O. Shimotoyodome A. Safety assessment of green tea based beverages and dried green tea extracts as nutritional supplements Toxicol. Lett. 2017 277 104 108 10.1016/j.toxlet.2017.06.008 28655517 45. Oketch-Rabah H.A. Roe A.L. Rider C.V. Bonkovsky H.L. Giancaspro G.I. Navarro V. Paine M.F. Betz J.M. Marles R.J. Casper S. United States Pharmacopeia (USP) comprehensive review of the hepatotoxicity of green tea extracts Toxicol. Rep. 2020 7 386 402 10.1016/j.toxrep.2020.02.008 32140423 46. Matsuyama T. Tanaka Y. Kamimaki I. Nagao T. Tokimitsu I. Catechin safely improved higher levels of fatness, blood pressure, and cholesterol in children Obes. Silver Spring 2008 16 1338 1348 10.1038/oby.2008.60 47. Huang J. Wang Y. Xie Z. Zhou Y. Zhang Y. Wan X. The anti-obesity effects of green tea in human intervention and basic molecular studies Eur. J. Clin. Nutr. 2014 68 1075 1087 10.1038/ejcn.2014.143 25074392 48. Chen I.J. Liu C.Y. Chiu J.P. Hsu C.H. Therapeutic effect of high-dose green tea extract on weight reduction: A randomized, double-blind, placebo-controlled clinical trial Clin. Nutr. 2016 35 592 599 10.1016/j.clnu.2015.05.003 26093535 49. Xu X.Y. Zhao C.N. Li B.Y. Tang G.Y. Shang A. Gan R.Y. Feng Y.B. Li H.B. Effects and mechanisms of tea on obesity Crit. Rev. Food Sci. Nutr. 2021 1 18 10.1080/10408398.2021.1992748 34704503 50. Asbaghi O. Fouladvand F. Gonzalez M.J. Aghamohammadi V. Choghakhori R. Abbasnezhad A. Effect of Green Tea on Anthropometric Indices and Body Composition in Patients with Type 2 Diabetes Mellitus: A Systematic Review and Meta-Analysis Complement. Med. Res. 2021 28 244 251 10.1159/000511665 33207344 51. Lin Y. Shi D. Su B. Wei J. Găman M.A. Sedanur Macit M. Borges do Nascimento I.J. Guimaraes N.S. The effect of green tea supplementation on obesity: A systematic review and dose-response meta-analysis of randomized controlled trials Phytother. Res. 2020 34 2459 2470 10.1002/ptr.6697 32372444 52. Phung O.J. Baker W.L. Matthews L.J. Lanosa M. Thorne A. Coleman C.I. Effect of green tea catechins with or without caffeine on anthropometric measures: A systematic review and meta-analysis Am. J. Clin. Nutr. 2010 91 73 81 10.3945/ajcn.2009.28157 19906797 53. Haghighatdoost F. Nobakht M.G.B.F. Hariri M. Effect of green tea on plasma leptin and ghrelin levels: A systematic review and meta-analysis of randomized controlled clinical trials Nutrition 2018 45 17 23 10.1016/j.nut.2017.06.022 29129232 54. Haghighatdoost F. Nobakht M.G.B.F. Hariri M. Effect of Green Tea on Plasma Adiponectin Levels: A Systematic Review and Meta-analysis of Randomized Controlled Clinical Trials J. Am. Coll. Nutr. 2017 36 541 548 10.1080/07315724.2017.1333470 28853999 55. Mansour-Ghanaei F. Hadi A. Pourmasoumi M. Joukar F. Golpour S. Najafgholizadeh A. Green tea as a safe alternative approach for nonalcoholic fatty liver treatment: A systematic review and meta-analysis of clinical trials Phytother. Res. 2018 32 1876 1884 10.1002/ptr.6130 29947156 56. Zheng X.X. Xu Y.L. Li S.H. Hui R. Wu Y.J. Huang X.H. Effects of green tea catechins with or without caffeine on glycemic control in adults: A meta-analysis of randomized controlled trials Am. J. Clin. Nutr. 2013 97 750 762 10.3945/ajcn.111.032573 23426037 57. Liu K. Zhou R. Wang B. Chen K. Shi L.Y. Zhu J.D. Mi M.T. Effect of green tea on glucose control and insulin sensitivity: A meta-analysis of 17 randomized controlled trials Am. J. Clin. Nutr. 2013 98 340 348 10.3945/ajcn.112.052746 23803878 58. Wang X. Tian J. Jiang J. Li L. Ying X. Tian H. Nie M. Effects of green tea or green tea extract on insulin sensitivity and glycaemic control in populations at risk of type 2 diabetes mellitus: A systematic review and meta-analysis of randomised controlled trials J. Hum. Nutr. Diet. 2014 27 501 512 10.1111/jhn.12181 24206044 59. Yu J. Song P. Perry R. Penfold C. Cooper A.R. The Effectiveness of Green Tea or Green Tea Extract on Insulin Resistance and Glycemic Control in Type 2 Diabetes Mellitus: A Meta-Analysis Diabetes. Metab. J. 2017 41 251 262 10.4093/dmj.2017.41.4.251 28868822 60. Xu R. Bai Y. Yang K. Chen G. Effects of green tea consumption on glycemic control: A systematic review and meta-analysis of randomized controlled trials Nutr. Metab. 2020 17 56 10.1186/s12986-020-00469-5 32670385 61. Asbaghi O. Fouladvand F. Gonzalez M.J. Ashtary-Larky D. Choghakhori R. Abbasnezhad A. Effect of green tea on glycemic control in patients with type 2 diabetes mellitus: A systematic review and meta-analysis Diabetes. Metab. Syndr. 2021 15 23 31 10.1016/j.dsx.2020.11.004 33285391 62. Kim A. Chiu A. Barone M.K. Avino D. Wang F. Coleman C.I. Phung O.J. Green tea catechins decrease total and low-density lipoprotein cholesterol: A systematic review and meta-analysis J. Am. Diet. Assoc. 2011 111 1720 1729 10.1016/j.jada.2011.08.009 22027055 63. Onakpoya I. Spencer E. Heneghan C. Thompson M. The effect of green tea on blood pressure and lipid profile: A systematic review and meta-analysis of randomized clinical trials Nutr. Metab. Cardiovasc. Dis. 2014 24 823 836 10.1016/j.numecd.2014.01.016 24675010 64. Xu R. Yang K. Li S. Dai M. Chen G. Effect of green tea consumption on blood lipids: A systematic review and meta-analysis of randomized controlled trials Nutr. J. 2020 19 48 10.1186/s12937-020-00557-5 32434539 65. Yuan F. Dong H. Fang K. Gong J. Lu F. Effects of green tea on lipid metabolism in overweight or obese people: A meta-analysis of randomized controlled trials Mol. Nutr. Food Res. 2018 62 1601122 10.1002/mnfr.201601122 66. Asbaghi O. Fouladvand F. Moradi S. Ashtary-Larky D. Choghakhori R. Abbasnezhad A. Effect of green tea extract on lipid profile in patients with type 2 diabetes mellitus: A systematic review and meta-analysis Diabetes. Metab. Syndr. 2020 14 293 301 10.1016/j.dsx.2020.03.018 32289742 67. Li D. Wang R. Huang J. Cai Q. Yang C.S. Wan X. Xie Z. Effects and Mechanisms of Tea Regulating Blood Pressure: Evidences and Promises Nutrients 2019 11 1115 10.3390/nu11051115 68. Xu R. Yang K. Ding J. Chen G. Effect of green tea supplementation on blood pressure: A systematic review and meta-analysis of randomized controlled trials Med. Baltim. 2020 99 e19047 10.1097/MD.0000000000019047 69. Khalesi S. Sun J. Buys N. Jamshidi A. Nikbakht-Nasrabadi E. Khosravi-Boroujeni H. Green tea catechins and blood pressure: A systematic review and meta-analysis of randomised controlled trials Eur. J. Nutr. 2014 53 1299 1311 10.1007/s00394-014-0720-1 24861099 70. Li G. Zhang Y. Thabane L. Mbuagbaw L. Liu A. Levine M.A. Holbrook A. Effect of green tea supplementation on blood pressure among overweight and obese adults: A systematic review and meta-analysis J. Hypertens. 2015 33 243 254 10.1097/HJH.0000000000000426 25479028 71. Marcason W. What is green coffee extract? J. Acad. Nutr. Diet. 2013 113 364 10.1016/j.jand.2012.12.004 23351637 72. Lu H. Tian Z. Cui Y. Liu Z. Ma X. Chlorogenic acid: A comprehensive review of the dietary sources, processing effects, bioavailability, beneficial properties, mechanisms of action, and future directions Compr. Rev. Food Sci. Food Saf. 2020 19 3130 3158 10.1111/1541-4337.12620 33337063 73. Naveed M. Hejazi V. Abbas M. Kamboh A.A. Khan G.J. Shumzaid M. Ahmad F. Babazadeh D. FangFang X. Modarresi-Ghazani F. Chlorogenic acid (CGA): A pharmacological review and call for further research Biomed. Pharm. 2018 97 67 74 10.1016/j.biopha.2017.10.064 74. Farah A. de Paula Lima J. Consumption of Chlorogenic Acids through Coffee and Health Implications Beverages 2019 5 11 10.3390/beverages5010011 75. Olthof M.R. Hollman P.C. Buijsman M.N. van Amelsvoort J.M. Katan M.B. Chlorogenic acid, quercetin-3-rutinoside and black tea phenols are extensively metabolized in humans J. Nutr. 2003 133 1806 1814 10.1093/jn/133.6.1806 12771321 76. Integrated Laboratory Systems Chlorogenic Acid and Caffeine Acid: Review of Toxicological Literature Available online: https://ntp.niehs.nih.gov/ntp/htdocs/chem_background/exsumpdf/chlorogenicacid_508.pdf (accessed on 15 December 2021) 77. Soni & Associates Inc. Evaluation of the Generally Recognized as Safe (GRAS) Status of Coffeeberry (R) Coffee Fruit Extract as a Food Ingredient Available online: https://www.fda.gov/media/135526/download (accessed on 15 December 2021) 78. Asbaghi O. Sadeghian M. Nasiri M. Khodadost M. Shokri A. Panahande B. Pirouzi A. Sadeghi O. The effects of green coffee extract supplementation on glycemic indices and lipid profile in adults: A systematic review and dose-response meta-analysis of clinical trials Nutr. J. 2020 19 71 10.1186/s12937-020-00587-z 32665012 79. Morvaridi M. Rayyani E. Jaafari M. Khiabani A. Rahimlou M. The effect of green coffee extract supplementation on cardio metabolic risk factors: A systematic review and meta-analysis of randomized controlled trials J. Diabetes. Metab. Disord. 2020 19 645 660 10.1007/s40200-020-00536-x 32550217 80. Nikpayam O. Najafi M. Ghaffari S. Jafarabadi M.A. Sohrab G. Roshanravan N. Effects of green coffee extract on fasting blood glucose, insulin concentration and homeostatic model assessment of insulin resistance (HOMA-IR): A systematic review and meta-analysis of interventional studies Diabetol. Metab. Syndr. 2019 11 91 10.1186/s13098-019-0489-8 31709016 81. Roshan H. Nikpayam O. Sedaghat M. Sohrab G. Effects of green coffee extract supplementation on anthropometric indices, glycaemic control, blood pressure, lipid profile, insulin resistance and appetite in patients with the metabolic syndrome: A randomised clinical trial Br. J. Nutr. 2018 119 250 258 10.1017/S0007114517003439 29307310 82. Blum J. Lemaire B. Lafay S. Effect of a green decaffeinated coffee extract on glycaemia. A pilot prospective clinical study Nutrafoods 2007 6 13 17 83. Mansour A. Mohajeri-Tehrani M.R. Samadi M. Qorbani M. Merat S. Adibi H. Poustchi H. Hekmatdoost A. Effects of supplementation with main coffee components including caffeine and/or chlorogenic acid on hepatic, metabolic, and inflammatory indices in patients with non-alcoholic fatty liver disease and type 2 diabetes: A randomized, double-blind, placebo-controlled, clinical trial Nutr. J. 2021 20 35 10.1186/s12937-021-00694-5 33838673 84. Asbaghi O. Sadeghian M. Rahmani S. Mardani M. Khodadost M. Maleki V. Pirouzi A. Talebi S. Sadeghi O. The effect of green coffee extract supplementation on anthropometric measures in adults: A comprehensive systematic review and dose-response meta-analysis of randomized clinical trials Complement. Med. 2020 51 102424 10.1016/j.ctim.2020.102424 32507437 85. Chen Y. Zhao Y. Wang Y. Nazary-Vannani A. Clark C.C.T. Sedanur Macit M. Khani V. Zhang Y. The influence of green coffee bean extract supplementation on blood glucose levels: A systematic review and dose-response meta-analysis of randomized controlled trials Phytother. Res. 2020 34 2159 2169 10.1002/ptr.6667 32159261 86. Han B. Nazary-Vannani A. Talaei S. Clark C.C.T. Rahmani J. Rasekhmagham R. Kord-Varkaneh H. The effect of green coffee extract supplementation on blood pressure: A systematic review and meta-analysis of randomized controlled trials Phytother. Res. 2019 33 2918 2926 10.1002/ptr.6481 31429515 87. Bosso H. Barbalho S.M. de Alvares Goulart R. Otoboni A. Green coffee: Economic relevance and a systematic review of the effects on human health Crit. Rev. Food Sci. Nutr. 2021 1 17 10.1080/10408398.2021.1948817 88. Arias N. Arboleya S. Allison J. Kaliszewska A. Higarza S.G. Gueimonde M. Arias J.L. The Relationship between Choline Bioavailability from Diet, Intestinal Microbiota Composition, and Its Modulation of Human Diseases Nutrients 2020 12 2340 10.3390/nu12082340 89. Wiedeman A.M. Barr S.I. Green T.J. Xu Z. Innis S.M. Kitts D.D. Dietary Choline Intake: Current State of Knowledge Across the Life Cycle Nutrients 2018 10 1513 10.3390/nu10101513 90. Institute of Medicine Dietary Reference Intakes for Thiamin, Riboflavin, Niacin, Vitamin B6, Folate, Vitamin B12, Pantothenic Acid, Biotin, and Choline The National Academies Press Washington, DC, USA 1998 10.17226/6015 91. EFSA NDA Panel EFSA Panel on Dietetic Products, Nutrition and Allergies. Scientific Opinion on Dietary Reference Values for Choline EFSA J. 2016 14 4484 4554 10.2903/j.efsa.2016.4484 92. Amazon Best Sellers in Choline Vitamin Supplements Available online: https://www.amazon.com/gp/bestsellers/hpc/6939007011 (accessed on 13 December 2021) 93. Elsawy G. Abdelrahman O. Hamza A. Effect of choline supplementation on rapid weight loss and biochemical variables among female taekwondo and judo athletes J. Hum. Kinet. 2014 40 77 82 10.2478/hukin-2014-0009 25031675 94. Killgore W.D. Ross A.J. Kamiya T. Kawada Y. Renshaw P.F. Yurgelun-Todd D.A. Citicoline affects appetite and cortico-limbic responses to images of high-calorie foods Int. J. Eat. Disord. 2010 43 6 13 10.1002/eat.20658 19260039 95. Gao X. Wang Y. Randell E. Pedram P. Yi Y. Gulliver W. Sun G. Higher Dietary Choline and Betaine Intakes Are Associated with Better Body Composition in the Adult Population of Newfoundland, Canada PLoS ONE 2016 11 e0155403 10.1371/journal.pone.0155403 27166611 96. Jafari A. Jalilpiran Y. Suitor K. Bellissimo N. Azadbakht L. The association of dietary choline and betaine and anthropometric measurements among Iranian children: A cross-sectional study BMC Pediatr. 2021 21 213 10.1186/s12887-021-02677-1 33931062 97. Leermakers E.T. Moreira E.M. Kiefte-de Jong J.C. Darweesh S.K. Visser T. Voortman T. Bautista P.K. Chowdhury R. Gorman D. Bramer W.M. Effects of choline on health across the life course: A systematic review Nutr. Rev. 2015 73 500 522 10.1093/nutrit/nuv010 26108618 98. Nakamura T. Nakamura S. Karoji N. Aikawa T. Suzuki O. Hepatic function tests in heavy drinkers among workmen Tohoku. J. Exp. Med. 1967 93 219 226 10.1620/tjem.93.219 5587662 99. Buchman A.L. Ament M.E. Sohel M. Dubin M. Jenden D.J. Roch M. Pownall H. Farley W. Awal M. Ahn C. Choline deficiency causes reversible hepatic abnormalities in patients receiving parenteral nutrition: Proof of a human choline requirement: A placebo-controlled trial JPEN J. Parenter. Enter. Nutr. 2001 25 260 268 10.1177/0148607101025005260 100. Guerrerio A.L. Colvin R.M. Schwartz A.K. Molleston J.P. Murray K.F. Diehl A. Mohan P. Schwimmer J.B. Lavine J.E. Torbenson M.S. Choline intake in a large cohort of patients with nonalcoholic fatty liver disease Am. J. Clin. Nutr. 2012 95 892 900 10.3945/ajcn.111.020156 22338037 101. Yu D. Shu X.O. Xiang Y.B. Li H. Yang G. Gao Y.T. Zheng W. Zhang X. Higher dietary choline intake is associated with lower risk of nonalcoholic fatty liver in normal-weight Chinese women J. Nutr. 2014 144 2034 2040 10.3945/jn.114.197533 25320186 102. Corbin K.D. Zeisel S.H. Choline metabolism provides novel insights into nonalcoholic fatty liver disease and its progression Curr. Opin. Gastroenterol. 2012 28 159 165 10.1097/MOG.0b013e32834e7b4b 22134222 103. Sherriff J.L. O’Sullivan T.A. Properzi C. Oddo J.L. Adams L.A. Choline, Its Potential Role in Nonalcoholic Fatty Liver Disease, and the Case for Human and Bacterial Genes Adv. Nutr. 2016 7 5 13 10.3945/an.114.007955 26773011 104. Naghipour S. Cox A.J. Peart J.N. Du Toit E.F. Headrick J.P. Trimethylamine N-oxide: Heart of the microbiota-CVD nexus? Nutr. Res. Rev. 2021 34 125 146 10.1017/S0954422420000177 32718365 105. Lemaitre R.N. Jensen P.N. Wang Z. Fretts A.M. McKnight B. Nemet I. Biggs M.L. Sotoodehnia N. de Oliveira Otto M.C. Psaty B.M. Association of Trimethylamine N-Oxide and Related Metabolites in Plasma and Incident Type 2 Diabetes: The Cardiovascular Health Study JAMA Netw. Open 2021 4 e2122844 10.1001/jamanetworkopen.2021.22844 34448864 106. Rashvand S. Mobasseri M. Tarighat-Esfanjani A. Effects of Choline and Magnesium Concurrent Supplementation on Coagulation and Lipid Profile in Patients with Type 2 Diabetes Mellitus: A Pilot Clinical Trial Biol. Trace Elem. Res. 2020 194 328 335 10.1007/s12011-019-01802-7 31385202 107. Wallace J.M.W. McCormack J.M. McNulty H. Walsh P.M. Robson P.J. Bonham M.P. Duffy M.E. Ward M. Molloy A.M. Scott J.M. Choline supplementation and measures of choline and betaine status: A randomised, controlled trial in postmenopausal women Br. J. Nutr. 2012 108 1264 1271 10.1017/S000711451100674X 22172554 108. Olthof M.R. van Vliet T. Verhoef P. Zock P.L. Katan M.B. Effect of homocysteine-lowering nutrients on blood lipids: Results from four randomised, placebo-controlled studies in healthy humans PLoS Med. 2005 2 e135 10.1371/journal.pmed.0020135 15916468 109. Devaraj R.D. Reddy C.K. Xu B. Health-promoting effects of konjac glucomannan and its practical applications: A critical review Int. J. Biol. Macromol. 2019 126 273 281 10.1016/j.ijbiomac.2018.12.203 30586587 110. Tester R.F. Al-Ghazzewi F.H. Beneficial health characteristics of native and hydrolysed konjac (Amorphophallus konjac) glucomannan J. Sci. Food Agric. 2016 96 3283 3291 10.1002/jsfa.7571 26676961 111. FDA FDA Grants Citizen Petition on Glucomannan as a Dietary Fiber Available online: https://www.fda.gov/food/cfsan-constituent-updates/fda-grants-citizen-petition-glucomannan-dietary-fiber (accessed on 12 February 2022) 112. Hornick B. Birkett A. Liska D.J. The Fiber Deficit, Part 3—Beyond Traditional Fiber Sources Nutr. Today 2013 48 168 173 10.1097/NT.0b013e31829da149 113. EFSA Panel on Food Additives Nutrient Sources added to Food (ANS) Mortensen A. Aguilar F. Crebelli R. Di Domenico A. Frutos M.J. Galtier P. Gott D. Gundert-Remy U. Lambre C. Re-evaluation of konjac gum (E 425 i) and konjac glucomannan (E 425 ii) as food additives EFSA J. 2017 15 e04864 10.2903/j.efsa.2017.4864 32625526 114. Onakpoya I. Posadzki P. Ernst E. The efficacy of glucomannan supplementation in overweight and obesity: A systematic review and meta-analysis of randomized clinical trials J. Am. Coll. Nutr. 2014 33 70 78 10.1080/07315724.2014.870013 24533610 115. Zalewski B.M. Chmielewska A. Szajewska H. The effect of glucomannan on body weight in overweight or obese children and adults: A systematic review of randomized controlled trials Nutrition 2015 31 437 442 10.1016/j.nut.2014.09.004 25701331 116. Martino F. Martino E. Morrone F. Carnevali E. Forcone R. Niglio T. Effect of dietary supplementation with glucomannan on plasma total cholesterol and low density lipoprotein cholesterol in hypercholesterolemic children Nutr. Metab. Cardiovasc. Dis. 2005 15 174 180 10.1016/j.numecd.2004.04.004 15955465 117. Vido L. Facchin P. Antonello I. Gobber D. Rigon F. Childhood obesity treatment: Double blinded trial on dietary fibres (glucomannan) versus placebo Padiatr. Padol. 1993 28 133 136 8247594 118. Zalewski B.M. Szajewska H. No Effect of Glucomannan on Body Weight Reduction in Children and Adolescents with Overweight and Obesity: A Randomized Controlled Trial J. Pediatr. 2019 211 85 91.e1 10.1016/j.jpeds.2019.03.044 31036412 119. Henry D.A. Mitchell A.S. Aylward J. Fung M.T. McEwen J. Rohan A. Glucomannan and risk of oesophageal obstruction Br. Med. J. Clin. Res. Ed. 1986 292 591 592 10.1136/bmj.292.6520.591-a 120. FDA 21 CFR 201.319. Water-Soluble Gums, Hydrophilic Gums, and Hydrophilic Mucilloids (Including, but not Limited to Agar, Alginic Acid, Calcium Polycarbophil, Caeboxymethylcellulose Sodium, Carrageenan, Chondrus, Glucomannas ((B-1,4 Linked) Polymannose Acetate), Guar Gum, Karaya Gum, Kelp, Methylcellulose, Plantago Seed (Psyllium), Polycarbophil Tragacanth, and Xanthan Gum) as Active Ingredients; Requried Warnings and Directions Available online: https://www.ecfr.gov/current/title-21/chapter-I/subchapter-C/part-201/subpart-G/section-201.319 (accessed on 12 February 2022) 121. EFSA Panel on Dietetic Products, Nutrition and Allergies (NDA) Scientific Opinion on the substantiation of health claims related to konjac mannan (glucomannan) and reduction of body weight (ID 854, 1556, 3725), reduction of post-prandial glycaemic responses (ID 1559), maintenance of normal blood glucose concentrations (ID 835, 3724), maintenance of normal (fasting) blood concentrations of triglycerides (ID 3217), maintenance of normal blood cholesterol concentrations (ID 3100, 3217), maintenance of normal bowel function (ID 834, 1557, 3901) and decreasing potentially pathogenic gastro-intestinal microorganisms (ID 1558) pursuant to Article 13(1) of Regulation (EC) No 1924/2006 EFSA J. 2010 8 1798 10.2903/j.efsa.2010.1798 122. Bessell E. Maunder A. Lauche R. Adams J. Sainsbury A. Fuller N.R. Efficacy of dietary supplements containing isolated organic compounds for weight loss: A systematic review and meta-analysis of randomised placebo-controlled trials Int. J. Obes. 2021 45 1631 1643 10.1038/s41366-021-00839-w 123. Calatayud M. Van den Abbeele P. Ghyselinck J. Marzorati M. Rohs E. Birkett A. Comparative Effect of 22 Dietary Sources of Fiber on Gut Microbiota of Healthy Humans in vitro Front. Nutr. 2021 8 700571 10.3389/fnut.2021.700571 34277691 124. Li Y. Kang Y. Du Y. Chen M. Guo L. Huang X. Li T. Chen S. Yang F. Yu F. Effects of Konjaku Flour on the Gut Microbiota of Obese Patients Front. Cell Infect. Microbiol. 2022 12 771748 10.3389/fcimb.2022.771748 35300378 125. Kaats G.R. Bagchi D. Preuss H.G. Konjac Glucomannan Dietary Supplementation Causes Significant Fat Loss in Compliant Overweight Adults J. Am. Coll. Nutr. 2015 1 7 10.1080/07315724.2015.1009194 26492494 126. Behera S.S. Ray R.C. Konjac glucomannan, a promising polysaccharide of Amorphophallus konjac K. Koch in health care Int. J. Biol. Macromol. 2016 92 942 956 10.1016/j.ijbiomac.2016.07.098 27481345 127. Sood N. Baker W.L. Coleman C.I. Effect of glucomannan on plasma lipid and glucose concentrations, body weight, and blood pressure: Systematic review and meta-analysis Am. J. Clin. Nutr. 2008 88 1167 1175 10.1093/ajcn/88.4.1167 18842808 128. Ho H.V.T. Jovanovski E. Zurbau A. Blanco Mejia S. Sievenpiper J.L. Au-Yeung F. Jenkins A.L. Duvnjak L. Leiter L. Vuksan V. A systematic review and meta-analysis of randomized controlled trials of the effect of konjac glucomannan, a viscous soluble fiber, on LDL cholesterol and the new lipid targets non-HDL cholesterol and apolipoprotein B Am. J. Clin. Nutr. 2017 105 1239 1247 10.3945/ajcn.116.142158 28356275 129. Ueno H. Haraguchi N. Azuma M. Shiiya T. Noda T. Ebihara E. Uehira Y. Uchida T. Sasaba K. Nakamura M. Active Consumption of Konjac and Konjac Products Improves Blood Glucose Control in Patients with Type 2 Diabetes Mellitus J. Am. Coll. Nutr. 2021 1 7 10.1080/07315724.2021.2002739 130. Yoshida A. Kimura T. Tsunekawa K. Araki O. Ushiki K. Ishigaki H. Shoho Y. Suda I. Hiramoto S. Murakami M. Glucomannan Inhibits Rice Gruel-Induced Increases in Plasma Glucose and Insulin Levels Ann. Nutr. Metab. 2020 76 259 267 10.1159/000508674 32659777 131. Basith S. Cui M. Hong S. Choi S. Harnessing the Therapeutic Potential of Capsaicin and Its Analogues in Pain and Other Diseases Molecules 2016 21 966 10.3390/molecules21080966 132. Huang X.F. Xue J.Y. Jiang A.Q. Zhu H.L. Capsaicin and its analogues: Structure-activity relationship study Curr. Med. Chem. 2013 20 2661 2672 10.2174/0929867311320210004 23627937 133. Torres-Ugalde Y.C. Romero-Palencia A. Román-Gutiérrez A.D. Ojeda-Ramírez D. Guzmán-Saldaña R.M.E. Caffeine Consumption in Children: Innocuous or Deleterious? A Systematic Review Int. J. Environ. Res. Public Health 2020 17 2489 10.3390/ijerph17072489 134. Whiting S. Derbyshire E. Tiwari B.K. Capsaicinoids and capsinoids. A potential role for weight management? A systematic review of the evidence Appetite 2012 59 341 348 10.1016/j.appet.2012.05.015 22634197 135. Kwon Y. Estimation of Dietary Capsaicinoid Exposure in Korea and Assessment of Its Health Effects Nutrients 2021 13 2461 10.3390/nu13072461 34371974 136. Lopez-Carrillo L. Lopez-Cervantes M. Robles-Diaz G. Ramirez-Espitia A. Mohar-Betancourt A. Meneses-Garcia A. Lopez-Vidal Y. Blair A. Capsaicin consumption, Helicobacter pylori positivity and gastric cancer in Mexico Int. J. Cancer 2003 106 277 282 10.1002/ijc.11195 12800206 137. EFSA Panel on Dietetic Products, Nutrition and Allergies (NDA) Scientific Opinion on Dihydrocapsiate EFSA J. 2012 10 2812 10.2903/j.efsa.2012.2812 138. FDA NDI 739, Dihydrocapsiate from Ajinomoto North America Available online: https://www.regulations.gov/document/FDA-2012-S-1178-0020 (accessed on 18 February 2021) 139. EFSA Panel on Dietetic Products, Nutrition and Allergies (NDA) Safety of phenylcapsaicin as a novel food pursuant to Regulation (EU) 2015/2283 EFSA J. 2019 17 5718 10.2903/j.efsa.2019.5718 140. Ludy M.J. Moore G.E. Mattes R.D. The effects of capsaicin and capsiate on energy balance: Critical review and meta-analyses of studies in humans Chem. Senses 2012 37 103 121 10.1093/chemse/bjr100 22038945 141. Shin K.O. Moritani T. Alterations of autonomic nervous activity and energy metabolism by capsaicin ingestion during aerobic exercise in healthy men J. Nutr. Sci. Vitam. 2007 53 124 132 10.3177/jnsv.53.124 142. Shin K.O. Moritani T. Capsaicin supplementation fails to modulate autonomic and cardiac electrophysiologic activity during exercise in the obese: With variants of UCP2 and UCP3 polymorphism J. Sports Sci. Med. 2008 7 365 370 24149904 143. Lejeune M.P. Kovacs E.M. Westerterp-Plantenga M.S. Effect of capsaicin on substrate oxidation and weight maintenance after modest body-weight loss in human subjects Br. J. Nutr. 2003 90 651 659 10.1079/BJN2003938 13129472 144. Jang H.H. Lee J. Lee S.H. Lee Y.M. Effects of Capsicum annuum supplementation on the components of metabolic syndrome: A systematic review and meta-analysis Sci. Rep. 2020 10 20912 10.1038/s41598-020-77983-2 33262398 145. Deshpande J. Jeyakodi S. Juturu V. Tolerability of Capsaicinoids from Capsicum Extract in a Beadlet Form: A Pilot Study J. Toxicol. 2016 2016 6584649 10.1155/2016/6584649 27066073 146. Pabalan N. Jarjanazi H. Ozcelik H. The impact of capsaicin intake on risk of developing gastric cancers: A meta-analysis J. Gastrointest. Cancer 2014 45 334 341 10.1007/s12029-014-9610-2 24756832 147. Luo L. Yan J. Wang X. Sun Z. The correlation between chili pepper consumption and gastric cancer risk: A meta-analysis Asia Pac. J. Clin. Nutr. 2021 30 130 139 10.6133/apjcn.202103_30(1).0016 33787049 148. Du Y. Lv Y. Zha W. Hong X. Luo Q. Chili Consumption and Risk of Gastric Cancer: A Meta-Analysis Nutr. Cancer 2021 73 45 54 10.1080/01635581.2020.1733625 32241189 149. Shi Z. Riley M. Taylor A.W. Page A. Chilli consumption and the incidence of overweight and obesity in a Chinese adult population Int. J. Obes. Lond. 2017 41 1074 1079 10.1038/ijo.2017.88 28360431 150. EFSA Panel on Dietetic Products, Nutrition and Allergies (NDA) Scientific Opinion on the substantiation of heatlh claims related to capsaicin and maintenance of body weight after loss (ID 2039, 2041, 2042), increase in carbohydrate oxidation (ID 2040), and contribution to normal hair growth (ID 2044) pursuant to Article 13 (1) of Regulation (EC) No. 1924/2006 EFSA J. 2011 9 2210 151. Zsiborás C. Mátics R. Hegyi P. Balaskó M. Pétervári E. Szabó I. Sarlós P. Mikó A. Tenk J. Rostás I. Capsaicin and capsiate could be appropriate agents for treatment of obesity: A meta-analysis of human studies Crit. Rev. Food Sci. Nutr. 2018 58 1419 1427 10.1080/10408398.2016.1262324 28001433 152. Fuse S. Endo T. Tanaka R. Kuroiwa M. Ando A. Kume A. Yamamoto A. Kuribayashi K. Somekawa S. Takeshita M. Effects of Capsinoid Intake on Brown Adipose Tissue Vascular Density and Resting Energy Expenditure in Healthy, Middle-Aged Adults: A Randomized, Double-Blind, Placebo-Controlled Study Nutrients 2020 12 2676 10.3390/nu12092676 153. Nirengi S. Homma T. Inoue N. Sato H. Yoneshiro T. Matsushita M. Kameya T. Sugie H. Tsuzaki K. Saito M. Assessment of human brown adipose tissue density during daily ingestion of thermogenic capsinoids using near-infrared time-resolved spectroscopy J. Biomed. Opt. 2016 21 091305 10.1117/1.JBO.21.9.091305 27135066 154. Whiting S. Derbyshire E.J. Tiwari B. Could capsaicinoids help to support weight management? A systematic review and meta-analysis of energy intake data Appetite 2014 73 183 188 10.1016/j.appet.2013.11.005 24246368 155. van Avesaat M. Troost F.J. Westerterp-Plantenga M.S. Helyes Z. Le Roux C.W. Dekker J. Masclee A.A. Keszthelyi D. Capsaicin-induced satiety is associated with gastrointestinal distress but not with the release of satiety hormones Am. J. Clin. Nutr. 2016 103 305 313 10.3945/ajcn.115.123414 26718419 156. Liang W. Lan Y. Chen C. Song M. Xiao J. Huang Q. Cao Y. Ho C.T. Lu M. Modulating effects of capsaicin on glucose homeostasis and the underlying mechanism Crit. Rev. Food. Sci. Nutr. 2021 1 19 10.1080/10408398.2021.1991883 34657531 157. Qin Y. Ran L. Wang J. Yu L. Lang H.D. Wang X.L. Mi M.T. Zhu J.D. Capsaicin Supplementation Improved Risk Factors of Coronary Heart Disease in Individuals with Low HDL-C Levels Nutrients 2017 9 1037 10.3390/nu9091037 28930174 158. Urbina S.L. Roberts M.D. Kephart W.C. Villa K.B. Santos E.N. Olivencia A.M. Bennett H.M. Lara M.D. Foster C.A. Purpura M. Effects of twelve weeks of capsaicinoid supplementation on body composition, appetite and self-reported caloric intake in overweight individuals Appetite 2017 113 264 273 10.1016/j.appet.2017.02.025 28235621 159. Shirani F. Foshati S. Tavassoly M. Clark C.C.T. Rouhani M.H. The effect of red pepper/capsaicin on blood pressure and heart rate: A systematic review and meta-analysis of clinical trials Phytother. Res. 2021 35 6080 6088 10.1002/ptr.7217 34676607 160. Amini M.R. Sheikhhossein F. Bazshahi E. Hajiaqaei M. Shafie A. Shahinfar H. Azizi N. Eghbaljoo Gharehgheshlaghi H. Naghshi S. Fathipour R.B. The effects of capsinoids and fermented red pepper paste supplementation on blood pressure: A systematic review and meta-analysis of randomized controlled trials Clin. Nutr. 2021 40 1767 1775 10.1016/j.clnu.2020.10.018 33129596 161. Daniells S. CRN Survey: 80% of Americans are Now Using Dietary Supplements Available online: https://www.nutraingredients-usa.com/Article/2021/10/22/CRN-survey-80-of-Americans-are-now-using-dietary-supplements# (accessed on 18 February 2022)

PMC009xxxxxx/PMC9099656.txt

==== Front Polymers (Basel) Polymers (Basel) polymers Polymers 2073-4360 MDPI 10.3390/polym14091686 polymers-14-01686 Article Potential Use of Chitosan-TiO2 Nanocomposites for the Electroanalytical Detection of Imidacloprid Castillo Blanca Estela https://orcid.org/0000-0003-2611-2521 Prokhorov Evgen * https://orcid.org/0000-0002-1162-6928 Luna-Bárcenas Gabriel https://orcid.org/0000-0002-1991-246X Kovalenko Yuriy Marin Luminita Academic Editor Cinvestav del IPN, Unidad Querétaro, Queretaro 76230, Mexico; becr_iq@yahoo.com.mx (B.E.C.); gabriel.luna@cinvestav.mx (G.L.-B.); kovalenko.yuriy@gmail.com (Y.K.) * Correspondence: prokhorov@cinvestav.mx; Tel.: +52-442-211-9921 21 4 2022 5 2022 14 9 168631 1 2022 23 2 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). The detection of toxic insecticides is a major scientific and technological challenge. In this regard, imidacloprid is a neonicotinoid that is a systemic insecticide that can accumulate in agricultural products and affect human health. This work aims to study the properties of chitosan–TiO2 nanocomposites in which nanoparticles with high surface area serve as molecular recognition sites for electroanalytical imidacloprid detection. We show that the best sensitivity to imidacloprid was obtained using a modified electrode with a chitosan–TiO2 nanocomposite with a 40 wt.% of TiO2 nanoparticles. By using a three-phase effective permittivity model which includes chitosan, TiO2, an interface layer between nanoparticles and a matrix, we showed that nanocomposites with 40 wt.% of TiO2 the interface volume fraction reaches a maximum. At higher nanoparticle concentration, the sensitivity of the sensor decreases due to the decreasing of the interface volume fraction, agglomeration of nanoparticles and a decrease in their effective surface area. The methodology presented can be helpful in the design and optimization of polymer-based nanocomposites for a variety of applications. chitosan titanium dioxide interface layer dielectric spectroscopy cyclic voltametric imidacloprid ==== Body pmc1. Introduction Chitosan (CS)–titanium dioxide nanoparticles (TiO2 NPs) composites are one of the most attractive materials which have interesting technological properties and applications. Nanocomposites based upon CS–TiO2 NPs are widely used for: the development of different antibacterial package materials [1], wound healing applications [2,3], photocatalytic applications [4], wastewater treatment [5], different sensors [6], etc. The most important properties of CS are the presence of reactive amino and hydroxyl side groups which through hydrogen or covalent bond can interact with TiO2 NPs and form nanocomposites with high concentration without NPs agglomeration [6,7]. It is worth noting that most nanocomposite applications are based upon the high effective surface area of the NPs. However, it is difficult to obtain a uniform dispersion of nanoparticles owing to their strong tendency of agglomeration for their high surface energy. Agglomeration of NPs worsens the properties of nanocomposites due to the restriction of surface area [8,9]. In composites based on TiO2 NPs, agglomeration affected their photocatalytic activity [3,10], the conductivity of ion batteries [11], the biological activity of nanocomposites with TiO2 NPs [12], the antimicrobial capacity [1], the adsorption capacities [5], etc. In recent years, it has become clear that the interface layer formed around nanoparticles in polymer nanocomposites is critical for controlling their dispersion (see, for example, [13,14,15,16,17]). This layer appears when strong molecular interactions between nanoparticles and polymer matrix are present; it occupies a significant volume fraction of the polymer matrix and it exhibits properties that differ from both polymer matrix and NPs [8,10,18]. As a result of these strong molecular interactions, most properties of the nanocomposite significantly depend upon the interphase layer (see, for example [8,9,18,19,20,21,22,23,24,25,26,27]). The experimental and theoretical investigations in the recent years provided a large amount of information about interphase and interfacial interactions in polymer nanocomposites; nevertheless, they focused on dynamic of polymer chains [22,23,24,25], glass transition temperature [26], mechanical properties [8,20,21], adsorption of polymer chains on the nanoparticle surface [27]. However, to the best of our knowledge, there are no publications in the literature that optimizes the concentration of NPs before they agglomerate (or nanocomposites with the largest NP surface area) by considering the effect of the interphase layer. Additionally, there are no reported investigations between of the interface layer in CS–TiO2 nanocomposites. It is worth noting that chitosan is dielectric material and TiO2 is a high resistivity semiconductor (ca. 1015 Ohm cm) with dielectric constant ca. 100 which dependents upon the polymorphism of NPs [28]. Therefore, dielectric spectroscopy can be advantageously used to determine relevant parameters of the interface layer and the highest concentration of NPs before they agglomerate (decreasing the effective surface area). Recently, some publications on the use of TiO2 NPs for the development of an electrochemical sensor for the detection of pesticides, including imidacloprid (IMD) are reported [29,30]. IMD is one of the most used neonicotinoids for the crop protection worldwide due to its low soil persistence and high insecticidal activity at a very low application rate [31]. IMD is absorbed by plants via either their roots or leaves and then is transported throughout the tissues of the plants [32]; it may be present in various foods that ultimately affect the human health [29]. Different methods or techniques have been used for the detection and quantification of this insecticide in wastewater. High performance liquid chromatography, gas chromatography–mass spectrometry, liquid chromatography mass spectrometry and optical technique are commonly used methods for the detection of IMD. These instrumental methods are accurate but expensive and time consuming, requiring lengthy sample extraction and cleanup procedures [29,33,34]. On the other hand, nitro-group of IMD can be reduced electrochemically at a negative potential. Therefore, due to electro-activity of IMD, several electrochemical methods were reported for the detection of IMD using different electrodes modified by polymer–NPs composites [30]. The electrochemical sensor and biosensor platforms have emerged as powerful analytical methods to detect pesticides due to ease of detection and appreciable sensitivity [35]. However, to the best of our knowledge, there are no reports about the electroanalytical determination of IMD using electrodes modified by CS–TiO2 NPs nanocomposites. Due to interaction of CS with TiO2 NPs and the formation of interface layer which prevents the agglomeration of NPs at relative high concentration, we propose that such modified electrode will offer high surface area and enhance the accessibility of IMD to the recognition NPs sites. Thus, we expect that the agglomeration will be affected on sensibility of CS–TiO2 sensor for detection of imidacloprid. Based upon the above, this work aims to investigate the properties of CS–TiO2 films as a function of the concentration of NPs by considering an interface layer formation, and to shed light about the development of a sensor for electroanalytical detection of IMD. 2. Materials and Methods Chitosan (CS), medium molecular weight (ca. 350 kDa), deacetylation ca. 72%, acetic acid (99.7%), and TiO2 NPs with dimension between 20–40 nm was purchased from Chemours Co. (Mexico City, Mexico), and used as received. 2.1. Preparation of Films The films were cast by dissolving 1% w/w of chitosan in 1% w/w aqueous acetic acid solution with continuous magnetic stirring for 24 h. Different concentration of TiO2 nanoparticles were dispersed within the chitosan solution for 3 h. Then, each solution with different concentration of TiO2 nanoparticles was placed on the ultrasonic tip for 10 min with intervals of 3 min and pulses of 5 s to avoid heating the solution. This solution was poured into a Petri dish and allowed to evaporate at 60 °C for 18 h to obtain the chitosan acetate films. The neutralization of acetate film was done with aqueous ammonia solution 2 mol/dm3. The initial pH was ca. 3.7–4. For impedance measurements, CS–TiO2 films were gold-sputtered on both sides to serve as contacts. 2.2. Preparation of Working Electrode The glassy carbon electrodes (GCE) were hand polished with 0.3 and 0.1 µm alumina slurries, washed under ultrasound for 5 min with ethanol, and dried in air oven for 2 h at 60 °C. An aliquot of 7 µL of each chitosan–TiO2 solution prepared for films was taken. This aliquot was deposited on the surface of each GCE. The electrodes were dried in air for 3 h and a uniform film was formed over each electrode surface. In total, 6 electrodes were modified with CS–TiO2 at each TiO2 concentration, 1 electrode modified with CS, and the GCE electrode used as a target. 2.3. Characterization The crystalline structure of TiO2 NPs and CS–TiO2 films were tested by an X-ray diffractometer (Rigaku Dmax 2100, Austin, TX, USA) with Cu Kα radiation (λ = 0.154 nm). The interaction between CS functional groups with TiO2 NPs was analyzed by FTIR measurements on a Perkin Elmer Spectrum GX spectrophotometer using ATR (MIRacle™, Madison, WI, USA) sampling technique, with a diamond tip, in the range from 4000 to 1000 cm−1 at room temperature. The amount of free water was determined by thermogravimetric analysis (TGA) (TGA 4000—PerkinElmer; PerkinElmer, Inc., Waltham, MA, USA). Measurements were made in the dry air with a heating rate of 10 °C/min. Impedance measurements were carried out using Agilent 4249 A in the frequency range 40 Hz–100 MHz with an amplitude of AC voltage 100 mV at room temperature. The dielectric constant of nanocomposites has been calculated from impedance spectra using ZView program from the following relationship: ε = (C·d)/(ε0·S), where C is capacitance measured at the frequency 100 Hz, d and S are the thickness and area of samples, respectively. Film thickness was measured in each sample using micrometer Mitutoyo with resolution 1 mkm. Voltammetry cyclic measurements were performed using a potentiostat–galvanostat VoltaLab PGZ301 (Radiometer analytical), and a conventional three-electrode cell at room temperature. The CS/TiO2 modified GCE electrodes were used for electrochemical experiments as working electrode, Ag/AgCl electrode was used as reference electrode and platinum plaque was used as a counter electrode. The cell electrochemical was covered with black tape to avoid photocatalytic effect. Na2SO4 0.05 M was used as electrolyte and the imidacloprid concentrations were of 10, 50, 100, 250 and 500 ppm, this last concentration according with the imidacloprid solubility. The potential range was established between −1.8 and −0.1 V, scan rate of 150 mV/s and, prior to use, the working electrode was stabilized through 10 continuous repetitive cyclic voltammograms running, to obtain a stable and reproducible background current. All experiments were carried out with solutions previously deaerated with a stream of N2 gas bubbled in the solution for 10 min. 3. Results and Discussion 3.1. Morphology According to SEM measurements the most of NPs have dimension between 18 and 38 nm. SEM images recorded from of CS–TiO2 nanocomposites films with 10 wt.% and 30 wt.% show homogeneous distribution of NPs (Figure 1a,b) in CS matrix. In contrast, the films with 50 wt.% agglomeration is observed (Figure 1c). This behavior plays an important role in the explaining structural properties of nanocomposites and their application. 3.2. XRD Figure 2 shows XRD patterns of TiO2 NPs and CS–TiO2 NPs composite film with 40 wt.% of NPs. The XRD pattern of pure TiO2 NPs showed diffraction peaks and confirm the presence of anatase (JCPDS PDF#21-1272) and rutile phases (JCPDS PDF#21-1276) which often observed in TiO2 powder (see, for example [36,37]). In the pattern of CS–TiO2 films, additionally to diffraction peaks of anatase and rutile observed broad peak related to amorphous CS (between 8–18 degree). The quantification and refinement of the volume fraction of every crystalline phase was carried out by the Rietveld method implemented in the Fullprof software; this allows to calculate the volume fraction of every phase (68.5% of anatase and 31.5% of rutile). This estimation is important for the appropriate interpretation of dielectric measurements (see Section 3.5). Figure 2 XRD pattern of TiO2 NPs and CS–TiO2 film with 40 wt.% of NPs. 3.3. FTIR FTIR spectroscopy was used to observe the interactions between the chitosan and TiO2 NPs. In the case of neat CS, the spectrum shows the characteristics bands reported by other authors [38,39]: the broadband characteristic peak centered at 3322 cm−1 correspondents to the overlap of stretching vibration of -NH and -OH groups of CS, and peak at 1562 cm−1 related to bending vibration of NH2 group. In CS–TiO2 films, these peaks shift from 3322 cm−1 and 1562 cm−1 to 3269 and 1547 cm−1 wavenumbers (Figure 3). The shift of characteristic peaks in FTIR measurements are due to the interaction between CS reactive amino and hydroxyl groups with TiO2 NPs. Other studies have proposed a different mechanism of interaction including chelation [7,40,41]; the formation multiple coordination bonds between organic molecules and metals; interaction of CS side groups with Ti ions on TiO2 surface [42]; covalent interaction between CS and TiO2 [43]; electrostatic interaction between negative charge of CS carboxyl groups and TiO2 positive charge [44]. Another possible mechanism of the interaction between CS and NPs can be related to the dissociation of water on the surface sites or defects of TiO2 and the formation of two different OH functional groups at the surface of NPs [45,46]. These groups are responsible to interaction of TiO2 with lateral groups of CS. The most important thing for us is the presence of interaction between CS and TiO2 NPs, which is responsible for the formation of an interface layer surrounding NPs. The properties of this interphase layer will be discussed in the next section. In addition, a decrease in the intensity of the band centered at 3322–3269 cm−1 demonstrates a decrease in the water absorption capacity with an increase in the concentration of NPs in the films (decrease in intensity). Figure 3 FTIR spectra of neat CS film and CS–TiO2 films with 40 and 60 wt.% of NPs. 3.4. TGA The interaction between CS side groups with TiO2 NPs precludes water absorption capacity and it is supported by TGA measurements (Figure 3). Water absorption in CS is closely linked to the availability of amino and hydroxyl groups of CS interact via hydrogen bond with water molecules [47,48]. It was shown that the water content depends upon TiO2 NPs concentration (Figure 4) and decreases with increasing weight % of NPs (10.9% in neat CS and 7.1% in CS–TiO2 film with 60 wt.% of NPs, at the temperature 150 °C). As it was shown by FTIR analysis, CS lateral groups can be bond with OH groups on the surface of NPs that is responsible for decreasing of water content in the nanocomposite films. 3.5. Dielectric Measurements Due to the difference between dielectric constant of CS matrix and TiO2 NPs, the dielectric spectroscopy measurements can be used for the determination of parameters of the interface layer and the highest concentration of NPs before they agglomerate (decreasing effective surface area). In the literature have been proposed a three-phase model to describe the dielectric properties of polymer–dielectric NPs composites [49,50]. Here, the effective dielectric constant of such composite materials depends upon the εm of the polymer matrix, the εNPs of NPs, and the εint of interface layer between filler and the dielectric matrix. To describe such three-phase system, Refs. [49,50] introduce a parameter K, termed the interface volume constant, which accounted for the matrix–filler interaction strength as: (1) ∅int=K∅NPs∅m where Φint, ΦNPs and Φm are the volume fractions of interface phase, dielectric particles and polymer matrix, respectively. K depends upon the degree of particle clustering. A value of zero for K indicates that there is no interracial phase region between the NPs and polymer matrix. Positive values for K are related to an interaction between the polymer and NPs. This model has demonstrated that the dependence of the dielectric constant on NPs concentration is nonmonotonic and in dependence of interface volume fraction as function of NPs concentration can observed extremum due to an overlap of interface layers and NPs agglomeration. Before fitting, experimental dependence of effective dielectric constant of nanocomposite is necessary to recalculate the weight fraction of TiO2 NPs (Wt) in their volume fraction ΦNPs using next equation [51]:(2) ∅NPs=WtWt+ρTiO2ρCS1−Wt where, ρTiO2 and ρCS denote the TiO2 and CS density. The density of CS films is ca. 1.5 g cm−3 [52,53]. TiO2 NPs, according to XRD measurements, have two phases (68.5% of anatase with density ρanat = 3.78 g/cm3 and 31.5% of rutile with density ρrut = 4.23 g/cm3) [54]. Using simple mixture rule, the density of TiO2 NPs can be calculated [51]: (3) ρTiO2=0.685∗ρanat+0.315ρrut=3.92 g/cm3 Next important value for fitting experimental dependence is effective dielectric constant of TiO2 NPs (εNPs). This effective dielectric constant of TiO2 NPs can be estimated using Lichtenecher logarithmic model [55,56]:(4) logϵNPs=∅anatlogεanat+∅rutlogεrut where Φanat and Φrut represent the volume fraction of anatase and rutile, respectively. Taking data εanat = 75 and εrut = 105 [56], the effective dielectric constant of TiO2 NPs is equal 82.5. This value well correlates with dielectric constant (ca. 100) measures at the frequency 100 Hz on pellets consist of TiO2 NPs [57,58,59]. Therefore, dielectric constant of nanocomposites has been measured at the frequency 100 Hz. In this work, we fit the model of effective dielectric constant (5) ε=fεNPs,εint,εm,K,ϕNPs proposed in Refs. [35,36] to experimentally obtained values of effective dielectric constant of CS–TiO2 composite (ϵi)i=1N. The least squares fitting is performed using a standard function of genetic algorithm optimization in the Scilab [60]. The fitting error is specified as the sum of the squares of the differences between the dielectric constants predicted by the model and obtained in result of N measurements. (6) EK,εint =∑i=1NfεNPs,εint,εm,K,ϕNPs,i − ϵi2,   i=1…N The dependence of the fitting error E(K, εint) on interphase volume constant K and interphase dielectric constant εint is shown in Figure 5. The graph shows that the function is not unimodal. This fact helped determine the selection of the genetic algorithm as an optimization method. Dielectric constant of CS has been obtained from measurements on neat CS films. Dielectric constant of TiO2 was taken 82.5 (calculated using Equation (4)). Only the values of K and εint are the adjustable parameters. In the process of fitting the program found the optimum values of K and εint at the minimum value of fitting error. As a result of optimization, the fitted values are K = 15, εint = 19.7 and fitting error E(K, εint) = 1.25. Figure 6 shows the dependencies of the dielectric constant obtained in CS–TiO2 films at the frequency 100 Hz with different NPs concentration (points). Results of the referred fittings are shown on Figure 5 as continuous line. One can observe that this three-phase model fits well with the experimental results. Positive values of K mean that there is significant interface in CS–TiO2 interactions. The value of interface dielectric constant is less than ε of neat CS. The low value of εint can be related to the interaction of CS lateral groups with OH groups on the surface of NPs that is responsible for the decreasing of water content in the interface layer. However, the most important result of fitting is shown on the insert of Figure 6: dependence of interface volume fraction of nanocomposite on TiO2 NPs volume fraction. This dependence demonstrates maximum at the concentration of TiO2 at ca. 39 wt.%. At higher concentration of NPs, interface volume fraction decreases due to agglomeration of NPs, which is well observed in SEM measurements (Figure 1c). Thus, it can be assumed that an electrode with CS–TiO2 films with 40 wt.% NPs will demonstrate better sensitivity to imidacloprid, since at a higher concentration, due to NPs agglomeration, the surface area of NPs will decrease. 3.6. Cyclic Voltammetry Measurements Figure 7 shows the redox behavior voltamperograms of imidacloprid measured on the electrodes modified by CS–TiO2 nanocomposites films with different concentration of TiO2 NPs. For all redox behavior, the plot of peak current (Ip) versus the respective concentration of IMD was found to be linear in the range 10–500 ppm (3.9 × 10−5–2 × 10−3 mol/L) as shown in the insets in Figure 5. The cathodic peak can be observed at a value of −1.42 ± 0.048 V for GCE and −1.36 ± 0.03 V for the GCE/CS electrode. It can be observed that in the electrode of low concentration of TiO2 nanoparticles at 10 wt.%, it is evident that, during the cathodic scan, a single reduction peak was observed at a potential of −1.3 ± 0.025 V. The noted cathodic process is derived from the nitro-group irreversible reduction of imidacloprid [33,61,62,63]. In the case of the electrode GCE/CS–TiO2 20 wt.% only the reduction of the IMD to 500 ppm was observed in −1.42 V and its oxidation value was presented for the lowest concentrations in −1.52 V. For electrodes with CS–TiO2 with 30, 40, 50 and 60 wt.% the reduction and oxidation of the imidacloprid was observed, this indicates the reversibility of the electrode in the behavior redox measurement. However, for electrodes with CS–TiO2 30 wt.%, CS–TiO2 40% wt.% and CS–TiO2 50 wt.%, one can see the appearance of two peaks of reduction: one at −1.663 V and another at −1.351 V. This the second peak is clearly observed at concentrations of 50, 100 and 250 ppm of IMD. At the electrode with 40 wt.% of TiO2 nanoparticles, the second peak of reduction is more visible except for the low concentration of 10 ppm. This second reduction peak had already been reported by Navalon et al. [64] through of differential pulse polarographic method. Thus, it was able to propose an overall reduction mechanism for each of the two imidacloprid reduction peaks. For the first peak, the nitro group of the imidacloprid molecule takes four electrons to give the corresponding hydroxylamine derivative and then in the second reduction peak this compound takes two electrons to be transformed in the corresponding amine derivative. The limit of detection (LOD) and the limit of quantitation (LOQ) (Table 1) were calculated from using the equations: LOD = 3 sd/m and LOQ = 10 sd/m, where sd is the standard deviation of the intercept and m is the slope of the calibration curve [65]. The reported detection limit was in the same order for all electrodes and compared to other modified electrodes reported in the literature [33,65,66,67,68]. However, the electrode modified by CS–TiO2 film with 40 wt.% presents the best sensitivity to imidacloprid compare with another investigate (Table 2) modified electrodes because it detects two peaks of reduction, and its current values are higher above −150 µA. Hg(Ag) FE, Silver-amalgam film electrode; BDD, Boron-Doped Diamond; CPE, carbon-paste electrode; GCE, glassy carbon electrode; nAgnf/nTiO2nf/GCE, nanosilver Nafion®/nanoTiO2 Nafion® modified glassy carbon electrode; PCz/CRGO/GCE, poly(carbazole)/chemically reduced graphene oxide modified glassy carbon electrode; BiFE, Bismuth-film electrode. 4. Conclusions In this paper, for the first time, we propose the potential application of chitosan–TiO2 nanocomposites for the development of an electroanalytical sensor to detect imidacloprid with high sensitivity. Preliminary studies of imidacloprid detection show that the best results are obtained on a modified electrode with CS–TiO2 NPs with 40 wt.% of NPs. According to dielectric spectroscopy measurements on the nanocomposite, the dependence of the dielectric constant on NPs concentration is nonmonotonic. By using a three-phase model which includes: (1) CS matrix, (2) TiO2 NPs and (3) interface layer between NPs and CS matrix and by fitting dielectric spectroscopy measurements, the interface volume fraction was calculated. The highest calculated interface volume fraction is 39 wt.% of NPs which is practically the same value at which the best performance of the nanocomposite is observed. Our results strongly suggest that the interface layer is responsible for the good dispersion of NPs in chitosan matrix without agglomeration. At higher NPs concentration, the interface volume fraction decreases due to their overlap that leads to NPs agglomeration which ultimately decreases their effective surface area. TiO2 nanoparticles serve as molecular recognition sites for electroanalytical imidacloprid detection; therefore, their agglomeration decreases the sensitivity of imidacloprid detection. The construction of such a sensor requires additional investigation. However, the methodology presented in this work, which allows to determine the optimum NPs concentration, may prove useful in the design and optimization of polymer-based nanocomposites for development nanocomposite for different applications. Acknowledgments The authors are grateful to J.A. Muñoz Salas for technical assistance in electrical measurements, R.A. Mauricio-Sánchez for assistance in FTIR measurements, M.A. Hernandez Landaverde for assistance in XRD measurements, E. Urbina Alvarez for assistance in SEM measurements and Ma. Carmen Delgado Cruz for assistance in TGA measurements. Author Contributions B.E.C.: investigation, writing—original draft preparation; E.P.: conceptualization, investigation, writing—original draft preparation, review and editing; G.L.-B.: writing—review and editing; Y.K.: software, simulation. All authors have read and agreed to the published version of the manuscript. Funding This research was funded and supported by CONACYT, Mexico (grant A1-S-9557). Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 SEM images of CS–TiO2 nanocomposites films with (a) 10 wt.% of TiO2 NPs, (b) 30 wt.% of TiO2 NPs and (c) 50 wt.% of TiO2 NPs. Insert on Figure 1c shows NPs distribution. Figure 4 Dependence of weight loss in CS–TiO2 films on NPs wt.% measured at the temperature 150 °C. Insert shows TGA measurements of pure CS and CS–TiO2 NPs films with 10 and 30 wt.% of NPs. Figure 5 Plot of fitting error E(K, εint) versus interphase volume constant K and interphase dielectric constant εint. Figure 6 Dependence of the dielectric constant obtained in CS–TiO2 films at the frequency 100 Hz with different NPs concentration (points). Continuous line shows the results of fitting. Bars represent the standard deviation calculated from measurements on 4 samples. Insert shows the dependence of interface volume fraction on TiO2 volume fraction calculated using models proposed in [49,50]. Figure 7 Cyclic voltammograms obtained for IMD on electrodes: (a) GCE; (b) GCE modified by CS; (c) GCE modified by CS-TiO2 film with 10 wt.% of NPs; (d) GCE modified by CS-TiO2 film with 20 wt.% of NPs; (e) GCE modified by CS–TiO2 film with 30 wt.% of NPs; (f) GCE modified by CS-TiO2 film with 40 wt.% of NPs; (g) GCE modified by CS–TiO2 film with 50 wt.% of NPs; (h) GCE modified by CS-TiO2 film with 60 wt.% of NPs. Measurements were carried out at 150 mVs−1 in 0.05 M Na2SO4 as supporting electrolyte, pH = 7. Each cycle corresponding to the different concentrations: a-10, b-50, c-100, d-250, e-500 ppm. Inserts show relationship between peak current of reduction and/or oxidation vs IMD concentration. polymers-14-01686-t001_Table 1 Table 1 Analytical parameters for the determination of IMD on different electrodes. Electrode Cathodic Peak (V) Anodic Peak (V) Linear Range (mol/L) Cathodic Anodic LOD (mol/L) LOQ (mol/L) R2 LOD (mol/L) LOQ (mol/L) R2 GCE −1.42 ± 0.048 ---- 3.9 × 10−5–2 × 10−3 1.82 × 10−4 6.08 × 10−4 98.5 --- --- --- GCE/CS −1.36 ± 0.030 --- 3.9 × 10−5–2 × 10−3 2.76 × 10−4 9.20 × 10−4 96.7 --- --- --- GSE/CS-TiO2 10% −1.37 ± 0.026 --- 3.9 × 10−5–2 × 10−3 2.60 × 10−4 8.68 × 10−4 97 --- --- --- GSE/CS-TiO2 20% −1.42 ± 0.030 −1.52 ± 0.032 3.9 × 10−5–2 × 10−3 --- --- --- 2.96 × 10−4 9.85 × 10−4 90.5 GSE/CS-TiO2 30% −1.34 ± 0.036 −1.51 ± 0.046 3.9 × 10−5–2 × 10−3 2.04 × 10−4 6.79 × 10−4 98.2 1.42 × 10−4 4.75 × 10−4 97.7 GSE/CS-TiO2 40% −1.34 ± 0.043 −1.52 ± 0.027 3.9 × 10−5–2 × 10−3 2.74 × 10−4 9.14 × 10−4 96.7 6.2 × 10−4 2.07 × 10−3 84.7 GSE/CS-TiO2 50% −1.32 ± 0.028 −1.54 ± 0.022 3.9 × 10−5–2 × 10−3 2.53 × 10−4 8.44 × 10−4 97.2 1.67 × 10−4 5.58 × 10−4 98.8 GSE/CS-TiO2 60% −1.30 ± 0.030 −1.52 ± 0.046 3.9 × 10−5–2 × 10−3 1.46 × 10−4 4.88 × 10−4 99.1 1.44 × 10−4 4.79 × 10−4 97.6 polymers-14-01686-t002_Table 2 Table 2 Comparison of different electrodes in the detection of IMD. Method Electrode Linear Range (mmol L–1) LOD (mmol L–1) LOQ (mmol L–1) References SWV Hg(Ag) FE 3.55–185.6 1.05 3.6 [69] SWV BDD 30–200 8.6 28.6 [33] DPV CPE 6.7–117.4 2.04 6.8 [70] DPV nAgnf/nTiO2nf/GCE 0.5–3.5 0.25 0.8 [66] DPV PCz/CRGO/GCE 3–10 0.44 1.5 [71] DPV BiFE 9.5–200 2.9 – [72] CV GCE 10.9–1956 30.1 101.6 [62] CV CPE 1–7 0.63 2.1 [70] CV PCz/CRGO/GCE 3–10 0.22 0.7 [72] CV GCE/CS-TiO2 40% 0.039–2 0.6 2.1 This work Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Babaei-Ghazvini A. Acharya B. Korber D.R. Antimicrobial Biodegradable Food Packaging Based on Chitosan and Metal/Metal-Oxide Bio-Nanocomposites: A Review Polymers 2021 13 2790 10.3390/polym13162790 34451327 2. Bui V.K.H. Park D. Lee Y.C. Chitosan Combined with ZnO, TiO2 and Ag Nanoparticles for Antimicrobial Wound Healing Applications: A Mini Review of the Research Trends Polymers 2017 9 21 10.3390/polym9010021 30970696 3. Gobi R. Ravichandiran P. Babu R.S. Yoo D.J. Biopolymer and Synthetic Polymer-Based Nanocomposites in Wound Dressing Applications: A Review Polymers 2021 13 1962 10.3390/polym13121962 34199209 4. Razzaz A. Ghorban S. Hosayni L. Irani M. Aliabadi M. Chitosan nanofibers functionalized by TiO2 nanoparticles for the removal of heavy metal ions J. Taiwan Inst. Chem. Eng. 2016 58 333 343 10.1016/j.jtice.2015.06.003 5. Wang X. Li Z. Wu Y. Guo H. Zhang X. Yang Y. Mu H. Duan J. Construction of a Three-Dimensional Interpenetrating Network Sponge for High-Efficiency and Cavity-Enhanced Solar-Driven Wastewater Treatment ACS Appl. Mater. Interfaces 2021 13 10902 10915 10.1021/acsami.0c21690 33629587 6. Khan R. Dhayal M. Nanocrystalline bioactive TiO2–chitosan impedimetric immunosensor for ochratoxin-A Electrochem. Commun. 2008 10 492 495 10.1016/j.elecom.2008.01.013 7. Ananpattarachai J. Kumket P. Tung T.V. Kajitvichyanukul P. Chromium (VI) removal using nano-TiO2/chitosan film in photocatalytic system Int. J. Environ. Waste Manag. 2015 16 55 70 10.1504/IJEWM.2015.070481 8. Ashraf M.A. Peng W. Zare Y. Rhee K.Y. Effects of Size and Aggregation/Agglomeration of Nanoparticles on the Interfacial/Interphase Properties and Tensile Strength of Polymer Nanocomposites Nanoscale Res. Lett. 2018 13 214 10.1186/s11671-018-2624-0 30019092 9. Zare Y. Rhee K.Y. Hui D. Influences of nanoparticles aggregation/agglomeration on the interfacial/interphase and tensile properties of nanocomposites Compos. B 2017 122 41 46 10.1016/j.compositesb.2017.04.008 10. Wongaree M. Chiarakorn S. Chuangchote S. Photocatalytic Improvement under Visible Light in TiO2 Nanoparticles by Carbon Nanotube Incorporation J. Nanomater. 2015 2015 689306 10.1155/2015/689306 11. Qin G. Zhang H. Wang C. Ultrasmall TiO2 nanoparticles embedded in nitrogen doped porousgraphene for high rate and long life lithium ion batteries J. Power Sour. 2014 272 491 500 10.1016/j.jpowsour.2014.08.105 12. Jastrzębska A. Kurtycz P. Olszyna A. Karwowska E. Miaśkiewicz-Pęska E. Załęska-Radziwiłł M. Doskocz N. Basiak D. The Impact of Zeta Potential and Physicochemical Properties of TiO2-Based Nanocomposites on Their Biological Activity Int. J. Appl. Ceram. Technol. 2015 12 1157 1173 10.1111/ijac.12340 13. Pandey J.C. Singh M. Dielectric polymer nanocomposites: Past advances and future prospects in electrical insulation perspective SPE Polym. 2021 2 236 256 10.1002/pls2.10059 14. Sattar M.A. Interface Structure and Dynamics in Polymer-Nanoparticle Hybrids: A Review on Molecular Mechanisms Underlying the Improved Interfaces ChemistrySelect 2021 6 5068 5096 10.1002/slct.202100831 15. Pitsa D. Danicas D. Interfaces features in polymer nanocomposites: A review of proposed models Nano 2011 6 497 508 10.1142/S1793292011002949 16. Sun W. Mao J. Wang S. Zhang L. Cheng Y. Review of recent advances of polymer based dielectrics for high-energy storage in electronic power devices from the perspective of target applications Front. Chem. Sci. Eng. 2021 15 18 34 10.1007/s11705-020-1939-4 17. Nioua Y. Aribou N. Boukheir S. Achour M.E. Costa L.C. Interphase approach for modeling the DC conductivity of diverse allotropic types of carbon-reinforced polymer composites J. Reinf. Plast. Comp. 2020 1 10.1177/0731684420973264 18. Peng W. Rhim S. Zare Y. Rhee K.Y. Effect of “Z” Factor for Strength of Interphase Layers on the Tensile Strength of Polymer Nanocomposites Polym. Compos. 2019 40 1117 1122 10.1002/pc.24813 19. Zare Y. Rhee K.Y. Dependence of Z Parameter for Tensile Strength of Multi-Layered Interphase in Polymer Nanocomposites to Material and Interphase Properties Nanoscale Res. Lett. 2017 12 42 10.1186/s11671-017-1830-5 28097595 20. Zare Y. Rhee K.Y. Simulation of Young’s modulus for clay-reinforced nanocomposites assuming mechanical percolation, clay-interphase networks and interfacial linkage J. Mater. Res. Technnol. 2020 9 12473 12483 10.1016/j.jmrt.2020.08.097 21. Zare Y. Rhee K.Y. Simulation of tensile modulus of polymer carbon nanotubes nanocomposites in the case of incomplete interfacial bonding between polymer matrix and carbon nanotubes by critical interfacial parameters Polymer 2020 191 122260 10.1016/j.polymer.2020.122260 22. Ginzburg V.V. Hall L.M. Theory and Modeling of Polymer Nanocomposites Springer New York, NY, USA 2021 23. Kashfipour M.A. Mehra N. Zhu J. A review on the role of interface in mechanical, thermal, and electrical properties of polymer composites Adv. Comp. Hybrid Mat. 2018 1 415 439 10.1007/s42114-018-0022-9 24. Popov I. Carroll B. Bocharova V. Genix A.C. Cheng S. Khamzin A. Kisliuk A. Sokolov A.P. Strong Reduction in Amplitude of the Interfacial Segmental Dynamics in Polymer Nanocomposites Macromolecules 2020 53 4126 4135 10.1021/acs.macromol.0c00496 25. Bailey E.J. Winey K.I. Dynamics of polymer segments, polymer chains, and nanoparticles in polymer nanocomposite melts: A review Prog. Polym. Sci. 2020 105 101242 10.1016/j.progpolymsci.2020.101242 26. Cheng S. Carroll B. Bocharova V. Carrillo J.M.Y. Sumpter B.G. Sokolov A.P. Focus: Structure and dynamics of the interfacial layer in polymer nanocomposites with attractive interactions J. Chem. Phys. 2017 146 203201 10.1063/1.4978504 28571333 27. Nikfar N. Esfandiar M. Shahnazari M.R. Mojtahedi N. Zare Y. The reinforcing and characteristics of interphase as the polymer chains adsorbed on the nanoparticles in polymer nanocomposites Colloid Polym. Sci. 2017 295 2001 2010 10.1007/s00396-017-4164-z 28. Ali I. Suhail M. Alothmanc Z.A. Alwarthanc A. Recent advances in syntheses, properties and applications of TiO2 nanostructures RSC Adv. 2018 8 30125 30147 10.1039/C8RA06517A 29. Ghodsi J. Rafati A.A. A novel molecularly imprinted sensor for imidacloprid pesticide based on poly(levodopa) electro-polymerized/TiO2 nanoparticles composite Anal. Bioanal. Chem. 2018 410 7621 7633 10.1007/s00216-018-1372-4 30267274 30. Mahmoudpour M. Torbati M. Mousavi M.M. de la Guardia M. Dolatabadi J.E.N. Nanomaterial-based molecularly imprinted polymers for pesticides detection: Recent trends and future prospects TrAC Trends Anal. Chem. 2020 129 115943 10.1016/j.trac.2020.115943 31. Chao S.L. Casida J.E. Interaction of Imidacloprid Metabolites and Analogs with the Nicotinic Acetylcholine Receptor of Mouse Brain in Relation to Toxicity Pestic. Biochem. Physiol. 1997 58 77 88 10.1006/pest.1997.2284 32. Sumon K.A. Ritika A.K. Peeters E.T. Rashid H. Bosma R.H. Rahman M.S. Fatema M.K. Van den Brink P.J. Effects of imidacloprid on the ecology of sub-tropical freshwater microcosms Environ. Pollut. 2018 236 432 441 10.1016/j.envpol.2018.01.102 29414368 33. Brahim B. Determination of ultra-trace amounts of neonicotinoid insecticide imidacloprid by cyclic and square wave voltammetric methods using pretreated glassy carbon electrode Glob. Nest J. 2018 20 628 636 10.30955/gnj.002509 34. Umapathi R. Sonwal S. Lee M.J. Rani G.M. Lee E.-S. Jeon T.-J. Kang S.-M. Oh M.-H. Huh Y.S. Colorimetric based on-site sensing strategies for the rapid detection of pesticides in agricultural foods: New horizons, perspectives, and challenges Coord. Chem. Rev. 2021 446 214061 10.1016/j.ccr.2021.214061 35. Umapathi R. Ghoreishian S.M. Sonwal S. Rani G.M. Huh Y.S. Portable electrochemical sensing methodologies for on-site detection of pesticide residues in fruits and vegetables Coord. Chem. Rev. 2022 453 214305 10.1016/j.ccr.2021.214305 36. Anjali K.K. Maheswari A.U. Sivakumar M. Size Dependent Dielectric Properties of TiO2 Nanoparticles Key Eng. Mat. 2020 833 147 151 10.4028/www.scientific.net/KEM.833.147 37. Mallakpour S. Madania M. Effect of Functionalized TiO2 on Mechanical, Thermal and Swelling Properties of Chitosan-Based Nanocomposite Films Polym. Plast. Technol. Eng. 2015 54 1035 1042 10.1080/03602559.2014.974194 38. Mauricio-Sánchez R.A. Salazar R. Luna-Bárcenas G. Mendoza-Galván A. FTIR spectroscopy studies on the spontaneous neutralization of chitosan acetate films by moisture conditioning Vib. Spectrosc. 2018 94 1 6 10.1016/j.vibspec.2017.10.005 39. Lu S. Song X. Cao D. Chen Y. Yao K. Preparation of Water-Soluble Chitosan J. Appl. Polym. Sci. 2004 91 3497 3503 10.1002/app.13537 40. Karthikeyana K.T. Nithya A. Jothivenkatachalam K. Photocatalytic and antimicrobial activities of chitosan-TiO2nanocomposite Int. J. Biol. Macromol. 2017 104 1762 1773 10.1016/j.ijbiomac.2017.03.121 28342756 41. Palacio-Marquez A. Ramírez-Estrada C.A. Gutierrez-Ruelas N.J. Sanchez E. Ojeda-Barrios D.L. Chavez-Mendoza C. Sida-Arreola J.P. Efficiency of foliar application of zinc oxide nanoparticles versus zinc nitrate complexed with chitosan on nitrogen assimilation, photosynthetic activity, and production of green beans (Phaseolus vulgaris L.) Sci. Hortic. 2021 288 110297 10.1016/j.scienta.2021.110297 42. Ramasamy S. Organic photosensitizers containing fused indoleimidazole ancillary acceptor with triphenylamine donor moieties for efficient dye-sensitized solar cells Int. J. Hydrog. Energy 2021 46 3475 3483 10.1016/j.ijhydene.2020.10.248 43. Cheng L. Hao Y. Song X. Chang Q. Effects of Chitosan and Nano Titanium Dioxide on the Mechanical, Physicochemical and Antibacterial Properties of Corn Starch Films J. Macromol. Sci. Part B 2021 60 616 630 10.1080/00222348.2021.1889123 44. Hussein E.M. Desoky W.M. Hanafy M.F. Guirguis O.W. Effect of TiO2 nanoparticles on the structural configurations and thermal, mechanical, and optical properties of chitosan/TiO2 nanoparticle composites J. Phys. Chem. Solids 2021 152 109983 10.1016/j.jpcs.2021.109983 45. Orlando F. Artiglia L. Yang H. Kong X. Roy K. Waldner A. Chen S. Bartels-Rausch T. Ammann M. Disordered Adsorbed Water Layers on TiO2 Nanoparticles under Subsaturated Humidity Conditions at 235 K J. Phys. Chem. Lett. 2019 10 7433 7438 10.1021/acs.jpclett.9b02779 31725306 46. Soria F.A. Di Valentin C. Reactive molecular dynamics simulations of hydration shells surrounding spherical TiO2 nanoparticles: Implications for proton-transferreactions Nanoscale 2021 13 4151 4166 10.1039/D0NR07503E 33576363 47. Gonzalez-Campos B. Prokhorov E. Luna-Barcenas G. Sanchez I.C. Lara-Romeo J. Mendoza-Durante M.E. Villaseñor F. Guevara-Olvera L. Chitosan/Silver Nanoparticles Composite: Molecular Relaxations Investigation by Dynamic Mechanical Analysis and Impedance Spectroscopy J. Polym. Sci. B 2010 48 739 748 10.1002/polb.21941 48. Corazzari I. Nistico R. Turci F. Faga M.G. Franzoso F. Tabasso S. Magnacca G. Advanced physico-chemical characterization of chitosan means of TGA coupled on-line with FTIR and GCMS: Thermal degradation and water adsorption capacity Polym. Degrad. Stab. 2015 112 1121 1129 10.1016/j.polymdegradstab.2014.12.006 49. Vo H.T. Shi F.G. Towards model-based engineering of optoelectronic package materials: Dielectric constant modeling Microelectron. J. 2002 33 409 415 10.1016/S0026-2692(02)00010-1 50. Todd M.G. Shi F.G. Characterizing the interphase dielectric constant of polymer composite materials: Effect of chemical coupling agents J. Appl. Phys. 2003 94 4551 4557 10.1063/1.1604961 51. Rafiee M.A. Rafiee J. Wang Z. Song H. Yu Z.Z. Koratkar N. Enhanced Mechanical Properties of Nanocomposites at Low Graphene Content ACS Nano 2009 3 3884 3890 10.1021/nn9010472 19957928 52. Nunthanid J. Laungtana-anan M. Sriamornsak P. Limmatvapirat S. Puttipipatkhachorn S. Lim L.Y. Khor E. Characterization of chitosan acetate as a binder for sustained release tablets J. Control. Release 2004 99 15 26 10.1016/j.jconrel.2004.06.008 15342177 53. Mezina E.A. Lipatova I.M. Losev N.V. Effect of mechanical activation on rheological and film-forming properties of suspensions of barium sulfate in chitosan solutions Russ. J. Appl. Chem. 2011 84 486 490 10.1134/S1070427211030268 54. Haynes W.M. CRC Handbook of Chemistry and Physics 96th ed. CRC Press Boca Raton, FL, USA 2015 55. Kim K.S. Shim S.H. Kim S. Yoon S.O. Low temperature and microwave dielectric properties of TiO2/ZBS glass composites Ceram. Int. 2010 36 1571 1575 10.1016/j.ceramint.2010.02.030 56. Houng B. Wu S.-J.J. Lu S.H. Chien W.C. Microstructure and properties evaluation of TiO2 ceramics with multi-oxides glass additions Ceram. Int. 2014 40 3731 3736 10.1016/j.ceramint.2013.09.049 57. Wypych A. Bobowska I. Tracz M. Opasinska A. Kadlubowski S. Krzywania-Kaliszewska A. Grobelny J. Wojciechowski P. Dielectric Properties and Characterisation of Titanium Dioxide Obtained by Different Chemistry Methods J. Nanomater. 2014 2014 124814 10.1155/2014/124814 58. Sagadevan S. Synthesis and electrical properties of TiO2 nanoparticles using a wet chemical technique Am. J. Nanosci. Nanotechnol. 2013 1 27 30 10.11648/j.nano.20130101.16 59. Kumar N.S. Naveen Kumar S.K. Yesappa L. Structural, optical and conductivity study of hydrothermally synthesized TiO2 nanorods Mater. Res. Express 2020 7 015071 10.1088/2053-1591/ab691f 60. Campbell S.L.V. Chancelier J.P. Nikoukhah R. Modeling and Simulation in Scilab/Scicos Springer Science and Business Media New York, NY, USA 2006 61. Ben Brahim M. Ammar H.B. Abdelhédi R. Samet Y. Electrochemical behavior and analytical detection of Imidacloprid insecticide on a BDD electrode using square-wave voltammetric method Chin. Chem. Lett. 2016 27 666 672 10.1016/j.cclet.2015.12.032 62. Guzsvány V. Gaál F. Bjelica L. Ökrész S. Voltammetric determination of imidacloprid and thiamethoxam J. Serb. Chem. Soc. 2005 70 735 743 10.2298/JSC0505735G 63. Paiva W.D.A. Oliveira T.M.B.F. Sousa C.P. Neto P.D.L. Correia A.N. Morais S. Silva D.R. Castro S.S.L. Electroanalysis of Imidacloprid Insecticide in River Waters Using Functionalized Multi-Walled Carbon Nanotubes Modified Glassy Carbon Electrode J. Electrochem. Soc. 2018 165 B431 B435 10.1149/2.1541809jes 64. Navalón A. El-Khattabi R. González-Casado A. Vilchez J.L. Differential-pulse polarographic determination of the insecticide imidacloprid in commercial formulations Microchim. Acta 1999 130 261 265 Available online: https://link.springer.com/article/10.1007%2FBF01242914 (accessed on 28 January 2022) 10.1007/BF01242914 65. Ahuja S. Scypinski S. Validation of analytical test methods Separation Science and Technology Academic Press Cambridge, MA, USA 2011 Volume 10 429 457 10.1016/B978-0-12-375680-0.00011-5 66. Kumaravel A. Chandrasekaran M. Electrochemical determination of imidacloprid using nanosilver Nafion®/nanoTiO2 Nafion® composite modified glassy carbon electrode Sens. Actuators B Chem. 2011 158 319 326 10.1016/j.snb.2011.06.028 67. Li X. Kan X. A ratiometric strategy -based electrochemical sensing interface for the sensitive and reliable detection of imidacloprid Analyst 2018 143 2150 2156 10.1039/C8AN00111A 29667693 68. Ajermoun N. Aghris S. Farahi A. Lahrich S. Saqrane S. Bakasse M. El Mhammedi M.A. Electrochemical reduction of neonicotinoids insecticides catalysed by metallic silver: Case of the detection of imidacloprid in tomato and orange juices Int. J. Environ. Anal. Chem. 2019 101 585 597 10.1080/03067319.2019.1669586 69. Guzsvany V. Petrovic J. Krstic J. Papp Z. Putek M. Bjelica L. Bobrowski A. Abramović B. Renewable silver-amalgam film electrode for voltammetric monitoring of solar photodegradation of imidacloprid in the presence of Fe/TiO2 and TiO2 catalysts J. Electroanal. Chem. 2013 699 33 39 10.1016/j.jelechem.2013.04.003 70. Papp Z. Švancara I. Guzsvány V. Vytřas K. Gaál F. Voltammetric determination of Imidacloprid insecticide in selected samples using a carbon paste electrode Microchim. Acta 2009 166 169 175 10.1007/s00604-009-0181-z 71. Lei W. Wu Q.J. Si W.M. Electrochemical determination of imidacloprid using poly(carbazole)/chemically reduced graphene oxide modified glassy carbon electrode Sens. Actuators B Chem. 2013 183 102 109 10.1016/j.snb.2013.03.118 72. Guzsvány V. Kádár M. Papp Z. Bjelica L. Gaál F. Tóth K. Monitoring of photo-catalytic degradation of selected neonicotinoid insecticides by cathodic voltam-metry with a bismuth film electrode Electroanalysis 2008 20 291 300 10.1002/elan.200704057

PMC009xxxxxx/PMC9099657.txt

==== Front Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods11091229 foods-11-01229 Article Lactobacillus fermentum Stimulates Intestinal Secretion of Immunoglobulin A in an Individual-Specific Manner Mei Liya 12 Chen Ying 12 Wang Jialiang 12 Lu Jian 3* Zhao Jianxin 124 https://orcid.org/0000-0002-9192-4684 Zhang Hao 12456 https://orcid.org/0000-0003-2007-2143 Wang Gang 124* Chen Wei 125 Recio Isidra Academic Editor 1 State Key Laboratory of Food Science and Technology, Jiangnan University, Wuxi 214122, China; 6190112082@stu.jiangnan.edu.cn (L.M.); Manju_ying@163.com (Y.C.); wjl120587@126.com (J.W.); zhaojianxin@jiangnan.edu.cn (J.Z.); zhanghao61@jiangnan.edu.cn (H.Z.); chenwei66@jiangnan.edu.cn (W.C.) 2 School of Food Science and Technology, Jiangnan University, Wuxi 214122, China 3 Department of Gastroenterology, Affiliated Wuxi No. 2 People’s Hospital of Nanjing Medical University, Wuxi 214122, China 4 (Yangzhou) Institute of Food Biotechnology, Jiangnan University, Yangzhou 225004, China 5 National Engineering Center of Functional Food, Jiangnan University, Wuxi 214122, China 6 Wuxi Translational Medicine Research Center and Jiangsu Translational Medicine Research Institute Wuxi Branch, Wuxi 214122, China * Correspondence: ljhaimen@163.com (J.L.); wanggang@jiangnan.edu.cn (G.W.); Tel.: +86-510-85912155 (G.W.) 25 4 2022 5 2022 11 9 122905 3 2022 23 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Immunoglobulin A (IgA), as the most secreted immunoglobulin in the intestine, plays an irreplaceable role in mucosal immunity regulation. Previous studies have indicated that Lactobacillus showed strain specificity in stimulating the secretion of IgA through intestinal mucosal lymphocytes. The reason for this phenomenon is not clear. The current studies have been aimed at exploring the effect of a strain on the secretion of IgA in the host’s intestine, but the mechanism behind it has not been seriously studied. Based on this, we selected five strains of Lactobacillus fermentum isolated from different individuals to determine whether there are intraspecific differences in stimulating the secretion of IgA from the intestinal mucosa. It was found that IgA concentrations in different intestinal segments and faeces induced by L. fermentum were different. 12-1 and X6L1 strains increased the secretion of IgA by the intestine significantly. In addition, different strains of L. fermentum were also proven to have different effects on the host gut microbiota but no significant effects on IgA-coated microbiota. Besides, it was speculated that different strains of L. fermentum may act on different pathways to stimulate IgA in a non-inflammatory manner. By explaining the differences of IgA secretion in the host’s intestine tract stimulated by different strains of L. fermentum, it is expected to provide a theoretical basis for the stimulation of intestinal secretion of IgA by Lactobacillus and a new direction for exploring the relationship between Lactobacillus and human immunity. probiotics Immunoglobulin A immunity gut microbiota Immunoglobulin A-coated bacteria ==== Body pmc1. Introduction There are approximately 1014 microorganisms in the intestinal cavity, including bacteria, fungi, viruses and protozoa [1]. The gut microbiota promotes the development and response of the host mucosal immune system, enhances the close association between intestinal epithelial cells and antagonises pathogens [2]. The intestinal mucus layer is composed of mucin secreted by goblet cells, antimicrobial peptides secreted by intestinal epithelial cells and secretory immunoglobulin A (IgA) produced by B cells. IgA can effectively inhibit the adhesion and colonization of some bacteria in the intestinal epithelium [3]. The intestine is not only the site of food digestion and absorption but is also an important immune organ [4]. Through coevolution, the mammalian host and microorganism have formed a mutually beneficial symbiotic relationship. The host provides nutrition and a living environment for microorganisms, while microorganisms promote the steady-state balance of the intestine by participating in and regulating a series of physiological activities of the host [5]. Evidence suggests that a balance between intestinal bacteria and the mucosal immune system is an important factor in the development of intestinal mucosal inflammation [6,7]. The mucosal immune system elicits an effective defence against pathogens that invade the mucosa. Furthermore, proper tolerance needs to be developed against microbes that live in symbiosis with organisms in the mucosa [8]. IgA is the most commonly produced immunoglobulin in mammals [9]. The specific immune response in the gut is mainly through the production of IgA, IgM and IgG. Among them, IgA can inhibit the colonization and proliferation of some microorganisms in the mucus layer. Also, IgA can combine with some microorganisms to varying degrees to regulate the gut microbiota. There is no substitute for the role of IgA in intestinal immunity. At present, there have been many reports of probiotics that can regulate the level of IgA in the intestinal tract. It is known that L. plantarum YU could stimulate the secretion of IgA in Peyer’s patches by stimulating Th1 immune responses [10]. Pediococcus acidilactici K15 promoted the secretion of IgA in the oral mucosa, and this response was mainly induced via IL-10 [11]. L. fermentum CECT5716 enhanced IgA secretion in rats during pregnancy [12]. L. fermentum UCO-979C increased intestinal IgA levels while activating TLR4 [13]. The current studies mainly focus on how pathogenic bacteria stimulate host immunity and then affect the production of IgA. There is a lack of in-depth mechanism studies on how Lactobacillus promotes IgA secretion. In addition, it has been reported that the content of IgA in human faeces can be increased by diets containing high concentrations of acetic acid [14]. However, it is not known whether the SCFAs produced by the host’s gut microbiota are responsible for stimulating the host’s gut to secrete IgA. Most studies on the effect of Lactobacillus on intestinal IgA secretion remain on inter-species differences, and there are few studies on whether there are intra-species differences. In the previous study, we used forty strains of different species of Lactobacillus to conduct in vitro cell experiments and animal experiments (unpublished data). The results showed that among the various species of Lactobacillus, L. fermentum had the best potential for stimulating intestinal secretion of IgA. In this study, we selected five strains of L. fermentum derived from different human donors to investigate whether the hosts had differences in their corresponding response to IgA. 2. Materials and Methods 2.1. Bacterial Treatment The five strains of L. fermentum (L. fermentum X6L1 [CCFM1225], L. fermentum 12-1 [CCFM1226], L. fermentum S24-1, L. fermentum 24M1 and L. fermentum 20M5) used in this study were isolated from faeces of healthy individuals and stored at the culture collection of food microorganisms in Jiangnan University (Wuxi, Jiangsu, China). The sources of these strains are shown in Table 1. The study did not involve human experiments, and the faecal samples containing the isolated strains were collected from healthy volunteers and did not cause any foreseeable risk or discomfort to the participants. The volunteers signed a written informed consent form or obtained the consent of their legal guardian. This study used previously isolated and preserved strains. Strain isolation was not involved in this study. All bacterial strains were cultured in modified de Man, Rogosa and Sharpe broth and incubated at 37 °C under anaerobic conditions (400TG; Electrotek, West Yorkshire, UK). Bacterial cells were collected by centrifugation at 6000× g for 5 min and suspended in a small volume of 30% glycerol solution to prepare a stock, and were stored at −80 °C for further treatment. The stock was diluted with sterilised phosphate-buffered saline (PBS) to yield a final concentration of 109 colony-forming units (CFU)/mL for oral administration. 2.2. Animal Experiments The animal experiments involved in this study were carried out in the Animal Centre of Jiangnan University under environmentally controlled conditions (12-h light–dark cycle at 22 °C ± 3 °C and humidity of 55% ± 10%). All experiments were approved by the Animal Ethics Committee of Experimental Animal at Jiangnan University (JN. No:20210315c1080601 [012]). Three-week-old C57BL/6J female mice were purchased from the Model Animal Research Centre of Vital River (Shanghai, China). Standard food and sterile water were provided ad libitum. After acclimatisation for 7 days, 36 mice were randomly divided into six groups (n = 6/group). The control group was gavaged with 0.2 mL PBS solution. The X6L1, S24-1, 12-1, 20M5 and 24M1 groups were gavaged with 0.2 mL bacterial suspension (109 CFU/mL) every day for three weeks. The experimental procedure timeline is shown in Figure 1. Mice were randomly assigned to the control group and different experimental condition groups using simple randomisation. 2.3. Tissue Processing On the 7th, 14th and 21st days of the experiment, the faeces of mice were collected, placed on ice immediately and transferred to a freezer at −80 °C for storage. On the last day of the experiment, the mice were intraperitoneally injected with isoflurane after fasting for 12 h. One hour after intraperitoneal injection, the mice were sacrificed and their blood was collected. The blood was centrifuged at 2000× g for 15 min. The supernatant was collected as serum and stored in a freezer at −80 °C for later use. The mice were dissected, and the duodenum, jejunum, ileum, caecum, colon and other tissues were collected. Part of the intestinal tissues were stored in 40 g/L paraformaldehyde for later immunofluorescence staining. Of the remaining intestinal tissue, the caecal and colon contents were scraped, and the remaining tissue samples and caecal and colon contents were immediately placed into liquid nitrogen for rapid freezing. These samples were then stored at −80 °C for later use [15]. 2.4. Real-Time Polymerase Chain Reaction Approximately 20 mg of mouse colon that was preserved at −80 °C were placed together with high-temperature inactivated zirconia beads into an enzyme-free centrifuge tube. We added 1 mL of TRIzol and fully crushed the sample with a high-throughput crusher (SCIENTZ-48, Ningbo, China). The supernatant was collected, protein impurities and DNA were precipitated with chloroform and isopropanol was added for static precipitation. After discarding the supernatant, 75% pre-cooled ethanol was added to clean the extracted RNA twice, the solvent was left to air dry and the sample was re-dissolved with diethylpyrocarbonate-treated water. The purity and integrity of the extracted RNA were tested by measuring its optical density at 260/280 nm. In accordance with the instructions provided by the manufacturer of the Vazyme kit, complementary DNA was synthesised with the extracted total RNA as a template for real-time fluorescence quantitative polymerase chain reaction (qPCR) detection (Bio-Rad, Berkeley, CA, USA). The PCR system was prepared in accordance with the instructions of the qPCR mix, and the real-time qPCR program was run [16]. The transcription levels of genes encoding the polymeric immunoglobulin receptor (pIgR), myeloid differentiation factor 88 (MyD88), B-cell activating factor receptor (BAFFR), epidermal growth factor receptor (EGFR) and activation induced cytidine deaminase (ACIDA) in the mouse colon were measured by real-time qPCR. The primer sequences of mouse pIgR, MyD88, BAFFR, EGFR and ACIDA were found on the PrimerBank website (accessed on 19 January 2022, https://pga.mgh.harvard.edu/primerbank/index.html). Sonny Biotechnology Co., Ltd. (Shanghai, China) synthesised these primers. The specific primer information is shown in Table 2. 2.5. Determination of Short Chain Fatty Acids Freeze-dried stool samples (50 mg) were homogenised in 500 μL of saturated NaCl solution and acidified with 40 μL of 10% sulphuric acid. Diethyl ether (1 mL) was added to the samples to extract short-chain fatty acids, after which the samples were centrifuged at 14,000× g for 15 min at 4 °C. Each supernatant (1 μL) was injected into an Rtx-WAX capillary column for gas chromatography–mass spectrometry analysis (QP2010 Ultra; Shimadzu, Kyoto, Japan). The initial oven temperature (100 °C) was increased to 140 °C at a rate of 7.5 °C min−1. The temperature was then further increased to 200 °C at a rate of 60 °C min−1 and maintained for 3 min. Helium was used as the carrier gas (flow rate: 0.89 mL min−1; column head pressure: 62.7 kPa). The injector temperature was set at 240 °C. The mass spectrometer was set at an ion source temperature of 220 °C, an interface temperature of 250 °C and a scan range of 2–100 m/z [17]. 2.6. Enzyme-Linked Immunosorbent Assay The faeces, duodenum, jejunum, ileum and colon, which were preserved at −80 °C, were thawed and the adipose tissue was removed. Pre-cooled sterile PBS solution was added according to a weight ratio of 1:9. The samples were then placed into a centrifuge tube together with cleaned and sterilised zirconia beads, broken down with a high-throughput crusher and then centrifuged at 4 °C for 2000× g for 10 min [18]. The supernatant was collected for assay measurement. Interleukin (IL)-1β, IL-6 and IL-17 concentrations were measured using mouse enzyme-linked immunosorbent assay kits according to the manufacturer’s protocol (R&D, Minneapolis, MN, USA). IgA and IgG concentrations were measured using mouse enzyme-linked immunosorbent assay kits according to the manufacturer’s protocol (Elabscience Biotechnology Co., Ltd., Wuhan, China). Protein was measured in intestinal tissue and faeces using the bicinchoninic acid assay (Beyotime Biotechnology, Shanghai, China). 2.7. Immunofluorescence of IgA in Plasma Cells in Mouse Intestinal Segments After sacrificing the mice, approximately 1 cm of mouse intestinal tissue was carefully cut to avoid mechanical damage to the colon tissue as much as possible. After removing other tissues attached to the intestinal tissue, the residual faeces were removed by rinsing with pre-cooled normal saline, and then 4% paraformaldehyde solution was immediately added for fixation. The samples were embedded within 48 h: after washing the samples with clean water to remove excess fixative, they were successively treated with an ethanol gradient and xylene for dehydration and immersed in paraffin wax. The wax-soaked colon tissue was embedded using a Leica tissue embedding machine. After cooling, sections were cut with a hand wheel Leica tissue slicer. The sections were spread on slides and dried for subsequent staining [19]. The sections were stained overnight using the following reagents for immunofluorescence analysis: goat anti-mouse IgA alpha chain antibody (1:500, ab97231, Abcam, Shanghai, China), an FITC immunofluorescence detection kit (E670007, BBI, Shanghai, China) and 4′,6-diamidino-2-phenylindol dihydrochloride (Fcmacs, Nanjing, China). 2.8. Isolation and Identification of IgA-Coated Bacteria IgA-coated bacteria were collected from faeces using magnetic bead-based enrichment. Briefly, faeces were suspended at 20% in pre-reduced PBS containing 0.5% Tween 20 (PBST) and protease inhibitors (1 mg/mL leupeptin, 1.6 mg/mL aprotinin; Sigma-Aldrich, St. Louis, MO, USA). The faeces were homogenised and centrifuged at 400× g to remove large debris. The supernatant was then centrifuged at 8000× g to pellet bacteria and washed with PBST three times. The bacterial pellet was resuspended in pre-reduced PBS supplemented with 0.25% bovine serum albumin, 5% goat serum and biotinylated goat anti-mouse IgA (ab97233, Abcam, Shanghai, China). After being washed, the biotinylated samples were mixed with streptavidin-coated magnetic beads to allow crosslinking (D110557; BBI, Shanghai, China). IgA-coated bacteria were separated from the suspension with the aid of a magnet. The collected bacteria were washed three times with PBST. DNA was quantified and pooled at equal concentrations by following the instructions of the Qubit dsDNA Assay Kit (Life Technologies, Carlsbad, CA, USA). All isolated strains were further typed by 16S rDNA high-throughput sequencing [20,21]. 2.9. 16 S rDNA High-Throughput Sequencing The total DNA of bacteria in fresh stool was extracted using the Fast DNA Stool Kit (MP Biomedicals, Carlsbad, CA, USA). PCR amplification of the 16S rDNA gene was performed using universal primers (341 forward: 5′-CCTAYGGGRBGCASCAG-3′ and 806 reverse: 5′-GGACTACNNGGGTATCTAAT-3′). The PCR products were purified using the TIANgel Mini Purification Kit (Tiangen, Beijing, China). DNA was quantified and pooled at equal concentrations by following the instructions of the Qubit dsDNA Assay Kit (Life Technologies, Carlsbad, CA, USA). Samples were barcoded and finally paired-end sequenced on the Illumina MiSeq PE300 platform by following the manufacturer’s protocol. A gene sequencing analysis using 16S rDNA was performed with Quantitative Insights into Microbial Ecology version 2 (QIIME2) [22]. Amplicon sequence variants were rarefied to 10,000 according to the sampling depth. The composition of Lactobacillus species was analysed and modified to run on QIIME2 [22]. The amplicon sequence variants were rarefied to 3000 according to the sampling depth. Subsequent processing was similar to that for 16S rDNA gene analysis unless otherwise noted. 2.10. Statistical Analysis All data in this study are expressed as the mean ± standard deviation and were plotted with Prism 7 and Origin Pro 2021. We used one-way analysis of variance or Welch’s t test. The Spearman correlation coefficient was calculated. The results were corrected by False Discovery Rate (FDR) and with p less than 0.05 were retained. p < 0.05 indicated significance of the data (compared with the control group), and 95% confidence intervals are shown. CaseViewer was used to intercept the visual field of photographs after a slice scanner was used. Image J was used for immunofluorescence analysis, Xcalibur was used for gas chromatography–mass spectrometry off-machine data analysis and Origin Pro 2021 was used for correlation analysis. Statistical Analysis of Metagenomic (and other) Profiles (STAMP) was used for the Welch’s t test [23]. 3. Results 3.1. L. fermentum Affects Intestinal Secretion of IgA After gavage for 1 week, compared with the control group, IgA concentrations in mouse faeces changed by varying degrees according to the Lactobacillus strain. The 12-1 and X6L1 strains significantly increased IgA concentrations in the faeces, but there were no significant changes in the S24-1, 20M5 or 24M1 groups (Figure 2a). During continuous gavage for three weeks, IgA concentrations in the faeces showed a downward trend and gradually approached those in the control group. IgA concentrations in each mouse intestinal segment were detected after three weeks of gavage. There were differences in IgA concentrations in the duodenum and ileum between the Lactobacillus strain groups (Figure 2b,d). IgA concentrations in the X6L1, S24-1 and 20M5 groups were significantly higher than those in the control group in the duodenum. However, IgA concentrations in the 12-1, X6L1 and 20M5 groups were significantly lower than those in the control group in the ileum. After gavage for three weeks, IgA concentrations were significantly higher in the X6L1 group than in the control group in the colonic contents, but this difference was not significant in the colon (Figure 2e). An immunofluorescence assay on paraffin sections of mouse intestinal segments showed that IgA in plasma cells was sporadically distributed in the duodenum and colon, compared with the control group. IgA was also observed in the jejunum and ileum, but the number of plasma cells was low (Figure 3a,b). There was no difference in the fluorescence intensity of IgA and plasma cells in the jejunum. The IgA content in the intestinal segments was different among the groups and decreased with the direction of intestinal peristalsis. As shown in Figure 3b, compared with the control group, the fluorescence intensity of IgA in different intestinal segments of Lactobacillus-treated mice were different. The IgA plasma cell content in the duodenum was significantly higher than that in the other intestinal segments. X6L1 and 20M5 significantly increased the immunofluorescence intensity of IgA plasma cells in the duodenum. In the ileum, 12-1 and X6L1 decreased the fluorescence intensity. X6L1, S24-1, 20M5 also reduced the immunofluorescence intensity of IgA plasma cells in the colon. 3.2. L. fermentum Affects the Host Gut Microbiota and IgA-Coated Bacteria The microbiota of the mouse colon was examined by IgA immunomagnetic beads and 16S rDNA analysis. Firmicutes and Bacteroidetes comprised the main part of the microbiota on the whole. However, there were some differences in Actinomycetes, Tenericutes and Verrucous in the microbiota. Most IgA-binding bacteria were Proteobacteria and Firmicutes (Figure 4a,b). The overall changes in each group tended to be similar, but the proportion of IgA-bound Actinomycetes was significantly higher in the 12-1 group (Figure 4c). The Chao1 index and operational taxonomic unit calculation and analysis showed that the 12-1 strain reduced the overall diversity of the host colon intestinal microbiota, but it did not affect the diversity of IgA-coated bacteria (Figure 5a,b). At the level of Lactobacillus species, the relative abundance of L. johnsonii, L. reuteri, L. acidophilus, L. murinus and L. fermentum was high. Different strains of L. fermentum had different effects on the host genus and species levels (Figure 5c,d). Welch’s t-test showed no significant difference in the S24-1 strain at the genus level compared with control group (data not shown; only groups with significant differences at the genus level are shown in Figure 6). Gavage with the 24M1 strain significantly increased the proportion of Lactobacillus species. Gavage with the 12-1 strain reduced the relative abundance of Bacteroides, Escherichia and Shigella. Gavage with the 20M5 strain increased the proportions of Eubacterium xylanophilum and Turicibacter, but decreased the proportion of Bilophila. Gavage with the X6L1 strain improved the relative abundance of Prevotellaceae UCG-001, Candidatus gastranaerophilales bacteriom zag_ 1 and Turicibacter. (Figure 6a,b). In the correlation analysis between the genus level and IgA, Negativibacillus, Oscillibacter, Ruminiclostridium 5 and UBA1819 were significantly correlated with IgA (Figure 7a). The proportions of Escherichia and Shigella in the 24M1 strain were higher and the abundance of Bacillus was lower than those in the control group. The 12-1 strain only reduced the abundance of Bacillus. There were no significant differences among the other species in the S24-1 strain except Brevundimonas increased. At the genus level, the X6L1 and 20M5 strains did not affect the relative abundance of IgA-bound bacteria. All strains decreased the isobutyrate and isovalerate concentrations in the intestines. The 12-1 strain reduced the acetate, propionate and butyrate concentrations. The 24M1 strain also reduced the propionate and butyrate concentrations (Figure 7b). 3.3. L. fermentum May Stimulate Intestinal Production of IgA in Different Ways The expression of genes encoding several reported IgA-related proteins in the mouse colon is shown in Figure 8. Of the two strains of L. fermentum, 12-1 and X6L1, that can increase the IgA content in faeces, only 12-1 upregulated BAFFR expression in the colon. The 20M5 strain upregulated pIgR expression, but the IgA concentration did not significantly change. There was no difference in gene expression related to IgA antibody class conversion among the five strains of Lactobacillus (Figure 8a). 3.4. L. fermentum May Produce IgA in a Non-Inflammatory Manner To examine the effects of different strains of L. fermentum on IgA secretion from intestinal mucosa, we measured IgG concentrations in the faeces after gavage of one week and colon contents after gavage of three weeks, when IgA concentrations were high. Only the X6L1 strain increased IgG concentrations (Figure 8b). There were no differences in IL-6, IL-17 or IL-1β concentrations among the strains (Figure 8c). 4. Discussion Host intestinal mucosal immunity is related to intestinal symbiotic bacteria [24]. IgA is the most secreted immunoglobulin in the intestine. However, there have been few studies on the mechanism by which common bacteria, especially lactic acid bacteria, affect host intestinal IgA [25]. Three problems need to be solved to determine how an antibody reaction is related to common microorganisms in the intestine. First, whether the IgA response is specific to the symbiotic bacteria that stimulate it is unknown. Second, how the IgA reaction adapts to the current bacteria in the intestine is unclear. Third, whether all bacteria can effectively induce IgA in an equal manner is unknown [13,26,27]. In this study, we used five strains of L. fermentum from different sources to examine whether there are intraspecific differences in the effect of inducing IgA. We found that not all L. fermentum strains stimulated IgA production in the gut. The increase in free IgA content in faeces may be due to (1) an increase in IgA plasma cells (2) and an increase in IgA transported from the intestinal lamina propria to the intestinal cavity. With regard to an increase in the free IgA content in faeces, Lactobacillus may promote the proliferation of B cells or stimulate antibody class switch recombination (CSR) in the B cells. Either of these possibilities could increase the number of IgA plasma cells at the corresponding effector site in mice. We performed a series of experiments to investigate this possibility. As shown in Figure 2 and Figure 3, compared with the control group, the concentrations of IgA in different intestinal segments of Lactobacillus-treated mice were different, suggesting that different L. fermentum strains had different effector sites in the intestine of mice. Immunofluorescence staining showed that the duodenum and colon had the most plasma cells. There were differences in the duodenum and ileum. Considering the results of IgA content in various intestinal segments in Figure 2b–f, it is speculated that X6L1, S24-1, and 20M5 first induced IgA-specific immunity in the duodenum and stimulated the differentiation of IgA plasma cells in the lamina propria. However, there was no difference in the fluorescence intensity of IgA plasma cells in the jejunum, so the effect sites of X6L1 and 12-1 may not be in the jejunum. The fluorescence intensity of X6L1 and 12-1 in the ileum were significantly lower than that of the control group and other Lactobacillus intervention groups. In addition, the ileum was the site with the lowest IgA content in both groups. X6L1, S24-1, and 20M5 decreased the fluorescence intensity of IgA plasma cells in the colon, while there was no difference in IgA concentration in colon tissue. However, from the content of IgA in colonic contents, it was significantly increased in the X6L1 group, significantly decreased in S24-1 group, and not significantly changed in the 20M5 group. This further indicates that different strains of L. fermentum have different effects on intestinal stimulating IgA. In the faeces, the IgA levels of the 12-1, X6L1 group were significantly higher than that of the control group, which means that the two groups of L. fermentum increased the content of IgA from the overall level of the intestinal tract. Notably, as shown in Figure 3b, 12-1 did not increase the number of IgA plasma cells in the gut. Therefore, it is speculated that 12-1 might alter the transport efficiency of IgA from the lamina propria, rather than promote the proliferation of IgA-producing cells. This indicated that different strains of Lactobacillus of the same species had differences in the way of stimulating the intestinal secretion of IgA. The small intestine may play an important role in the production of IgA. In Figure 2a, during continuous gavage for 3 weeks, IgA concentrations in the faeces showed a downward trend and gradually approached those in the control group. With extended contact time between the intestinal mucosa and bacteria, IgA concentrations could not be maintained at a high level. Whether the immunogenicity of lactic acid bacteria is related to this issue is unclear [28]. After a period of adaptation in this study, the gut microbiota reached a new balance. The mucosal immunity developed tolerance to the experimental strains and no longer secreted too much IgA [29]. As a result, IgA concentrations in the faeces decreased with an increase in the experimental duration. Considering an increase in IgA transported from the intestinal lamina propria to the intestinal cavity, Lactobacillus may stimulate B cell associated pathways protein gene expression in mouse intestinal epithelial cells [30]. This stimulation could result in a higher rate of transport of IgA produced in the lamina propria out of the intestinal cavity, which could improve the transport efficiency [31]. BAFF is involved in B cell homeostasis and survival including promoting B cell to plasma cell transformation, making it an important pathway for regulating CSR and antibody production [32]. As a BAFF receptor, the content of BAFFR directly reflects the number of plasma cells [32]. In Figure 8a, the up-regulation of BAFFR levels suggested that L. fermentum 12-1 may increase the number of plasma cells by stimulating colonic secretion of BAFF. As a transmembrane receptor, EGFR can widely regulate cell proliferation, differentiation, migration and survival, and participate in various physiological and biochemical reactions [33]. However, L. fermentum treatments showed no effect on EGFR levels. It is suggested that this species of Lactobacillus does not activate the EGFR pathway. MyD88 acts as a linker molecule in the TLR pathway. It is involved in mediating the activation of the NF-κB pathway, stimulating the secretion of cytokines, and transmitting inflammatory signals [34]. In this study, the gene transcription level of this protein in the colon did not change, which indicated that L. fermentum may not stimulate colon inflammation. pIgR located on the basal surface of intestinal epithelial cells can bind to IgA in the lamina propria, allowing it to pass through the epithelial cells to the mucosal surface to exert its protective effect [35]. Only the expression of pIgR was up-regulated in the 20M5 group, but there was no significant change in the concentration of IgA. It is speculated that other subclasses of antibodies such as IgG (Figure 8b) may be transported from the lamina propria to the intestinal lumen. As a subclass of immunoglobulin, IgA needs to complete complex biochemical reactions through a special antibody CSR in order to exert its corresponding efficacy. Antigens are transported to B cells through antigen-presenting cells [36]. After B cells are stimulated, activation-induced cytidine deaminase (AID) is activated accordingly and begins to transcribe specific gene regions on the immunoglobulin heavy chain [36]. The expression level of ACIDA is an important indicator to detect whether antibody CSR occurs [36]. The ACIDA transcription levels were up-regulated in the colon tissue of mice treated with S24-1, 24M1 and 20M5. But the data showed that these three strains did not up-regulate the level of IgA in the intestine. Therefore, it is presumed that B cells at this site undergo antibody CSR but have not matured into IgA+ plasma cells. They may be transformed into plasma cells that secrete other subclasses of antibodies such as IgG. From this, it is not difficult to infer that Lactobacillus that affects different levels of intestinal secretion of IgA may cause IgA-specific immunity in different ways. These pathways are not limited to BAFFR, MyD88, EGFR, pIgR, or ACIDA-related pathways. An increase in IgA concentrations in the faeces may not only be the result of the direct action of L. fermentum but also could be due to a new balance of the intestinal microbiota. The relative abundance of Lactobacillus and L. fermentum did not increase at the genus or species level in this study (Figure 6a,b). However, L. fermentum affected the relative abundance of other members of the intestinal microbiota. The change in IgA concentrations in the intestine is a dynamic and comprehensive result. To further study the effect of stimulation of a single bacterium on host intestinal IgA, germ-free mice need to be colonised with single species of bacteria. IgA plays different roles by binding to different bacteria [29]. IgA combines with pathogenic bacteria to achieve the effect of immune exclusion, and with symbiotic bacteria to help colonisation [22,25,26]. However, how IgA identifies pathogenic bacteria and symbiotic bacteria is unknown. Using the immunomagnetic beads combined with 16S rDNA analysis in this study, L. fermentum was showed to have different effects on the host colonic microbiota and IgA-coated bacteria. There were differences in Escherichia, Shigella, Bacillus and Brevundimonas at the genus level (Figure 6a,b). The effect of IgA on these genera may be different [27]. IgA can specifically target lipopolysaccharide in Escherichia and Shigella species to limit it to the intestinal mucus layer and reduce intestinal inflammation [35]. However, IgA induced by Bacteroides fragilis helps the bacterium cluster and anchor on the surface of the intestinal epithelium, thus providing it with a competitive advantage [28]. In this study, differences in IgA-coated bacteria were not related to IgA concentrations. It is doubtful whether the difference of microbiota caused by different L. fermentum has any effect on IgA concentration. At present, whether this result is a two-way effect caused by IgA and the gut microbiota is unclear. Exploring the causes of this phenomenon will require more in-depth research. In recent years, there has been a debate on whether short-chain fatty acids are related to IgA. As a metabolite of gut microbiota, acetic acid has been formally known to stimulate the secretion of IgA [37]. L. fermentum is a heterologous fermentation strain which produces lactic acid and large amounts of acetic acid, ethanol and CO2 after fermentation [37]. However, acetic acid concentrations in the caecum did not increase in this study (Figure 7b). There was no correlation between IgA and short-chain fatty acids in the caecum. Short-chain fatty acids are mainly produced in the large intestine [38]. Indeed, IgA has been confirmed in the existing studies to be produced in the small intestine and enriched in the colon [38]. However, this conclusion is mainly proposed for pathogenic bacteria or invasive and strong bacteria that can cause IgA-T cell-dependent-PPs pathways, such as Segmented filamentous bacteria [4]. Whether all the bacteria only stimulate the small intestine to secrete IgA is unknown. Therefore, the causal relationship between IgA and short-chain fatty acids cannot be determined. Short-chain fatty acids concentrations in the small intestine might be strongly correlated with IgA concentrations. Initially, we would like to explore the effector site of L. fermentum to stimulate the intestinal secretion of IgA. It was not clear from the previous experiments that L. fermentum would stimulate IgA secretion in the small intestine, since the bacteria is mainly concentrated in the colon. Therefore, we detected the IgA-coated bacteria in the colon in order to explore the effect of L. fermentum on the gut microbiota. The path through which the 12-1 and X6L1 strains stimulate intestinal lymphocytes to secrete IgA will be the focus of a future study. However, we speculate that these two strains affect different reaction pathways to a great extent. Intestinal IgA may provide immune protection and rejection in a non-inflammatory manner, which promotes the establishment of the host microbial interaction mechanism [39]. Interleukins are a class of cytokines that are produced by cells and have direct or indirect stimulatory effects on immune cells [39]. It plays a special role in innate and adaptive immunity. Macrophages, Th2 cells, vascular endothelial cells, and fibroblasts are the main sources of interleukins [39]. IL-6 can stimulate the proliferation and differentiation of B cells, and regulate the inflammatory response by releasing antibodies [39]. IL-17 is an early initiator of T cell-induced inflammatory response and can release pro-inflammatory cytokines to amplify the inflammatory response [40]. IL-1β is a cytokine and regulator secreted by macrophages to activate cellular immunity [40]. However, the contents of IL-6, IL-17 and IL-1β in the colon of all L. fermentum treated mice did not change significantly (Figure 8c). X6L1 and 12-1, which induced IgA, might affect the intrinsic and specific immunity of intestinal mucosa through their special epitopes or secondary metabolites, and then promote the secretion of IgA. Intestinal IgA might provide immune protection for the host in a non-inflammatory manner, promoting the interaction between the host’s gut microbiota and mucosal immunity, and maintaining immune balance. The cell walls of different strains of L. fermentum need to be investigated in further studies to determine whether the difference in antigenic determinants in the cell wall lead to a difference in the IgA response [41], whether different surface antigens stimulate IgA in the same manner and whether other types of lactic acid bacteria also have intraspecific differences in stimulating the secretion of IgA from the intestinal mucosa [41]. At present, the two-way effect between symbiotic bacteria and IgA is not well understood. IgA may have an undiscovered response mechanism in humoral immunity. Notably, among the five tested strains of L. fermentum, two were from infants and three were from older people (Table 1). The strains that increased IgA concentrations in faeces were from the infants. Whether the bacteria that cause changes in IgA in the intestine are related to age and the physical condition of the provider are unknown [42]. An infant’s intestinal environment might be different from that of an adult, which may lead to differences in genes of the same bacteria and thus to different characteristics of the bacteria [42]. The intestinal microbiota and immune system in infants are in a process of development [43]. During this process, the production of IgA may be different from that in adults, which is stimulated by a stable intestinal microbiota [43]. Additionally, the induction of IgA by bacteria in different intestinal environments may be different between infants and adults. This difference may be at the levels of the gut microbiota and of different genotypes and physiological characteristics of the same species. This possibility would indicate a complicated interaction between the host gut microbiota and IgA [44]. There have been no relevant reports on whether the age or sex of the provider affects their intestinal IgA concentration. Therefore, this lack of knowledge could point to a new direction of IgA research. 5. Conclusions Different strains of L. fermentum show differences in stimulating the secretion of IgA in the host intestine. Some of these strains increase the number of IgA plasma cells in the duodenum, which increases IgA concentrations, while some stimulate BAFFR expression to induce high IgA concentrations. Furthermore, IgA concentrations are not significantly correlated with short-chain fatty acid concentrations in the caecum. L. fermentum, from different sources, has different effects on the host intestinal microbiota, but it does not affect the diversity of IgA-coated microbiota. Additionally, L. fermentum may stimulate intestinal secretion of IgA through different reaction pathways in a non-inflammatory manner. Acknowledgments Thanks to Yihan Li, Mengshu Xu, Luyao Wang and Zheng Wang for their help in the experiment. Author Contributions L.M.: investigation, writing—original draft, visualization; Y.C.: methodology, investigation; J.W.: software, formal analysis; J.Z.: validation, resources, data curation; H.Z.: methodology, validation, resources; J.L. and G.W.: conceptualization, validation, writing—review and editing, supervision, project administration, funding acquisition; W.C.: resources, funding acquisition. All authors have read and agreed to the published version of the manuscript. Funding This work was financially supported by the National Natural Science Foundation of China (No. 31972052, 32021005, 31820103010), the Fundamental Research Funds for the Central Universities (JUSRP22006, JUSRP51501), “Light of Taihu Lake” science and technology research project sponsored by Science and Technology Bureau of Wuxi City (Y20212030), the Program of Collaborative Innovation Centre of Food Safety and Quality Control in Jiangsu Province. Institutional Review Board Statement The study was conducted according to the Guidelines for Care and Use of Laboratory Animals of Jiangnan University (SYXK 2016-0045) and approved by the Animal Ethics Committee of Experimental Animals at Jiangnan University (qualified number: JN. No:20210315c1080601 [012]). Informed Consent Statement Not applicable. Conflicts of Interest The authors declare that they have no conflict of interest. Figure 1 Schematic of the animal experiment. Five strains were included in the lactic acid bacteria (LAB) groups, Lactobacillus strains X6L1, S24-1, 12-1, 24M1 and 20M5. Figure 2 Effect of L. fermentum on IgA concentration in faeces and intestinal tract of normal mice (a) IgA in one to three week faeces; (b–f) Duodenum, jejunum, ileum, colon and colon contents were detected after three weeks by ELISA, “*” indicates p < 0.05, and “**” indicates p < 0.01, “***” indicates p < 0.001, “****” indicates p < 0.0001 (compared with control group). Figure 3 Effect of L. fermentum on IgA+ plasma cell in intestinal tract of normal mice.(a) The IgA immunofluorescence map of each intestinal segment of mice was obtained from the visual field of the photos intercepted by the CaseViewer (10×, green-IgA, blue-DAPI); (b) The average fluorescence intensity of immunofluorescence images of mouse intestinal tissues treated by Image J “*” indicates p < 0.05, and “**” indicates p < 0.01, “***” indicates p < 0.001, “****” indicates p < 0.0001 (compared with the control group). Figure 4 Effect of L. fermentum on gut microbiota in mice at phylum level. (a) Mouse colon contents 16S rDNA sequencing, phylum level relative abundance; (b) Mouse colon contents IgA−coated bacteria 16S rDNA sequencing, phylum level relative abundance; (c) Results of comparison between IgA−coated bacteria and colon contents microbiota on gate level (“#”indicates p < 0.05, “###” indicates p < 0.001, “####” indicates p < 0.0001 (compared with 16S)). Figure 5 Effect of L. fermentum on gut microbiota in mice at α diversity and species level. (a) Mouse colon contents bacterial α-diversity calculated by Chao1 and observed OTUs; “*” indicates p < 0.05. (Compared with control group). (b) IgA-coated bacterial α−diversity calculated by Chao1 and observed OTUs; (c) 16S rDNA sequencing of mouse colon contents, relative abundance of Lactobacillus species level; (d) IgA−coated bacteria 16S rDNA sequencing of mouse colon contents, relative abundance of Lactobacillus species level. Figure 6 Effect of L. fermentum on gut microbiota in mice at genera level. (a) Relative abundance of bacteria in mouse colon and bacteria with differences by Welch’s t test at genus level; (b) Relative abundance of IgA-coated bacteria in mouse colon and bacteria with differences by Welch’s t test at genus level. Figure 7 Analysis of L. fermentum on gut microbiota in mice and cecum SCFAs. (a) The correlation between IgA, IgG and colonic microbiota. “*” indicates p < 0.05, and “**” indicates p < 0.01 (compared with control group); (b) The content of short chain fatty acids in cecum. “*” indicates p < 0.05, and “**” indicates p < 0.01, “***” indicates p < 0.001 (compared with control group). Figure 8 Effect of L. fermentum on IgA related protein gene expression and inflammatory factor in the colon of normal mice. (a) The relative gene expression of IgA related proteins, “**” indicates p < 0.01(compared with control group); (b) The content of IgG in feces and colonic contents in one week, “***” indicates p < 0.001 (compared with control group); (c) The content of colitis indicators. foods-11-01229-t001_Table 1 Table 1 Strains used in this study. Strain Species Origin Age 12-1 Lactobacillus fermenti Human feces, male 0 X6L1 Lactobacillus fermenti Human feces, female 0 S24-1 Lactobacillus fermenti Human feces, female 100 20M5 Lactobacillus fermenti Human feces, male 83 24M1 Lactobacillus fermenti Human feces, female 75 foods-11-01229-t002_Table 2 Table 2 Primer sequence. Gene Primer (5′-3′) pIgR F-AGGCAATGACAACATGGGG R-ATGTCAGCTTCCTCCTTGG MyD88 F-CACCTGTGTCTGGTCCATTG R-CTGTTGGACACCTGGAGACA BAFFR F-GAAACTGCGTGTCCTGTGAG R-CTGAGGCTGCAGAGCTGTC EGFR F-GCCATCTGGGCCAAAGATACC R-GTCTTCGCATGAATAGGCCAAT ACIDA F-CGTGGTGAAGAGGAGAGATAGTG R-CAGTCTGAGATGTAGCGTAGGAA GAPDH F-TGTGTCCGTCGTGGATCTGA R-CCTGCTTCACCACCTTCTTGAT Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Feifei X. Newby J.M. Schiller J.L. Schroeder H.A. Wessler T. Chen A. Forest M.G. Lai S.K. Modeling Barrier Properties of Intestinal Mucus Reinforced with IgG and Secretory IgA against Motile Bacteria ACS Infect. Dis. 2019 5 1570 1580 10.1021/acsinfecdis.9b00109 31268295 2. Steinhoff U. Who controls the crowd? New findings and old questions about the intestinal microflora Immunol. Lett. 2005 99 12 16 10.1016/j.imlet.2004.12.013 15894105 3. Maldonado Galdeano C. Cazorla S.I. Beneficial Effects of Probiotic Consumption on the Immune System Ann. Nutr. Metab. 2019 74 115 124 10.1159/000496426 30673668 4. Fagarasan S. Kawamoto S. Kanagawa O. Suzuki K. Adaptive Immune Regulation in the Gut: T Cell–Dependent and T Cell–Independent IgA Synthesis Annu. Rev. Immunol. 2010 28 243 273 10.1146/annurev-immunol-030409-101314 20192805 5. Liu H.Y. Seignez C. Ahl D. Giraud A. Guo F. Sedin J. Walden T. Roos S. Holm L. Bertilsson S. Bacteria-immune cell interactions at the intestinal Peyer’s Patches Free. Radic. Biol. Med. 2019 139 S3 6. Yoda K. Miyazawa K. Hosoda M. Hiramatsu M. Yan F. He F. Lactobacillus GG-fermented milk prevents DSS-induced colitis and regulates intestinal epithelial homeostasis through activation of epidermal growth factor receptor Eur. J. Nutr. 2014 53 105 115 10.1007/s00394-013-0506-x 23468308 7. Shroff K.E. Meslin K. Cebra J.J. Commensal enteric bacteria engender a self-limiting humoral mucosal immune response while permanently colonizing the gut Infect. Immun. 1995 63 3904 3913 10.1128/iai.63.10.3904-3913.1995 7558298 8. Macpherson A.J. Hunziker L. McCoy K. Lamarre A. IgA responses in the intestinal mucosa against pathogenic and non-pathogenic microorganisms Microbes Infect. 2001 3 1021 1035 10.1016/S1286-4579(01)01460-5 11580989 9. Gupta S. Basu S. Bal V. Rath S. George A. Gut IgA abundance in adult life is a major determinant of resistance to dextran sodium sulfate-colitis and can compensate for the effects of inadequate maternal IgA received by neonates Immunology 2019 158 19 34 10.1111/imm.13091 31215020 10. Kawashima T. Hayashi K. Kosaka A. Kawashima M. Igarashi T. Tsutsui H. Tsuji N.M. Nishimura I. Hayashi T. Obata A. Lactobacillus plantarum strain YU from fermented foods activates Th1 and protective immune responses Int. Immunopharmacol. 2011 11 2017 2024 10.1016/j.intimp.2011.08.013 21893216 11. Kawashima T. Ikari N. Kouchi T. Kowatari Y. Kubota Y. Shimojo N. Tsuji N.M. The molecular mechanism for activating IgA production by Pediococcus acidilactici K15 and the clinical impact in a randomized trial Sci. Rep. 2018 8 5065 10.1038/s41598-018-23404-4 29567956 12. Azagra-Boronat I. Tres A. Lactobacillus fermentum CECT5716 supplementation in rats during pregnancy and lactation affects mammary milk composition J. Dairy Sci. 2020 103 2982 2992 10.3168/jds.2019-17384 32008776 13. Garcia-Castillo V. Komatsu R. Evaluation of the Immunomodulatory Activities of the Probiotic Strain Lactobacillus fermentum UCO-979C Front. Immunol. 2019 10 1376 10.3389/fimmu.2019.01376 31263467 14. Takeuchi T. Miyauchi E. Acetate differentially regulates IgA reactivity to commensal bacteria Nature 2021 595 560 564 10.1038/s41586-021-03727-5 34262176 15. Ashraf R. Shah N.P. Immune system stimulation by probiotic microorganisms Crit. Rev. Food Sci. Nutr. 2014 54 938 956 10.1080/10408398.2011.619671 24499072 16. Geuking M.B. McCoy K.D. The function of secretory IgA in the context of the intestinal continuum of adaptive immune responses in host-microbial mutualism Semin. Immunol. 2012 24 36 42 10.1016/j.smim.2011.11.005 22138187 17. Caixin N. Xing J. Evaluation of antioxidant activity of Lactobacillus in vivo using Caenorhabditis elegans and comparison with its antioxidant parameters in vitro Food Ferment. Ind. 2019 3 21 27 18. Choi Y. Jeon B.Y. Shim T.S. Jin H. Cho S.N. Lee H. Development of a highly sensitive one-tube nested real-time PCR for detecting Mycobacterium tuberculosis Diagn. Microbiol. Infect. Dis. 2014 80 299 303 10.1016/j.diagmicrobio.2014.08.009 25241639 19. Wang L. Wang L. A randomised, double-blind, placebo-controlled trial of Bifidobacterium bifidum CCFM16 for manipulation of the gut microbiota and relief from chronic constipation Food Funct. 2022 13 1628 1640 10.1039/D1FO03896F 35079761 20. Mattioli C.A. Tomasi T.B. The life span of IgA plasma cells from the mouse intestine J. Exp. Med. 1973 138 452 460 10.1084/jem.138.2.452 4578299 21. Sun J. Qi C. IgA-Targeted Lactobacillus jensenii Modulated Gut Barrier and Microbiota in High-Fat Diet-Fed Mice Front. Microbiol. 2019 10 1179 10.3389/fmicb.2019.01179 31178854 22. Qi C. Ding M. Sex-dependent modulation of immune development in mice by secretory IgA-coated Lactobacillus reuteri isolated from breast milk J. Dairy Sci. 2021 104 3863 3875 10.3168/jds.2020-19437 33612242 23. Tian P. Chen Y. Pediococcus acidilactici CCFM6432 mitigates chronic stress-induced anxiety and gut microbial abnormalities Food Funct. 2021 12 11241 11249 10.1039/D1FO01608C 34704999 24. Huus K.E. Bauer K.C. Commensal Bacteria Modulate Immunoglobulin A Binding in Response to Host Nutrition Cell Host Microbe 2020 27 909 921.e5 10.1016/j.chom.2020.03.012 32289261 25. Macpherson A.J. Yilmaz B. IgA Function in Relation to the Intestinal Microbiota Annu. Rev. Immunol. 2018 36 359 381 10.1146/annurev-immunol-042617-053238 29400985 26. Yang C. Mogno I. Fecal IgA Levels Are Determined by Strain-Level Differences in Bacteroides ovatus and Are Modifiable by Gut Microbiota Manipulation Cell Host Microbe 2020 27 467 475.e6 10.1016/j.chom.2020.01.016 32075742 27. Daniel H. Markus A. Diard M. Loverdo C. Slack E. Growing, evolving and sticking in a flowing environment: Understanding IgA interactions with bacteria in the gut Immunology 2020 159 52 62 31777063 28. Bunker J.J. Flynn T.M. Innate and Adaptive Humoral Responses Coat Distinct Commensal Bacteria with Immunoglobulin A Immunity 2015 43 541 553 10.1016/j.immuni.2015.08.007 26320660 29. Donaldson G.P. Ladinsky M.S. Gut microbiota utilize immunoglobulin A for mucosal colonization Science 2018 360 795 800 10.1126/science.aaq0926 29724905 30. Fadlallah J. El Kafsi H. Sterlin D. Juste C. Parizot C. Dorgham K. Autaa G. Gouas D. Almeida M. Lepage P. Microbial ecology perturbation in human IgA deficiency Sci. Transl. Med. 2018 10 eaan1217 10.1126/scitranslmed.aan1217 29720448 31. Guzman-Bautista E.R. Suzuki K. Bacteria-immune cells dialog and the homeostasis of the systems Curr. Opin. Immunol. 2020 66 82 89 10.1016/j.coi.2020.05.010 32673977 32. Liang Z.-Q. Tu P. Gu-Ben-Fang-Xiao attenuates allergic airway inflammation by inhibiting BAFF-mediated B cell activation Biomed. Pharmacother. 2020 132 110801 10.1016/j.biopha.2020.110801 33049582 33. Yan F. Cao H.W. Colon-specific delivery of a probiotic-derived soluble protein ameliorates intestinal inflammation in mice through an EGFR-dependent mechanism J. Clin. Investig. 2011 121 2242 2253 10.1172/JCI44031 21606592 34. Ramanan D. Sefik E. An Immunologic Mode of Multigenerational Transmission Governs a Gut Treg Setpoint Cell 2020 181 1276 1290 10.1016/j.cell.2020.04.030 32402238 35. Ma N. Guo P. Zhang J. He T. Kim S.W. Zhang G. Ma X. Nutrients Mediate Intestinal Bacteria-Mucosal Immune Crosstalk Front. Immunol. 2018 9 5 10.3389/fimmu.2018.00005 29416535 36. Bunker J.J. Bendelac A. IgA Responses to Microbiota Immunity 2018 49 211 224 10.1016/j.immuni.2018.08.011 30134201 37. Schwab C. Ganzle M. Lactic acid bacteria fermentation of human milk oligosaccharide components, human milk oligosaccharides and galactooligosaccharides FEMS Microbiol. Lett. 2011 315 141 148 10.1111/j.1574-6968.2010.02185.x 21175746 38. Dalile B. Van Oudenhove L. Vervliet B. The role of short-chain fatty acids in microbiota-gut-brain communication Nat. Rev. Gastroenterol. Hepatol. 2019 16 461 478 10.1038/s41575-019-0157-3 31123355 39. Fujihashi K. McGhee J.R. Cytoking specific elispot assay single cell analysis of IL-2, IL-4 and IL-6 production cells J. Immunol. Methods 1993 160 181 189 10.1016/0022-1759(93)90176-8 8459105 40. Isailovic N. Daigo K. Interleukin-17 and innate immunity in infections and chronic inflammation J. Autoimmun. 2015 60 1 11 10.1016/j.jaut.2015.04.006 25998834 41. Le Doare K. Bellis K. Faal A. Birt J. SIgA, TGF-beta1, IL-10, and TNF alpha in Colostrum Are Associated with Infant Group B Streptococcus Colonization Front. Immunol. 2017 8 1269 10.3389/fimmu.2017.01269 29109718 42. Pannaraj P.S. Li F. Association between Breast Milk Bacterial Communities and Establishment and Development of the Infant Gut Microbiome JAMA Pediatr. 2017 171 647 654 10.1001/jamapediatrics.2017.0378 28492938 43. Simon A.K. Hollander G.A. Evolution of the immune system in humans from infancy to old age Proc. Biol. Sci. 2015 282 20143085 10.1098/rspb.2014.3085 26702035 44. Goulet O. Potential role of the intestinal microbiota in programming health and disease Nutr. Rev. 2015 73 (Suppl. S1) 32 40 10.1093/nutrit/nuv039 26175488

PMC009xxxxxx/PMC9099658.txt

==== Front Animals (Basel) Animals (Basel) animals Animals : an Open Access Journal from MDPI 2076-2615 MDPI 10.3390/ani12091183 animals-12-01183 Article Changes in the Diversity and Composition of Gut Microbiota of Red-Crowned Cranes (Grus japonensis) after Avian Influenza Vaccine and Anthelmintic Treatment Zhao Xinyi 1† Ye Wentao 1† Xu Wei 1 Xu Nan 1 Zheng Jiajun 1 Chen Rong 2 Liu Hongyi 1* Li Robert Academic Editor 1 The Co-Innovation Center for Sustainable Forestry in Southern China, College of Biology and the Environment, Nanjing Forestry University, Nanjing 210037, China; zxy19981104@njfu.edu.cn (X.Z.); yewentao@njfu.edu.cn (W.Y.); xuwei2001@njfu.edu.cn (W.X.); 18871854847@sina.cn (N.X.); z2657244667@163.com (J.Z.) 2 Nanjing Hongshan Forest Zoo, Nanjing 210028, China; haze876@163.com * Correspondence: hongyi_liu@njfu.edu.cn † These authors contributed equally to this work. 05 5 2022 5 2022 12 9 118308 4 2022 02 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Simple Summary Gut microbiota homeostasis is significant to host health and well-being. Treatment of red-crowned cranes with avian influenza vaccines and anthelmintics has played pivotal roles in therapeutic management in zoos. This study provides evidence concerning if and how the avian influenza vaccine and anthelmintic treatment impact the gut microbiota of red-crowned cranes. For the first time, we revealed that the gut microbiota of red-crowned cranes is resilient to the avian influenza vaccine and anthelmintic, which may be affected to disorder in the short term but can recover to homeostasis over time. In addition, it is proposed that more controlled experiments should be performed properly to contribute to animal disease control and optimal management in the zoo. Abstract Gut microbiota homeostasis is important for host health and well-being; however, drugs may affect the composition and function of the gut microbiota. Red-crowned cranes are a vulnerable species. Treatment of red-crowned cranes with avian influenza vaccines and anthelmintics has played pivotal roles in therapeutic management in zoos. To investigate the changes in the diversity and composition of gut microbiota after the avian influenza vaccine and anthelmintic treatment, we used 16S rRNA sequencing to obtain and compare the bacterial community composition before and after the treatment. The alpha diversity of the gut microbiota of red-crowned cranes decreased on the day of the treatment and then fluctuated over time. The composition of gut microbiota tended to be similar in the short term after the treatment, as supported by the beta diversity hierarchical cluster analysis. Only 3, 8, and 72 operational taxonomic units (OTUs) of the three individuals were shared among the five groups before and after treatment. The relative abundance of Firmicutes significantly increased to 99.04% ± 0.28% on the day of the treatment, in which the relative abundance of Lactobacillus was 93.33% ± 5.85%. KEGG pathways analysis indicated that the main function of the gut microbiota is involved in metabolism, and the present study indicates that the gut microbiota of red-crowned cranes is resilient to the avian influenza vaccine and anthelmintic, even disordered in the short term, and could recover over time. More individual experimentation and functional potential in metabolism are needed in the future to support animal disease control and optimal management in the zoo. red-crowned crane gut microbiota diversity and composition avian influenza vaccine anthelmintic National Natural Science Foundation of China31800453 This research was funded by the National Natural Science Foundation of China (No. 31800453). ==== Body pmc1. Introduction Gut microbiota is often considered essential for host health and well-being because its ecological stability ensures that beneficial symbionts and their associated functions are maintained over time [1,2]. Correspondingly, in the case of dysbiosis, disruption of the normal balance between the gut microbiota and host [3], the host is more susceptible to various diseases such as obesity [4], inflammatory bowel diseases [5,6], and cancer [7]. In general, although the composition and taxonomy of gut microbiota vary markedly across biotic (such as host age, host sex, and host disease) and abiotic (such as diet composition, geographical environment, and seasonal variation) factors [8,9], the gut microbiota can recover from external perturbations to adapt to changes and maintain stability [10]. However, the gut microbiota can be dramatically altered by short-term drug interventions. Microbial regulatory changes in drug metabolism and drug-mediated changes in intestinal microbiota may have beneficial or harmful effects on the host [11,12]. For example, the microbial composition of hens fed a diet supplemented with Astragalus, a common herbal medicine with anti-inflammatory, immunostimulatory, antioxidative, and antiviral activities, was greater than that of the control group, which could play a vital role in the modulation of fecal microbiota [13]. Many studies have shown that treatment with drugs can result in changes in the function and composition of the gut microbiota and even disrupt the normal population and function of bacteria [11,14,15,16,17]. Firmicutes were found to be significantly depleted (p < 0.05) in the day 7 and day 14 monensin/virginiamycin-treated groups and the day 14 monensin/tylosin-treated group of the chicken cecal microbiome, as compared to the control group for each respective timepoint [17]. Additionally, changes in the quantity and composition of the gut microbiota are affected by gastric acid, malnutrition, and parasitic infections [18,19,20]. A study on Langya hens showed that excessively high stocking density drove the abundance of bacteria and fungi connected with health problems, and the gut microecology gradually reached a mature and balanced status with age [21]. The alpha diversity of the gut microbiota in rabbits infected with nematodes and mice infected with flagellates decreased significantly during the acute stage of infection [22,23]. Therefore, the gut microbiota is dynamic and continually changes in composition to adapt to changes in the internal and external environments, which is closely related to the health and disease of organisms. Since the development of veterinary medicine, avian influenza vaccines, and anthelmintic treatment have played pivotal roles in the management of avian diseases caused by viruses and parasites [24,25]. Vaccines are the cornerstone of preparing for and combating potential pandemics [26]. The use of anthelmintic treatment to control parasitic infections has long been standard practice. The need to respond quickly to potential influenza pandemics and parasite prevention in birds is necessary for zoos. The red-crowned crane (Grus japonensis) is one of the largest wading birds and is listed as vulnerable (VU) according to the IUCN Red List (https://www.iucnredlist.org/, accessed on 7 April 2022) in 2021. This species was also classified as a national first-grade protected animal in China (http://www.forestry.gov.cn/, accessed on 7 April 2022). Measures including the establishment of biosphere reserves and artificial breeding programs [27] have been developed to maintain populations of red-crowned cranes in China, Japan, South Korea, and other countries where red-crowned cranes have been kept in captivity in many zoos or nature reserves [28,29,30]. A previous study on red-crowned cranes found that both captivity and artificial breeding influenced the structure and diversity of the gut microbiota, potentially due to changes in diet, vaccination, antibiotics, and living conditions [29]. With the development of high-throughput sequencing technology, the characteristics of changes in gut microbiota induced by various factors have been further revealed [31,32,33,34]. However, how the avian influenza vaccine, anthelmintic treatment, and stress during these processes affect the diversity, composition, and function of the gut microbiota of red-crowned cranes kept in captivity in the zoo needs to be further investigated. Considering the above, and with little available information on the animals’ disease control and optimal management in the zoo, this study aimed to provide an elementary comparison and assessment of gut microbiota of red-crowned cranes before and after the avian influenza vaccine and anthelmintic treatment by high-throughput sequencing of the V3-V4 region of the 16S rRNA gene. Moreover, the effects of the avian influenza vaccine and anthelmintic treatment on the structure and function of the gut microbiota of red-crowned cranes were anticipated to be evaluated. 2. Materials and Methods 2.1. Animals This study was approved by the Laboratory Animal Welfare and Ethics Committee of Nanjing Forestry University. Three healthy red-crowned cranes, two adults, and one subadult (Sample 1: S1; Sample 2: S2; Sample 3: S3), kept in Nanjing Hongshan Forest Zoo (32°11 N; 118°83 E) in Jiangsu Province, China, were enrolled in this study. They were subjected to routine physical examinations and disease prevention measures during the autumn. Albendazole (0.2 g/tablet, Sino-American Tianjin Smithkline Pharmaceutical, Tianjin, China), a high-efficiency and low-toxicity broad-spectrum anthelmintic drug, was used for deworming. Reassortant avian influenza virus (H5 + H7) trivalent vaccine (Zhaoqing Dahuanong Biological Pharmaceutical, Zhaoqing, China) was used to prevent avian influenza caused by the H5 subtype (clade 2.3.4.4d and clade 2.3.2.1d) and H7 subtype avian influenza viruses. Each crane was administered one tablet and injected with a 1 mL vaccine. Before that, they had been in good health and had received no other drugs or treatments in the past 6 months. 2.2. Sample Collection, DNA Extraction, and PCR Amplification Fecal samples were collected before (10 d: B) and after (0 d: D0; 5 d: D5; 10 d: D10; 15 d: D15) the avian influenza vaccine and anthelmintic treatment. All samples were snap-frozen in liquid nitrogen and stored at −80 °C in the laboratory. Total DNA was extracted from the samples using an OMEGA Soil DNA Kit (D5625-01) (Omega Bio-Tek, Norcross, GA, USA). DNA concentrations were assessed using a UV spectrophotometer (NC2000, Thermo Scientific, Shanghai, China), and purity was monitored using 0.8% agarose gel electrophoresis. DNA was diluted to 10 mM for PCR amplification of the V3-V4 region of the 16S rRNA gene using the specific primers F: ACTCCTACGGGAGGCAGCA and R: GGACTACHVGGGTWTCTAAT. For a final volume of 25 μL, 2 μL of template DNA was added to a solution containing 8.75 μL of water, 0.25 μL of Q5 high-fidelity DNA polymerase, 5 μL 5× reaction buffer, 5 μL 5× high GC buffer, 2 μL dNTP (10 mM), and 1 μL of each primer (10 μM). The reaction conditions were as follows: initial denaturation at 98 °C for 30 s, followed by 27 cycles of denaturation at 98 °C for 15 s, annealing at 50 °C for 30 s, elongation at 72 °C for 30 s, and finally, 72 °C for 5 min. Amplification was subjected to 2% agarose gel electrophoresis, and the target fragments were cut and recovered using the Axygen gel recovery kit. 2.3. Gene Library Construction and Sequencing Sequencing libraries were constructed using the TruSeq Nano DNA LT Library Prep Kit (Illumina, NEB, Ipswich, MA, USA) following the manufacturer’s recommendations. Library quality was assessed by Agilent High Sensitivity DNA Kit using Agilent Bioanalyzer and quantified on the Promega QuantiFluor Quant-iT PicoGreen dsDNA Assay Kit (Invitrogen, Carlsbad, CA, USA). Paired-end reads sequencing at 2× 250 bp was then performed using the NovaSeq 6000 SP Reagent Kit (500 cycles) using Illumina NovaSeq at Shanghai Personal Biotechnology Co., Ltd. (Shanghai, China). 2.4. Analysis of Sequences All data analyses were performed at the Genescloud Platform (Shanghai Personal Biotechnology Co., Ltd., Shanghai, China) (https://www.genescloud.cn/home, accessed on 7 April 2022), using the R Programming Language (version 4.1.3) and QIIME2 (2019.4) software. Sequence denoising or clustering was performed using QIIME2. The primer fragments of the sequence were excised with the qiime cutadapt trim-paired, and the unmatched primer sequences were discarded. DADA2 [35] was used for quality control, denoising, splicing, and de-chimerization using qiime dada2 denoise-paired. The remaining high-quality sequences were clustered into operational taxonomic units (OTUs) at 97% sequence identity, with a representative sequence selected from each OTU using default parameters [36]. OTU taxonomic classification was performed by blasting a representative sequence set from the Greengenes database. After obtaining representative OTU sequences, statistics on the length distribution of high-quality sequences in all samples were performed. OTU leveling was performed using rarefaction. A certain number of sequences were randomly selected from each sample to reach a uniform depth to predict the observed OTUs and their relative abundances in each sample at this sequencing depth [37,38]. To calculate alpha diversity, we rarified the OTU table and calculated two metrics: the Chao1 index [39] to estimate species richness and the Simpson index [40] to describe community diversity. Additionally, we used non-metric multidimensional scaling (NMDS) analysis and hierarchical clustering in beta diversity analysis. The NMDS analysis is not affected by the numerical value of the sample distance, and only the size relationship between them is considered. For data with complex structures, the sorting results may be more stable. The Bray–Curtis distance matrix was used for NMDS analysis. Clustering analysis displayed the similarity between samples in the form of a hierarchical tree and measured the quality of the clustering effect based on the branch length of the clustering tree. The Bray–Curtis distance matrix adopts the UPGMA algorithm (average) for cluster analysis. Species and community composition analyses are represented by Venn diagrams and bar plots. Venn diagrams were used to count the number of species according to the uniqueness of each group and OTUs shared between the groups. According to the bar plots of the community, two aspects of the results are visualized intuitively: (i) what specific composition of microbial communities are contained in each sample at the taxonomic level (phylum and genus); (ii) the relative abundance of microbial phyla and genera in the fecal samples of red-crowned cranes are shown visually to understand the composition of the community structure of different samples at each taxonomic level. The relative abundance is represented as the mean ± standard deviation. Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt2) [41] was used to predict the functional abundance of samples and explore functional profiles [42]. PICRUSt2 can predict 16S rRNA gene sequences in the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database, in which metabolic pathways are classified into six categories: metabolism, genetic information processing, environmental information processing, cellular processes, organismal systems, and human diseases. We used the normalized pathway abundance table to calculate the average abundance or the total number of primary and secondary pathways depending on the selected samples. 3. Results 3.1. DNA Sequence Data A total of 956,460 raw paired-end reads were obtained from the Illumina HiSeq of 16S rRNA gene amplicons from 15 samples. A total of 639,123 (66.82%) high-quality reads remained for analysis after chimeras were filtered out and low-quality sequences were removed, with an average read length of 424 bp, ranging from 25,685 to 73,802 reads per sample. After subsampling reads to the same sequencing depth (34,558 sequences) between samples, 13855, OTUs clustered at 97% sequence identity (Table S1). 3.2. Alpha Diversity and Beta Diversity Analyses The alpha diversity of the samples at the OTU level was estimated using the Chao1 and Simpson indices (Figure 1). Both the Chao1 and Simpson indices of S2 and S3 dropped on the day after the avian influenza vaccine and anthelmintic treatment (D0), whereas they fluctuated between 5 and 15 days. The alpha diversity of these samples suggested that the richness and community diversity of the gut microbiota of red-crowned cranes could be quickly influenced by avian influenza vaccine and anthelmintic treatment and gradually recovered over time. Beta diversity was focused on the comparison of diversity among different samples. From the NMDS results (Table S2), the stress was 0.106 (stress < 0.2), indicating that the NMDS analysis could reliably reflect the degree of difference between the samples. According to beta diversity hierarchical cluster analysis (Figure 2), a multivariate statistical analysis method, samples in accordance with their degree of affinity in genera composition were classified, with an abundance ranking stacked histogram of the top 10 genera. The shorter the number of branches between the samples, the more similar the genus composition of the two samples. The bar plot shows that the genera composition of most samples in groups D0, D5, and D10 were similar, with the maximum abundance of Lactobacillus. Additionally, three samples in Group B clustered together, and two samples in Group D15 clustered together. These results showed that the gut microbiota of red-crowned cranes formed a steady state before the avian influenza vaccine and anthelmintic treatment. Fifteen days after the treatment, the gut microbiota recovered to another steady-state, which was different from that before the treatment. 3.3. Microbial Community Composition of Samples at Different Taxonomic Levels We analyzed the community composition of the fecal samples of red-crowned cranes in different groups (B, D0, D5, D10, and D15) (Figure 3) at the phylum level (Figure 4a, Table S3) and genus level (Figure 4b, Table S3). The bar plots show the percentage of community abundance at different taxonomic levels. The dominant microbial phyla (top 10) and genera (top 15) are shown. As shown in Figure 4a, comparing the microbial phyla of samples in groups B and D0, although the composition of samples in group B was not similar, the relative abundance of Firmicutes significantly increased to 99.04% ± 0.28% in group D0, which meant that the bacteria of other phyla were instantly reduced after avian influenza vaccine and anthelmintic treatment. In groups D5, D10, and D15, Firmicutes was still the most abundant phylum, accounting for 91.15% ± 3.82%; 66.39% ± 22.18%; 57.76% ± 16.67%, respectively, whereas the relative abundance of Proteobacteria and Actinobacteria also increased. At the genus level (Figure 4b), in group D0, the relative abundance of Lactobacillus was 93.33% ± 5.85%. The relative abundance of Lactobacillus showed a decreasing trend in the samples after the treatment, whereas the other microbial genera in S1 and S2 increased. However, S3 was a subadult, younger than S1 and S2, which may lead to dissimilar disorder and recovery state and ecological stability of microbial communities affected by the treatment. 3.4. Functional Gene Analysis Six primary metabolic pathways and secondary metabolic pathways (top three) of all fecal samples of red-crowned cranes were annotated using KEGG pathways. The maximum percentage of primary metabolic pathways (78.92%) was found to be involved in metabolism, in which carbohydrate metabolism (Figure 5A), amino acid metabolism (Figure 5B), and the metabolism of cofactors and vitamins (Figure 5C) were the most abundant secondary metabolic pathways. Furthermore, 14.40% of the primary metabolic pathways were involved in genetic information processing, with the top three being replication and repair (Figure 5L); translation (Figure 5M); and folding, sorting, and degradation (Figure 5N). 4. Discussion The present study employed high-throughput sequencing of the V3-V4 region of the 16S rRNA gene to explore the effects of the avian influenza vaccine and anthelmintic treatment on the gut microbiota of red-crowned cranes. We aimed to evaluate whether and how the treatment might lead to changes in the diversity and composition of gut microbiota. The results showed that alpha diversity indices were reduced by the avian influenza vaccine and anthelmintic treatment, even though the reduction was limited. Additionally, there were few common OTUs in the five groups, indicating that the composition of the gut microbiota of red-crowned cranes was greatly affected. The effects of antibiotics and anthelmintics on the alpha diversity of other animals are similar to our results [9,43]. Moreover, the gut microbiota of the recovery group of brown frogs after 7 days was not significantly different from that of the gentamicin group [43], which also proved that alpha diversity gradually recovered over time. High bacterial diversity benefits the overall animal health and productivity [44]. The use of drugs reduces the alpha diversity of the gut bacterial community, which may be a manifestation of either the adverse side effects of drugs or disorders of the gut microbiota [45]. However, the beta diversity hierarchical cluster analysis of our study demonstrated that the gut microbiota of red-crowned cranes was resilient to the treatment in the short term. The organism has a certain ability to reshape gut microbiota homeostasis, but the microbial compositions of the samples in group D15 were different from the compositions of the samples in group B. So, further studies are needed to verify whether the microbial composition is likely to finally recover to the composition similar to that before the treatment. It is significant to study its dynamic ecological stability over prolonged periods [46]. According to the results of microbial community composition of samples at the phylum and genus levels, we found that Firmicutes had a high proportion in almost all samples, with the highest abundance of Lactobacillus in group D0. Interestingly, some gut anti-microbial proteins may affect the composition of the gut microbiota, typically increasing the proportion of Lactobacillus, demonstrating that increased Lactobacillus may promote gut mucus-layer homeostasis [47]. Additionally, Lactobacillus can strengthen epithelial defense functions, modulate the gut microbiota, and maintain the gut barrier [47,48], which is similar to the findings of our study. Thus, Lactobacillus may help maintain gut microbiota stability. Previous studies have shown that the gut microbiota of wild, captive adolescent, captive adult, artificially bred adolescent and artificially bred adult cranes are dominated by three phyla: Firmicutes (62.9%), Proteobacteria (29.9%), and Fusobacteria (9.6%) [29]. However, Fusobacteria existed in S1 and S2 before the avian influenza vaccine and anthelmintic treatment, and the relative abundance of Fusobacteria was quite low after the treatment in our study. We may infer that the treatment influenced this phylum. In contrast, the relative abundance of Fusobacteria was low in all S3 samples. Whether the subadult, age, and dietary preferences have such effects on the Fusobacteria of intestinal bacteria requires further investigation. Immunity develops approximately 14 days after the reassortant avian influenza virus (H5 + H7) trivalent vaccine, and the immune period of chickens is approximately 6 months [49]. Albendazole is a benzimidazole derivative with an efficacy profile [50]. Approximately 2–2.5 h after oral administration, the blood drug concentration of animals reaches a peak. Albendazole and its metabolites (including albendazole sulfoxide and albendazole sulfone) were discharged from the urine and feces within 24 h, without accumulation in the body [50,51]. We collected fecal samples of red-crowned cranes after 3 h on the day after the avian influenza vaccine and anthelmintic treatment (D0). Therefore, we inferred that the effects on the gut microbiome were largely driven by anthelmintics instead of vaccines. However, due to the limited number of red-crowned cranes in the Nanjing Hongshan Forest Zoo, our inference deserves more individual experimentation and consideration, and functional potential in metabolism requires further metagenomic sequencing. Furthermore, gut bacterial homeostasis appeared to recover with a daily diet over time in our study. The gut microbiome can be affected by diet, and different dietary interventions are used by poultry producers or zookeepers to enhance avian growth and reduce the risk of enteric infection by pathogens [52]. A comprehensive understanding of the effects of disease control measures or treatments, diets, and stress will help develop new dietary or management interventions to promote the growth of red-crowned cranes, maximize the use of host feed, and protect birds from intestinal diseases caused by avian influenza, parasites, and pathogenic bacteria. 5. Conclusions Overall, the present research provides evidence concerning if and how the avian influenza vaccine and anthelmintic treatment impact the gut microbiota of red-crowned cranes. The results suggest that the gut microbiota of red-crowned cranes is resilient to avian influenza vaccine and anthelmintic, which may be affected to disorder in the short term but can recover to homeostasis over time. Additionally, more controlled experiments should be performed properly to understand the effects on the gut microbiome driven by anthelmintics and vaccines, respectively. Finally, since the gut microbiota is a very dynamic ecosystem, we hope to make a contribution to understanding the factors leading to the microbiota dysbiosis and the relationship between species forming gut microbiota. Acknowledgments Thanks to the zookeepers of Nanjing Hongshan Forest Zoo for their help in sampling. We wish to thank Hongjian Chen for advice on formal analysis. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ani12091183/s1, Table S1: Statistical table of sequencing quantity per sample; Table S2: Bray–Curtis distance matrix for NMDS analysis; Table S3: Statistics of the number of microbial taxa at the phylum and genus levels. Click here for additional data file. Author Contributions Conceptualization, H.L.; methodology, X.Z.; resources, W.X., J.Z. and R.C.; software, X.Z., W.Y. and N.X.; formal analysis, X.Z. and W.Y.; writing—original draft preparation, X.Z.; writing—review and editing, W.Y. and H.L.; project administration, H.L. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was approved by Animal Welfare and Ethics Committee of Nanjing Forestry University (No. 2022002). Data Availability Statement The nucleotide sequence data reported are available in the GenBank databases under the accession number PRJNA823535. Conflicts of Interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Figure 1 Alpha diversity of the intestinal bacterial community of the samples of red-crowned cranes. Comparison of the alpha diversity (Chao1 and Shannon) of the intestinal bacterial community of red-crowned cranes before and after the avian influenza vaccine and anthelmintic treatment. Figure 2 Beta diversity hierarchical cluster analysis of microbial genera of the fecal samples in different groups (B, D0, D5, D10, D15) of red-crowned cranes. The panel on the left is a hierarchical clustering tree diagram, in which samples are clustered according to their similarity with each other. The panel on the right is the abundance ranking Stacked histogram of the top 10 genera. Figure 3 Venn diagrams. Venn diagrams show the numbers of OTUs (97% sequence identity) that were shared or not shared of the fecal samples of each individual (S1, S2, and S3) and shared OTUs between different groups (B and D0; D0 and D5; D5 and D10; D10 and D15). Figure 4 Differences in relative abundance of microbial phyla and genera of the fecal samples of red-crowned cranes in different groups (B, D0, D5, D10, D15). Differences in abundance of dominant microbial phyla (a: top 10 are shown) and genera (b: top 15 are shown) of five groups (B, D0, D5, D10, D15) of each individual (S1, S2, and S3). Figure 5 Distributions of primary and secondary metabolic pathways of all fecal samples of red-crowned cranes annotated by KEGG pathways. Pie chart represents primary metabolic pathways and numbers indicate relative % representation. Bar charts represent secondary metabolic pathways (top 3 are shown). A: Carbohydrate metabolism; B: Amino acid metabolism; C: Metabolism of cofactors and vitamins; D: Signal transduction; E: Membrane transport; F: Cell growth and death; G: Cell motility; H: Cellular community prokaryotes; I: Infectious diseases; J: Neurodegenerative diseases; K: Cancers; L: Replication and repair; M: Translation; N: Folding, sorting, and degradation; O: Environmental adaptation; P: Endocrine system; Q: Immune system. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Dethlefsen L. Relman D.A. Incomplete recovery and individualized responses of the human distal gut microbiota to repeated antibiotic perturbation Proc. Natl. Acad. Sci. USA 2011 108 4554 4561 10.1073/pnas.1000087107 20847294 2. Coyte K.Z. Schluter J. Foster K.R. The ecology of the microbiome: Networks, competition, and stability Science 2015 350 663 666 10.1126/science.aad2602 26542567 3. Lozupone C.A. Stombaugh J.I. Gordon J.I. Jansson J.K. Knight R. Diversity, stability and resilience of the human gut microbiota Nature 2012 489 220 230 10.1038/nature11550 22972295 4. Ley R.E. Turnbaugh P.J. Klein S. Gordon J.I. Microbial ecology: Human gut microbes associated with obesity Nature 2006 444 1022 1023 10.1038/4441022a 17183309 5. Frank D.N. St Amand A.L. Feldman R.A. Boedeker E.C. Harpaz N. Pace N.R. Molecular-phylogenetic characterization of microbial community imbalances in human inflammatory bowel diseases Proc. Natl. Acad. Sci. USA 2007 104 13780 13785 10.1073/pnas.0706625104 17699621 6. Dicksved J. Halfvarson J. Rosenquist M. Järnerot G. Tysk C. Apajalahti J. Engstrand L. Jansson J.K. Molecular analysis of the gut microbiota of identical twins with Crohn’s disease ISME J. 2008 2 716 727 10.1038/ismej.2008.37 18401439 7. Lupton J.R. Microbial degradation products influence colon cancer risk: The butyrate controversy J. Nutr. 2004 134 479 482 10.1093/jn/134.2.479 14747692 8. Jami E. Israel A. Kotser A. Mizrahi I. Exploring the bovine rumen bacterial community from birth to adulthood ISME J. 2013 7 1069 1079 10.1038/ismej.2013.2 23426008 9. Hu X. Xu Y. Liu G. Hu D. Wang Y. Zhang W. Zheng Y. The impact of anthelmintic treatment on gut bacterial and fungal communities in diagnosed parasite-free sika deer Cervus Nippon Appl. Microbiol. Biot. 2020 104 9239 9250 10.1007/s00253-020-10838-y 10. Faith J.J. Guruge J.L. Charbonneau M. Subramanian S. Seedorf H. Goodman A.L. Clemente J.C. Knight R. Heath A.C. Leibel R.L. The long-term stability of the human gut microbiota Science 2013 341 1237439 10.1126/science.1237439 23828941 11. Walsh J. Griffin B.T. Clarke G. Hyland N.P. Drug-gut microbiota interactions: Implications for neuropharmacology Br. J. Pharmacol. 2018 175 4415 4429 10.1111/bph.14366 29782640 12. Ljubojević D. Pelić M. Puvača N. Milanov D. Resistance to tetracycline in Escherichia coli isolates from poultry meat: Epidemiology, policy and perspective World’s Poult. Sci. J. 2017 73 409 417 10.1017/S0043933917000216 13. Qiao H. Zhang L. Shi H. Song Y. Bian C. Astragalus affects fecal microbial composition of young hens as determined by 16S rRNA sequencing AMB Express 2018 8 70 10.1186/s13568-018-0600-9 29713833 14. Arnold C.E. Isaiah A. Pilla R. Lidbury J. Coverdale J.S. Callaway T.R. Lawhon S.D. Steiner J. Suchodolski J.S. The cecal and fecal microbiomes and metabolomes of horses before and after metronidazole administration PLoS ONE 2020 15 e0232905 10.1371/journal.pone.0232905 32442163 15. Jiménez B.L.M. Elghandour M.M.M.Y. Adegbeye M.J. Tirado González D.N. Tirado Estrada G. Salem A.Z.M. Pacheco E.B.F. Pliego A.B. Use of antibiotics in equines and their effect on metabolic health and cecal microflora activities J. Equine Vet. Sci. 2021 105 103717 10.1016/j.jevs.2021.103717 34607682 16. Puvača N. Milenković J. Galonja Coghill T. Bursić V. Petrović A. Tanasković S. Pelić M. Ljubojević Pelić D. Miljković T. Antimicrobial activity of selected essential oils against selected pathogenic bacteria: In vitro study Antibiotics 2021 10 546 10.3390/antibiotics10050546 34066788 17. Danzeisen J.L. Kim H.B. Isaacson R.E. Tu Z.J. Johnson T.J. Modulations of the chicken cecal microbiome and metagenome in response to anticoccidial and growth promoter treatment PLoS ONE 2011 6 e27949 10.1371/journal.pone.0027949 22114729 18. Schwarz R.S. Moran N.A. Evans J.D. Early gut colonizers shape parasite susceptibility and microbiota composition in honey bee workers Proc. Natl. Acad. Sci. USA 2016 113 9345 9350 10.1073/pnas.1606631113 27482088 19. Knutie S.A. Relationships among introduced parasites, host defenses, and gut microbiota of galapagos birds Ecosphere 2018 9 e02286 10.1002/ecs2.2286 20. Zhang Y.K. Zhang Q. Wang Y.L. Zhang W.Y. Hu H.Q. Wu H.Y. Sheng X.Z. Luo K.J. Zhang H. Wang M. A Comparison study of age and colorectal cancer-related gut bacteria Front. Cell. Infect. Microbiol. 2021 11 606490 10.3389/fcimb.2021.606490 33996615 21. Wang Y. Jin T. Zhang N. Li J. Wang Y. Kulyar M.F. Han Z. Li Y. Effect of stocking density and age on physiological performance and dynamic gut bacterial and fungal communities in Langya hens Microb. Cell Fact. 2021 20 218 10.1186/s12934-021-01707-y 34863176 22. Houlden A. Hayes K.S. Bancroft A.J. Worthington J.J. Wang P. Grencis R.K. Roberts I.S. Chronic trichuris muris infection in C57BL/6 mice causes significant changes in host microbiota and metabolome: Effects reversed by pathogen clearance PLoS ONE 2015 10 e0125945 10.1371/journal.pone.0125945 25938477 23. Cattadori I.M. Sebastian A. Hao H. Katani R. Albert I. Eilertson K.E. Kapur V. Pathak A. Mitchell S. Impact of helminth infections and nutritional constraints on the small intestine microbiota PLoS ONE 2016 11 e0159770 10.1371/journal.pone.0159770 27438701 24. Prichard R.K. Ranjan S. Anthelmintics Vet. Parasitol. 1993 46 113 120 10.1016/0304-4017(93)90052-O 8484204 25. Swayne D.E. Impact of vaccines and vaccination on global control of avian influenza Avian Dis. 2012 56 818 828 10.1637/10183-041012-Review.1 23402099 26. De Vries R.D. Herfst S. Richard M. Avian influenza A virus pandemic preparedness and vaccine development Vaccines 2018 6 46 10.3390/vaccines6030046 27. Xu H. Zhu G. Wang L. Bao H. Design of nature reserve system for red-crowned crane in China Biodivers. Conserv. 2005 14 2275 2289 10.1007/s10531-004-1663-2 28. Li J. Qi M. Chang Y. Wang R. Li T. Dong H. Zhang L. Molecular characterization of Cryptosporidium spp., Giardia duodenalis, and Enterocytozoon bieneusi in captive wildlife at Zhengzhou zoo China. J. Eukaryot. Microbiol. 2015 62 833 839 10.1111/jeu.12269 26384582 29. Xie Y. Xia P. Wang H. Yu H. Giesy J.P. Zhang Y. Mora M.A. Zhang X. Effects of captivity and artificial breeding on microbiota in feces of the red-crowned crane (Grus japonensis) Sci. Rep. 2016 6 33350 10.1038/srep33350 27628212 30. Si Y.J. Lee Y.N. Cheon S.H. Park Y.R. Baek Y.G. Kye S.J. Lee M.H. Lee Y.J. Isolation and characterization of low pathogenic H7N7 avian influenza virus from a red-crowned crane in a zoo in South Korea BMC Vet. Res. 2020 16 432 10.1186/s12917-020-02645-4 33167987 31. Dethlefsen L. Huse S. Sogin M.L. Relman D.A. The pervasive effects of an antibiotic on the human gut microbiota, as revealed by deep 16S rRNA sequencing PLoS Biol. 2008 6 e280 10.1371/journal.pbio.0060280 19018661 32. Costa M.C. Arroyo L.G. Allen-Vercoe E. Stämpfli H.R. Kim P.T. Sturgeon A. Weese J.S. Comparison of the fecal microbiota of healthy horses and horses with colitis by high throughput sequencing of the V3-V5 region of the 16S rRNA gene PLoS ONE 2012 7 e41484 10.1371/journal.pone.0041484 22859989 33. Manchester A.C. Webb C.B. Blake A.B. Sarwar F. Lidbury J.A. Steiner J.M. Suchodolski J.S. Long-term impact of tylosin on fecal microbiota and fecal bile acids of healthy dogs J. Vet. Intern. Med. 2019 33 2605 2617 10.1111/jvim.15635 31674054 34. Chen H.J. Hao D.J. Wei Z.Q. Wang L.J. Lin T. Acterial communities associated with the pine wilt disease insect vector Monochamus alternatus (Coleoptera: Cerambycidae) during the larvae and pupae stages Insects 2020 11 376 10.3390/insects11060376 32560536 35. Callahan B.J. Mcmurdie P.J. Rosen M.J. Han A.W. Johnson A.J. Holmes S.P. Dada2: High-resolution sample inference from illumina amplicon data Nat. Methods 2016 13 581 583 10.1038/nmeth.3869 27214047 36. Huang F. Zheng X. Ma X. Jiang R. Zhou W. Zhou S. Zhang Y. Lei S. Wang S. Kuang J. Theabrownin from Pu-erh tea attenuates hypercholesterolemia via modulation of gut microbiota and bile acid metabolism Nat. Commun. 2019 10 4971 10.1038/s41467-019-12896-x 31672964 37. Heck K.L. van Belle G. Simberloff D. Explicit calculation of the rarefaction diversity measurement and the determination of sufficient sample size Ecology 1975 56 1459 1461 10.2307/1934716 38. Kemp P.F. Aller J.Y. Bacterial diversity in aquatic and other environments: What 16S rDNA libraries can tell us FEMS Microbiol. Ecol. 2004 47 161 177 10.1016/S0168-6496(03)00257-5 19712332 39. Chao A. Nonparametric estimation of the number of classes in a population Scand. J. Stat. 1984 11 265 270 40. Simpson E.H. Measurement of diversity Nature 1949 163 688 10.1038/163688a0 41. Douglas G.M. Maffei V.J. Zaneveld J.R. Yurgel S.N. Brown J.R. Taylor C.M. Huttenhower C. Langille M.G.I. PICRUSt2 for prediction of metagenome functions Nat. Biotechnol. 2020 38 685 688 10.1038/s41587-020-0548-6 32483366 42. Langille M.G. Zaneveld J. Caporaso J.G. McDonald D. Knights D. Reyes J.A. Clemente J.C. Burkepile D.E. Vega Thurber R.L. Knight R. Predictive functional profiling of microbial communities using 16S rRNA marker gene sequences Nat. Biotechnol. 2013 31 814 821 10.1038/nbt.2676 23975157 43. Tong Q. Cui L.Y. Bie J. Han X.Y. Hu Z.F. Wang H.B. Zhang J.T. Changes in the gut microbiota diversity of brown frogs (Rana dybowskii) after an antibiotic bath BMC Vet. Res. 2021 17 333 10.1186/s12917-021-03044-z 34674716 44. Hildebrand F. Nguyen T.L. Brinkman B. Yunta R.G. Cauwe B. Vandenabeele P. Liston A. Raes J. Inflammation-associated enterotypes, host genotype, cage and inter-individual effects drive gut microbiota variation in common laboratory mice Genome Biol. 2013 14 R4 10.1186/gb-2013-14-1-r4 23347395 45. Pirker A. Stockenhuber A. Remely M. Harrant A. Hippe B. Kamhuber C. Adelmann K. Stockenhuber F. Haslberger A.G. Effects of antibiotic therapy on the gastrointestinal microbiota and the influence of Lactobacillus casei Food Agric. Immunol. 2013 24 315 330 10.1080/09540105.2012.689816 46. Palleja A. Mikkelsen K.H. Forslund S.K. Kashani A. Allin K.H. Nielsen T. Hansen T.H. Liang S. Feng Q. Zhang C. Recovery of gut microbiota of healthy adults following antibiotic exposure Nat. Microbiol. 2018 3 1255 1265 10.1038/s41564-018-0257-9 30349083 47. Qi H. Li Y. Yun H. Zhang T. Huang Y. Zhou J. Yan H. Wei J. Liu Y. Zhang Z. Lactobacillus maintains healthy gut mucosa by producing L-Ornithine Commun. Biol. 2019 2 171 10.1038/s42003-019-0424-4 31098404 48. Wang J. Ji H. Wang S. Liu H. Zhang W. Zhang D. Wang Y. Probiotic Lactobacillus plantarum promotes intestinal barrier function by strengthening the epithelium and modulating gut microbiota Front. Microb. 2018 9 1953 10.3389/fmicb.2018.01953 49. Zeng X.Y. Chen X.H. Ma S.J. Wu J.J. Bao H.M. Pan S.X. Liu Y.J. Deng G.H. Shi J.Z. Chen P.C. Protective efficacy of an H5/H7 trivalent inactivated vaccine produced from Re-11, Re-12, and H7-Re2 strains against challenge with different H5 and H7 viruses in chickens J. Integr. Agric. 2020 19 2294 2300 10.1016/S2095-3119(20)63301-9 50. Schulz J.D. Neodo A. Coulibaly J.T. Keiser J. Pharmacokinetics of albendazole, albendazole sulfoxide, and albendazole sulfone determined from plasma, blood, dried-Blood spots, and mitra samples of hookworm-infected adolescents Antimicrob. Agents Chemother. 2019 63 e02489-18 10.1128/AAC.02489-18 30745388 51. Adas G. Arikan S. Kemik O. Oner A. Sahip N. Karatepe O. Use of albendazole sulfoxide, albendazole sulfone, and combined solutions as scolicidal agents on hydatid cysts (in vitro study) World J. Gastroenterol. 2009 15 112 116 10.3748/wjg.15.112 19115476 52. Kelly D. Yang L. Pei Z. Gut microbiota, Fusobacteria, and colorectal cancer Diseases 2018 6 109 10.3390/diseases6040109 30544946

PMC009xxxxxx/PMC9099659.txt

==== Front Int J Environ Res Public Health Int J Environ Res Public Health ijerph International Journal of Environmental Research and Public Health 1661-7827 1660-4601 MDPI 10.3390/ijerph19095691 ijerph-19-05691 Article How Do Adolescents Use Social Networks and What Are Their Potential Dangers? A Qualitative Study of Gender Differences https://orcid.org/0000-0002-6993-0914 de Felice Giulio 12 https://orcid.org/0000-0002-3223-4421 Burrai Jessica 3 https://orcid.org/0000-0003-2367-3139 Mari Emanuela 3* Paloni Fabrizio 3 https://orcid.org/0000-0002-6676-2230 Lausi Giulia 3 https://orcid.org/0000-0002-0614-4457 Giannini Anna Maria 3 https://orcid.org/0000-0003-2341-1876 Quaglieri Alessandro 3 Tremolada Marta Academic Editor Bonichini Sabrina Academic Editor 1 Faculty of Literature and Philosophy, Sapienza University of Rome, 00185 Rome, Italy; giulio.defelice@uniroma1.it 2 Xenophon College London, University of Chichester, London PO19 6PE, UK 3 Department of Psychology, Sapienza University of Rome, 00185 Rome, Italy; jessica.burrai@uniroma1.it (J.B.); fabrizio.paloni@uniroma1.it (F.P.); giulia.lausi@uniroma1.it (G.L.); annamaria.giannini@uniroma1.it (A.M.G.); alessandro.quaglieri@uniroma1.it (A.Q.) * Correspondence: e.mari@uniroma1.it; Tel.: +39-06-49917534 07 5 2022 5 2022 19 9 569114 3 2022 05 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). The rapid development of software applications and the increasing use of the Internet have raised many questions about the impact of this technology on the lives of adolescents, especially on “digital natives.” The advent of social networks (SNs) restructures their relationships in various ways, affecting both adolescents’ development and mental health. The present study aims to investigate uses and dangers of SNs according to a sample of 296 (166 female and 130 male) Italian middle and high schools adolescents (age range 13–18) and build a model of how SNs can turn out to be dangerous. To achieve this, twenty-four audio-recorded focus groups of Italian male and female adolescents were investigated by a Grounded Theory approach, abstracting from the transcripts the main uses and dangers of SNs and proposing a final model for the interpretation of the whole set of categories. The results highlighted two main dangers of SNs: (a) the desperate search for popularity, and (b) the exhibition of violent or offensive behavior facilitated by the sense of protection and anonymity derived from being hidden behind a virtual account. Finally, a psychological model of how SNs can turn out to be dangerous is presented. This study could be useful in developing prevention procedures against the risks of SNs (e.g., cyberbullying, internet addiction) without demonizing the use of social media as such. social networks adolescents Grounded-Theory haters violence gender differences Dipartimento delle Pari Opportunità-Presidenza del Consiglio dei MinistriCUP B86C18004130001 This research was funded by “Dipartimento delle Pari Opportunità-Presidenza del Consiglio dei Ministri”, CUP B86C18004130001, with “#Hashtag Project: prevention of discrimination against women on line”. ==== Body pmc1. Introduction Mobile phone technology and the Internet have become an integral part of the daily interactions and activities of individuals, especially adolescents [1,2]. This generation, often referred to as “digital natives,” has been exposed to technology from birth [3]. The idea that digital devices and the Internet have a lasting influence on the way humans develop, socialize, and thrive seems to have great relevance [4,5], especially as the time spent online by young people has doubled in the last decade [6]. The rapid development of software applications and the increasing use of the Internet raise many questions about the impact of this technology on the lives of adolescents [7]. Subrahmanyam et al. [8] hypothesized that adolescents bring into social media the problems of their “offline” lives with regard to identity construction, peer group relations, sexuality, sensation seeking, and risk taking, aspects typically faced during this period of psychophysical development [9]. Recent studies [10,11] have indicated that social media use among 13–17-year-olds is around 93–97%, and since the introduction of social media, such as Facebook and Instagram, the scientific community has questioned whether such massive use affects an adolescent’s wellbeing and health. However, results to date have been dissenting [12,13,14,15,16]. According to Strasbuger [17] and his “Super peer” theory, social media, unlike face-to-face interactions, seems to exert a strong influence and pressure on adolescents, forcing them toward risky behaviors that are, instead, represented as normative. A further theory, the “Facebook Influence Model” [18], describes social media, in contrast to face-to-face situations or “traditional” media, as a mechanism for amplifying peer influence in which a behavioral reinforcement is manifested in the form of “likes” and/or comments as well as in the possibility of interacting directly or indirectly with a wider network of people outside one’s peer group. One aspect that should not be underestimated, closely linked to the use of SNs and their ability to “cross borders,” is related to online violence. We refer here to “hate speech” (or “cyberhate”), a form of online aggression toward individuals based on their race, gender, nationality, sexual orientation, ethnicity, religion, or disability with the aim of promoting hostility, discrimination, and/or violence [19]. Hate speech can manifest itself through highly offensive online posts, comments, messages, videos, or images [20,21]. The difference between hate speech and cyberbullying, cyber-harassment, or cyberstalking is that these latter forms are typically directed against an individual or a small group of individuals [22], while hate speech is directed toward specific social subgroups or a group of people representative of that given subgroup [23,24]. Several studies have shown that adolescent online experiences with cyberhate are relatively common. Cyberhate can be offensive, cruel, or threatening, and can be expressed through degrading online writing or speech, such as posts, comments, text messages, videos, or images [20,21]. Violent and discriminatory behavior performed “online” by young people generates a rush of excitement related to the aspects of recognition and approval and, additionally, the negative consequences of the actions committed often go unpunished, thus favoring a process of de-responsibilization [25]. Although there are no direct consequences associated with violent actions online, there are harmful consequences in the use of SNs including health problems [26,27], emotional problems, Internet addiction [28,29,30], and self-harm, including suicide [31]. The SNs may also provide benefits, such as a perceived greater connection with others when, for instance, in-person social interactions are severely limited (e.g., the COVID-19 pandemic; [32,33]. In addition, positive relationships among peers [34], family, neighborhood, and school apparently mitigate some negative outcomes such as delinquency and risky behavior [35]. Thus, the potentials and risks of SNs for adolescent development and mental health should still be clarified. For example, it is unclear whether the number of accounts owned, or the frequency with which they are checked, can affect psychosocial functioning [36]. There are gender differences in the use of the Web, particularly in the use of SNs. In fact, several studies have shown that interaction with SNs is greater with females than males. In particular, girls spend more time on social media, smartphones, and computers for social networking [37], while boys are more likely to use the Web for online gaming [37,38,39,40,41]. Furthermore, males can be more heavily influenced by marketing strategies and are at higher risk of developing behavioral addiction [42]. By contrast, girls are more likely to use social media as a platform for social comparison and feedback on their appearance and personal value [43,44]. Arguably, this attitude could be the result of girls becoming more self-objectified (i.e., placing more emphasis on how their physical bodies appear to others; [45]. Indeed, the use of SNs is linked to specific concerns about body weight, especially among adolescent girls [46]. The relationship between time spent on SNs and symptoms of anxiety and depression is stronger for girls than boys [47]. This is consistent with other contributions indicating that girls may be more sensitive than boys to feedback given to them on SNs, thus, reacting with higher levels of distress [48,49]. Study Aim The present study can be included in the above-mentioned scientific literature. Its main aim is to study the uses and potential dangers of SNs among a sample of Italian adolescents. In particular: RQ1: What specific characteristics of use and dangerousness do SNs have? RQ2: Are there any differences in the use and dangerousness of SNs among boys versus girls? RQ3: Is it possible to create a model that could explain how SNs can turn out to be dangerous and the mechanism of use and negative reinforcement in the use of SNs? 2. Materials and Methods 2.1. Study Design This paper is based on Grounded Theory: a qualitative approach to the analysis of interviews or focus groups ideal for abstracting the main categories in which they are organized [50]. Its purpose is to generate a model that explains the variability of the data rather than testing an a priori hypothesis. This approach is mainly based on three steps: open coding, axial coding, and selective coding. Researchers must constantly review the connections between abstract categories and data until the latter are completely saturated. In other words, the research team’s work ends when the abstract model explains the greater amount of data variance. At this point the new complete theoretical framework has been developed [51,52]. 2.2. Ethical Issues All study procedures were carried out in accordance with the Declaration of Helsinki. The Institutional Review Board of the Department of Psychology, Sapienza University of Rome (protocol number 1450/2021) approved the procedures and the accompanying consent forms. 2.3. Participants The dataset we analyzed consists of twenty-four focus groups carried out and recorded within Italian middle and high schools (age range 13–18). The focus groups were carried out in 6 middle schools and 6 high schools, respectively. Each class (consisting of about 24 students) was divided in half in order to conduct 24 focus groups. Each focus group was led by a psychologist experienced in leading groups of adolescents and young adults and included approximately twelve participants. The total number of participants was 296, comprising 166 girls and 130 boys. 2.4. Procedural Section Data were collected by using twenty-four focus groups in twelve schools (i.e., Italian middle and high schools), consisting of open-ended questions covering violence and aggression in SNs (for more detailed information see Section 2.4.1). All the focus groups were conducted at the beginning of the pre-pandemic period, from September to December 2019. Participants were interviewed in their classrooms, and each interview lasted about 60 to 90 min. The interviews were conducted without the presence of the teacher, with the aim of making students feel freer to express their opinions. The collection of data was concluded when a saturation point was reached and new interviews did not seem to make any substantial contribution to the model previously generated on early data. The interviews were taped and transcribed verbatim and analyzed according to the Grounded Theory. Data were examined line by line in order to identify students’ opinions, feelings, and actions related to the themes mentioned in the interviews. Codes maintained words used by the scholars in order to maintain the semantic of the data, to verify their descriptive content, and to confirm that they were “grounded” in the data. 2.4.1. Focus Group Questions The core theme of the focus groups was violence and aggression in SNs. The focus groups were conducted using thirteen open questions and leaving the participants’ discussion free until no one had anything more to add. The open questions were the following: 1. How do you communicate with your friends, classmates, and family members when you are not physically together? 2. What digital communication tools do you use? (e.g., WhatsApp, Instagram, Facebook, Snapchat, Twitter, WeChat, Viber, TikTok, others?) 3. Which ones do you enjoy the most? Why? 4. How much time do you spend daily on electronic devices (e.g., Smartphone, tablet, computer, play station, X-box)? For what kind of activities, mainly? 5. Have you ever come into direct contact with content that you consider inappropriate and/or violent? Targeting you personally or others? What was it about? What were your reactions? 6. Is it different for you to communicate aggressively online versus face to face? What do you think is different? 7. In your opinion, are there specific categories of people who can be particularly vulnerable to this inappropriate behavior? (Do gender-related issues come up? If they don’t come up spontaneously, ask question 7a) 7a. “There are many statistics indicating that on social networks there is a prevalence of insulting communications, with words and/or images, directed towards women and girls. Do you have any feedback or experience in relation to that?” 8. Are there behaviors and/or attitudes that when displayed on SNs are more likely to lead to being attacked? Which ones? Why? 9. Who, in your opinion, most frequently communicates aggressively on SNs? What effect does this have on the targeted people, groups, and specific social categories? 10. What are the reasons for this aggressive behavior on SNs? 11. The following is a case that actually happened: a boy created a closed group on Facebook where he posted photos of some of his friends (girls) taken from social media. He then invited other friends (boys) to make sexually explicit comments about the girls. Participants enthusiastically commented, but at a certain point one of the girls discovered what happened. How do you think the girl reacted? 12. Do you know how to recognize an aggression against a person, a group, or a social category on SNs? How do you notice it? 13. In your opinion, can these online aggressions have consequences? What kind of consequences? 2.5. Data Analysis The transcripts of the focus groups were analyzed, highlighting the similarities and differences in the responses of male and female adolescents. In particular, 75% of the transcripts were analyzed, leaving 25% for the theoretical saturation test. The first step of analysis, open coding, is based on dividing the entire transcript into analyzable text segments. Through continuous comparison and brainstorming, the categories that can best represent the meaning of a specific segment of text are abstracted [53]. The second step of Grounded Theory is called axial coding. In this further step, the categories that emerged during the open coding phase are reanalyzed and the relationships between them are highlighted. The group of researchers, therefore, abstracts second-level categories which can group the categories that emerged from the open coding process. The last step in Grounded Theory is selective coding. The goal of selective coding is to connect together, in a single model, the relationships between the second-level categories that emerged during the axial coding phase. This process culminates in showing the reader a theoretical framework suitable to explain most of the variability of the data. The data (the first-level categories and the second-level categories) are finally reanalyzed making sure that new possible categories would not improve the accuracy of the proposed model. Finally, the theoretical saturation test was performed: an analysis of the remaining 25% of the transcript using the first and second level categories that emerged. 3. Results In the open coding phase, after multiple reviews, the group of researchers extracted 10 categories. Examples of the process of forming these categories are shown in Table 1. In the axial coding phase, the group of researchers extracted two main second-level categories that contain the categories of the open coding phase of analysis. These are: (a) Uses of SNs and (b) Dangers of SNs. Table 2 below shows the contents of each category. Included in the “Uses of SNs” we highlight “the use of SNs to know my personal value” and “the use of SNs to be successful” as the riskiest attitudes towards SNs, exposing the subject to potential dangers. During the selective coding phase, a complete theoretical framework was proposed to explain most of the data variability. The following figure shows the relationships existing between the first and second-level categories, highlighting similarities and differences between male and female adolescents (Figure 1). In Figure 1, we have included the categories common to males and females in the center, with the categories belonging to males on the left and the categories belonging to females on the right. Two common categories (“SNs to meet each other” and “Offence”) convey a different meaning depending on whether it is a male or a female adolescent who is speaking. These differences were shown through arrows. The theoretical saturation test applied to the remaining 25% of the data did not generate any further possible category and showed the accuracy of the proposed classification in describing the entire dataset. Finally, a psychological model on how SNs can turn out to be dangerous tools is proposed. We highlighted the interactions among the riskiest uses of SNs and how they mutually reinforce each other (Figure 2). 4. Discussion The present study aimed to investigate uses and dangers of SNs according to a sample of Italian adolescents and build a model for how SNs can turn out to be dangerous. The first two research questions were: (1) what specific characteristics of use and dangerousness do SNs have? (RQ1); and (2) are there any differences in the use and dangerousness of SNs among boys versus girls? (RQ2). As we can evince from Figure 1, the majority of the categories belong to both male and female adolescents. In fact, both genders use social networks to communicate with peers (“SNs to communicate”) and to share images or information (“SNs to share”). Both genders also use SNs to meet each other and meet new people (“SNs to meet each other”), but while for males this means meeting new girls for a possible relationship, for girls it is intended as meeting other girls to support each other. Furthermore, it is mainly males who use social networks to play online games (“SNs to play”), while it is mainly females who use social networks to get feedback on their own value (“SNs to know my personal value”). In this case, the personal value is assigned on the basis of likes and comments received—in other words, on the basis of popularity. In addition, both males and females use social networks to try to have money and success (“SNs to be successful”). This sometimes happens by uploading comments or extreme videos aimed at obtaining as many likes as possible. Examples described by the participants concern a boy who wants to show his ability to set fire to a sofa and eventually the fire spreads throughout the whole house, or a girl who films herself trying to throw herself under a train and is seriously injured, or boys who heavily insult people who have many followers in order to get a lot of comments, and finally, a group of boys who record a homeless man having a seizure and upload the image online instead of calling an ambulance. It is this equation (i.e., “popularity” = “personal value”) that we believe to be a potentially very dangerous ingredient for adolescent wellbeing. In fact, regarding the major dangers that adolescents highlight in the use of SNs, we find four categories: “Violence”; verbal “offences” or offences shown through videos and photos; “pornography”; and receiving an “unreal self-image.” Here we must highlight a difference in how male and female adolescents perceive offences committed in SNs. For males, these are often perceived as camaraderie. The atmosphere of SNs for males often mimics that of the team sports locker room in which the teammates offend each other to make the group laugh. For females, on the other hand, the offences are perceived exclusively as violent acts. However, consistent with the existing literature [5,6,7,12], all participants highlight how frequently violent or offensive content is present in SNs. While the existing literature often highlights the problem of cyberbullying, our results also shed light on another phenomenon. In fact, the females in our study emphasize a further potentially dangerous aspect: coming into contact via SNs with an image of an ideal girl which turns out to be very frustrating for them. Here are the words of a participant: “I uninstalled social media from my phone, and I was much calmer because I no longer saw the stereotype of a girl I am supposed be every day. Since I took it off, it seems strange to say, but my self-esteem has risen more because I always saw types of girls that I knew I couldn’t be.” The constant exposure to an “ideal model” who is perceived to be very distant from their current condition results in a very pronounced feeling of frustration, especially in girls. Furthermore, adolescents point out a final ingredient that characterizes SNs: the lack of persecutory anxiety and sense of guilt. In the words of one participant: “Hiding yourself behind an account, perhaps sometimes fake, causes the insults and offences to start very quickly, and the person has no worries about what can happen after this behavior.” There are numerous contributions in the literature that underline the importance of the sense of guilt and persecutory anxiety as a key mechanism in the formation of a less sadistic and more adequate-to-reality “Super-Ego” [54,55]. In other words, the lack of persecutory anxiety and sense of guilt generates a violent and offensive vicious circle that is sometimes difficult to break (Figure 2). This leads to the third research question of this study: is it possible to create a model that could explain how SNs can turn out to be dangerous? (RQ3). Some adolescents sign into SNs eager for recognition of their personal value and therefore are sensitive to the judgments of others. Some people present in SNs often exhibit violent or offensive behavior due to (a) the sense of protection they perceive by being hidden behind a virtual account and (b) the effort directed toward being successful and popular. These two ingredients, taken together, cause the lack of persecutory anxiety and sense of guilt in those groups exhibiting such behaviors. This type of conduct is particularly frustrating for those adolescents needing recognition of their personal value. In fact, these latter can react in a depressive way, feeling sadness and dejection, or in an aggressive way, populating the group of persons who exhibit violent or offensive behaviors, increasingly in search of popularity, increasingly with the perception of being protected behind a virtual account, increasingly with a lack of persecutory anxiety and sense of guilt (Figure 2). To our knowledge, this is the first time that a model has been proposed in the literature for how SNs can become dangerous for adolescents. The existing literature often focuses on cyberbullying (e.g., 7, 12), more rarely on the problems of envy and popularity (e.g., 6), almost never on the relational ingredients of SNs that promote their emergence. Finally, we emphasize how that vicious circle (Figure 2) is fed by two different types of frustration: frustration coming from violent or offensive behavior, and frustration coming from the constant exposure to an ideal and unattainable self-image (i.e., what in this study we have called “Unreal self-image” and in the literature goes under the name of “Ego-Ideal”). In fact, in the literature it has been highlighted how the indirect frustration coming from the “Ego-Ideal” can be as unbearable as the direct frustration coming from the “Super-Ego” [56,57,58,59,60]. We have called the frustration deriving from the Ego-Ideal “indirect” because it comes from the perceived distance between the ideal self-image and the real self, and we have named the frustration deriving from the Super-Ego “direct” because it represents the introjected moral rules. SNs, therefore, as well as being useful tools in the service of social relations and knowledge, can unfortunately also represent a global, and very frustrating, Ego-Ideal (for the gap between the ideal self and the real self) and/or Super-Ego (for the exhibition of violent and offensive behaviors). This study, while showing important characteristics of the uses and dangers of SN, is certainly not devoid of limitations. The most important lies in the sample size, which cannot be considered representative of Italy as a whole. However, we intend to replicate the work by enlarging the study sample. Finally, another important limitation lies in the lack of quantitative measures to complement the results obtained by means of a purely qualitative methodology. We believe that a quali-quantitative approach can be used in the replication of the present study. 5. Conclusions This study shows important characteristics of the uses and dangers of social networks. Our findings highlighted two main dangers of SNs (the desperate search for popularity and the exhibiting of violent or offensive behavior) which seem to be facilitated by the sense of protection derived from being hidden behind a virtual account. The effort to be popular and the sense of protection cause a lack of persecutory anxiety or sense of guilt, which allows violent and offensive behaviors to be implemented. In fact, it is much easier to exhibit such behaviors without the fear of being persecuted or hurt by others. To our knowledge, this is the first time that a model has been proposed in the literature for how SNs can become dangerous for adolescents. The model has various clinical implications; among them, we highlight: (a) in the case of a patient who suffered of cyberbullying, the health professional (e.g., general practitioner, psychologist, psychiatrist, or psychotherapist) should pay attention to the patient’s internal fragilities that sometimes facilitate such a violent relational dynamic. Specifically, the use of SNs to increase one’s popularity and to know or confirm one’s personal value can be grounded on scarce self-esteem (i.e., scarce benign narcissism), which should be addressed in order to make the patient less dependent on others’ judgments; (b) in the case of a patient who committed cyberbullying, the health professional should instead pay attention to the suffering embedded in such behavior. Sometimes, with the extreme suffering due to the patient’s needs not being adequately seen, the patient’s thoughts and emotions are systematically discarded and derided. In summary, the violent or offensive act often represents an extreme attempt to seek help, to seek an environment capable of catering to the patient’s care. Social media is a pervasive part of modern social life and represents an artificial world in which there is a growing desire to present ideal representations of oneself as an extension of offline identity in which users present relatively authentic versions of themselves. This allows users to create a “virtual self” by performing or editing the content that is presented to others. The artificiality of these platforms highlights a constant tendency toward self-idealization that could be harmful to individual wellbeing. While a growing body of research suggests a mixed effect (i.e., both positive and negative) of social media use on wellbeing [5,61,62,63], our findings, in line with the study by Bailey et al., 2020 [64], suggest that social media can in a sense help individual wellbeing. However, this depends on the way these platforms are used; indeed, while creating an “enhanced” self-figure can be beneficial on the one hand, an authentic self-expression is always preferable and psychologically beneficial on the other. Future research could assess the role of individual differences in self-expression on social media in order to avoid the development of social media addiction. This study aims to highlight the uses and dangers of social networks according to a sample of Italian adolescents for developing prevention procedures against the risks of SNs (e.g., cyberbullying, internet addiction) without demonizing the use of social media as such. Author Contributions Conceptualization, J.B., E.M., F.P., G.L. and A.M.G.; Data curation, G.d.F., J.B., E.M. and A.Q.; Formal analysis, G.d.F., F.P. and A.Q.; Funding acquisition, A.M.G.; Methodology, G.d.F., J.B., E.M., F.P. and A.Q.; Project administration, A.M.G.; Software, A.Q.; Supervision, A.M.G.; Writing—original draft, G.d.F., J.B. and E.M.; Writing—review & editing, G.L. and A.Q. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of the Department of Psychology, Sapienza University of Rome (protocol number 1450/2021) for studies involving humans. Informed Consent Statement Informed consent was obtained from all the subjects involved in the study. Data Availability Statement He data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy considerations. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Relations between the first and second-level categories and the gender of the participants. Note. Items with * represent the riskiest uses of SNs, exposing the subject to potential dangers. Figure 2 A psychological model on how SNs can turn out to be dangerous. Note. Items with * represent the riskiest uses of SNs, exposing the subject to potential dangers. ijerph-19-05691-t001_Table 1 Table 1 Examples of the open coding analysis. Number Category Original Transcript 1 SNs to Communicate I use social media to talk to friends. Sometimes I ask them if they want to play on PlayStation and while we play, we talk. 2 SNs to meet each other I use social media to meet my friends. I send a message to our group, and everyone sees it. In this way it is much easier to meet them. 3 SNs to play Sometimes I spend hours and hours playing because there is nothing to do. When I play with my cousins with a special server that gives us the possibility to play all together, I stay connected a lot. If I am alone at home I am bored. 4 SNs to share We don’t use social media a lot, but they are useful for sharing homework. Sometimes I don’t make them all, there are often a lot of them and after school I’m tired. 5 SNs to know my personal value In SNs I include my photos and photos with my friends. I wait to see the reactions and if others like me. This also happens with comments. I always wait for the reactions of others to see if they appreciate what I write. 6 SNs to be successful I remember when I signed up, I wanted to see if I was successful with others, if they liked my posts and photos. I didn’t expect to be popular but just that others liked me. Since I got a girlfriend, I’ve been online a lot less, connected a lot less. 7 Violence A while ago there was a video of a homeless man at the station who either took drugs or had a seizure. Some guys were filming him on the phone, others were kicking him. I felt bad looking at it. Almost vomiting. There are tons of videos like that. 8 Offence There are people that I don’t even know who comment on my photos saying that “I suck.” I don’t say anything, I’m sad. 9 Pornography Well, it happened to me that some older men sent me pictures of their penis. I blocked them immediately and made the report. 10 Unreal self-image I uninstalled social media from my phone, and I was much calmer because I don’t see the stereotype of a girl I should be every day. Since I took it off, it seems strange to say, but my self-esteem has risen more because I always saw types of girls that I knew I couldn’t be. ijerph-19-05691-t002_Table 2 Table 2 Content of axial coding categories. Category Number Open Coding Category Axial Coding Category 1 SNs to communicate Use of SNs 2 SNs to meet each other 3 SNs to play 4 SNs to share 5 SNs to know my personal value * 6 SNs to be successful * 7 Violence Dangers of SNs 8 Offence 9 Pornography 10 Unreal self-image Note. Items with * represent the riskiest uses of social networks, exposing the subject to potential dangers. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. George M.J. Russell M.A. Odgers C.L. Using mobile technologies to advance the study of psychopathology among children and adolescents The Wiley Handbook of Developmental Psychopathology Centifanti L.C. Williams D.M. Wiley London, UK 2017 45 60 2. Manago A.M. Brown G. Lawley K.A. Anderson G. Adolescents’ Daily Face-to-Face and Computer-Mediated Communication: Associations with Autonomy and Closeness to Parents and Friends Dev. Psychol. 2020 56 153 10.1037/dev0000851 31670551 3. Palfrey J. Gasser U. Born Digital: Understanding the First Generation of Digital Natives 2011 Available online: https://www.readhowyouwant.com/ (accessed on 15 December 2021) 4. Bell V. Bishop D.V. Przybylski A.K. The Debate over Digital Technology and Young People Bmj 2015 351 h3064 10.1136/bmj.h3064 26268481 5. Orben A. Przybylski A.K. The Association between Adolescent Well-Being and Digital Technology Use Nat. Hum. Behav. 2019 3 173 182 10.1038/s41562-018-0506-1 30944443 6. Steers M.-L.N. It’s Complicated: Facebook’s Relationship with the Need to Belong and Depression Curr. Opin. Psychol. 2016 9 2 26 10.1016/j.copsyc.2015.10.007 7. Guan S.-S.A. Subrahmanyam K. Youth Internet Use: Risks and Opportunities Curr. Opin. Psychiatry 2009 2 351 356 10.1097/YCO.0b013e32832bd7e0 8. Subrahmanyam K. Smahel D. Greenfield P. Connecting Developmental Constructions to the Internet: Identity Presentation and Sexual Exploration in Online Teen Chat Rooms Dev. Psychol. 2006 42 395 10.1037/0012-1649.42.3.395 16756432 9. Gerwin R.L. Kaliebe K. Daigle M. The Interplay between Digital Media Use and Development Child Adolesc. Psychiatr. Clin. 2018 7 345 355 10.1016/j.chc.2017.11.002 10. Barry C.T. Sidoti C.L. Briggs S.M. Reiter S.R. Lindsey R.A. Adolescent Social Media Use and Mental Health from Adolescent and Parent Perspectives J. Adolesc. 2017 61 1 11 10.1016/j.adolescence.2017.08.005 28886571 11. Anderson M. Jiang J. Teens, Social Media & Technology 2018 Pew Res. Cent. 2018 31 1673 1689 12. Best P. Manktelow R. Taylor B. Online Communication, Social Media and Adolescent Wellbeing: A Systematic Narrative Review Child. Youth Serv. Rev. 2014 41 7 36 10.1016/j.childyouth.2014.03.001 13. James T.L. Lowry P.B. Wallace L. Warkentin M. The Effect of Belongingness on Obsessive-Compulsive Disorder in the Use of Online Social Networks J. Manag. Inf. Syst. 2017 34 560 596 10.1080/07421222.2017.1334496 14. McCrae N. Gettings S. Purssell E. Social Media and Depressive Symptoms in Childhood and Adolescence: A Systematic Review Adolesc. Res. Rev. 2017 315 330 10.1007/s40894-017-0053-4 15. Sarmiento I.G. Olson C. Yeo G. Chen Y.A. Toma C.L. Brown B.B. Bellmore A. Mares M.-L. How Does Social Media Use Relate to Adolescents’ Internalizing Symptoms? Conclusions from a Systematic Narrative Review Adolesc. Res. Rev. 2020 5 381 404 10.1007/s40894-018-0095-2 16. Orben A. Teenagers, Screens and Social Media: A Narrative Review of Reviews and Key Studies Soc. Psychiatry Psychiatr. Epidemiol. 2020 55 407 414 10.1007/s00127-019-01825-4 31925481 17. Strasburger V. Super-Peer Theory Encyclopedia of Children, Adolescents, and the Media Sage Publications Thousand Oaks, CA, USA 2007 18. Moreno M.A. Whitehill J.M. Influence of Social Media on Alcohol Use in Adolescents and Young Adults Alcohol Res. Curr. Rev. 2014 36 91 19. Agustina J.R. Montiel I. Gamez-Guadix M. Cibercriminología y Victimización Online (Cibercriminology and Online Victimization) Piramide Madrid, Spain 2020 20. Costello M. Rukus J. Hawdon J. We Don’t like Your Type around Here: Regional and Residential Differences in Exposure to Online Hate Material Targeting Sexuality Deviant Behav. 2019 40 385 401 10.1080/01639625.2018.1426266 21. Wachs S. Wright M.F. The Moderation of Online Disinhibition and Sex on the Relationship between Online Hate Victimization and Perpetration Cyberpsychol. Behav. Soc. Netw. 2019 2 300 306 10.1089/cyber.2018.0551 22. Smith P.K. Cyberbullying and Cyber Aggression Handbook of School Violence and School Safety: International Research and Practice Jimerson S.R. Nickerson A.B. Mayer M.J. Furlong M.J. Routledge Oxforshire, UK 2012 23. Blaya C. Audrin C. Toward an Understanding of the Characteristics of Secondary School Cyberhate Perpetrators Frontiers 2019 4 46 10.3389/feduc.2019.00046 24. Wachs S. Wright M.F. Vazsonyi A.T. Understanding the Overlap between Cyberbullying and Cyberhate Perpetration: Moderating Effects of Toxic Online Disinhibition Crim. Behav. Ment. Health 2019 9 179 188 10.1002/cbm.2116 25. Navarro J.I. The Psychology of Hatred Open Criminol. J. 2013 6 10 17 10.2174/1874917801306010010 26. Surís J.-C. Akre C. Piguet C. Ambresin A.-E. Zimmermann G. Berchtold A. Is Internet Use Unhealthy? A Cross-Sectional Study of Adolescent Internet Overuse Swiss Med. Wkly. 2014 144 w14061 10.4414/smw.2014.14061 25474244 27. Guinta M.R. John R.M. Social Media and Adolescent Health Pediatr. Nurs. 2018 44 196 201 28. Starcevic V. Aboujaoude E. Internet Addiction: Reappraisal of an Increasingly Inadequate Concept CNS Spectr. 2017 2 7 13 10.1017/S1092852915000863 26831456 29. Starcevic V. Billieux J. Does the Construct of Internet Addiction Reflect a Single Entity or a Spectrum of Disorders? Clin. Neuropsychiatry 2017 14 5 10 30. Quaglieri A. Biondi S. Roma P. Varchetta M. Fraschetti A. Burrai J. Lausi G. Martí-Vilar M. González-Sala F. Di Domenico A. From Emotional (Dys) Regulation to Internet Addiction: A Mediation Model of Problematic Social Media Use among Italian Young Adults J. Clin. Med. 2021 11 188 10.3390/jcm11010188 35011929 31. Mok K. Jorm A.F. Pirkis J. Suicide-Related Internet Use: A Review Aust. N. Z. J. Psychiatry 2015 49 697 705 10.1177/0004867415569797 25698810 32. Buzzi C. Tucci M. Ciprandi R. Brambilla I. Caimmi S. Ciprandi G. Marseglia G.L. The Psycho-Social Effects of COVID-19 on Italian Adolescents’ Attitudes and Behaviors Ital. J. Pediatr. 2020 46 69 10.1186/s13052-020-00833-4 32448323 33. Burrai J. Barchielli B. Cricenti C. Borrelli A. D’amato S. Santoro M. Vitale M. Ferracuti S. Giannini A.M. Quaglieri A. Older Adolescents Who Did or Did Not Experience COVID-19 Symptoms: Associations with Mental Health, Risk Perception and Social Connection Int. J. Environ. Res. Public. Health 2021 18 5006 10.3390/ijerph18095006 34065093 34. Lansford J.E. Dodge K.A. Fontaine R.G. Bates J.E. Pettit G.S. Peer Rejection, Affiliation with Deviant Peers, Delinquency, and Risky Sexual Behavior J. Youth Adolesc. 2014 43 1742 1751 10.1007/s10964-014-0175-y 25150986 35. Bernat D.H. Resnick M.D. Connectedness in the Lives of Adolescents Adolescent Health: Understanding and Preventing Risk Behaviors DiClemente R.J. Santelli J.S. Crosby R.A. Wiley Hoboken, NJ, USA 2009 36. Underwood M.K. Ehrenreich S.E. The Power and the Pain of Adolescents’ Digital Communication: Cyber Victimization and the Perils of Lurking Am. Psychol. 2017 72 144 10.1037/a0040429 28221066 37. Twenge J.M. Martin G.N. Gender Differences in Associations between Digital Media Use and Psychological Well-Being: Evidence from Three Large Datasets J. Adolesc. 2020 79 91 102 10.1016/j.adolescence.2019.12.018 31926450 38. Willoughby T. A Short-Term Longitudinal Study of Internet and Computer Game Use by Adolescent Boys and Girls: Prevalence, Frequency of Use, and Psychosocial Predictors Dev. Psychol. 2008 44 195 10.1037/0012-1649.44.1.195 18194017 39. Booker C.L. Skew A.J. Kelly Y.J. Sacker A. Media Use, Sports Participation, and Well-Being in Adolescence: Cross-Sectional Findings from the UK Household Longitudinal Study Am. J. Public Health 2015 105 173 179 10.2105/AJPH.2013.301783 25494209 40. Witt E.A. Massman A.J. Jackson L.A. Trends in Youth’s Videogame Playing, Overall Computer Use, and Communication Technology Use: The Impact of Self-Esteem and the Big Five Personality Factors Comput. Hum. Behav. 2011 7 763 769 10.1016/j.chb.2010.10.025 41. Brodersen N.H. Steptoe A. Williamson S. Wardle J. Sociodemographic, Developmental, Environmental, and Psychological Correlates of Physical Activity and Sedentary Behavior at Age 11 to 12 Ann. Behav. Med. 2005 9 11 10.1207/s15324796abm2901_2 42. Yu L. Shek D.T.L. Internet Addiction in Hong Kong Adolescents: A Three-Year Longitudinal Study J. Pediatr. Adolesc. Gynecol. 2013 6 S10 S17 10.1016/j.jpag.2013.03.010 43. Nesi J. Prinstein M.J. Using Social Media for Social Comparison and Feedback-Seeking: Gender and Popularity Moderate Associations with Depressive Symptoms J. Abnorm. Child Psychol. 2015 43 1427 1438 10.1007/s10802-015-0020-0 25899879 44. Yau J.C. Reich S.M. It’s Just a Lot of Work: Adolescents’ Self-Presentation Norms and Practices on Facebook and Instagram J. Res. Adolesc. 2019 9 196 209 10.1111/jora.12376 45. Fredrickson B.L. Roberts T. Objectification Theory: Toward Understanding Women’s Lived Experiences and Mental Health Risks Psychol. Women Q. 1997 1 173 206 10.1111/j.1471-6402.1997.tb00108.x 46. Tiggemann M. Slater A. Facebook and Body Image Concern in Adolescent Girls: A Prospective Study Int. J. Eat. Disord. 2017 50 80 83 10.1002/eat.22640 27753130 47. Thorisdottir I.E. Sigurvinsdottir R. Asgeirsdottir B.B. Allegrante J.P. Sigfusdottir I.D. Active and Passive Social Media Use and Symptoms of Anxiety and Depressed Mood among Icelandic Adolescents Cyberpsychol. Behav. Soc. Netw. 2019 2 535 542 10.1089/cyber.2019.0079 48. Asgeirsdottir B.B. Gudjonsson G.H. Sigurdsson J.F. Sigfusdottir I.D. Protective Processes for Depressed Mood and Anger among Sexually Abused Adolescents: The Importance of Self-Esteem Personal. Individ. Differ. 2010 49 402 407 10.1016/j.paid.2010.04.007 49. Haggerty R.J. Garmezy N. Sherrod L.R. Rutter M. Stress, Risk, and Resilience in Children and Adolescents: Processes, Mechanisms, and Interventions Cambridge University Press Cambridge, UK 1996 50. Glaser B.G. Strauss A.L. The Discovery of Grounded Theory: Strategies for Qualitative Research Routledge Oxfordshire, UK 2017 51. Charmaz K. Grounded Theory: Objectivist and Constructivist Methods Handb. Qual. Res. 2000 2 509 535 52. Dey I. Grounded Theory Qualitative Research Practice Seale C. Gobo G. Gubrium J. Silverman D. Sage London, UK 2004 80 93 53. Strauss A. Corbin J. Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory Sage Publications London, UK 1998 0-8039-5940-0 54. Klein M. Criminal Tendencies in Normal Children Psychoanal. Rev. 1931 18 17 55. Klein M. The Early Development of Conscience in the Child Psychoanal. Today 1933 149 162 56. Groot J.L.-D. Ego Ideal and Superego Psychoanal. Study Child 1962 17 94 106 10.1080/00797308.1962.11822840 57. Hanly C. Ego Ideal and Ideal Ego Int. J. Psychoanal. 1984 65 53 261 58. Halfon S. Çavdar A. Orsucci F. Schiepek G.K. Andreassi S. Giuliani A. De Felice G. The Non-Linear Trajectory of Change in Play Profiles of Three Children in Psychodynamic Play Therapy Front. Psychol. 2016 7 1494 10.3389/fpsyg.2016.01494 27777561 59. Halfon S. Cavdar A. Paoloni G. Andreassi S. Giuliani A. Orsucci F.F. de Felice G. Monitoring Nonlinear Dynamics of Change in a Single Case of Psychodynamic Play Therapy Nonlinear Dyn. Psychol. Life Sci. 2019 23 113 135 60. de Felice G. Palmiero N. Mirabella A. Belli A. Nese A. Nese G. De Vita G. A Psychoanalytic Contribution to the Understanding of Criminal Tendencies Psychoanal. Psychother. 2020 34 309 324 10.1080/02668734.2021.1875026 61. Gentile B. Twenge J.M. Freeman E.C. Campbell W.K. The effect of social networking websites on positive self-views: An experimental investigation Comput. Hum. Behav. 2012 8 1929 1933 10.1016/j.chb.2012.05.012 62. Allcott H. Braghieri L. Eichmeyer S. Gentzkow M. The Welfare Effects of Social Media (No. w25514) Natl. Bur. Econ. Res. 2019 110 629 676 63. Heffer T. Good M. Daly O. MacDonell E. Willoughby T. The longitudinal association between social-media use and depressive symptoms among adolescents and young adults: An empirical reply to Twenge et al. (2018) Clin. Psychol. Sci. 2019 7 462 470 10.1177/2167702618812727 64. Bailey E.R. Matz S.C. Youyou W. Iyengar S.S. Authentic self-expression on social media is associated with greater subjective well-being Nat. Commun. 2020 11 4889 10.1038/s41467-020-18539-w 33024115

PMC009xxxxxx/PMC9099660.txt

==== Front Int J Environ Res Public Health Int J Environ Res Public Health ijerph International Journal of Environmental Research and Public Health 1661-7827 1660-4601 MDPI 10.3390/ijerph19094978 ijerph-19-04978 Systematic Review Pharmacological Treatment of Acute Psychiatric Symptoms in COVID-19 Patients: A Systematic Review and a Case Series Carmassi Claudia 1 Pacciardi Bruno 1 Gravina Davide 1* Fantasia Sara 1 De Pascale Gennaro 2 https://orcid.org/0000-0001-8135-6284 Cutuli Salvatore Lucio 2 https://orcid.org/0000-0003-1752-0339 Bertelloni Carlo Antonio 1 Dell’Osso Liliana 1 Bravaccio Carmela Academic Editor 1 Psychiatric Unit, Department of Clinical and Experimental Medicine, AOUP, University of Pisa, 56126 Pisa, Italy; claudia.carmassi@unipi.it (C.C.); pacciardib@libero.it (B.P.); dr.fantasiasara@gmail.com (S.F.); carlo.ab@hotmail.it (C.A.B.); liliana.dellosso@unipi.it (L.D.) 2 Department of Emergency, Catholic University of the Sacred Heart, 00168 Rome, Italy; gennaro.depascale@unicatt.it (G.D.P.); sl.cutuli@gmail.com (S.L.C.) * Correspondence: davide.gravina@hotmail.it 20 4 2022 5 2022 19 9 497805 1 2022 12 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Delirium and psychomotor agitation are relevant clinical conditions that may develop during COVID-19 infection, especially in intensive care unit (ICU) settings. The psychopharmacological management of these conditions is receiving increasing interest in psychiatry, considering hyperkinetic delirium as one of the most common neuropsychiatries acute consequences in COVID-19 recovery patients. However, there are no actual internationally validated guidelines about this topic, due to the relatively newly introduced clinical condition; in addition, a standardized psychopharmacologic treatment of these cases is a complex goal to achieve due to the risk of both drug–drug interactions and the vulnerable conditions of those patients. The aim of this systematic review and case series is to evaluate and gather the scientific evidence on pharmacologic handling during delirium in COVID-19 patients to provide practical recommendations on the optimal management of psychotropic medication in these kinds of patients. The electronic databases PubMed, Embase and Web of Science were reviewed to identify studies, in accordance with the PRISMA guidelines. At the end of the selection process, a total of 21 studies (n = 2063) were included. We also collected a case series of acute psychomotor agitation in COVID-19 patients hospitalized in ICU. Our results showed how the symptom-based choice of the psychotropic medication is crucial, and even most of the psychotropic drug classes showed good safety, one must not underestimate the possible drug interactions and also the possible decrease in vital functions which need to be strictly monitored especially during treatment with some kinds of molecules. We believe that the evidence-based recommendations highlighted in the present research will enhance the current knowledge and could provide better management of these patients. psychomotor agitation delirium acute psychosis restlessness COVID-19 (coronavirus disease 2019) SARS-CoV-2 ==== Body pmc1. Introduction Coronavirus disease 2019 (COVID-19) is a pandemic infection caused by a novel strain of coronavirus named Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2). Coronaviruses are single-stranded RNA viruses and several subtypes affecting humans have been identified, most of which cause upper respiratory tract infections in immunocompetent individuals [1]. Already known strains of coronavirus caused the severe acute respiratory syndrome (SARS) outbreak, starting in 2002, and the Middle East’s respiratory syndrome (MERS) outbreak, starting in 2012 [2]. The outbreak of SARS-CoV-2 was declared a pandemic by WHO in March 2020. Although most prominently associated with pulmonary manifestations, COVID-19 is increasingly implicated in neuropsychiatric complications, including delirium and psychosis [3]. This seems to be due to the neurotropic properties of the virus [4] leading to a direct infiltration damage during the central nervous system (CNS) invasion and an indirect damage due to the cytokine spread caused by the dysregulation of the inflammatory responses who could exacerbate the neurocognitive impairment [5,6,7]. Given the rapid diffusion of the COVID-19 virus and the increase in social isolation, it has become necessary to implement the at-distance consultation and tele-psychiatry assessment for mild clinical cases [8,9] and, on the other hand, to take care of acute psychopathological symptoms in patients admitted to hospitals where the COVID-19 infection is managed and treated in specific wards [10]. Delirium and psychomotor agitation are significant clinical conditions that may develop during COVID-19 infection, especially in intensive care unit (ICU) settings, in patients with acute respiratory distress and in isolation environments [11]. Delirium is defined as a state of acute confusion presenting with a change in mental status, associated with an altered level of consciousness, impaired attention and concentration, and disorganized thinking or perceptual disturbance until hallucinations, illusions, and misinterpretation of senses [12,13]. Episodes of psychomotor agitation or hyperkinetic delirium during COVID-19 infections are among the most frequent psychiatric conditions, with an incidence of approximately 65–80% in the ICU [14], and in these cases the patient may be particularly at risk if acute psychiatric symptoms are not immediately treated, and their medical treatment promptly restored. Psychiatric assessment and psychopharmacological treatment of patients hospitalized for complications of COVID-19 infection may be necessary when acute psychiatric conditions interfere or, in some cases, make impossible to move forward with medical assistance. Most severe cases of COVID-19-related pneumonia are treated in anesthesiologic and reanimation wards or in the ICU, where acute psychopathological symptoms may complicate the course of the infective disorder; in fact, delirium is associated with prolonged hospitalization, long-term cognitive and functional impairment, and increased mortality [14,15,16]. The goal of the psycho-pharmacological treatment of the acute psychopathological symptomatology is a first stabilization on a psychic and behavioral level, that will allow the prosecution of the normal therapeutic process. The general medical condition of COVID-19-infected patients with symptomatic pneumonia and the necessary medical treatment, do not allow psychiatrists to use all available psychotropic drugs to treat acute psychiatric symptoms. Severe impairment of respiratory function, QTc prolongation, and drug–drug interactions are just some of the many issues that clinicians must consider when they need to use psychotropic compounds in hospitalized patients with COVID-19 infections. Given our cursory understanding of the pathophysiology of delirium in patients with COVID-19, treatment decisions must be based on symptom presentation, underlying medical comorbidities, and consideration of medication interactions. In people with COVID-19, psychotropic medications may interact with the medical treatments, and some of their adverse effects may worsen the course and outcome of the underlying medical condition [17]. It is also worth noticing that available recommendations for drug treatment of delirium in COVID-19 patients are for off-label use and that they have been extrapolated from literature and reflecting the general practice patterns of the different workgroups [18]. All classes of psychotropic medications have potentially relevant safety issues for people with COVID-19. Unavoidably, in clinical practice, the risk of unfavorable outcomes needs to be carefully weighed on a case-by-case basis, considering several co-existing risk factors. Moreover, although different safety issues have been explored separately, they are actually broadly overlapping (i.e., respiratory function might be impaired by both the sedative effect of medications and the increased risk for respiratory infections). All antipsychotics have warnings or explicit contraindications for the use in people with risk of QTc prolongation and for the association with some of the commonly used anti-COVID medical treatments. From a regulatory standpoint, only a few medications have a marketing authorization for at least one of the conditions considered [19]. Regulatory data indicated that most of the medications considered are off-label in people with COVID-19 and delirium, and their prescription should therefore strictly follow the medico-legal procedures for off-label prescribing, being particularly alert of any unexpected safety issues [20]. This situation applies particularly to people with COVID-19, considering that many medical treatments are similarly being used off-label or compassionately [21]. In its complexity, COVID-19 infection provides a paradigmatic example of how standard treatment procedures, being designed around “average” patients, are hardly applicable as complexity increases. Given the circumstances, clinicians must be particularly vigilant when initiating psychotropic agents in patients receiving medical drugs for COVID-19 [17]. According to Ostuzzi et al., few medications showed potential benefits for the treatment of delirium, and possible benefits only emerged for Quetiapine and Dexmedetomidine in ICU settings [19]. The risk of sedation, and potentially associated respiratory impairment, appears to be higher for first-generation antipsychotics, Benzodiazepines, Dexmedetomidine, and some antidepressants (Mirtazapine and Trazodone). According to some authors, Quetiapine, Risperidone, and Aripiprazole are potentially effective medications for the short-term treatment of hyperactive delirium, and might represent an alternative to conventional treatments, such as Haloperidol in COVID-19 patients [19]. Few studies also suggest that Dexmedetomidine and other second-generation antipsychotics are potentially effective medications for the short-term treatment of hyperactive delirium, and might represent an alternative to conventional treatments, such as Haloperidol [19]. Taking into account anecdotal observations, some patients with COVID-19 delirium appear to have increased rates of myoclonus, rigidity, alogia, and abulia, suggesting a dopamine-depletion state. When there are no absolute contraindications, other authors report the use of low-potency antipsychotics to manage behavioral disturbance [18]. Additionally, in these kinds of low dopamine level-related symptoms, Pimavanserin could be a potential candidate [22]. Nowadays, there is a lack of specific studies on COVID-19-related psychopathology and medical treatments of COVID-19-related psychiatric symptoms are necessarily empirical and based on non-specific evidence. Even less is the available evidence about acute psychopathology and psychotropic use in COVID-19-infected patients hospitalized in ICU. Therefore, there are no guidelines for the management of delirium in patients with COVID-19, and the evidence base for treatment is exceedingly thin [18]. Based on these premises, to contribute to the assistance of this very specific and only recently defined population of patients, we carried out a systematic review of the scientific literature and we also collected four cases of acute psychopathology in patients hospitalized in an intensive care unit because of COVID-19 infection complications. 2. Materials and Methods 2.1. Literature Search This systematic search was conducted in accordance with the Preferred Reporting Items for Reviews and Meta-Analyses Statement (PRISMA) guidelines [23]. We have examined, from 1 September 2021 to 28 March 2022, the following databases: PubMed, Embase, and Web of Science. We developed database-specific search strategies including a combination of controlled vocabulary terms and keywords. The basic search string used towards this review was: (“Psychomotor Agitation” [Mesh] OR “Psychomotor Agitation” [Title/Abstract] OR “Delirium” [Mesh] OR “Delirium” [Title/Abstract] OR “Acute Psychosis” [Title/Abstract]) AND (“COVID-19” [Mesh] OR “COVID-19” [Title/Abstract]). The full search strategies for each of the databases included and the relative database-specific string utilized are provided and viewable in the Supplementary Materials. 2.2. Eligibility Criteria The criteria used to include articles in this review were as follows:Case report or studies including human subjects with a diagnosis of COVID-19 and psychomotor agitation or hyperkinetic delirium; Case report or studies that only included individuals of ages > 17; Articles available in English. Because we aimed to investigate the therapeutic side effects of psychopharmacological therapy of hyperkinetic delirium in patients with COVID-19, studies or case reports in which the type of psychotropic drug or outcome of the patient was not specified, were excluded. 2.3. Study Selection and Data Extraction The publication selection process and data extraction were independently conducted by two of the authors D.G (Davide Gravina) and S.F. (Sara Fantasia). The reviewers analyzed the title and abstract of the articles excluding duplicates and those that did not meet the inclusion criteria or were not available. Subsequently, full text articles were evaluated and those that did not meet the inclusion criteria were removed. The following information was extracted: the name of the first author, year of publication, country of research, and study type. For case reports, we collected patient characteristics (gender, age, preexisting pathological condition), treatment, side effects, and outcome. For the other types of studies, sample size, delirium group size, type of population, previous psychiatric history, mean age, treatment, and results, were extracted. Any discrepancy that emerged during the process was discussed and consensus reached. Any disagreements were discussed and resolved by a third author C.C. (Claudia Carmassi). 2.4. Quality Assessment The quality of the case reports was assessed by a standardized tool adapted from Murad et al. [24]. Furthermore, we used the Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies (QATOCCSS) to assess the quality of the other types of studies [25]. Each study was scored as either “good,” “fair,” or “poor” (see Table 1 and Table 2). The quality assessment was performed by two independent reviewers (D.G. and S.F.) and a third reviewer (C.C.) cross-checked the quality assessment results. Disagreements were discussed and resolved with the research team. 3. Results A total of 2063 results was produced in the primary database research. After that, 1835 articles were removed after titles because they were duplicates (n = 1006) or not relevant (n = 829), and 172 were removed because of other publication type (n = 162) or full text was not available or was not in English (n = 10). Subsequently, 35 publications were excluded because pharmacological therapy was not specified or because of a lack of other eligibility criteria. No suitable articles emerged from manually screening the references cited in the selected publications and in the review. Finally, 21 articles were included in the present review. Assessment for inclusion or exclusion is summarized in a flow diagram according to PRISMA recommendations [23]. The study selection process is outlined in a flow diagram (Figure 1). In total, 21 publications were provided by the search, including 16 case reports, 3 case series, 1 cross-sectional study, and 2 single-center cohort study, ranging from 2020 to 2021. Details of each study included in the review are reported in Table 1 and Table 2. 3.1. Case Report and Case Series 3.1.1. Sample In the present search, the sample is equally divided between male and female (50%) and an average age of 55.96 years was reported; in two articles, the age of the subject was not available. In 53.57% of the cases, the subjects suffered from a pre-existing pathological condition: in nine cases, psychiatric disorders (one PTSD, four schizoaffective disorder, one history of acute psychotic reaction during febrile state, one paranoid personality disorder, one alcohol use disorder, one schizophrenia), and in eleven cases other systemic diseases (such as hypertension, hypothyroidism, diabetes type 2, coronary artery disease and aortic stenosis, asthma and atopic dermatitis, epilepsy, colon cancer, atrial fibrillation, dementia, congestive heart failure, osteoarthritis, chronic obstructive pulmonary disease, hyperlipidemia, right bundle branch block, pre-eclampsia at30 weeks pregnant). In ten cases, there were comorbidities between the previous disorders and in thirteen cases the subject was previously healthy. 3.1.2. Psychopharmacological Treatment In 96.43% of cases, at least one antipsychotic was used, specifically in six cases a single antipsychotic was used, in seven cases a combination of antipsychotics, in eight cases a combination of antipsychotics and BDZ, in one case a combination of antipsychotics with other psychopharmacologic drugs, and in one case a single BDZ. The main types of antipsychotics (42.86%) used were Haloperidol (twelve case reports) and other types of antipsychotics were Risperidone (in nine cases), Quetiapine (in seven cases), Olanzapine (in five cases), Clozapine (in two cases), Chlorpromazine (in two cases), Aripiprazole (in two cases), and Ziprasidone (in one case). Lorazepam was chosen in six of the nine cases in which a BDZ was used; Clonazepam, Midazolam or Diazepam were preferred in the other three. 3.1.3. Side Effects In 22 cases from the 28 included in this review, no side effects were reported. In the other six cases, side effects were reported: one pneumothorax, one prolonged QTc, one neuroleptic malignant syndrome, one rash of the right lower extremity, one diffuse rash surrounding injection site, and one case of fever associated with tachycardia, rise in white blood cells, and decline in CK value. 3.1.4. Outcome A good outcome is reported in 17 cases from the 28 selected, specifically in 5 case a single antipsychotic was used, in 4 cases a combination of antipsychotics, in 6 cases a combination of antipsychotics and BDZ, and in 2 cases a combination of antipsychotics and other psychopharmacologic drugs (Valproic Acid or Gabapentin). Considering the only case about a pregnant patient, the therapy was based on a combination of antipsychotic (Haloperidol) and BDZ (Midazolam) and the patient had remission of acute psychiatric symptoms; however, neonatal death was reported on day 31. In eight of the remaining ten cases, a worse outcome has been reported: in four of them the psychopharmacological treatment (one with a single BDZ, one with a single antipsychotic, one with a combination of antipsychotics, one with a single antipsychotic plus Trazodone) was insufficient to resolve the delirium. In the other five cases (three cases with a single antipsychotic plus valproic acid and two cases with a combination of antipsychotics), a prolonged hospitalization and the progressive worsening of general conditions were reported. In the last case, the patient was transferred to the ICU and intubated for airway protection and to facilitate sedation; two weeks later, the patient was extubated, and delirium was improved. 3.2. Other Type of Studies 3.2.1. Type In this review were also included one cross-sectional study and two single-center cohort studies. 3.2.2. Sample In two studies, the samples included male and female patients admitted because of COVID-19; in the other one, the sample includes older patients admitted because of COVID-19-developing delirium. The mean age of all patients of the samples was 61.5. 3.2.3. Psychopharmacological Treatment In both studies with patients with previous psychiatric history and previous psychopharmacological treatment, a therapeutic change has been required; for the other patients, a new therapy was introduced. In two studies, psychopharmacological therapy was based on antidepressants, BDZ, antipsychotic (mostly Haloperidol), or anticonvulsant; however, in the other studies, Propofol, BDZ (Midazolam or Lorazepam) or Dexmedetomidine were used. 3.2.4. Outcome The cross-sectional study does not show any statistically significant difference in the length of hospitalization or mortality between patients experiencing delirium and the non-delirium group. In one of the single-center cohort studies, the length of hospitalization, ICU length of stay, and duration of mechanical ventilation were higher in the delirium group; however, in the other single-center cohort study, a mortality rate of 71% has been reported. 4. Case Series We collected four case reports of acute psychomotor agitation in COVID-19 patients hospitalized in the ICU. These patients treated with first-generation (low potency) antipsychotics and Trazodone in order to provide further information about the choice of medications for the management of delirium in people with COVID-19. These patients also received complex medical treatment that included medications used for the COVID-19 treatment (dexamethasone, etc.) as well as anesthetics (Propofol, Remifentanil, Midazolam), first-generation low-potency antipsychotics (Chlorpromazine), and antidepressants (Trazodone) because of a condition of hyperkinetic delirium that made it impossible to proceed with medical treatment. Psychotropic treatment led to different outcomes improving the medical conditions of the patients; however, in two instances, part of the treatment was not effective and lead to complications. Based on the reported cases, some clinical considerations about psycho-pharmacological treatment of COVID-19-infected patients are discussed. 4.1. Case Scenario 1 This report is about a 59-year-old female patient, who was affected by hypertension and obesity (55.8 kg/m2). She was admitted to the ICU for acute respiratory failure due interstitial pneumonia caused by SARS-CoV-2 infection. Accordingly, she received daily dexamethasone 6 mg/dL for 10 days, that caused hyperglycemia and required insulin administration. The patient was intubated and mechanically ventilated for 31 days, during which she was sedated with Propofol (5 mg/min) and Remifentanil (4.95 mcg/min) continuous intravenous infusion. Sedation was titrated to maintain Bispectral Index (BIS) values between 40–60 and Richmond Agitation-Sedation Scale at −5 while paralyzed, in order to prevent undesirable awareness. On day 11, Propofol was shifted to Midazolam (15 mg/h) intravenous infusion because of high triglyceride serum concentration (450 mg/dL). The patient was tracheotomized after three weeks of oro-tracheal intubation. Although pulmonary function recovered after 20 days of invasive mechanical ventilation, weaning from the latter was slow and complicated by patient agitation when intravenous sedation was stopped. Confusion assessment method for the ICU (CAM-ICU) revealed hyperkinetic delirium. In light of this view, Chlorpromazine (8.3 mg/h) intravenous infusion was started for five days, which was associated with delirium resolution. The patient was discharged alive from the ICU after 45 days of stay. 4.2. Case Scenario 2 This report is about a 52-year-old male patient, whose past medical history did not reveal comorbidities. He was admitted to the ICU for acute respiratory failure due to interstitial pneumonia caused by SARS-CoV-2 infection. Accordingly, he received daily dexamethasone 6 mg/dL for 10 days. The patient was intubated and mechanically ventilated for 52 days. Moreover, he required ECMO for 27 days due to severe hypoxemia. The patient was sedated with Propofol (3 mg/min) and Remifentanil (4.5 mcg/min) continuous intravenous infusion. Sedation was titrated to maintain Bispectral Index (BIS) values between 40–60 and Richmond Agitation-Sedation Scale at −5 while paralyzed, in order to prevent undesirable awareness. The patient was tracheotomized after three weeks of oro-tracheal intubation. Although pulmonary function recovered after 45 days of invasive mechanical ventilation, weaning from the latter was slow and complicated by patient agitation when intravenous sedation was stopped. Confusion assessment method for the ICU (CAM-ICU) revealed hyperkinetic delirium. In light of this view, Trazodone (50 mg tid for 23 days) and Chlorpromazine (50 mg bid for 5 days) intravenous infusion were started with significant agitation improvement. The patient was discharged alive from the ICU after 56 days of stay. 4.3. Case Scenario 3 This report is about a 54-year-old male patient, whose past medical history did not reveal comorbidities. He was admitted to the ICU for acute respiratory failure due to interstitial pneumonia caused by SARS-CoV-2 infection. The patient was intubated and mechanically ventilated for 20 days, during which he was sedated with Propofol (5 mg/min) and Remifentanil (4.5 mcg/min) continuous intravenous infusion. Sedation was titrated to maintain Bispectral Index (BIS) values between 40–60 and Richmond Agitation-Sedation Scale at −5, while paralyzed, in order to prevent undesirable awareness. Although pulmonary function recovered after 15 days of invasive mechanical ventilation, weaning from the latter was slow and complicated by patient agitation when intravenous sedation was stopped. Confusion assessment method for the ICU (CAM-ICU) revealed hyperkinetic delirium. In light of this view, Trazodone (100 mg bid for 22 days) and Chlorpromazine (6.3 mg/h for 13 days) intravenous infusion were started with significant agitation improvement. The patient was discharged alive from the ICU after 34 days of stay. 4.4. Case Scenario 4 This report is about a 72-year-old male patient, whose past medical history did not reveal comorbidities. He was admitted to the ICU for acute respiratory failure due to interstitial pneumonia caused by SARS-CoV-2 infection. Accordingly, he received daily dexamethasone 6 mg/dL for 10 days. Initially, respiratory failure was managed using noninvasive respiratory supports. However, he developed hyperkinetic delirium (RASS 3, CAM-ICU positive) that was managed with intravenous Midazolam (6 mg) administration. Such intervention ended up in further agitation (RASS 4) in a few minutes, that required profound sedation by Propofol and Remifentanil continuous infusion to target RASS-1. Consequently, he received oro-tracheal intubation and invasive mechanical ventilation to improve severe hypoxemia. Since then, 53 days has passed, and the patient is still admitted to the ICU due to severe clinical conditions. 5. Discussion The aim of this review is to summarize and integrate the existing knowledge on pharmacological treatment of acute psychiatric symptoms in COVID-19 patients. At the actual time of writing, there are no other systematic reviews in literature which collect direct data on these kinds of samples. The challenge is due to both the lack of studies on this topic and the particularly complex management of episodes of hyperkinetic delirium or other forms of acute psychopathology in this peculiar population of patients, given the vulnerability of these subjects to side effects of psychotropic treatments. These premises were confirmed by this search, in fact in just three observational studies we met our inclusion criteria, apart from the case reports. Moreover, of these three studies, one of them showed prolonged length of hospitalization, ICU length of stay, and duration of mechanical ventilation in the delirium group compared to patients without psychiatric acute symptoms [6]. In the second, although it does not have statistically significant results, we found higher mortality in patients experiencing delirium [14]. Finally, in the third one, which included older patients admitted because of COVID-19-developing delirium, a mortality rate of 70% is reported [42]. This vulnerability derives from the poor general medical condition of COVID-19-infected patients (severe impairment of respiratory function, etc.), high chances of comorbidity with other general medical conditions, an often-reduced functionality of the excretory organs and the frequent necessity of medical treatment leading to QTc prolongation and with high probability of drug–drug interactions. Pharmacologic management should be reserved for patients with severe agitation which would result in the interruption of essential medical therapies (such as mechanical ventilation or dialysis catheters) or result in self-harm, or for patients with extremely distressing psychotic symptoms (such as hallucinations or delusions). Intramuscular injections of typical antipsychotics and BDZ, given alone or in combination, have been the treatment of choice for psychomotor agitation over the past few decades, with Haloperidol and Lorazepam being among the most widely used agents [43]. The data obtained from this systematic review also confirmed choice in patients affected by COVID-19, although these drugs have questionable tolerability profiles, and their use may be particularly problematic in patients with COVID-19 infections. In COVID-19 patients receiving pharmacological treatment, a QTc prolongation is possible with several of the drugs commonly used in this disorder (chloroquine/hydroxychloroquine, antibiotics and anti-virals) and QTc prolongation may be worsened by adding psychotropics to the treatment. Some of the drugs commonly used to treat COVID-19 infection (such as Lopinavir/Ritonavir, Chloroquine/Hydroxychloroquine, Doxorubicin, Ceftriaxone, and other antibiotics) may have clinically significant interactions with commonly used psychotropics and their interactions with them should be accurately taken into consideration in treatment planning of acute psychiatric conditions. Particularly, Haloperidol is involved in a potentially lethal cardiac arrhythmia named “torsade de pointes.” This finding led European drug authorities to issue a black box warning. After that, a prolongation of QTc interval in ECG of some individuals treated with Haloperidol was related with fatal torsade de pointes and the use of this drug became strictly regulated, especially in its parenteral formulation [44,45,46]. In COVID-19 patients receiving pharmacological treatment, a QTc prolongation is possible with several of the drugs commonly used in this disorder and QTc prolongation may be worsened by adding psychotropics to the treatment. Among the case reports included in this search, a case of prolongation of QTc interval was reported in a single case after introduction of Haloperidol in psychopharmacological treatment [3]. Therefore, the first tendency would be preferable to avoid Haloperidol and to use, if possible, drugs with a lower risk of QTc prolongation or reserve it to most severe or treatment resistant cases. Among typical antipsychotics, Chlorpromazine may be used in this population given its indication including delirium, low propensity for cardiovascular or respiratory complications, and low propensity to pharmacokinetic interactions with drugs commonly used in COVID-19 infection treatment. A significant risk of severe hypotension should be considered and, therefore, QTc and blood pressure should be monitored before and during treatment. As described above, our case series would suggest its possible use in COVID-19 patients, especially when no systemic comorbidities are detectable. During the systematic search, these data were confirmed; in fact, the use of Chlorpromazine in combination with Haloperidol in a previous healthy patient, did not report side effects and had a good outcome [26]. Despite this, another case included in the review showed how its use in combination with Trazodone in a patient with severe systemic diseases could have an insufficient effect on symptom resolution [41]. Benzodiazepines (BDZs) may cause respiratory depression or acute paradox reaction and their use should be limited as much as possible in patients with respiratory complications and even more in COVID-19 patients who are at a high risk of respiratory impairment. When BDZ use is considered clinically necessary, then short-lived molecules should be preferred. Benzodiazepines maintain their role in the treatment of patients with alcohol and substance withdrawal deliriums. This search would be in line with the available data on this topic. Trazodone, an atypical antidepressant, has been used successfully in the treatment of agitation, even in elderly patients and in cases of delirium where general medical conditions precluded the use of other compounds, as described above. There are extremely interesting preliminary data regarding its efficacy and tolerability in the treatment of behavioral and psychological disorders related to organic pathologies of the nervous system [47,48]. Atypical antipsychotics may also be particularly useful in this population of patients. Literature data begin to report the efficacy and tolerability of atypical antipsychotics in COVID-19 subjects with psychomotor agitation but nowadays there are no specific studies about their use in COVID-19-infected patients. Preliminary evidence supports the use of Quetiapine, Risperidone, and Aripiprazole in COVID-19 patients [19]. These data were also confirmed by the present search, even if efficacy was mainly reported when atypical antipsychotics were used in polytherapy except for Risperidone. In fact, we reported six cases in which antipsychotics were used in monotherapy: a good outcome after treatment with Risperidone was reported in five of them [4,34,35]; otherwise, no resolution of psychiatric acute symptoms was observed with monotherapy using Olanzapine [28], as well as monotherapy with Quetiapine showing an insufficient effect, making it necessary to apply an add-on therapy based on Lorazepam [29]. Valproate should be used as third-line medication in add-on therapy for cases where response to treatment is only partial [32]. Dexmedetomidine is an anesthetic increasingly used for sedation in intensive care patients. Evidence suggests that its use may prevent delirium, but also that postoperative administration of this drug may reduce the incidence of delirium [49,50] and may reduce agitation due to delirium, with better effectiveness and safety than Haloperidol [51]. Even though the evidence of delirium treatment with Dexmedetomidine is limited and needs further investigation, it may be a promising treatment option. When discussing our results some limitations had to be taken into account. Firstly, during the selection process, we only included articles in the English language, so there is a risk of missing relevant articles. Another potential limitation of our research is that, at the date of the last research study, the literature predominantly included case reports, which are anecdotal and inherently biased. 6. Conclusions Even if evidence-based indications are slowly making headway, there are not current validate guidelines for the management of acute agitation in COVID-19-infected patients. As delirium is a complex manifestation, key triggers and pathophysiology pathways might significantly differ in different populations, according to many factors (e.g., age, underlying medical and psychiatric conditions, altered states of consciousness). Thus, there are notable limitations in comparing data from such different populations. However, from a pragmatic standpoint, similar medications are generally prescribed to COVID-19 patients with hyperkinetic delirium irrespective of the underlying etiology, probably because they target final common pathways (e.g., dysregulations of dopaminergic, serotonergic, noradrenergic, and GABAergic systems) [52,53]. In this particular and vulnerable group of patients, we suggest an individualized and step-based therapy, considering the symptom presentation and the medical comorbidities of the patient. Therefore, this systematic review and case reports might generate useful insights for future research on promising interventions. Randomized head-to-head studies enrolling clinical patients are urgently needed to test promising medications with safe profiles for COVID-19-infected patients with delirium. Supplementary Materials The following are available online at https://www.mdpi.com/article/10.3390/ijerph19094978/s1. Click here for additional data file. Author Contributions C.C.: conceptualization, methodology, investigation, writing—original draft, writing—review and editing, and supervision; B.P.: conceptualization, methodology, investigation, writing—original draft, and provision of clinical case report; D.G.: conceptualization, methodology, investigation, writing—original draft, and writing—review and editing; S.F.: conceptualization, methodology, investigation, writing—original draft, and writing—review; G.D.P.: provision of clinical case report; S.L.C.: provision of clinical case report; C.A.B.: conceptualization and methodology; L.D.: conceptualization, supervision, and final manuscript version revision. All authors have read and agreed to the published version of the manuscript. Funding This research received no external funding. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement All data generated or analyzed during this study are included in this published article. Conflicts of Interest The authors declare no conflict of interest. No specific grant from any funding agency in the public, commercial, or not-for-profit sectors exist. Register and Protocol The study protocol was not registered. Figure 1 PRISMA flow diagram of the study selection process. PRISMA, Preferred Reporting Items for Systematic reviews and Meta-Analyses. ijerph-19-04978-t001_Table 1 Table 1 Characteristics of the case report and case series included in the systematic review. PTSD: post-traumatic stress disorder; DM2: diabetes mellitus type 2; HBP: high blood pressure; PPD: paranoid personality disorder; VPA: valproic acid; WBC: white blood cells; CK: creatine kinase; CAD: coronary artery disease; CHF: congestive heart failure; DLB: dementia with Lewy bodies; RBBB: right bundle branch block; COPD: chronic obstructive pulmonary disease. Study Date Country Type N° Case Quality Rating Previous History Age Gender Treatment Side Effects Outcome Clouden et al. [4] 2020 USA Case report 1 Good PTSD 46 F Risperidone 3 mg None Good Duyan et al. [26] 2021 Turkey Case report 1 Good Previously healty 31 F Haloperidol 10 mg; Chlorpromazine 25 mg None Good Espiridion et al. [27] 2021 USA Case report 1 Good Schizoaffective disorder 46 F Lorazepam 2 mg, Ziprasidone 10 mg; Clozapine 300 mg; Risperidone 8 mg Neuroleptic Malignant Syndrome ICU and intubated for 2 weeks Parker et al. [3] 2021 USA Case report 1 Good DM2, HBP 57 M Haloperidol 5 mg; Lorazepam 2 mg; Aripiprazole 5 mg QTc Prolungation Good Saje et al. [28] 2020 Slovenia Case report 1 Fair Acute psychotic reaction during febrile state Middle-age M Olanzapine None Insufficient effect Elfil et al. [29] 2021 USA Case report 1 Poor Asthma, atopic dermatitis 20 F Quetiapine Lorazepam None Good Mahajan et al. [30] 2021 India Letter to editor/Case report 1 Good Pre-eclampsia 37 F Midazolam Haloperidol None Agitation remission, neonatal death on day 31 Anmella et al. [31] 2020 Spain Case report 1 Poor PPD and DP 68 M Haloperidol 7.5/24 h IV; Quetiapine 200 mg None Good Sher et al. [32] 2020 USA Case report 1 Good Previously healty 70 F Quetiapine 250 mg; Melatonin 5 mg; VPA IV 1250 mg; Haloperidol IV 8 mg Pneumo-thorax Good Amouri et al. [33] 2020 USA Case report 1 Good Previously healty 70 F Lorazepam IM 0.5 mg None Improved catatonia symptoms but no effect on delirium Mawhinney et al. [34] 2020 UK Case report 1 Fair Previously healty 41 M Olanzapine 10 mg; Clonazepam 2 mg None Good Alonso-Sànchez et al. [35] 2021 Spain Case series 6 Fair Previously healty 63 M Aripiprazole; Quetiapine; Risperidone 6 mg None Good Previously healty 61 F Risperidone 6 mg None Good Previously healty 65 M Risperidone 6 mg None Good Previously healty 76 F Risperidone 6 mg; Gabapentin 900 mg None Good Previously healty 51 M Risperidone 4 mg None Good Previously healty 62 F Risperidone 3 mg None Good Syed et al. [36] 2021 USA Case series 4 Good DM2, schizoaffective disorder, bipolar type 52 F VPA 1000 mg; Haloperidol IV 2 mg Tachycardia; Fever; Rise in WBC; Decline in CK value Prolonged hospitalization and physical deconditioning HBP, hyperlipidemia, schizoaffective disorder, bipolar type 61 F Haloperidol 15 mg; VPA 1000 mg; Benztropine 2 mg None Prolonged hospitalization and physical deconditioning Colon cancer, DM2, atrial fibrillation, schizoaffective disorder, bipolar type, epilepsy 54 M Risperidone 2 mg; VPA 500 mg None Prolonged hospitalization Schizophrenia, hyperlipidemia, hypothyroidism 63 M Haloperidol 10 mg; Clozapine 350 mg None Prolonged hospitalization Los et al. [37] 2021 Polans Casereport 1 Good Previously healthy 39 M Haloperidol 5 mg IM; Lorazepam 2.5 mg; Olanzapine 10 mg; None Good Gillet et al. [38] 2020 UK Case report 1 Fair Previously healthy 37 M Diazepam; Olanzapina None Good Khatib et al. [39] 2020 Qatar Case report 1 Good Epilepsy 52 M Quetiapine; Haloperidol IM None Good Haddad et al. [40] 2021 Qatar Case report 1 Good Previously healthy Late 30s F Lorazepam 2 mg; Quetiapine 300 mg None Good Beach et al. [41] 2020 USA Case series 3 Good Dementia, alcohol use disorder, CAD, HBP, atrial fibrillation, CHF 76 M Olanzapine 2.5 mg; Olanzapine 10 mg; IM, Haloperidol 4 mg IV Rash of the right lower extremity Insufficient effect DLB, osteoarthritis, HBP 70 M Trazodone 25 mg; Chlorpromazine 25 mg IV Diffuse rash surrounding injection site Insufficient effect COPD, DM2, dementia, atrial fibrillation, RBBB, CAD, aortic stenosis, CHF 87 F Olanzapine 10 mg IM, Haloperidol 1–2.5 mg IV Quetiapine 25–50 mg None Improvement in delirium; however, worsening general conditions ijerph-19-04978-t002_Table 2 Table 2 Characteristics of the other types of studies included in the systematic review. Study Date Country Type Quality Rating Sample Delirium Population Previous Psychiatric History (Delirium Group) Mean Age (Delirium Group) Treatment Results Arbelo et al. [6] 2020 Spain Cross-sectional study Good 71 25 Admitted because of COVID-19 53 (12) 64 (69) Antidepressant in 8 pts BDZ in 7 pts Antipsychotic in 8 pts Anticonvulsant in 8 pts Not statistically significant difference Ragheb et al. [14] 2021 USA Single-center cohort study Good 148 108 Admitted because of COVID-19 17 (11) 59 (58) Propofol, Midazolam, Dexmedetomidine, Lorazepam ↑the median length of stay in delirium group Rozzini et al. [42] 2020 Italy Single-center cohort study Fair 14 14 Older patients admitted because of COVID-19-developing delirium None 78.2 None in 2 pts Antidepressant in 1 pts BDZ in 4 pts Antipsychotic in 4 ptsUndefined in 3 pts Mortality rate was 71% Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Rogers J.P. Chesney E. Oliver D. Pollak T.A. McGuire P. Fusar-Poli P. Zandi M.S. Lewis G. David A.S. Psychiatric and neuropsychiatric presentations associated with severe coronavirus infections: A systematic review and meta-analysis with comparison to the COVID-19 pandemic Lancet Psychiatry 2020 7 611 627 10.1016/S2215-0366(20)30203-0 32437679 2. Desforges M. Le Coupanec A. Dubeau P. Bourgouin A. Lajoie L. Dube M. Talbot P.J. Human Coronaviruses and Other Respiratory Viruses: Underestimated Opportunistic Pathogens of the Central Nervous System? Viruses 2019 12 14 10.3390/v12010014 31861926 3. Parker C. Slan A. Shalev D. Critchfield A. Abrupt Late-onset Psychosis as a Presentation of Coronavirus 2019 Disease (COVID-19): A Longitudinal Case Report J. Psychiatr. Pract. 2021 27 131 136 10.1097/PRA.0000000000000533 33656820 4. Clouden T.A. Persistent Hallucinations in a 46-Year-Old Woman After COVID-19 Infection: A Case Report Cureus 2020 12 e11993 10.7759/cureus.11993 33437548 5. Wu Y. Xu X. Chen Z. Duan J. Hashimoto K. Yang L. Liu C. Yang C. Nervous system involvement after infection with COVID-19 and other coronaviruses Brain Behav. Immun. 2020 87 18 22 10.1016/j.bbi.2020.03.031 32240762 6. Arbelo N. López-Pelayo H. Sagué M. Madero S. Pinzón-Espinosa J. Gomes-Da-Costa S. Ilzarbe L. Anmella G. Llach C.-D. Imaz M.-L. Psychiatric Clinical Profiles and Pharmacological Interactions in COVID-19 Inpatients Referred to a Consultation Liaison Psychiatry Unit: A Cross-Sectional Study Psychiatr. Q. 2021 92 1021 1033 10.1007/s11126-020-09868-6 33411128 7. Tong Y. Fu H. Xia C. Song W. Li Y. Zhao J. Zhang X. Gao X. Yong J. Liu Q. Astragalin Exerted Antidepressant-like Action through SIRT1 Signaling Modulated NLRP3 Inflammasome Deactivation ACS Chem. Neurosci. 2020 11 1495 1503 10.1021/acschemneuro.0c00156 32364698 8. Menculini G. Tortorella A. Albert U. Carmassi C. Carrà G. Cirulli F. Dell’Osso B. Luciano M. Nanni M.G. Pompili M. Access to Mental Health Care during the First Wave of the COVID-19 Pandemic in Italy: Results from the COMET Multicentric Study Brain Sci. 2021 11 1413 10.3390/brainsci11111413 34827412 9. Carmassi C. Bertelloni C.A. Dell’Oste V. Barberi F.M. Maglio A. Buccianelli B. Cordone A. Dell’Osso L. Tele-Psychiatry Assessment of Post-traumatic Stress Symptoms in 100 Patients With Bipolar Disorder During the COVID-19 Pandemic Social-Distancing Measures in Italy Front. Psychiatry 2020 11 580736 10.3389/fpsyt.2020.580736 33343419 10. Carmassi C. Foghi C. Dell’Oste V. Cordone A. Bertelloni C.A. Bui E. Dell’Osso L. PTSD symptoms in healthcare workers facing the three coronavirus outbreaks: What can we expect after the COVID-19 pandemic Psychiatry Res. 2020 292 113312 10.1016/j.psychres.2020.113312 32717711 11. Qiu H. Tong Z. Ma P. Hu M. Peng Z. Wu W. Du B. China Critical Care Clinical Trials Group (CCCCTG) Intensive care during the coronavirus epidemic Intensive Care Med. 2020 46 576 578 10.1007/s00134-020-05966-y 32077996 12. American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders, DSM-5 American Psychiatric Press Washington, DC, USA 2013 9783801728038 13. Martinotti G. Bonanni L. Barlati S. Miuli A. Sepede G. Prestia D. Trabucco A. Palumbo C. Massaro A. Olcese M. Delirium in COVID-19 patients: A multicentric observational study in Italy Neurol. Sci. 2021 42 3981 3988 10.1007/s10072-021-05461-2 34318364 14. Ragheb J. McKinney A. Zierau M. Brooks J. Hill-Caruthers M. Iskander M. Ahmed Y. Lobo R. Mentz G. Vlisides P.E. Delirium and neuropsychological outcomes in critically Ill patients with COVID-19: A cohort study BMJ Open 2021 11 e050045 10.1136/bmjopen-2021-050045 34535480 15. Brummel N.E. Jackson J.C. Pandharipande P. Thompson J.L. Shintani A.K. Dittus R.S. Gill T. Bernard G.R. Ely E.W. Girard T. Delirium in the ICU and Subsequent Long-Term Disability Among Survivors of Mechanical Ventilation Crit. Care Med. 2014 42 369 377 10.1097/CCM.0b013e3182a645bd 24158172 16. Ely E.W. Shintani A. Truman B. Speroff T. Gordon S.M. Harrell F.E. Jr. Inouye S.K. Bernard G.R. Dittus R.S. Delirium as a Predictor of Mortality in Mechanically Ventilated Patients in the Intensive Care Unit JAMA 2004 291 1753 1762 10.1001/jama.291.14.1753 15082703 17. Ostuzzi G. Papola D. Gastaldon C. Schoretsanitis G. Bertolini F. Amaddeo F. Cuomo A. Emsley R. Fagiolini A. Imperadore G. Safety of psychotropic medications in people with COVID-19: Evidence review and practical recommendations BMC Med. 2020 18 215 10.1186/s12916-020-01685-9 32664944 18. Baller E.B. Hogan C.S. Fusunyan M.A. Ivkovic A. Luccarelli J.W. Madva E. Nisavic M. Praschan N. Quijije N.V. Beach S.R. Neurocovid: Pharmacological Recommendations for Delirium Associated With COVID-19 J. Psychosom. Res. 2020 61 585 596 10.1016/j.psym.2020.05.013 19. Ostuzzi G. Gastaldon C. Papola D. Fagiolini A. Dursun S. Taylor D. Correll C.U. Barbui C. Pharmacological treatment of hyperactive delirium in people with COVID-19: Rethinking conventional approaches Ther. Adv. Psychopharmacol. 2020 10 2045125320942703 10.1177/2045125320942703 32733668 20. Wittich C.M. Burkle C.M. Lanier W.L. Ten Common Questions (and Their Answers) About Off-label Drug Use Mayo Clin. Proc. 2012 87 982 990 10.1016/j.mayocp.2012.04.017 22877654 21. Kalil A.C. Treating COVID-19—Off-Label Drug Use, Compassionate Use, and Randomized Clinical Trials During Pandemics JAMA 2020 323 1897 1898 10.1001/jama.2020.4742 32208486 22. Hacksell U. Burstein E.S. McFarland K. Mills R.G. Williams H. On the Discovery and Development of Pimavanserin: A Novel Drug Candidate for Parkinson’s Psychosis Neurochem. Res. 2014 39 2008 2017 10.1007/s11064-014-1293-3 24682754 23. Page M.J. McKenzie J.E. Bossuyt P.M. Boutron I. Hoffmann T.C. Mulrow C.D. Shamseer L. Tetzlaff J.M. Akl E.A. Brennan S.E. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews BMJ 2021 372 105906 10.1136/bmj.n71 24. Murad M.H. Sultan S. Haffar S. Bazerbachi F. Methodological quality and synthesis of case series and case reports BMJ Evid. Based Med. 2018 23 60 63 10.1136/bmjebm-2017-110853 29420178 25. National Heart, Lung, and Blood Institute (NHLBI) Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies Available online: https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools (accessed on 23 February 2022) 26. Duyan M. Ozturan I.U. Acute Psychosis in COVID-19: Is It Due to Favipiravir Treatment or Acute Viral Illness? SN Compr. Clin. Med. 2021 3 1627 1629 10.1007/s42399-021-00952-x 33997622 27. Espiridion E.D. Mani V. Oladunjoye A.O. Neuroleptic Malignant Syndrome After Re-introduction of Atypical Antipsychotics in a COVID-19 Patient Cureus 2021 13 e13428 10.7759/cureus.13428 33763319 28. Saje M. Prebil K. Plesnicar B.K. Time for lessons from the first wave of SARS-CoV-2 epidemic. A case report of a patient with acute psychosis and COVID-19 and later virus reactivation Eur. J. Psychiatry 2021 35 197 199 10.1016/j.ejpsy.2020.12.002 33518880 29. Elfil M. Selby L. Van Schooneveld T.C. Fadul N. Acute psychosis associated with recent SARS-CoV-2 infection: A case report IDCases 2021 24 e01140 10.1016/j.idcr.2021.e01140 33936950 30. Mahajan N.N. Gajbhiye R.K. Pednekar R.R. Pophalkar M.P. Kesarwani S.N. Bhurke A.V. Mahale S.D. Delirium in a pregnant woman with SARS-CoV-2 infection in India Asian J. Psychiatry 2021 55 102513 10.1016/j.ajp.2020.102513 31. Anmella G. Arbelo N. Fico G. Murru A. Llach C. Madero S. Gomes-Da-Costa S. Imaz M. López-Pelayo H. Vieta E. COVID-19 inpatients with psychiatric disorders: Real-world clinical recommendations from an expert team in consultation-liaison psychiatry J. Affect. Disord. 2020 274 1062 1067 10.1016/j.jad.2020.05.149 32663933 32. Sher Y. Rabkin B. Maldonado J.R. Mohabir P. COVID-19–Associated Hyperactive Intensive Care Unit Delirium With Proposed Pathophysiology and Treatment: A Case Report J. Psychosom. Res. 2020 61 544 550 10.1016/j.psym.2020.05.007 33. Amouri J. Andrews P.S. Heckers S. Ely E.W. Wilson J.E. A Case of Concurrent Delirium and Catatonia in a Woman With Coronavirus Disease 2019 J. Acad. Consult.-Liaison Psychiatry 2021 62 109 114 10.1016/j.psym.2020.09.002 33069380 34. Mawhinney J.A. Wilcock C. Haboubi H. Roshanzamir S. Neurotropism of SARS-CoV-2: COVID-19 presenting with an acute manic episode BMJ Case Rep. 2020 13 e236123 10.1136/bcr-2020-236123 32540882 35. Alonso-Sánchez M. Delgado-Parada E. Ayuso-Mateos J.L. Delirium in hospitalized COVID-19 patients: A case series Psychiatry Res. 2021 305 114245 10.1016/j.psychres.2021.114245 34678522 36. Syed S. Couse M. Ojha R. Management challenges in patients with comorbid COVID-19 associated delirium and serious mental illness—A case series Int. J. Psychiatry Med. 2021 56 255 265 10.1177/0091217420985979 33517809 37. Łoś K. Kulikowska J. Waszkiewicz N. First-Time Psychotic Symptoms in a Patient After COVID-19 Infection—A Case Report Front. Psychiatry 2021 12 726059 10.3389/fpsyt.2021.726059 34721104 38. Gillett G. Jordan I. Severe psychiatric disturbance and attempted suicide in a patient with COVID-19 and no psychiatric history BMJ Case Rep. 2020 13 e239191 10.1136/bcr-2020-239191 39. Khatib M.Y. Mahgoub O.B. Elzain M. Ahmed A.A. Mohamed A.S. Nashwan A.J. Managing a patient with bipolar disorder associated with COVID-19: A case report from Qatar Clin. Case Rep. 2021 9 2285 2288 10.1002/ccr3.4015 33936680 40. Haddad P.M. Alabdulla M. Latoo J. Iqbal Y. Delirious mania in a patient with COVID-19 pneumonia BMJ Case Rep. 2021 14 e243816 10.1136/bcr-2021-243816 34725060 41. Beach S.R. Praschan N.C. Hogan C. Dotson S. Merideth F. Kontos N. Fricchione G.L. Smith F.A. Delirium in COVID-19: A case series and exploration of potential mechanisms for central nervous system involvement Gen. Hosp. Psychiatry 2020 65 47 53 10.1016/j.genhosppsych.2020.05.008 32470824 42. Rozzini R. Bianchetti A. Mazzeo F. Cesaroni G. Bianchetti L. Trabucchi M. Delirium: Clinical Presentation and Outcomes in Older COVID-19 Patients Front. Psychiatry 2020 11 586686 10.3389/fpsyt.2020.586686 33262713 43. Battaglia J. Pharmacological Management of Acute Agitation Drugs 2005 65 1207 1222 10.2165/00003495-200565090-00003 15916448 44. Tisdale J.E. Rasty S. Padhi I.D. Sharma N.D. Rosman H. The Effect of Intravenous Haloperidol on QT Interval Dispersion in Critically Ill Patients: Comparison with QT Interval Prolongation for Assessment of Risk of Torsades de Pointes J. Clin. Pharmacol. 2001 41 1310 1318 10.1177/00912700122012896 11762558 45. Hassaballa H.A. Balk R.A. Torsade de pointes associated with the administration of intravenous haloperidol:a review of the literature and practical guidelines for use Expert Opin. Drug Saf. 2003 2 543 547 10.1517/14740338.2.6.543 14585064 46. Meyer-Massetti C. Cheng C.M. Sharpe B.A. Meier C.R. Guglielmo B.J. The FDA extended warning for intravenous haloperidol and torsades de pointes: How should institutions respond? J. Hosp. Med. 2010 5 E8 E16 10.1002/jhm.691 20394022 47. Martinón-Torres G. Fioravanti M. Evans J.G. Trazodone for agitation in dementia Cochrane Database Syst. Rev. 2004 18 CD004990 10.1002/14651858.CD004990 15495135 48. Sultzer D.L. Gray K.F. Gunay I. Berisford M.A. Mahler M.E. A Double-Blind Comparison of Trazodone and Haloperidol for Treatment of Agitation in Patients With Dementia Am. J. Geriatr. Psychiatry 1997 5 60 69 10.1097/00019442-199705010-00008 9169246 49. Pavone K.J. Cacchione P.Z. Polomano R.C. Winner L. Compton P. Evaluating the use of dexmedetomidine for the reduction of delirium: An integrative review Heart Lung 2018 47 591 601 10.1016/j.hrtlng.2018.08.007 30266265 50. Carrasco G. Baeza N. Cabré L. Portillo E. Gimeno G. Manzanedo D. Calizaya M. Dexmedetomidine for the Treatment of Hyperactive Delirium Refractory to Haloperidol in Nonintubated ICU Patients: A Nonrandomized Controlled Trial Crit. Care Med. 2016 44 1295 1306 10.1097/CCM.0000000000001622 26925523 51. Yeo Q.M. Wiley T.L. Smith M.N. Hammond D.A. Oral Agents for the Management of Agitation and Agitated Delirium in Critically Ill Patients Crit. Care Nurs. Q. 2017 40 344 362 10.1097/CNQ.0000000000000172 28834857 52. Wilson M.P. Pepper D. Currier G.W. Holloman G.H. Jr. Feifel D. The psychopharmacology of agitation: Consensus statement of the American Association for Emergency Psychiatry Project BETA Psychopharmacology Workgroup West. J. Emerg. Med. 2012 13 26 34 10.5811/westjem.2011.9.6866 22461918 53. Lindenmayer J.P. The pathophysiology of agitation J. Clin. Psychiatry 2000 61 (Suppl. 14) 5 10 11154018

PMC009xxxxxx/PMC9099661.txt

==== Front Sensors (Basel) Sensors (Basel) sensors Sensors (Basel, Switzerland) 1424-8220 MDPI 10.3390/s22093179 sensors-22-03179 Article Indoor Positioning with CNN and Path-Loss Model Based on Multivariable Fingerprints in 5G Mobile Communication System Wang Yuhang 12 https://orcid.org/0000-0002-8381-3714 Zhao Kun 12* Zheng Zhengqi 12 Ji Wenqing 12 Huang Shuai 12 Ma Difeng 12 Ghorashi Seyed Ali Academic Editor Noureldin Aboelmagd Academic Editor 1 Engineering Center of SHMEC for Space Information and GNSS, East China Normal University, Shanghai 200241, China; 51205904050@stu.ecnu.edu.cn (Y.W.); zqzheng@ee.ecnu.edu.cn (Z.Z.); 51191214011@stu.ecnu.edu.cn (W.J.); 51191214028@stu.ecnu.edu.cn (S.H.); 51205904080@stu.ecnu.edu.cn (D.M.) 2 Shanghai Key Laboratory of Multidimensional Information Processing, East China Normal University, Shanghai 200241, China * Correspondence: kzhao@ce.ecnu.edu.cn 21 4 2022 5 2022 22 9 317909 3 2022 19 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Many application scenarios require indoor positioning in fifth generation (5G) mobile communication systems in recent years. However, non-line of sight and multipath propagation lead to poor accuracy in a traditionally received signal strength-based fingerprints positioning system. In this paper, we propose a positioning method employing multivariable fingerprints (MVF) composed of measurements based on secondary synchronization signals (SSS). In the fingerprint matching, we use MVF to train the convolutional neural network (CNN) location classification model. Moreover, we utilize MVF to train the path-loss model, which indicates the relationship between the distance and the measurement. Then, a hybrid positioning model combining CNN and path-loss model is proposed to optimize the overall positioning accuracy. Experimental results show that all three positioning algorithms based on machine learning with MVF achieve accuracy improvement compared with that of Reference Signal Receiving Power (RSRP)-only fingerprint. CNN achieves best performance among three positioning algorithms in two experimental environments. The average positioning error of hybrid positioning model is 1.47 m, which achieves 9.26% accuracy improvement compared with that of CNN alone. indoor positioning multivariable fingerprints convolutional neural network path-loss model 5G ==== Body pmc1. Introduction With the rapid development of Internet of Things technology, location-based services such as navigation and positioning are receiving extensive attention. The global navigation satellite system (GNSS) works well in an outdoor environment [1]. However, in a complex indoor environment, the signal of satellite is weakened due to blockage and multipath propagation. GNSS often has poor accuracy or even fails to work. In order to obtain satisfactory indoor location, many methods have been proposed, such as Ultra-Wide Band (UWB), Wi-Fi technology, Bluetooth and so on [2,3,4]. However, they are not widely deployed and not affordable for personnel positioning during epidemic control for COVID-19 or large-scale exhibition, such as Import Expo. The existing cellular network is suitable for these location-based services adjusting to the new times, and there is no need to lay a large number of special hardware. From the first generation (1G) to the fifth generation (5G) mobile communication system, each generation has relevant research contributions for positioning [5]. 2G and 3G have supported standard positioning methods including timing advance (TA), enhanced observed time difference (EOTD), uplink time difference of arrival (UTDOA) and GNSS-based method [6]. In the 4G standard for both the control plane and the user plane, many positioning methods are available, such as Cell ID (CID), user equipment (UE)-assisted and network-based enhanced CID (E-CID), UE-based and UE-assisted A-GNSS, UE-assisted observed time difference of arrival (OTDOA) and UTDOA [7]. With the evolution of communication technology and the deployment of 5G base stations (BS), the radio signal is easy to collect. People can easily connect mobile devices such as smart phones to 5G systems in indoor environments. At the same time, 5G has the advantages of larger bandwidth and flexible subcarrier spacing, which leads to higher positioning accuracy. Many researchers have proposed a wide range of machine learning (ML)-based indoor localization approaches using fingerprints [8]. Random forest (RF)-based location awareness is used to improve the execution time and accuracy [9]. They utilize both the received signal strength indication (RSSI) and basic service set identifier (BSSID) measurements. Ref. [10] combines grid search-based kernel support vector machine (SVM) and principle component analysis (PCA) to improve the localization accuracy. Ref. [11] calculates the position of UE in 5G system by combining Kalman filter (KF), Universal Kriging (UK) spatial interpolation algorithm and k-Nearest Neighbors (KNN). However, traditional ML algorithms usually learn the shallow features of the data, which makes them unable to extract all reliable features from the complex received signal strength (RSS) fingerprints. In recent years, the indoor fingerprinting positioning system based on deep learning shows better performance than the traditional methods. Deep-Fi [12] uses a deep neural network (DNN) with four hidden layers in the Wi-Fi system to train the channel state information (CSI) of all subcarriers or antennas. DNN is very sensitive to the changes of input data, as when the data set is not sufficient, the accuracy is not high. ConFi [13] proposed the first Wi-Fi location algorithm based on convolutional neural network (CNN). The CNN is composed of three convolution layers and two fully connected layers, which transforms the location problem into an image classification problem [14]. They group 30 CSI measurements for 30 subcarriers for one antenna at the same reference point to construct a 30 × 30 matrix. Due to the limitation of hardware, not all base stations support CSI reporting. Ref. [15] uses CNN and Bluetooth RSS to classify the floor and location. The RSS of 144 Bluetooth APs collected at each reference point are converted into 12 × 12 eigenvector image. However, the fingerprinting database based on RSS images needs lots of base stations to achieve satisfactory positioning accuracy. For single base station, RSS fingerprint is vulnerable to non-line of sight (NLOS) and multipath propagation. The features of different positions may be similar and results in poor accuracy. Therefore, multiple measurements fingerprint is necessary. The reference signal reception power (RSRP) is combined with the reference signal reception quality (RSRQ) in the 4G cellular network. This physical layer information of the signal is used to build a fingerprinting database to improve the positioning accuracy [16]. Ref. [17] presents a localization method employing a Hybrid Wireless fingerprint (HW-fingerprint) based on CNN in Wi-Fi systems. Ref. [18] proposed random forest variable selection (RFVS) to sort variable importance and combinations for establishing multivariable fingerprinting database in 5G cellular network to improve the robustness of the positioning system. Indoor positioning methods include trilateration and fingerprinting. Trilateration obtains the positioning result by calculating the intersection between the geometry, such as circle or hyperbola. RSS is commonly used to fit the radio propagation path-loss model [19,20]. It is suitable for large open scenes because of sufficient line of sight (LOS) path information. Fingerprinting uses the features of the scene to estimate the target position. The position of the target device is usually determined as the reference point with the most similar features, such as RSS, delay or channel delay extension [21]. Fingerprints take the advantages of multipath propagation. Therefore, a combination of localization algorithms is implemented to improve the overall performance. Adaptive Enhanced Cell-ID (AECID) adjusts the similarity of signal power fingerprinting according to roundtrip time (RTT), and uses weighted KNN (WKNN) algorithm to calculate the final position [22]. Ref. [23] combines the fingerprinting system with CSI model under LOS environment to improve the robustness and accuracy of multidimensional scaling (MDS)-KNN system. In this paper, we study the indoor positioning problem based on fingerprint in 5G systems. The fingerprint is established by using four measurements in secondary synchronization signal (SSS). We use CNN to transform the indoor positioning problem into an image classification problem. At the same time, the path-loss model is trained to improve the overall positioning accuracy. The method in this paper only uses a single base station, but can also be extended to multiple base stations. The main work of this paper is as follows:We combine four measurements in SSS to construct 5G physical layer multivariable fingerprints (MVF), and use MVF to train a CNN location classification model for indoor positioning. We use MVF to train the path-loss model, which indicates the relationship between distance and radio signal measurement. A hybrid positioning model combining CNN and path-loss model is proposed to optimize the overall positioning results. We conduct experiments in the actual indoor environment to verify the effectiveness of the proposed method. 2. System Model The user equipment (UE) captures the SSS in 5G system and the MVF consist of four radio measurements of synchronization signal including reference signal received power (RSRP), reference signal received quality (RSRQ), received signal strength indication (RSSI) and signal-to-noise and interference ratio (SINR) [24]. The indoor positioning system based on MVF includes an offline stage and online stage, as shown in the Figure 1. In the offline stage, the positioning area is evenly divided into two-dimensional rectangular grid reference points. At each reference point, UE is used to capture SSS containing multiple radio measurements. The measurements are used to construct observation matrix. Then, the observation matrix is preprocessed, and the data packets with missing values are eliminated. We use Kalman filter algorithm to smooth the data, reduce the noise and obtain the MVF. With the sliding window, the preprocessed observation matrix is transformed into the observation image. The fingerprinting database is obtained by combining the observation image and the coordinates of the corresponding reference points. The fingerprinting database is used to train a CNN location classification model. The MVF and the distance between reference points and base station are used to train the path-loss model. In the online stage, the test points are evenly selected in the positioning area. UE is used to capture SSS containing multiple radio measurements and construct observation matrix at each test point. Similarly, the data are preprocessed and obtain the MVF. With the sliding window, the preprocessed observation matrix is transformed into the observation image. Then, the observation image is provided to the CNN location classification model trained in the offline stage for pattern matching. The output of the CNN model is the probability that the test point belongs to each reference point. Use the criterion to judge the test point whether it fits the path-loss model. If the test point satisfy the criterion, we will combine the path-loss model and CNN to obtain the positioning results. Otherwise, the CNN model will work alone. 3. Algorithm and Methods 3.1. Construction of 5G Observation Image Suppose there are a total of B reference and test points in the positioning area, and the UE captures C sampling data packets at each point. We express the MVF as the following 5G observation matrix: (1) Mb=V11V21V31V41V12V22V32V42…………V1cV2cV3cV4c,b=1,2,3,…B,c=1,2,3,…,C where b is the index of point while B is total number of points, and c is the index of data packets while C is total number of data packets. The columns represent four radio signal measurements RSRP, RSRQ, RSSI and SINR, respectively, which is expressed as V1,V2,V3,V4, and different rows represent the measurements at different sampling times. The original observation matrix needs to be enhanced when the training samples are not sufficient. Traditional data enhancement methods, such as image inversion and scaling [25], will damage the information contained in the feature for positioning. Instead, we use sliding window with small sliding step to enhance the data set to prevent the potential over fitting problem. As shown in the Figure 2, the observation matrix and sliding window are expressed as two different matrices. We use a heat map to represent the observation matrix and the color of heat map varies with the measurements. The sliding window matrix is expressed as a rectangle with a red border. CNN usually solves the problem of image classification, which means the traditional fingerprinting database is converted into image before convolution operation. The sliding window slides down the observation matrix to build the observation image. Applying CNN on a time-series of measurements is also expected to reduce the noise and randomness present in separate measurement, and hence improve the positioning accuracy. Suppose the sliding window has a size of T×4T≤C, and we group T sampling of the observation matrix to reconstruct a T×4 matrix, which we call 5G observation image: (2) Ib=V11V21V31V41V12V22V32V42…………V1tV2tV3tV4t,t=1,2,3,…,T,b=1,2,3,…B where t is the index of sampling while T is total row number of the observation image. Therefore, each point will have at least C/T observation images. The observation images collected at the same point are regarded as samples from the same category for training CNN. We set T to 16, so the size of the observation image is 16×4. If T is small, the observation image is too short to capture the time-domain correlation between sample eigenvalues. The larger the sliding window step size is, the less the observation images used for training and the lower the positioning accuracy. The smaller the sliding window step size is, the more the observation images used for training and the longer the training time. Consider a trade-off between the training time and positioning accuracy, we choose the step size of 8, which is half the length of the image. 3.2. CNN Location Classification Model The structure of CNN we use is developed from AlexNet, which has produced remarkable performance in image classification. As shown in the Figure 3, the CNN location classification model consists of four convolution layers, one fully connected layer and one softmax layer in turn. CNN is robust to noise by using convolution kernel and constructs a higher-level representation of the input image in the latter layer. A two-dimensional image I is used as input and the two-dimensional convolution kernel is defined as Kc, the convolution operation is expressed as: (3) S(i,j)=(Kc∗I)(i,j)=∑m∑nI(i−m,j−n)Kc(m,n) where the size of the convolution kernel Kc is m×n, the size of the image is i×j. The larger the convolution kernel, the larger the receptive field and the more information is obtained. A large convolution kernel may lead to a surge in computational complexity, which is not conducive to increasing the depth of the model and reducing the computational performance. The convolution kernel size is usually odd. For each input image, we employed 10 convolutional filters with 1×1 kernel size in the first convolutional layer. The 1×1 convolution layer increases the nonlinear characteristics while keeping the image size unchanged, which is conducive to feature extraction. We employed 10, 5, 5 convolutional filters with 3×3 kernel size in the following three convolutional layers. Due to the dimension of the observation image itself is not big enough, we pad the observation image to ensure that the size of the feature image remains 16×4 during forward propagation. We set the stride step of the convolution kernel to one to obtain the information in the time domain accurately, so that the dimension of the input observation image will not be reduced. We hope that the fully connected layer can get enough input features. We believe that each pixel on the observation image is a description of the location features, so we do not use the pooling layer for down sampling to avoid losing the information. The dropout layer is added and set to 0.2 after the first fully connected layer to reduce the influence of over fitting. The activation function in convolution layer introduces nonlinearity into the neural network, which is an important factor affecting the performance of the neural network. The function of activation function is to compress the result of convolution into a fixed range, so that the numerical range is controlled after multiple convolution layers. We choose the Rectified Linear Units (ReLUs) as the activation function, which achieves high computing speed because the resulting neural network has good sparsity. ReLUs are expressed as: (4) f(x)=max(0,x) The number of neurons in the output layer is equal to the number of reference points, so each output neuron corresponds to a reference point. Since the UE may appear near any reference point, we use softmax as the activation function of the output layer, which means the sum of all outputs is equal to one. Therefore, the output of neurons is interpreted as the probability that the UE belongs to the corresponding reference point. The softmax function is defined as follows: (5) pj=ewjTqi∑j=1JewjTqi where pj represents the output of the jth neuron in softmax layer. There are a total of J output neurons whose number is equal to the number of reference points in the positioning area. qi is the output of the neuron of the second last layer, wj is the weight connecting the second last layer and softmax layer, and T represents the transpose of the weight vector. In the online stage, the observation images of the test points are input into the CNN location classification model. Because the test point can appear at any position in positioning area, the estimated position L of the test point is obtained by using the probability weighted centroid method, which is expressed as: (6) L=∑k∈Ωpk(xk,yk)∑k∈Ωpk We sort the probabilities of all reference points in descending order, and Ω is the set of the first K reference points with high probability. pk is the probability of the kth reference point. xk,yk is the coordinates of the kth reference point. 3.3. Path-Loss Model with MVF In the ideal free space, the propagation of signal between the transmitter and receiver conforms to Friss model: (7) Pr(d)=PtGtGrλ24πd2 where Pr and Pt are the power of receiving antenna and transmitting antenna in mW, Gr and Gt are the gain of receiving antenna and transmitting antenna. d is distance between the transmitter and receiver, and λ is wavelength. The power in Equation (8) is expressed in dBm and the reference distance d0 is introduced: (8) Pr(d)=−10log10dd02+Pr(d0) where Pr(d0) is the receiving antenna power when the distance between the transmitter and receiver is d0. Multipath propagation is ubiquitous in indoor positioning environment, and logarithmic path-loss model with path-loss exponent is more suitable, which is expressed as: (9) Pr(d)=−10αlog10dd0+Pr(d0) where α represents the path-loss exponent, and its value varies with the environment. In the offline stage, the secondary synchronization signal data are captured by UE at each reference point, and the noise of the data is removed by Kalman filter algorithm. The filtered RSRP in MVF is associated with the distance to train the logarithmic path-loss model: (10) RSRP(d)=−10αlog10dd0+C where d0 is the reference distance, which is set to one meter in this paper, and C is the average RSRP value at d0. In online stage, we easily obtain the distance between the test point and the base station: (11) d′=d0×10C−RSRP10α where d′ represents the distance between the test point and the base station. The RSRP measurement in MVF will be affected by multipath propagation to varying degrees. If all reference points in the room are used to train the path-loss model, the positioning accuracy will be reduced. Therefore, we need to find a criterion to filter out those points in a severe multipath environment. According to 3GPP TS 38.215, RSRP, RSRQ and RSSI satisfy the following equation: (12) RSRQ=N×RSRPRSSI where N represents the number of resource blocks in the RSSI measurement bandwidth of the carrier. N relates the three measurements and reflects the degree that one point is affected by multipath, which will be verified in the experiment section. We calculate the N value of the reference point by Equation (12). If the N value of the point satisfies the criterion expressed in Equation (13), the point is in a slight multipath environment: (13) N−N0≤η where N0 is the theoretical true value of N, and usually N0=20 for SSS plus demodulation reference signal (DMRS) of physical broadcast channel (PBCH). η is a threshold varying with different environments. The greater the deviation of N from N0, the greater the point is affected by multipath. We will choose the point less affected by multipath to train the path-loss model. 3.4. Hybrid Positioning Model The hybrid positioning model combines CNN location classification model and path-loss model. In the online stage, we will use hybrid positioning model when the test points satisfy Equation (13). Otherwise, the CNN model will work alone. We first substitute the multiple measurements of the test point into the Equation (12) to obtain the N value. When the N value satisfies the Equation (13), we substitute the RSRP value of the test point into Equation (11) to obtain the estimated distance d′. We obtain the first K reference points with high probability by CNN model, and the distance between kth reference point and the base station is defined as dk. Absolute value of the difference between dk and d′ is calculated: (14) Δdk=dk−d′ The smaller Δdk indicates that the closer the reference point is to the real position of the test point. The proximity of the kth reference point and the real position of the test point is defined as sk: (15) sk=pk+∑k=1KΔdkΔdk,Δdk≤δpk,Δdk>δ where pk is the probability of the kth reference point. The threshold δ is determined by the positioning error of CNN location classification model. Then the probability of the kth reference point is updated as: (16) pk′=sk∑k=1Ksk where pk′ is the updated probability of the kth reference point and the sum of pk′ is equal to one. Finally, we use the probability weighted centroid method to obtain the optimized positioning coordinates of the test points: (17) L′=∑k∈Ωpk′xk,yk∑k∈Ωpk′ The detailed steps of hybrid positioning are shown in Algorithm 1: Algorithm 1 Hybrid positioning. Input:  Ω: The first K reference points set with high probability estimated by CNN; pk: The probability of the first K reference points; xk,yk: The coordinates of the first K reference points; dk: The distance between the first K reference points and the base station; Path-loss model curve and measurements of the test points; Output:  Estimated position of the test point;   1. Initialize parameters N0,η, δ;   2. Substitute the multiple measurements of the test point into the Equation (12) to obtain the N value; ifN−N0>ηthen     continue; else     Substitute the average RSRP of the test point into Equation (11) to obtain the estimated distance d′ between the test point and the base station;     for dk of the kth reference point do           Obtain the absolute value of the difference Δdk between dk and d′;     end for     if Δdk>δ then           sk=pk;     else           sk=pk+∑k=1KΔdkΔdk;     end if   Update the probability of the kth reference point pk′ using Equation (16) end if      3. Use the Equation (17) to obtain the estimated position of the test point. 4. Experiment Results and Analysis 4.1. Experiment Setup Our experimental system consists of multiple components, as shown in the Figure 4. The signal source is Gongjin 5G Sub-6GHz Small Cell base station, the user equipment is Huawei P40, and the PC is HP Laptop equipped with Inter (R) Core (TM) i5-8250U CPU @ 1.60 GHz processor. The PC installs Pilot Pioneer Tools version 10.5.8.32 and HI-SILICON driver software. The user equipment stores the original radio signal into the PC through the USB cable, and we use the data analysis tool Pioneer to export the required measurements for constructing fingerprinting database. In order to verify the validity of MVF and proposed method in various scenarios, we give the experimental results of two typical indoor environments. The positioning area is divided into several reference points and test points evenly. The black dots represent reference points and the red stars represent test points. Establishing a coordinate system with the first reference point as the origin, we measure and record the coordinates of other reference points and test points. UE is placed at each reference point and test point to capture SSS for establishing fingerprinting database. The two rooms are described as follows and as shown in Figure 5. Room A: a typical meeting room shown in Figure 5a. The size of room A is 7 m × 6 m with a table placed in the center, and there is a projector in the front of the room and a cabinet in the back. The 5G base station is located in the lower left corner of the room instead of the center to avoid isotropy. The base station towards the room with a height of 2 m. The UE is placed horizontally on a tripod with a height of 1 m, and most positions are in LOS environment. The positioning area is divided into 41 reference points, and each point is evenly deployed in a grid of 1 m × 1 m. 23 test points are evenly selected. We collect data packets at each point for 2 min and the fetch rate is 200 ms/samples. Room B: a typical office room shown in Figure 5b. The size of room B is also 7 m × 6 m and the room is crowded with tables and computers, which forms a complex radio transmission environment. The 5G base station is in room A, which locates out of the room B with a height of 2 m. The signal propagates in room B forming a pure NLOS environment. The UE is placed horizontally on a tripod with a height of 1 m. The positioning area is divided into 30 reference points, and each point is evenly deployed in a grid of 1 m × 1 m. 17 test points are evenly selected. We collect data packets at each point for 2 min and the fetch rate is 200 ms/samples. The average error ε and 80% quantile of cumulative distribution function (CDF) are used as the performance metric for different positioning algorithms. Assuming that the true position of the test point is xt,yt and the estimated position is xe,ye. The root square error (RSE) is calculated as: (18) RSE=xt−xe2+yt−ye2 For M locations, the average error is calculated as: (19) ε=∑m=1MRSEmM 4.2. Localization Performance of MVF We apply the popular deep learning platform Keras in Python to build CNN location classification model. We select Adam as the optimization function and the initial learning rate is 0.0001. The training epoch and batch size are set as 200 and 50. Cross-entropy is selected as the loss function. The parameter K in the weighted probability centroid method is set to 5. In order to explore the rationality of MVF, we tested CNN and several other commonly used machine learning and deep learning methods. We apply the popular machine learning platform Sklearn in Python to build KNN and MLP model. The number of nearest neighbors in KNN is set to 5 by default in sklearn module. The weights of neighbors are the same and Euclidean is selected as the distance measurement method. For MLP, the number of hidden layers and the number of neurons in each layer are both 16. The activation function is ReLUs. The optimization function is Adam and the initial learning rate is 0.001. The training epoch and batch size are both set as 200 by default. The data set used by the three algorithms are the same. The input of CNN is an image, so the data set is converted into 16 readings as a training sample. KNN and MLP still employ 1 reading as a training sample. Figure 6 shows the positioning error of KNN, MLP and CNN with and without MVF in Room A, respectively. Figure 7 shows the positioning error of KNN, MLP and CNN with and without MVF in Room B respectively. In the case without using MVF, we only use RSRP measurements to construct the fingerprints. As show in Figure 6 and Figure 7, in both room A and room B, the positioning error of the three localization algorithms is reduced when the MVF is used. The positioning error of KNN, MLP and CNN with or without MVF in room A are shown in Table 1. In room A, it is seen from the Table 1 that the average positioning error of KNN, MLP and CNN without MVF are 3.00 m, 2.28 m and 2.54 m respectively. The average positioning error of KNN, MLP and CNN with MVF are 2.10 m, 1.66 m and 1.62 m respectively. The average positioning accuracy of KNN, MLP and CNN has been improved by 30.00%, 27.19% and 36.22% respectively. The positioning error of KNN, MLP and CNN with MVF are less than 3.21 m, 2.41 m and 2.26 m for 80% test samples respectively. The positioning accuracy of CNN is 29.59% and 6.22% higher than that of KNN and MLP respectively. The positioning error of KNN, MLP and CNN with or without MVF in room B are shown in Table 2. In room B, it is seen from the Table 2 that the average positioning error of KNN, MLP and CNN without MVF are 2.25 m, 1.85 m and 2.62 m, respectively. The average positioning error of KNN, MLP and CNN with MVF are 1.87 m, 1.58 m and 1.41 m respectively. The average positioning accuracy of KNN, MLP and CNN has been improved by 16.89%, 14.59% and 46.18% respectively. The positioning error of KNN, MLP and CNN with MVF are less than 2.98 m, 2.17 m and 1.96 m for 80% test samples respectively. The positioning accuracy of CNN is 34.23% and 9.68% higher than that of KNN and MLP, respectively. The results show that the CNN model has better performance in data feature extraction and classification than KNN and MLP. 4.3. Path-Loss Model and Performance of Hybrid Positioning The N value of each reference point is calculated by employing Equation (12). We use the reference points in room A and room B to draw three-dimensional cubic interpolation diagrams of N values at different positions. The distribution of N in two rooms are shown in the Figure 8. As Figure 8 shows, the calculated N value varies at different positions in both room A and room B due to noise and multipath propagation. In room A, the N value of most positions is about 20, which verifies that N0=20 in Equation (13). A few positions have large calculated N value. These positions are located at the corner of the wall, near the door and window where has complex propagation paths. In room B, the signal of all points propagate through severe multipath propagation. The distribution of N value is very uneven, and there is no clear trend. The distribution of N value comprehensively and qualitatively reflect the influence of multipath propagation on the positioning area. Room A in Los condition is less affected by multipath propagation than room B in NLOS condition. The reference points in room B are too far away from the base station, which leads to low discrimination of the path-loss model. Therefore, we choose the reference points satisfying Equation (13) in room A to train the path-loss model and the fitting curve is as shown in Figure 9. As shown in Figure 9, we plotted distance vs. average RSRP and then used Python’s curve fitting function to estimate a curve for distance vs RSRP in Room A. The path-loss exponent α equals to 1.97 and C equals to −66.75. The hybrid positioning model combines CNN location classification model and path-loss model shown in Figure 9. We set the threshold η to one in room A and δ equals to the positioning error of CNN model in room A. The positioning error of hybrid positioning model compared with CNN model in room A is shown in Figure 10. As shown in Figure 10, the hybrid positioning model performs better than CNN model though the two curves overlap at the end. Some test points in room A do not satisfy Equation (13), and their positioning results are the same under the two models. The average positioning error of hybrid positioning model and CNN alone are 1.47 m and 1.62 m respectively. The positioning accuracy of hybrid positioning model is 9.26% higher than that of CNN alone. 4.4. Localization Performance of the Proposed Method The performance of positioning algorithm based on fingerprints is directly proportional to the quality of fingerprinting database. Refs. [13,15] use multiple base stations or subcarriers to enhance the robustness of fingerprint database in Wi-Fi or Bluetooth system, which has high requirements for hardware deployments. Refs. [13,23] use the CSI fingerprints at subcarrier level to obtain richer information than RSS fingerprints. CSI reporting is not supported in existing 5G base stations. The construction of fingerprinting database is also related to the environmental complexity of the positioning area. We compare the proposed method with some references using 5G system, which is shown in Table 3. As shown in Table 3, although Refs. [11,18] achieve a little higher positioning accuracy than the proposed method, their positioning area is smaller and has lower complexity. In this paper, we use single base station and multiple RSS related measurements to construct fingerprinting database, which reduces the requirements for equipment. The proposed method achieves similar positioning accuracy in a more complex environment. 5. Conclusions To avoid the problem that traditional RSS fingerprint is vulnerable to multipath propagation, we proposed a multivariable fingerprinting-base indoor localization algorithm in 5G system. We combine SS-RSRP, SS-RSRQ, SS-RSSI and SS-SINR to construct 5G physical layer multivariable fingerprints. We use the sliding window to transform the origin observation matrix into the observation images and achieve data enhancement. MVF are used to train CNN location classification model for indoor positioning. A hybrid positioning model combining CNN with path-loss model is proposed to optimize the overall positioning. Experimental results show that KNN, MLP and CNN with MVF achieve accuracy improvement in both two experimental scenarios. CNN achieves best performance among three positioning algorithms and shows 36.22% and 46.12% accuracy improvement with MVF in two experimental environments, respectively. The positioning accuracy of CNN is 29.59% and 6.22% higher than that of KNN and MLP, respectively, in room A. The positioning accuracy of CNN is 34.23% and 9.68% higher than that of KNN and MLP, respectively, in room B. The average positioning error of hybrid positioning model is 1.47 m, which achieves 9.26% accuracy improvement compared with that of CNN alone. In future work, we will build more appropriate path-loss model in indoor environment. Additionally, we will try to construct more robust fingerprints and find a better method to combine CNN model and path-loss model to further improve positioning accuracy. Author Contributions The authors confirm contribution to the paper as follows: Conceptualization, Y.W. and K.Z.; data curation, S.H.; methodology, Y.W.; resources, K.Z. and Z.Z.; software, Y.W. and W.J.; supervision, K.Z. and Z.Z.; validation, Y.W. and D.M.; visualization, W.J.; writing—original draft, Y.W.; writing—review and editing, K.Z., S.H. and D.M. All authors have read and agreed to the published version of the manuscript. Funding This research was funded by the National Natural Science Foundation of China (No. 61771197) and Science and Technology Commission of Shanghai Municipality (Grant no. 18DZ2270800). Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement The data presented in this study are available on request from the corresponding author. The data are not publicly available due to restrictions. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Indoor positioning system with CNN and path-loss model based on multivariable fingerprints. Figure 2 Observation matrix and sliding window. Figure 3 CNN location classification model. Figure 4 Experimental equipment: (a) 5G small base station. (b) PC and user equipment. Figure 5 Map of positioning area: (a) Room A. (b) Room B. Figure 6 The positioning error of KNN, MLP and CNN with and without MVF in Room A: (a) KNN. (b) MLP. (c) CNN. Figure 7 The positioning error of KNN, MLP and CNN with and without MVF in Room B: (a) KNN. (b) MLP. (c) CNN. Figure 8 The distribution of N in room A and room B: (a) Room A. (b) Room B. Figure 9 Path-loss model curve of room A. Figure 10 The positioning error of hybrid positioning model and CNN model in room A. sensors-22-03179-t001_Table 1 Table 1 The positioning error of KNN, MLP and CNN with or without MVF in room A. Methods Average Error (m) CDF = 80% (m) KNN without MVF 3.00 4.29 KNN with MVF 2.10 3.21 MLP without MVF 2.28 2.76 MLP with MVF 1.66 2.41 CNN without MVF 2.54 3.67 CNN with MVF 1.62 2.26 sensors-22-03179-t002_Table 2 Table 2 The positioning error of KNN, MLP and CNN with or without MVF in room B. Methods Average Error (m) CDF = 80% (m) KNN without MVF 2.25 3.56 KNN with MVF 1.87 2.98 MLP without MVF 1.85 2.45 MLP with MVF 1.58 2.17 CNN without MVF 2.62 4.20 CNN with MVF 1.41 1.96 sensors-22-03179-t003_Table 3 Table 3 The comparison of the proposed method and other references. System Signal Source Measure-Ments Positioning Area Positioning Error Ref. [11] 5G Single base station RSSI 3 m × 4 m average 1.16 m Ref. [18] 5G Single base station multivariable 3 m × 4 m CDF = 80% 1.18 m Proposed method 5G Single base station multivariable 7 m × 6 m average 1.47 m Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Jiménez-Martínez M.J. Farjas-Abadia M. Quesada-Olmo N. An Approach to Improving GNSS Positioning Accuracy Using Several GNSS Devices Remote Sens. 2021 13 1149 10.3390/rs13061149 2. Hua C. Zhao K. Dong D. Zheng Z. Yu C. Zhang Y. Zhao T. Multipath Map Method for TDOA Based Indoor Reverse Positioning System with Improved Chan-Taylor Algorithm Sensors 2020 20 3223 10.3390/s20113223 32517133 3. Oh S.H. Kim J.G. DNN based WiFi positioning in 3GPP indoor office environment Proceedings of the 2021 International Conference on Artificial Intelligence in Information and Communication (ICAIIC) Jeju Island, Korea 13–16 April 2021 302 306 4. Lie M.M.K. Kusuma G.P. A fingerprint-based coarse-to-fine algorithm for indoor positioning system using Bluetooth Low Energy Neural Comput. Appl. 2021 33 2735 2751 10.1007/s00521-020-05159-0 5. del Peral-Rosado J.A. Raulefs R. López-Salcedo J.A. Seco-Granados G. Survey of Cellular Mobile Radio Localization Methods: From 1G to 5G IEEE Commun. Surv. Tutor. 2018 20 1124 1148 10.1109/COMST.2017.2785181 6. Campos R.S. Evolution of positioning techniques in cellular networks, from 2G to 4G Wirel. Commun. Mob. Comput. 2017 2017 2315036 10.1155/2017/2315036 7. Giancarlo F. Andreas M. Linda S. Positioning in LTE. Handbook of Position Location—Theory, Practice and Advances 2nd ed. Zekavat S.A. Michal B.R. John Wiley & Sons Inc. Hoboken, NJ, USA 2019 1165 1219 978-1-119-43458-0 8. Singh N. Choe S. Punmiya R. Machine Learning Based Indoor Localization Using Wi-Fi RSSI Fingerprints: An Overview IEEE Access 2021 9 127150 127174 10.1109/ACCESS.2021.3111083 9. Lee S. Kim J. Moon N. Random forest and WiFi fingerprint-based indoor location recognition system using smart watch Hum. Centric Comput. Inf. Sci. 2019 9 1 14 10.1186/s13673-019-0168-7 10. Zhang L. Li Y. Gu Y. Yang W. An efficient machine learning approach for indoor localization China Commun. 2017 14 141 150 10.1109/CC.2017.8233657 11. Huang S. Zhao K. Zheng Z. Ji W. Li T. Liao X. Zhu F. An Optimized Fingerprinting-Based Indoor Positioning with Kalman Filter and Universal Kriging for 5G Internet of Things Wirel. Commun. Mob. Comput. 2021 2021 9936706 10.1155/2021/9936706 12. Wang X. Gao L. Mao S. Pandey S. CSI-Based Fingerprinting for Indoor Localization: A Deep Learning Approach IEEE Trans. Veh. Technol. 2017 66 763 776 10.1109/TVT.2016.2545523 13. Chen H. Zhang Y. Li W. Tao X. Zhang P. ConFi: Convolutional Neural Networks Based Indoor Wi-Fi Localization Using Channel State Information IEEE Access 2017 5 18066 18074 10.1109/ACCESS.2017.2749516 14. Krizhevsky A. Sutskever I. Hinton G.E. ImageNet classification with deep convolutional neural networks Commun. ACM 2017 60 84 90 10.1145/3065386 15. Sun D. Wei E. Yang L. Xu S. Improving Fingerprint Indoor Localization Using Convolutional Neural Networks IEEE Access 2020 8 193396 193411 10.1109/ACCESS.2020.3033312 16. Poosamani N. Rhee I. Towards a practical indoor location matching system using 4G LTE PHY layer information Proceedings of the 2015 IEEE International Conference on Pervasive Computing and Communication Workshops (PerCom Workshops) St. Louis, MO, USA 23–27 March 2015 284 287 17. Liu Z. Dai B. Wan X. Li X. Hybrid Wireless Fingerprint Indoor Localization Method Based on a Convolutional Neural Network Sensors 2019 19 4597 10.3390/s19204597 31652626 18. Ji W. Zhao K. Zheng Z. Yu C. Huang S. Multivariable Fingerprints with Random Forest Variable Selection for Indoor Positioning System IEEE Sens. J. 2021 22 5398 5406 10.1109/JSEN.2021.3103863 19. Abou-Shehada I.M. AlMuallim A.F. AlFaqeh A.K. Muqaibel A.H. Park K.-H. Alouini M.-S. Accurate Indoor Visible Light Positioning Using a Modified Pathloss Model with Sparse Fingerprints J. Light. Technol. 2021 39 6487 6497 10.1109/JLT.2021.3098005 20. Gentner C. Günther D. Kindt P.H. Identifying the BLE Advertising Channel for Reliable Distance Estimation on Smartphones IEEE Access 2022 10 9563 9575 10.1109/ACCESS.2022.3140803 21. Xia S. Liu Y. Yuan G. Zhu M. Wang Z. Indoor Fingerprint Positioning Based on Wi-Fi: An Overview ISPRS Int. J. Geo-Inf. 2017 6 135 10.3390/ijgi6050135 22. Wigren T. LTE Fingerprinting Localization with Altitude Proceedings of the 2012 IEEE Vehicular Technology Conference (VTC Fall) Quebec City, QC, Canada 3–6 September 2012 1 5 23. Song Q. Guo S. Liu X. Yang Y. CSI Amplitude Fingerprinting-Based NB-IoT Indoor Localization IEEE Internet Things J. 2018 5 1494 1504 10.1109/JIOT.2017.2782479 24. 3GPP Release 15, TS 38.215, Physical Layer Measurements 2017 7 10 Available online: https://portal.3gpp.org/ChangeRequests.aspx?q=1&specnumber=38.215 (accessed on 8 March 2022) 25. Lu Y. Zhang X. Jing L. Fu X. Data Enhancement and Deep Learning for Bone Age Assessment using The Standards of Skeletal Maturity of Hand and Wrist for Chinese Proceedings of the 43rd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC) Mexico City, Mexico 1–5 November 2021

PMC009xxxxxx/PMC9099662.txt

==== Front Sensors (Basel) Sensors (Basel) sensors Sensors (Basel, Switzerland) 1424-8220 MDPI 10.3390/s22093348 sensors-22-03348 Article Applying Hybrid Lstm-Gru Model Based on Heterogeneous Data Sources for Traffic Speed Prediction in Urban Areas https://orcid.org/0000-0002-3181-8118 Zafar Noureen 12 https://orcid.org/0000-0002-5142-3965 Haq Irfan Ul 1* Chughtai Jawad-ur-Rehman 1 Shafiq Omair 3 Schramm Dieter Academic Editor Sieberg Philipp Academic Editor 1 Pakistan Institute of Engineering and Applied Sciences, Islamabad 44000, Pakistan; noureen_zafar@uaar.edu.pk (N.Z.); jawadchughtai@gmail.com (J.-u.-R.C.) 2 University Institute of Information Technology, Pir Mehr Ali Shah University of Arid Agriculture, Rawalpindi 46000, Pakistan 3 School of Information Technology, Carleton University, Ottawa, ON K1S 5B6, Canada; OmairShafiq@cunet.carleton.ca * Correspondence: irfanulhaq@pieas.edu.pk 27 4 2022 5 2022 22 9 334807 10 2021 03 12 2021 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). With the advent of the Internet of Things (IoT), it has become possible to have a variety of data sets generated through numerous types of sensors deployed across large urban areas, thus empowering the notion of smart cities. In smart cities, various types of sensors may fall into different administrative domains and may be accessible through exposed Application Program Interfaces (APIs). In such setups, for traffic prediction in Intelligent Transport Systems (ITS), one of the major prerequisites is the integration of heterogeneous data sources within a preprocessing data pipeline resulting into hybrid feature space. In this paper, we first present a comprehensive algorithm to integrate heterogeneous data obtained from sensors, services, and exogenous data sources into a hybrid spatial–temporal feature space. Following a rigorous exploratory data analysis, we apply a variety of deep learning algorithms specialized for time series geospatial data and perform a comparative analysis of Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Convolutional Neural Network (CNN), and their hybrid combinations. The hybrid LSTM–GRU model outperforms the rest with Root Mean Squared Error (RMSE) of 4.5 and Mean Absolute Percentage Error (MAPE) of 6.67%. IoT LSTM GRU CNN ITS Technology Development Fund (TDF) program of Higher Education Commission of Pakistan (HEC)TDF02-261 This research was supported by the grant TDF02-261 under Technology Development Fund (TDF) program of Higher Education Commission of Pakistan (HEC). ==== Body pmc1. Introduction Recently, the explosive growth of sensors, the internet, and huge data generation have provided new avenues of storage, execution, and implementation opportunities for IoT-based applications. Where the real-time availability of heterogeneous data generated from a huge number of sensors has brought about novel possibilities, a new set of challenges has also emerged, which is primarily related with the need of methodologies and capabilities to optimally harness the power of heterogeneity of data. Unveiling new interpretations by first integrating various forms of heterogeneous data and then applying intelligent algorithms to it is one of the most exciting possibilities of the notion of the Connected World [1]. Intelligent Transportation System (ITS) plays an enabling role in the realization of the concept of smart cities [2]. ITS has a huge demand to integrate data directly from various sensors, services built upon sensors, and from a variety of other exogenous data sources. ITS heavily relies on the power of IoT to build multisource, multisensor, and multimodel service systems for the prediction of traffic speed [3,4]. Major challenges faced during the activities of the data integration pipeline include sparsity handling, anomaly detection and rectification, normalization, map matching, and data resolution. The resolution quality of the fused and integrated traffic data plays a significant role when prediction algorithms are applied to it. The spatial temporal nature of traffic data defines the two corresponding facets of its resolution on GIS maps. Spatial resolution identifies the length of the road segment on which prediction is computed whereas the temporal resolution means the minimum time interval during which the prediction is made. The spatial temporal structure is used to integrate traffic data from different sources, e.g., Floating Car Data and Google in our case. Data elements from different sources at specific road segments during a specific time interval are fused together, thus adding value to accuracy. Necessary transformations from Estimated Time of Arrival (ETA) to speed or vice versa are performed before the fusion. This may result in spatiotemporal gaps due to the varying sparsity of data sources that need to be addressed. Data from the exogenous sources, e.g., weather, holidays, or peak hours, are also weaved upon the same spatial temporal framework in a similar fashion. In this research, a detailed workflow is developed for the data integration pipeline that transforms raw data extracted from different heterogeneous data sources, passing through various preprocessing, and integration activities finally resulting into an integrated hybrid feature space. Open Street Maps (OSM) was selected as the GIS mapping technology. The integrated data are mapped on the road segments between every two adjacent OSM nodes within the time intervals of 15 min for the whole 24 h of the day. Keeping in mind the spatial temporal nature of the data, machine learning and deep learning algorithms specialized for time series data were applied, and their results were compared. The main contributions of the paper can be summarized as follows:We integrated the heterogeneous data sources of Intelligent transportation systems for data collected from a particular city in Pakistan and built the hybrid LSTM-GRU model. We predicted the traffic speed on the basis of heterogeneous traffic data sources including exogenous data sources, e.g., weather, event and, peak hours. The Hybrid LSTM-GRU model has been applied on time intervals varying from 15 min to 1 h and the effectiveness of the model has been evaluated. The remainder of the paper is organized as follows: Section 2 presents the related work. Section 3 presents the proposed methodology based on Hybrid LSTM–GRU model. Section 4 provides the results and discussions. Section 5 concludes the paper and identifies the future research direction. 2. Related Work The related work was organized in context with the integration of heterogeneous data sources obtained from a variety of sensors leading to deep learning techniques applied on the spatial–temporal traffic data. Salanova Grau et al. [5] presented a novel auto-DL approach for tuning the 145 errors and 22 warnings hyperparameter of the LSTM model in order to reduce time consumption. However, the author worked only on temporal features, not spatial features, and also did not focus on multidata sources. Tang et al. [6] improved the fuzzy neural network model in order to enhance the learning capability of the model. The proposed model consists of a supervised and unsupervised learning process. In unsupervised learning, a k-means is used to partition input samples, and in the case of supervised learning, a weighted recursive least squares estimator is used for the optimization of hyperparameters. However, spatial patterns were not considered, while periodic features were extracted from traffic flow data. The authors did not deal with heterogeneous data sources of the arterial road network. Ma, Tao, Wang, and Yu [7] proposed the novel approach LSTM-NN to overcome the issue of backpropagated error decay and automatically determine the optimal time window for the time series data set. This means the short-term prediction of speed deals with NN and long-term prediction of speed deals with LSTM. The authors used microwave detector data for the prediction of speed. However, they have only addressed the temporal features. There was a need to address the spatial features as well as a need to investigate with different data aggregated levels that would act as an additional input and contribute towards the prediction of speed in a more accurate consistent manner. Salavona Grau et al. [1] proposed the framework that collects data from conventional(cameras, radars, and loops) as well as innovative (FCD and Bluetooth devices) technologies and analyzes the ETA and traffic flow on road networks. They are focused on data collection, filtration, and fusion of big data. They analyzed the conventional and innovative types of data sources with the help of statistical approaches such as correlation coefficients, R-square, absolute value, and percentages ranges. However, there is a need to check the spatial–temporal features of big data on the road networks with the help of a machine and deep learning models that are specially proposed to handle the time series and spatial patterns. Bratsas et al. [8] applied different machine learning models and figured out the forecasting effectiveness of the machine learning model on randomly selected dates, randomly selected roads with duration over eight consecutive 15 min intervals, and the whole day. The experiments show that SVR performed well with stable conditions and MPL performed well on greater variations. However, the paper does not analyze the specialized deep learning models that are designed to deal with spatial–temporal trends of the data. Yang et al. [9] present the effects of heterogeneous data sources (parking meter transactions, traffic conditions, and weather conditions) and spatial–temporal nature of data on the parking occupancy by linking Graph CNN and LSTM deep learning models. However, the authors did not incorporate ETA, event, holiday, calendar, OSM, and count data sources in their work. The authors focused on single-scale occupancy prediction rather than multiple scales, i.e., parking meters to the level of aggregated zones. In order to cope with chronological deviations for traffic forecasting, spatial–temporal patterns have been extensively addressed [10,11,12,13,14,15,16,17,18,19,20,21,22,23] using dynamic tensor completion method, LSTM, temporal graph convolutional network, multivariate regression model, GraphCNN-LSTM model, attention graph convolutional sequence-to-sequence model, and spatial–temporal residual graph attention network. The authors, however, have not considered the effects of heterogeneous and nonrecurrent events data sources for forecasting road traffic. Zhang et al. [24] proposed the deep learning-based multitask learning model to predict traffic speed on the road networks. The authors used the hybrid approach to increase the performance of the proposed model on taxis GPS data. They extracted the spatial–temporal patterns with the help of the nonlinear Granger causality analysis method, and for hyperparameter tuning, they used the Bayesian optimization technique. However, the authors did not consider other modes of traffic and their influence on the prediction of speed. Kong et al. [25] proposed the recommendation system of intelligent traffic based on LSTM deep learning model. However, they ignored the positive impact of performance by using a multimodel on intelligent traffic information and also did not make a comparison of the famous time-recurrent neural network models, e.g., CNN [26], GRU, and hybrid models. Authors in [27,28,29] used hybrid graph convolutional neural networks, CNN-LSTM, and LSTM-NN models on real-time traffic speed. The hybrid model produced promising results; however, other data sources related to road traffic that have directly or indirectly influenced the prediction of speed on road network were not taken into account. Authors in [30,31,32] used k-means clustering, PCA, SOM, SVM, and SVR for prediction of traffic speed on regular or irregular intervals. However, their prediction models can be improved with a fusion of nonrecurrent events(e.g., calendar, special events, accidents, and weather) and traffic flow analysis. Li et al. [33] proposed the transfer-learning model to address the missing data, data insufficiency, and mitigate model overfitting problems and stack LSTM for considering the time series patterns. However, the authors did not analyze the exogenous data sources, traffic types, and different traffic modes data sets. Hybrid feature space may be helpful for the construction of rules for applying transfer learning on the specific area while considering spatial factors. Mena-Oreja et al. [34] discussed the formation of congestion by using state-of-the-art error recurrent and deep learning models. The author conducted a survey in the field of transportation and identified which deep learning and error recurrent models are helpful while considering spatial–temporal factors and other traffic conditions. The author applied the error recurrent and deep learning models on real traffic data sets in order to generate the common benchmark under traffic congestion conditions and demonstrated that the error recurrent model shows better accuracy as compared to deep learning models. However, the authors do not focus on the impact of statistical and ensemble deep and machine learning models on traffic congestion conditions using real traffic data sets and also do not discus hybrid feature space and its impact on traffic congestion prediction. Ren et al. [2] propose the hybrid integrated–DL model in order to capture both spatial–temporal dependencies on the prediction of citywide spatial–temporal flow volume. The authors proposed the hybrid LSTM and ResNet model in order to deal with spatial–temporal effects on traffic volume. However, the authors proposed that the model shows large prediction error in sparse spatial areas and sleeping hours. The authors do not focus on the applicability of other types of flows, e.g., passenger flow, bike flow, etc. Yu et al. [35] introduced the piece-wise correlation function and Jenks clustering method with dynamic programming to fix relationship of segment intervals. They considered heterogeneous data sources, e.g., speed, traffic flow, density, and road occupancy for short-term speed prediction on 5 and 10 min time intervals. They prepared the results on the basis of only three days, i.e., 1 February to 3 February 2015 (from Sunday to Tuesday). They did not discuss the effects of the weekdays and nonweekdays on road traffic. The amount of data is very low, and it is difficult to explore the correlation function across the whole week. Liu et al. [36] used attention CNN to forecast traffic speed. They used 29,952 records as a training set and 5760 records as the test set. The amount of data is very low due to coverage of single road and difficult to explore traffic trends on their adjacent roads. In an earlier work [37], we built a solution that used traffic and weather data to predict traffic congestion using Estimated Time of Arrival (ETA). We proposed a hybrid LSTM–GRU approach on heterogeneous data sources which comprises 7,343,362 records of September 2020. We made a comprehensive comparison among famous time-recurrent neural network advanced deep learning models, e.g., CNN, GRU, and their combinations. 3. Proposed Methodology Based on Hybrid LSTM–GRU Model We obtained Floating Car data from a regional trackers’ company. A data integration pipe-line was developed to clean, preprocess, and integrate this data with other data sources. Data integration pipe-line caters data pertaining to FCD, Google, weather, peak hour, holiday, and OSM data sources. We addressed the issues related to the FCD data sources—some of which include zero speed adjustments, outlier removal, and map matching. A feature in FCD data provides the reason for a signal generation from the tracker. The reason may be regular time interval, ignition on/off, turn, etc. A map-matching technique was used for node correction whereas threshold value, parking info, and ignition on/off details were used for zero speed adjustment. Finally, all the data sources were aggregated on the basis of wayid attribute of OSM and time attribute at a regular interval of 15 min. This was followed by the normalization of road speed through Permissible Speed Limit (SPL) attribute and Speed Performance Index (SPI). 3.1. Data Sources Data from the following data sources were obtained and integrated. 3.1.1. FCD Data Source FCD data were obtained from a regional tracker company. The data set contained events generated by 2895 unique tracker ids for the whole month of September 2020. The tracker units were mounted with the sensors GSM Modem(Quectel M95) and GPS Chipset(U-blox EVA-M8M). The key features of Floating Car Data (FCD) include latitude, longitude, date time, address, location, direction, speed, reason, and unit id. We faced multiple issues related to data preparation of FCD data source. The following issues were identified in FCD data:There was an off-road mapping of cars. This could be due to two reasons. Either the car appears offroad because of inherent GPS error or because the car was actually parked somewhere off the road. A large number of speed values generated by trackers were zero. This again could be due to two reasons: either the car is parked or stuck in congestion. The congestion data needed to be distinguished from the data related to the parked cars. There was duplication of tuples. There are missing values causing spatial sparsity. This is because the FCD does not cover all segments of roads of the road network. The issues pertaining to the missing values and off-road mapping of cars are resolved through the map-matching process described in Algorithm 1. When the nodes are placed correctly on the map, the in-between missing values can be suggested. The zero speed issue is resolved through the zero speed adjustment section described in Algorithm 1. Outliers are identified and rectified through the max speed attribute obtained from OSM. Any speed exceeding the max speed on a segment of the road is replaced with the max speed value. Speed Performance Index is used for the normalization of data. Algorithm 1 Preprocessing and Data Integration Algo 1: Read coordinate point [latitude, longitude] from data 2: Initialize nodes-seg with empty start-point and end-point 3: for t = 1 to m do 4:     Initialize way-id to zero 5:     nodes-seg[“start-point”,“end-point”]:=get_nearest_seg(latitude,longitude) 6:     if nodes-seg [“start-point”] is zero then 7:         Check end-point value and update the start-point value with previous node value 8:     else 9:         nodes-seg [“start-point”] and nodes-seg [“end-point”] represents to same road points 10:     end if 11: end for 12: Update way-id 13: data.append(way-id) 14: Update data 15: take input records from data 16: F = 0 17: for t = 1 to n do 18:     Initialize speed = data[“FCD-speed”] 19:     Initialize elapsed time = data[“FCD-elapsed-time”] 20:     if reason is “ignition-ON” then 21:         F = 1 22:         if next reason is “ignition-OFF” then 23:            if speed is zero then 24:                if elapsed time >threshold then 25:                    continue F = 1 26:                end if 27:                if next speed is zero then 28:                    F = 0 29:                    delete record 30:                else 31:                    keep record 32:                end if 33:            else 34:                F = 0 35:                keep record 36:            end if 37:         end if 38:         remove record 39:     end if 40: end for 41: Initialize the data to empty 42: for each seg do 43:     for each agg-min each day do 44:         Compute avg-speed = 1n∑xi 45:         Initialize s = length of the segment 46:         Compute eta = L ÷ avg-speed 47:         data[“avg-speed”].append(“avg-speed”) 48:         data[“eta”].append(“eta”) 49:     end for 50:     if current_day is “Sunday” or “Saturday” then 51:         data[“holiday”] = 1 52:     else 53:         data[“holiday”] = 0 54:     end if 55:     data[“weather”] = get weather of the current day with respect to current location 56: end for 3.1.2. ETA Data Source We decided to obtain Google Map’s data from more than 500 points of interest on important roads of Islamabad. We acquired data by sending start and endpoints to Google Maps API. The Google data source is an authentic data source and provides an estimated time of arrival information. Data obtained from Google can be easily mapped on OSM Maps. The key features of Google data are Source Latitude, Source Longitude, Max Speed, Date, Time, Destination Latitude, Destination Longitude, and Estimated Time Arrival. 3.1.3. OSM Data Source The road network attributes were fetched from OSM’s Turbo Overpass API. It provides the tags of Islamabad which include start node, end node, highway types, max speed, min speed, way_id, and max length. Max speed is a feature of special importance that is utilized in the normalization of data and outlier identification and removal. Some roads do not have a max speed feature associated with them; therefore, we had to insert it manually. 3.1.4. Calendar Data Source In order to identify the effect of traffic on holidays at a particular location, we need a calendar data source. Behaviors and patterns of traffic are highly dependent on holiday data. The calendar data source features include DateTime, Name, and Type. 3.1.5. Weather Data Source Real-time data source is gathered from Yahoo and Dark Sky API on the basis of time and latitude and longitude. Table 1 shows the most relevant features of the proposed hybrid model. The maxspeed-real attribute is used to detect anomalies and normalized the speed. From feature engineering process, we exploited the input feature set and then applied machine and deep learning models that are capable enough to consider spatial–temporal effects on transportation data sources. We also applied hybrid approaches such as LSTM-GRU, GRU-LSTM, CNN-LSTM, and CNN-GRU as these hybrid approaches enhance the performance of models and provide more accurate prediction of the speed of a specific road at a specific time. (1) SPI=(Sit/SPL)∗100 3.2. Data Integration Pipeline In the Integration Pipeline, we performed the transformation on FCD and Google data set. The map-matching process was used to obtain the data in spatial format. Hence, the first transformation in the pipeline is the map matching. Figure 1 illustrates the data integration procedure. Data are collected from different data sources such as Google, tracking company, osm road, weather data, holiday data, and peak hour data. In addition to the map-matching procedure, the following activities were required for data transformation before the machine learning algorithms could be applied:Map matching of GPS points Handling the abnormal behavior of data Data generalization and transformation Calculating the average speed of road section. In the Google data set, the data points needed to be mapped on the OSM nodes. In this way, we could divide long roads in smaller segments with each segment marked by two adjacent OSM nodes. For this purpose, the nearest API of OSRM was again used. We verified the mapping results by visualizing the nodes on OSM maps. The mapping of FCD data and Google traffic data on OSM nodes provided a mechanism to spatially unify both traffic data. For temporal aggregation, the data points of both data sources were aggregated for every 15 min for all the road segments defined on the OSM road network. The integrated data were then merged with holiday data based on the date field. The purpose of the abovementioned integration was to encounter the special effects of congestion on working hours and weekends during holidays. These merged data are further integrated with road attributes on the basis of wayid. Furthermore, to handle the environmental effects, we combined these integrated data with only those weather data parameters that affect the behavior of the traffic, i.e., rainy, visibility, etc. Algorithm 1 represents the procedure of map matching. The data contains the coordinate points with the latitude and longitude of the located geographical positions. Firstly, these coordinates were sent to the nearest API of OSRM server to obtain the pair of nodes of the segment containing the location of the driving vehicle that had already been determined from different sources. This was followed by the ordering of nodes on OSM maps to identify whether the road is incoming or outgoing. Occasionally, OSRM nearest API returns zero value in place of start node that might be due to multiple options for the nearest start node owing to the junctions on roads. For zero values, the end-node value was traced back on the OSM road information, and the immediate previous node was assigned to the start-node value. 3.3. Model Selection We have applied the existing machine learning, classical deep learning and advanced deep learning models on the integrated traffic data. We have applied the well-known models such as XGBoost [38] (with Decision Tree [39] as weak learner), Linear Regression [40], K-Nearest Neigbor (KNN) [41], Multilayer Perceptron (MLP) based on MLPRegressor implementation in sklearn library (available at https://scikit-learn.org, accessed on 1 October 2021) [8], Artificial Neural Network (ANN) based on KerasRegressor implementation in keras library (available at https://keras.io, accessed on 1 October 2021) [1], Long Short Term Memory (LSTM) [42], Gated Recurrent Unit (GRU) [43], Convolutional Neural Networks (CNN) [44], and well-known possible combinations of the models such as LSTM-CNN, CNN-LSTM, CNN-GRU, GRU-CNN, GRU-LSTM and LSTM-GRU models. Due to the spatial-temporal nature of the data models based on LSTM and GRU techniques generated better results. The Hybrid LSTM-GRU model has produced the most promising results. In the following subsections we briefly describe the existing and well-known LSTM and GRU techniques and then move on to describe our hybrid model based on combining these existing techniques. 3.3.1. LSTM Long Short Term Memory (LSTM) [42] is a variant of recurrent neural network (RNN) [5,45]. It is specialized for time series data. A generalized LSTM unit consists of three gates (i.e., input, output and a forget) and a cell. Cells are used to memorize the values of data and flow the information to output and forget gate. It is used to address vanishing gradient problem. 3.3.2. GRU Gated Recurrent Unit (GRU) [43] is an advanced and more improved version of LSTM. It is also the type of recurrent neural network. It uses less hyper parameters because of reset gate and update gate as contrast to three gates of LSTM. Update gate and reset gate are basically vectors and are used to decide which information should be passed to the output. 3.3.3. Hybrid LSTM-GRU Model Description Since, our dataset is time series and regression problem, we generated the results by using both classical as well as deep learning techniques. Our hybrid approach techniques yielded the low RMSE and is thus more effective as compared to classical regression techniques e.g., KNN, XGBoost, Linear Regressor, ANN and MLP. In our hybrid LSTM-GRU model, we first applied LSTM. LSTM is used to tackled the problem of vanishing gradient in backpropagation. LSTM contain three gates e.g., input gate (ig), forget gate (fg) and output gate (og). Gates are used to store information in memory. It stores the information in analog format. These gates are element-wise multiplied by sigmoid function ranges between 0–1. If the value of the gate is zero then this information is ignored or discarded else remained in memory. Tanh [46] is a well-known non-linear activation function and ranges between −1 to +1. In order to avoid information fading, a second derivative is used. The sigmoid function [47] is also well-known as a non-linear activation function. A sigmoid function contains values between 0 to 1. It is basically used to suggest which information should stay or drop from memory units known as gates. The mathematical Equations of input gate (ig), forget gate (fg) and output gate (og) are taken and adapted from the literature and are explained in Equations (2)–(4). Whereas, GRU contains two gates e.g., update gate (ug) and reset gate (rg). The output of the LSTM was passed to GRU during this approach. xit is the input feature set that contains hybrid feature space(start-node, end-node, way-id, day, hour, agg-minutes, quarter, holiday, peakhour, mazspeed-real) at specific time and location. The aggregate speed is the target or output label, similar to [48] in which authors used LSTM and Bi-Directional LSTM Models to predict stock price. The details and equations presented are taken and adapted from the literature such as [42,43]. Input Gate: ig→representsinputgate Forget Gate: fg→representsforgetgate Output Gate: og→representsoutputgate Update Gate: ug→representsupdategate Reset Gate: rg→representsresetgate σ→representssigmoidfunction wx→representsweightfortherespectivegate (x) ht−1→outputofthepreviouslstmblockattimestampt−1 xt→inputatcurrenttimestamp bx→biasesfortherespectivegates (x) Cell Output: Ct→memoryattimestamp (t) Cell Input: ∼Ct→representscandidateformemoryattimestamp (t) The Cell Input state is ∼Ct, Cell Output state is Ct, and LSTM consists of three gates ig, fg, and og. GRU consists of two gates ug and rg. The hidden layers of LSTM–GRU model are ∼Ct, ∼ht, and ht. The weights of LSTM are wi, wf, wo, and wc. The weights of GRU are wu, wr, wo, and wCt. LSTM–GRU model have biases bi, bf, bo, and bc. tanh is known as the hyperbolic tangent function. The ratio of the hyperbolic sine and corresponding hyperbolic cosine functions is defined in terms of tanh function. The scalar products of two vectors are represented as ∘. When xt is passed to the input network unit, it is multiplied by its own weight (wi), and ht−1 is also multiplied by its own weight (wi) and then added the bias (bi). A ht−1 holds the information of previous units t−1. It passes to the sigmoid function and converts values between 0 and 1 and updates the status of the cell. The details and equations presented are taken and adapted from the literature such as [42,48,49]. (2) ig=σ(wi[ht−1,xt]+bi) (3) fg=σ(wf[ht−1,xi]+bf) (4) og=σ(wo[ht−1,xt]+bo) Equations (5) and (6) describes how to produce the result between 0 and 1 using sigmoid activation function. ∼Ct and Ct are used to decide what information is kept in memory and what information is forgotten. ∼Ct is multiplied by the tanh function and decides which value is more significant. (5) ∼Ct=tanh(wc[ht−1,xt]+bc) (6) Ct=ft∗Ct−1+it∗∼Ct The details and equations presented are taken and adapted from the literature such as [43,48]. Equations (7) and (8) explain that Ct is passed as an input to the first layer of GRU (ug), whereas ug and ht−1 are multiplied to weight and this information is forwarded to reset gate (rg). (7) ug=σ(wu[Ct]+wu[ht−1]) (8) rg=σ(wr[Ct]+wr[ht−1]) ht decides information to be kept. The stayed information is then attached to the output layer. Same layer contains tanh as an activation function that is used to predict road traffic speed at specific time and location. This is also discussed in Equations (9)–(11). We used adam as an optimizer and mean squared as loss function in this regression problem. (9) ∼ht=tanh(wCt+rg∘wCt[ht−1]) (10) ht=ug∘ht−1+(1−ug)∘∼ht (11) ht=og∗tanh(ht) In this study, the proposed LSTM–GRU model was applied to the data collected from Google and FCD which comprises 7,343,362 records of September 2020. The traffic condition data were captured every fifteen minutes of arterial roads in Islamabad, Pakistan. The proposed stacked LSTM–GRU architecture consisted of four hidden layers with 256 hidden units each. tanh was used as an activation function in all hidden layers. In the dense layer, we used one unit. The linear activation function was used in the output layer. We employed holdout crossvalidation to split data set into training and test sets. Then, we trained the model on the training data set by a batch learning approach using batch size of 512. This was followed by checking generalization of model on test data set. To evaluate the performance of the proposed deep architecture, we adopted RMSE, MAE, and MAPE as performance measures. The final configuration of the proposed LSTM–GRU model is summarized in Table 2. In the proposed model LSTM–GRU, we used four hidden layers (two for LSTM and the rest for GRU). We tested various configurations of hidden units, i.e., 4, 16, 32, 64, 128, and 256. We achieved the optimal results with 256 hidden units. Likewise, tanh activation function in the hidden layers and linear activation function in the output layer gives optimal results. Similarly, we varied the number of epochs from 5 to 50 and choose 10 as the optimal value. Moreover, we used batch size 512, learning rate 0.001, and loss function adam as the optimal parameter values. 3.3.4. Performance Measures for Proposed Hybrid LSTM–GRU Model Well-known and existing performance evaluation metrics are used to judge or measure the model performance and pattern. It also indicates the best model in order to achieve the output label performance. To evaluate the solution, we have used the existing performance evaluation metrics [50] such as RMSE (Root Mean Square Error), MAPE (Mean Absolute Percentage Error) and MAE (Mean Absolute Error). The details and equations for the existing performance evaluation metrics presented below are taken and adapted from the literature such as [50,51,52]. RMSE is also known as RMSD ( root-mean-square deviation). It is the square root of the mean squared difference between desired output and predicted output. The same is explained in Equation (12). (12) RMSE=(∑i(Ydsi−Ypsi)2/number of observations) where, Ydsi = desired speed Ypsi = predicted speed n = number of observations MAPE (mean-absolute-percentage-error) is robust to large outliers. It eliminates the scaling factor and explains the error in the form of percentage. The formula of MAPE is explained in Equation (13). (13) MAPE=median(|(Ydsi−Ypsi)/Ydsi|) In our scenario, MAE (Mean-Absolute-Error) is used to define how far our predicted output speed is from the desired output speed. Mathematically, we can explain this from the following Equation (14). (14) MAE=1/n∑i|(Ydsi−Ypsi)| 4. Results and Discussion In this study, we worked on the regression data set with the multitimestamp and single label. Some features related to traffic patterns were derived in each segments at specific interval of time such as minimum estimated time and maximum speed per segment from integrated data. Data were captured at 15 min time resolution and taking the space resolution less than or equal to 1 km. 4.1. Exploratory Data Analysis We analyzed the influence of weekdays and weekend on traffic patterns and speed fluctuation. Speed Performance Index (SPI) [53] is a derived feature and calculated by using the following formula in the Equation (15):(15) SPI=((VO)/Vm)∗100 where, VO = the current speed for the road segment; Vm = the permissible max speed for the road segment. SPI also provides a normalized expected speed which prevents having extreme values. Figure 2 presents the average speed performance index (SPI) vs. time of day on weekdays. Different working days, i.e., Monday to Thursday, are shown with different color lines. In Figure 2, it can be visualized that there is a major change in the SPI over different hours of the day. In the morning rush hours (5:00–6:00 a.m., 9:00–10:00 a.m.,12:00–13:00 p.m.), the SPI is 66%, 64.6%, and 63%, respectively, which is higher than the average of morning hours. During the evening rush hour (around 8:00–9:00 p.m.), the SPI is 55%. The different trend of the SPI in the different time slots is not only compelling for model training but also in taking the average of the SPI for each time slot and for the prediction of speed on the basis of historical data. Various time slots including 5:00–6:00 a.m., 9:00–10:00 a.m., 12:00–13:00 p.m., and 8:00–9:00 p.m. show high traffic congestion. Friday’s SPI trend is different from other working days. On Friday, during the morning rush hour (10:00 a.m.–12:00 p.m.), the SPI is 65.7%, which is higher than the average of morning hours. Friday traffic is different from other weekdays due to the Friday prayer, which is offered during 1–2 p.m. During the evening rush hour (9:00–10:00 p.m.), SPI is 60%, which is higher than all the time of day. SPI variations on weekends are addressed in Figure 3. This figure depicts that weekend SPI trend is extremely opposite from weekdays SPI. On Saturday, rush hours are 10:00 a.m.–1:00 p.m. in the morning and 3:00–4:00 p.m., 9:00–10:00 p.m. in the evening, whereas on Sunday, rush hours are 12:00 p.m. to 1:00 p.m. in the morning and in the evening, 9:00 p.m. to 10:00 p.m., respectively. 4.2. Feature Selection Feature selection is a method that was used to explore the most relevant feature set by using correlation. The correlation of each feature was calculated and compared with the target variable. The correlation ranged between −1 and +1. +1 shows the positive correlation or perfect correlation, whereas −1 shows negative correlation and zero value means nonexistence of correlation. Basically, we were dealing with numeric input and target features. As the most relevant and famous techniques that deal with numeric data are correlation feature selection and mutual feature selection techniques, we used the same ones in our study. 4.2.1. Correlation Feature Selection Technique Correlation is a statistical measure used to identify how two features changed together. We used Pearson correlation coefficient(PCC) [54] in our scenario. Equation (16) explains that PCC is a standard measure of linear correlation between the two features. The formula of PCC is the ratio between the covariance of two features divided by their standard deviations. It deals between normalized data that range between −1 and +1. V[:,j] = feature of Vj in hybrid data set. j = is a variable. mean(V[:,j]) = mean of the Vj feature of hybrid data set. W = feature of W in hybrid data set. meanW = mean of the W feature of hybrid data set. std(V[:,J]= Standard deviation of V[:,j]. std(W) = Standard deviation of W. (16) ((V[:,j]−mean(V[:,j]))∗(W−meanW))/(std(V[:,J])∗std(W)) Table 3 depicts that start-node, way-id, hour, peak hour, and max speed-real are most relevant features as they are positively correlated, because in transportation domain start-node, way-id shows spatial impact, whereas hour and peak hour show temporal impact. Max speed-real shows the permissible speed of the road. We selected the 9 most correlated features from 29 combined features of all heterogeneous data sources. The correlation shows how all features are close to the target feature. 4.2.2. Mutual Information Regression Feature Selection Technique The mutual information feature selection technique is a method that works on information gain [55]. It works on the decision tree principle and calculates the entropy of each feature by calculating the information gain of each feature. Entropy helps the decision tree draw boundaries and measures disorder and uncertainty in the available data set. The most relevant feature has the highest information gain. Figure 4 indicates that start-node, end-node, way-id, and max speed have the highest information gain and are therefore the most relevant feature set. 4.2.3. Heat Map of Hybrid Feature Space Heat Map is a visualization style for analyzing the intensity, density, patterns, outliers, and variance of the feature set. It provides a correlation among all features. Figure 5 shows that maxspeed-real have the highest correlation with target agg-speed followed by way-id, day, and holiday features among the other heterogeneous features. Here, maxspeed-real denotes the permissible road speed limit. It helps to identify outliers and trends and patterns of road speed. In this way, we can tackle the missing speed and also normalize the data set in order to predict the true speed of a specific road. Figure 6 depicts that LSTM–GRU yields the lowest RMSE, i.e., 4.5 as compared to deep learning technique LSTM with a RMSE yield of 4.86 and classical regression techniques such as KNN with an RMSE yield of 6.03. Because the LSTM–GRU model handles both spatial–temporal effects, LSTM–GRU is specialized in time series data set. LSTM contains the temporal effects, and GRU contains the spatial effects. Because we have worked on transport data set and in the domain of transportation prediction of speed highly depends on specific time and location. 4.3. Hybrid LSTM–GRU Results Table 4 elaborates the three performance metrics, i.e., RMSE, MAE, and MAPE, with respect to lowest RMSE-generated model, i.e., hybrid LSTM–GRU and GRU–LSTM models. In Figure 7, the prediction horizon indicates the capturing of data intervals such as 15, 30, 45, and 60 min. This helps in analyzing the time resolution impact on the model training. As we increased the sliding window, our test data performance improves, which proves a direct proportion of sliding window with the performance metrics. Table 4 contains the different performance metrics behavior on multiple individuals as well as hybrid deep learning models. As per RMSE and MAE evaluation metrics, hybrid LSTM–GRU produced the lowest RMSE of 4.5, MAE of 2.03, and MAPE of 6.67% being the lowest error. 5. Conclusions This paper utilizes the speed performance index (SPI) as the road network state evaluation indicator. We integrated heterogeneous data, i.e., traffic, GPS, weather, special condition, and OSM obtained from a variety of sensors and services. We analyzed the behavior of transportation data sources with the help of different machine and deep learning algorithms. The LSTM–GRU model proved to be the most effective hybrid model among all time series deep learning and classical machine learning models with a net RMSE yield of 4.5. In the future, we plan to automatically label the classes using fuzzy logic and k-means clustering, followed by analyzing the results automatically by using optimization of hyper parameters and statistical models. Author Contributions Conceptualization, N.Z. and I.U.H.; methodology, N.Z.; software, N.Z.; validation, N.Z., J.-u.-R.C. and I.U.H.; formal analysis, N.Z.; investigation, I.U.H., J.-u.-R.C. and O.S.; resources, N.Z. and O.S.; data curation, N.Z.; writing—original draft preparation, N.Z. and J.-u.-R.C.; writing—review and editing, N.Z., I.U.H. and O.S. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement The data presented in this study are available upon request from the corresponding author. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Urban traffic speed prediction based on hybrid LSTM–GRU model. Figure 2 Speed Performance index variation on weekdays. Figure 3 Speed performance index variation on weekends. Figure 4 Feature selection using mutual information regression feature selection technique. Figure 5 Heat map of hybrid feature space. Figure 6 Evaluation metrics of prediction on test data. Figure 7 Evaluation metrics of prediction on test data. sensors-22-03348-t001_Table 1 Table 1 Hybrid Feature Space. Nature of Attributes Data Type Day integer Hour integer Startnode integer Endnode integer aggminutes 15 min time interval Weather char maxspeed-real integer aggSpeed integer Holiday boolean sensors-22-03348-t002_Table 2 Table 2 Hyperparameters configuration for XGBoost, ANN, KNN, MLP, and Hybrid LSTM–GRU. Model Hyperparameters Values XGBOOST objective linear n-estimators 4000   ANN input dimension 10 activation function relu loss function RMSE optimizere adam epoch 100 batch-size 512 KNN K 20 loss RMSE MLP activation function relu loss function RMSE hidden-layer-size 100 optimizer SGD learning rate 0.001 LSTM-GRU Batch Size 512 Learning Rate 0.001 No of epochs 10 No of Hidden Layers 04 Hidden Units 256 Dropout Ratio 0.2 Activation Function tanh Output-Units 1 Output-Type Single Label Output-Layer-Activation-Function linear Optimizer Adam Loss Function mean squared error sensors-22-03348-t003_Table 3 Table 3 Feature selection through correlation feature selection technique. Features Scores start−node 14,218.665540 end−node 52.788980 way−id 19,974.123749 day 487.561578 hour 24,238.112125 agg−minutes 0.380622 quarter 25.339737 holiday 620.959742 peakhour 9876.836458 maxspeed−real 3,227,692.593161 sensors-22-03348-t004_Table 4 Table 4 Performance metrics of deep learning models. Model RMSE MAE MAPE LSTM 4.86 2.13 6.95 GRU 5.05 2.29 7.7 CNN 30.3 25.96 64.10 LSTM−CNN 4.7 23.9 7.9 CNN−LSTM 5.89 3.53 11.47 GRU−LSTM 4.6 2.08 6.85 LSTM−GRU 4.5 2.03 6.67 CNN−GRU 5.1 2.4 8.4 GRU−CNN 7.8 4.5 14.6 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Salanova Grau J.M. Mitsakis E. Tzenos P. Stamos I. Selmi L. Aifadopoulou G. Multisource data framework for road traffic state estimation J. Adv. Transp. 2018 2018 9078547 10.1155/2018/9078547 2. Ren Y. Chen H. Han Y. Cheng T. Zhang Y. Chen G. A hybrid integrated deep learning model for the prediction of citywide spatio-temporal flow volumes Int. J. Geogr. Inf. Sci. 2020 34 802 823 10.1080/13658816.2019.1652303 3. Aitkenhead M. Cooper R. Neural network time series prediction of environmental variables in a small upland headwater in NE Scotland Hydrol. Process. Int. J. 2008 22 3091 3101 10.1002/hyp.6895 4. Liu X. Liu Y. Song H. Liu A. Big data orchestration as a service network IEEE Commun. Mag. 2017 55 94 101 10.1109/MCOM.2017.1700090 5. Yi H. Bui K.H.N. An automated hyperparameter search-based deep learning model for highway traffic prediction IEEE Trans. Intell. Transp. Syst. 2020 22 5486 5495 10.1109/TITS.2020.2987614 6. Tang J. Liu F. Zou Y. Zhang W. Wang Y. An improved fuzzy neural network for traffic speed prediction considering periodic characteristic IEEE Trans. Intell. Transp. Syst. 2017 18 2340 2350 10.1109/TITS.2016.2643005 7. Ma X. Tao Z. Wang Y. Yu H. Wang Y. Long short-term memory neural network for traffic speed prediction using remote microwave sensor data Transp. Res. Part Emerg. Technol. 2015 54 187 197 10.1016/j.trc.2015.03.014 8. Bratsas C. Koupidis K. Salanova J.M. Giannakopoulos K. Kaloudis A. Aifadopoulou G. A comparison of machine learning methods for the prediction of traffic speed in urban places Sustainability 2020 12 142 10.3390/su12010142 9. Yang S. Ma W. Pi X. Qian S. A deep learning approach to real-time parking occupancy prediction in transportation networks incorporating multiple spatio-temporal data sources Transp. Res. Part C Emerg. Technol. 2019 107 248 265 10.1016/j.trc.2019.08.010 10. Impedovo D. Dentamaro V. Pirlo G. Sarcinella L. TrafficWave: Generative deep learning architecture for vehicular traffic flow prediction Appl. Sci. 2019 9 5504 10.3390/app9245504 11. Li L. Lin H. Wan J. Ma Z. Wang H. MF-TCPV: A Machine Learning and Fuzzy Comprehensive Evaluation-Based Framework for Traffic Congestion Prediction and Visualization IEEE Access 2020 8 227113 227125 10.1109/ACCESS.2020.3043582 12. Chen X.M. Zahiri M. Zhang S. Understanding ridesplitting behavior of on-demand ride services: An ensemble learning approach Transp. Res. Part C Emerg. Technol. 2017 76 51 70 10.1016/j.trc.2016.12.018 13. Candel A. Parmar V. LeDell E. Arora A. Deep Learning with H2O H2O.ai Inc. Mountain View, CA, USA 2016 14. Fang M. Tang L. Yang X. Chen Y. Li C. Li Q. FTPG: A Fine-Grained Traffic Prediction Method With Graph Attention Network Using Big Trace Data IEEE Trans. Intell. Transp. Syst. 2021 10.1109/TITS.2021.3049264 15. Zhang S. Li S. Li X. Yao Y. Representation of traffic congestion data for urban road traffic networks based on pooling operations Algorithms 2020 13 84 10.3390/a13040084 16. Tan H. Wu Y. Shen B. Jin P.J. Ran B. Short-term traffic prediction based on dynamic tensor completion IEEE Trans. Intell. Transp. Syst. 2016 17 2123 2133 10.1109/TITS.2015.2513411 17. Shin D.H. Chung K. Park R.C. Prediction of Traffic Congestion Based on LSTM Through Correction of Missing Temporal and Spatial Data IEEE Access 2020 8 150784 150796 10.1109/ACCESS.2020.3016469 18. Liu H. Xu H. Yan Y. Cai Z. Sun T. Li W. Bus arrival time prediction based on LSTM and spatial-temporal feature vector IEEE Access 2020 8 11917 11929 10.1109/ACCESS.2020.2965094 19. Xu H. Ying J. Bus arrival time prediction with real-time and historic data Clust. Comput. 2017 20 3099 3106 10.1007/s10586-017-1006-1 20. Servos N. Liu X. Teucke M. Freitag M. Travel Time Prediction in a Multimodal Freight Transport Relation Using Machine Learning Algorithms Logistics 2020 4 1 10.3390/logistics4010001 21. Zhang Z. Li M. Lin X. Wang Y. He F. Multistep speed prediction on traffic networks: A deep learning approach considering spatio-temporal dependencies Transp. Res. Part C Emerg. Technol. 2019 105 297 322 10.1016/j.trc.2019.05.039 22. Bogaerts T. Masegosa A.D. Angarita-Zapata J.S. Onieva E. Hellinckx P. A graph CNN-LSTM neural network for short and long-term traffic forecasting based on trajectory data Transp. Res. Part C Emerg. Technol. 2020 112 62 77 10.1016/j.trc.2020.01.010 23. Cai P. Wang Y. Lu G. Chen P. Ding C. Sun J. A spatiotemporal correlative k-nearest neighbor model for short-term traffic multistep forecasting Transp. Res. Part C Emerg. Technol. 2016 62 21 34 10.1016/j.trc.2015.11.002 24. Zhang K. Zheng L. Liu Z. Jia N. A deep learning based multitask model for network-wide traffic speed prediction Neurocomputing 2020 396 438 450 10.1016/j.neucom.2018.10.097 25. Kong F. Li J. Lv Z. Construction of intelligent traffic information recommendation system based on long short-term memory J. Comput. Sci. 2018 26 78 86 10.1016/j.jocs.2018.03.010 26. Aljawarneh S. Aldwairi M. Yassein M.B. Anomaly-based intrusion detection system through feature selection analysis and building hybrid efficient model J. Comput. Sci. 2018 25 152 160 10.1016/j.jocs.2017.03.006 27. Cao M. Li V.O. Chan V.W. A CNN-LSTM model for traffic speed prediction Proceedings of the 2020 IEEE 91st Vehicular Technology Conference (VTC2020-Spring) Antwerp, Belgium 25–28 May 2020 IEEE Piscataway, NJ, USA 2020 1 5 28. Essien A. Petrounias I. Sampaio P. Sampaio S. Improving urban traffic speed prediction using data source fusion and deep learning Proceedings of the 2019 IEEE International Conference on Big Data and Smart Computing (BigComp) Kyoto, Japan 27 February–2 March 2019 IEEE Piscataway, NJ, USA 2019 1 8 29. Zhang T. Jin J. Yang H. Guo H. Ma X. Link speed prediction for signalized urban traffic network using a hybrid deep learning approach Proceedings of the 2019 IEEE Intelligent Transportation Systems Conference (ITSC) Auckland, New Zealand 27–30 October 2019 IEEE Piscataway, NJ, USA 2019 2195 2200 30. Asif M.T. Dauwels J. Goh C.Y. Oran A. Fathi E. Xu M. Dhanya M.M. Mitrovic N. Jaillet P. Spatiotemporal patterns in large-scale traffic speed prediction IEEE Trans. Intell. Transp. Syst. 2013 15 794 804 10.1109/TITS.2013.2290285 31. Ye Q. Szeto W.Y. Wong S.C. Short-term traffic speed forecasting based on data recorded at irregular intervals IEEE Trans. Intell. Transp. Syst. 2012 13 1727 1737 10.1109/TITS.2012.2203122 32. Yao B. Chen C. Cao Q. Jin L. Zhang M. Zhu H. Yu B. Short-term traffic speed prediction for an urban corridor Comput.-Aided Civ. Infrastruct. Eng. 2017 32 154 169 10.1111/mice.12221 33. Li J. Guo F. Sivakumar A. Dong Y. Krishnan R. Transferability improvement in short-term traffic prediction using stacked LSTM network Transp. Res. Part C Emerg. Technol. 2021 124 102977 10.1016/j.trc.2021.102977 34. Mena-Oreja J. Gozalvez J. A Comprehensive Evaluation of Deep Learning-Based Techniques for Traffic Prediction IEEE Access 2020 8 91188 91212 10.1109/ACCESS.2020.2994415 35. Yu D. Liu C. Wu Y. Liao S. Anwar T. Li W. Zhou C. Forecasting short-term traffic speed based on multiple attributes of adjacent roads Knowl.-Based Syst. 2019 163 472 484 10.1016/j.knosys.2018.09.003 36. Liu Q. Wang B. Zhu Y. Short-term traffic speed forecasting based on attention convolutional neural network for arterials Comput.-Aided Civ. Infrastruct. Eng. 2018 33 999 1016 10.1111/mice.12417 37. Zafar N. Ul Haq I. Traffic congestion prediction based on Estimated Time of Arrival PloS ONE 2020 15 e0238200 10.1371/journal.pone.0238200 33326460 38. Chen T. Guestrin C. Xgboost: A scalable tree boosting system Proceedings of the 22nd ACM Sigkdd International Conference on Knowledge Discovery and Data Mining ACM New York, NY, USA 2016 785 794 39. Song Y. Ying L. Decision tree methods: Applications for classification and prediction Shanghai Arch. Psychiatry 2015 27 130 26120265 40. Seber G. Lee A. Linear Regression Analysis John Wiley & Sons Hoboken, NJ, USA 2012 41. Peterson L. K-nearest neighbor Scholarpedia 2009 4 1883 10.4249/scholarpedia.1883 42. Zaccone G. Karim M.R. Menshawy A. Deep Learning with TensorFlow Packt Publishing Ltd. Birmingham, UK 2017 43. Gulli A. Pal S. Deep Learning with Keras Packt Publishing Ltd. Birmingham, UK 2017 44. Albawi S. Mohammed T. Al-Zawi S. Understanding of a convolutional neural network Proceedings of the 2017 International Conference On Engineering And Technology (ICET) Antalya, Turkey 21–23 August 2017 1 6 45. George S. Santra A. An improved long short-term memory networks with Takagi-Sugeno fuzzy for traffic speed prediction considering abnormal traffic situation Comput. Intell. 2020 36 964 993 10.1111/coin.12291 46. Weisstein E. Hyperbolic Functions 2003 Available online: https://mathworld.wolfram.com/ (accessed on 1 October 2021) 47. Han J. Moraga C. The influence of the sigmoid function parameters on the speed of backpropagation learning Proceedings of the International Workshop On Artificial Neural Networks Perth, Western Australia 27 November–1 December 1995 195 201 48. Sunny M. Maswood M. Alharbi A. Deep Learning-Based Stock Price Prediction Using LSTM and Bi-Directional LSTM Model Proceedings of the 2020 2nd Novel Intelligent And Leading Emerging Sciences Conference (NILES) Giza, Egypt 24–26 October 2020 87 92 49. Zhang R. Bai L. Zhang J. Tian L. Li R. Liu Z. Convolutional LSTM networks for vibration-based defect identification of the composite structure Proceedings of the 2020 International Conference On Sensing, Measurement & Data Analytics In The Era Of Artificial Intelligence (ICSMD) Xi’an, China 15–17 October 2020 390 395 50. Hyndman R. Koehler A. Another look at measures of forecast accuracy Int. J. Forecast. 2006 22 679 688 10.1016/j.ijforecast.2006.03.001 51. Chai T. Draxler R. Root mean square error (RMSE) or mean absolute error (MAE)?–Arguments against avoiding RMSE in the literature Geosci. Model Dev. 2014 7 1247 1250 10.5194/gmd-7-1247-2014 52. Su Y. Zhang Q. Quality Assessment of Speckle Patterns by Estimating RMSE International Digital Imaging Correlation Society Springer New York, NY, USA 2017 71 74 53. He F. Yan X. Liu Y. Ma L. A traffic congestion assessment method for urban road networks based on speed performance index Procedia Eng. 2016 137 425 433 10.1016/j.proeng.2016.01.277 54. Benesty J. Chen J. Huang Y. Cohen I. Pearson correlation coefficient Noise Reduction in Speech Processing Springer Berlin/Heidelberg, Germany 2009 1 4 55. Ponomarev S. Dur J. Wallace N. Atkison T. Evaluation of random projection for malware classification Proceedings of the 2013 IEEE Seventh International Conference On Software Security And Reliability Companion Gaithersburg, MD, USA 18–20 June 2013 68 73

PMC009xxxxxx/PMC9099663.txt

==== Front Int J Mol Sci Int J Mol Sci ijms International Journal of Molecular Sciences 1422-0067 MDPI 10.3390/ijms23094951 ijms-23-04951 Review Soluble Epoxide Hydrolase as a Therapeutic Target for Neuropsychiatric Disorders Shan Jiajing https://orcid.org/0000-0002-8892-0439 Hashimoto Kenji * Fornai Francesco Academic Editor Division of Clinical Neuroscience, Chiba University Center for Forensic Mental Health, Chiba 260-8670, Japan; shanjiajing0525@gmail.com * Correspondence: hashimoto@faculty.chiba-u.jp; Tel.: +81-43-226-2587 29 4 2022 5 2022 23 9 495124 3 2022 28 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). It has been found that soluble epoxide hydrolase (sEH; encoded by the EPHX2 gene) in the metabolism of polyunsaturated fatty acids (PUFAs) plays a key role in inflammation, which, in turn, plays a part in the pathogenesis of neuropsychiatric disorders. Meanwhile, epoxy fatty acids such as epoxyeicosatrienoic acids (EETs), epoxyeicosatetraenoic acids (EEQs), and epoxyeicosapentaenoic acids (EDPs) have been found to exert neuroprotective effects in animal models of neuropsychiatric disorders through potent anti-inflammatory actions. Soluble expoxide hydrolase, an enzyme present in all living organisms, metabolizes epoxy fatty acids into the corresponding dihydroxy fatty acids, which are less active than the precursors. In this regard, preclinical findings using sEH inhibitors or Ephx2 knock-out (KO) mice have indicated that the inhibition or deficiency of sEH can have beneficial effects in several models of neuropsychiatric disorders. Thus, this review discusses the current findings of the role of sEH in neuropsychiatric disorders, including depression, autism spectrum disorder (ASD), schizophrenia, Parkinson’s disease (PD), and stroke, as well as the potential mechanisms underlying the therapeutic effects of sEH inhibitors. autism spectrum disorder depression inflammation Parkinson’s disease schizophrenia soluble epoxide hydrolase stroke Japan Society for the Promotion of Science21H02846 21H00184 21H05612 This research was funded by Japan Society for the Promotion of Science, grant number 21H02846 (to K.H.), 21H00184 (to K.H.) and 21H05612 (to K.H.). ==== Body pmc1. Introduction Neuropsychiatric disorders, including depression, autism spectrum disorder (ASD), schizophrenia, and Parkinson’s disease (PD), are common brain diseases characterized by cognitive deficits, psychiatric symptoms, and somatoform symptoms [1]. Meanwhile, psychiatric disorders such as depression, ASD, and schizophrenia are cumulatively common and show a remarkable increase of prevalence in young people [2,3]. Moreover, according to the 2019 Global Burden of Diseases, Injuries, and Risk Factors Study (GBD), psychiatric disorders are among the leading causes of the global health-related burden [4]. This burden was generally high across the entire lifespan for both genders and across many locations [5]. In 2020, as COVID-19 spread throughout the world, the increasing SARS-CoV-2 infection rates and decreasing human mobility contributed to a significant increase in the prevalence of neuropsychiatric disorders [6]. Neurological disorders, including PD, Alzheimer’s disease, and stroke, have also become a major public health challenge, due to the decreased quality of life and increased burden for millions of patients and their caregivers [7]. Although several approved drugs have been used to eliminate some symptoms in patients with neuropsychiatric disorders, the development of new therapeutic drugs for such disorders must occur to meet the medical needs of individuals [8,9,10,11,12]. In general, fatty acids are the major products of lipid metabolism [13]. Additionally, polyunsaturated fatty acids (PUFAs) are essential dietary fats that include more than one double bond and are classified based on the number of carbon atoms located between the first double bond of the carbon chain and the terminal methyl end [14,15]. In particular, PUFAs in the n-3 and n-6 (omega-3 and omega-6) families play a prominent role in biological function as components of cell membranes, membrane fluidity regulation, membrane-associated proteins, and neurotransmission [16]. Increasing evidence has shown that inadequate diets or metabolic deficiencies can cause low levels of n-3 PUFAs, which are related to the etiologies of various neuropsychiatric disorders such as PD, schizophrenia, depression, and attention deficit hyperactivity disorder [14,17,18,19]. The biologically important long-chain PUFAs include docosahexaenoic acid (DHA or 22:6 n-3) (see Figure 1), which represents approximately 40% of central PUFAs. Such acids contribute to neural cell signaling, membrane fatty acid chain fluidity, ion permeability, and protein function [20,21]. At the cerebral level of the n-6 family, roughly 50% of PUFAs are represented by arachidonic acid (ARA, 20:4 n-6) (see Figure 1), which plays a role in signaling, memory, and learning modulation [22,23]. Another important lipid is eicosapentaenoic acid (EPA, 20:5 n-3) (see Figure 1), whose concentration is significantly lower in the central nervous system. It also stands out for its anti-inflammatory properties, which can have neuroprotective impacts on the brain [24,25]. Specifically, PUFAs are metabolized into bioactive derivatives by main enzymes such as cyclooxygenases (COXs), lipoxygenases (LOXs), and cytochrome P450s (CYPs) (see Figure 1) [26,27,28], whereas EPA is converted into hydroxyeicosapentaenoic acids (HEPEs) and prostaglandin E3 (PGE3) through the LOX and COX pathways, respectively. Moreover, neuroprotectin D1 (NPD1) and electrophile oxo-derivatives (EFOXs) are synthesized from DHA through the LOX and COX pathways, respectively. These mediators such as HEPEs, PGE3, NPD1, and EFOXs act as anti-apoptotic protein activators and suppress inflammatory gene expression (see Figure 1) [14]. EPA and DHA are also converted into epoxyeicosatetraenoic acids (EEQs) and epoxydocosapentaenoic acids (EDPs) through the CYP pathway, respectively. These epoxide fatty acids are then metabolized into their corresponding diols (DHETEs and DHDPAs) by soluble epoxide hydrolase (sEH: coded by the EPHX2 gene). For example, 19,20-epoxydocosapentaenoic acid (19,20-EDP) synthesized from DHA through the CYP pathway is metabolized into 19,20-dihydroxydocosapentaenoic acid (19,20-DHDPA) by sEH [29,30]. ARA is also metabolized by the COX and LOX pathways to create a class of compounds known as leukotrienes and prostaglandins, which are important signaling molecules that control pro-inflammatory actions (see Figure 1 and Figure 2) [26,27,31]. Furthermore, the CYP pathway generates epoxyeicosatrienoic acids (EETs) from ARA, and EETs are metabolized into dihydroxyeicosatrienoic acids (DHETs) by sEH (see Figure 1 and Figure 2) [26,27,32]. Since DHETs dramatically reduce biologic activity, sEH inhibitors have been extensively used to prolong the anti-inflammatory function of EETs in the ARA cascade [33,34]. Such inhibitors have also been shown to decrease sEH activity, with little to no toxicity in animal models [35,36]. Meanwhile, inflammatory cytokines and chemokines are found in various neurobiological pathways which are related to neuropsychiatric disorders [37]. Overall, it is likely that sEH plays a role in the pathogenesis of psychiatric and neurological disorders. Thus, this review discusses the role of sEH in neuropsychiatric disorders such as depression, autism spectrum disorder (ASD), schizophrenia, Parkinson’s disease (PD), and stroke as well as the potential mechanisms underlying the therapeutic effects of sEH inhibitors. 2. Depression According to the 2021 Global Health Data Exchange, depression is one of the most common psychiatric disorders in the world, with an estimated 3.8% of the population affected, including 5.0% of adults and 5.7% of adults 60 years and older [38]. Current antidepressants such as selective serotonin reuptake inhibitors or serotonin–norepinephrine reuptake inhibitors can take several weeks before they are effective. Since approximately one-third of patients with depression do not respond to current antidepressants, new antidepressants must be developed for such treatment-resistant patients [27,39,40]. Moreover, numerous studies have suggested a strong association between inflammatory processes and the pathophysiology of depression [41,42,43,44,45,46,47]. The cytochrome P450 epoxygenase CYP2J2 converts ARA into four regioisomeric EETs (see Figure 2), while systemic overexpression of human CYP2J2 reduces the increased plasma levels of inflammatory cytokines and decreased levels of the anti-inflammatory mediator interleukin-10 (IL-10) after injection of tumor necrosis factor-α (TNF-α) [48]. In addition, the increase of inflammatory protein in TNF-α treated human bronchi is suppressed by 14,15-EET [49]. These findings suggest that the decrease of EETs metabolized by sEH can aggravate inflammation in the brain. Hence, regarding inflammation in depression, sEH is likely to play a crucial role [27,39]. In a related study, the expression of sEH protein in the brain was higher in a sample of susceptible mice after chronic social defeat stress (CSDS) compared to the control mice [50]. Other findings from the study were as follows. First, the expression of sEH protein was higher in the postmortem brain samples of patients with depression compared to those of the controls [50]. Second, pretreatment with the sEH inhibitor TPPU [1-(1-propionylpiperidin-4-yl)-3-(4-(trifluoromethoxy) phenyl) urea] prevented the onset of depression-like behaviors after CSDS. Third, the sample of Ephx2 KO mice did not show depression-like behaviors after CSDS, suggesting stress resilience [50]. Interestingly, fecal microbiota transplantation from CSDS-susceptible mice with depression-like phenotype produced such a phenotype in antibiotic-treated Ephx2 KO mice, indicating that the administration of “depression-related microbes” can contribute to the conversion of resilient Ephx2 KO mice into KO mice with depression-like phenotype [51]. Altogether, these results suggest that sEH plays a key role in the pathophysiology of depression and that sEH inhibitors can be potential therapeutic or prophylactic drugs for depression [27,39,50]. The downregulation of hepatic sEH in mice caused a reduction in sucrose preference and coat deterioration compared with the control group [52]. Moreover, patients with depression showed higher levels of sEH protein in the parietal cortex and liver compared to those in the control group [53]. Thus, it is likely that the brain–liver axis plays a role in depression [52,53,54]. Immunoreactivity of sEH was also detected in astrocytes throughout the brain [55], whereas sEH activity in the astrocytes of the medial prefrontal cortex (PFC) of CSDS-susceptible mice was negatively correlated with depression-like behaviors [56]. Moreover, a recent study showed that sEH in the central nucleus of the amygdala regulates anxiety-related behaviors [57], while TPPU produced antidepressant-like effects in the lipopolysaccharide (LPS)-induced inflammation model of depression and in the CSDS model [50]. However, the antidepressant-like effects of TPPU were blocked by the tropomyosin receptor kinase B (TrkB) antagonist, indicating that brain-derived neurotrophic factor (BDNF)-TrkB signaling plays a certain role in the antidepressant-like effects of TPPU [58,59]. In other research, pretreatment with TPPU attenuated the increase of pro-inflammatory cytokine IL-1β and rescued neuronal and dendritic spine loss in the hippocampus by increasing the expression of the N-methyl-D-aspartate receptor, the extracellular-signal-regulated kinase (ERK)1/2, and the CREB (cAMP response element binding protein) [60]. In the LiCl-pilocarpine-induced post-status epilepticus rat model, TPPU attenuated spontaneous recurrent seizures and epilepsy-associated depression-like behaviors through anti-inflammatory effects [61], while co-treatment with TPPU, EPA, and DHA was more effective in preventing IL-1β, IL-6, and TNF-α-induced decreased neurogenesis and increased apoptosis [62]. Furthermore, the serum levels of sEH-derived fatty acid diols increased in depressed patients with type 2 diabetes mellitus, while depressive symptom severity was associated with the oxylipin profile [63], suggesting higher activity of sEH in these patients. In sum, increased activity of sEH most likely plays a role in the pathogenesis of depression and suggests that sEH inhibitors are potential antidepressants [27,39,64,65,66,67]. 3. ASD and Schizophrenia Accumulating evidence has suggested that maternal immune activation (MIA) such as maternal infection can increase the risk of neuropsychiatric disorders (e.g., ASD and schizophrenia) in offspring [68,69,70,71,72,73,74,75]. A meta-analysis also showed a strong relationship between maternal infection during pregnancy and the increased risk of ASD in offspring [76]. It was also pointed out that the COVID-19 pandemic may increase the risk of ASD and schizophrenia in offspring after maternal infection of SARS-CoV-2 [77,78,79]. Meanwhile, MIA using poly(I:C) has been widely used as animal models of ASD and schizophrenia [80,81]. For example, using rodents, there were higher levels of sEH and decreased levels of epoxy fatty acids (i.e., 10,11-EDP, 5,6-EET, 8,9-EET, 11,12-EET) in the PFC of juvenile offspring after MIA, indicating increased activity of sEH in the PFC of juvenile offspring after MIA [82]. The expression of EPHX2 mRNA in induced pluripotent stem cell-derived neurospheres was higher among schizophrenia patients than the controls [82]. There was also a higher expression of EPHX2 mRNA in the postmortem brain samples of ASD patients than that of the controls [82]. Additionally, the levels of sEH in the parietal cortex from schizophrenia patients were higher than those of the controls [50,53]. Collectively, neuroinflammation by the increased expression of sEH most likely plays a role in the pathogenesis of ASD and schizophrenia. In related research, repeated treatment with TPPU in juvenile offspring from prenatal days P28 to P56 prevented cognitive deficits and loss of parvalbumin (PV)-immunoreactivity in the medial PFC of adult offspring, especially after MIA [82]. Additionally, treatment with TPPU in pregnant mothers from E5 to P21 prevented cognitive deficits, social interaction deficits, and PV-immunoreactivity in the medial PFC of juvenile offspring after MIA. Altogether, increased activity of sEH in the brain can contribute to ASD (or schizophrenia)-like behavioral abnormalities in offspring after MIA. Epidemiological studies have suggested that exposure to herbicides (i.e., glyphosate) during pregnancy might increase the risk of ASD in offspring. For instance, a population-based case study in California (USA) reported that the risk of ASD was associated with the use of glyphosate (odds ratio = 1.16) [83]. It was also found that maternal glyphosate exposure during pregnancy and lactation caused ASD-like behavioral abnormalities, an increase of expression of sEH in the PFC, hippocampus, and striatum of juvenile offspring, and a decrease of PV-immunoreactivity in the prelimbic of the medial PFC of juvenile mice [84,85]. In addition, the levels of 8,9-EET in the blood and brain regions (i.e., PFC, hippocampus, and striatum) of juvenile offspring after maternal glyphosate exposure were lower than those of the control groups, indicating increased expression of sEH in the brain regions [84]. Moreover, oral administration of TPPU to pregnant mothers from E5 to P21 prevented ASD-like behaviors such as social interaction deficits and increased grooming time in the juvenile offspring after maternal glyphosate exposure. Collectively, increased sEH in the brain seems to play a role in the pathogenesis of ASD after maternal glyphosate exposure [84,85,86]. Finally, the potent sEH inhibitor AS2586114 improved schizophrenia-like behavioral abnormalities (e.g., hyperlocomotion and pre-pulse inhibition deficits) in a sample of phencyclidine (PCP)-treated mice, suggesting that sEH inhibitor might have antipsychotic-like activity [87]. Furthermore, TPPU in drinking water during the juvenile and adolescent stages of offspring can also prevent the onset of cognitive deficits and a reduction of PV-immunoreactivity in the medial PFC of adult offspring after MIA [82]. Repeated treatment with TPPU from pregnancy to weaning can also prevent the onset of cognitive deficits in juvenile offspring after MIA or maternal glyphosate exposure [82,84]. Overall, these findings indicate that sEH plays a key role in the development of ASD and schizophrenia in offspring after MIA and that sEH inhibitors can have prophylactic or therapeutic impacts on neuropsychiatric disorders [10,88,89]. 4. Parkinson’s Disease Parkinson’s disease (PD) is a neurodegenerative disease characterized by the deposition of the aggregates of α-synuclein (termed “Lewy bodies”) and the loss of dopaminergic neurons in the substantia nigra (SN), which results in motor dysfunction and non-motor dysfunction [90,91,92]. PD is the second most prevalent neurodegenerative disorder, only after Alzheimer’s disease. It is also predicted that the number of patients with PD will double over the next 20 years [93]. Meanwhile, L-DOPA (the precursor of dopamine) or dopamine (DA) receptor agonists have been used in the treatment of PD [94]. Although these treatments seem to alleviate symptoms, there are no disease-modifying or neuroprotective drugs for PD [95,96,97]. Multiple evidence has supported neuroinflammation-related oxidative stress in the pathogenesis of PD [66,98,99,100]. Related research has shown that activated astrocytes and microglia caused by brain immune response are involved in the development of neuroinflammatory features, leading to the exacerbation of DA neurons in the substantia nigra pars compacta (SNc) [101]. In addition, EETs, regulators of inflammation processes, can produce the neurotrophic role of astrocytes, increase the release of BDNF, reduce glutamatergic toxicity through the astrocytic metabotropic glutamate receptor mGluR5, prevent mitochondrial dysfunction and apoptosis, and protect synaptic function in the brain [100,102,103,104]. MPTP (1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine)-induced neurotoxicity in the striatum and SNc of a sample of rodents has been a well-established preclinical in vivo model to study the pathogenesis of PD [105,106]. Higher expression of the sEH protein in the striatum has also been found in MPTP-treated mice or postmortem brain samples from dementia patients with Lewy bodies [107]. Moreover, the expression of the EPHX2 mRNA in human PARK2 (Parkinson’s disease protein 2) iPSC-derived neurons were significantly higher compared to the control groups [107], while the levels of 8,9-EET in the striatum of MPTP-treated mice were lower than those of the control mice, suggesting higher activity of sEH in the striatum of the MPTP-treated mice [107]. Interestingly, sEH expression was positively correlated to the phosphorylation of α-synuclein in the striatum of the MPTP-treated mice [107]. Since PD may not display symptoms until approximately 80% of the striatal DA has been lost, a biomarker for earlier diagnosis is important for finding treatment options [108]. In this regard, peripheral detection of elevated levels of sEH in the gut, which is possible prior to the losses of dopaminergic neurons, may potentially provide an early biomarker of PD [95,108]. MPTP-induced neurotoxicity in the striatum and SN, including the loss of the DA transporter (DAT), loss of tyrosine hydrolase (TH)-positive cells, and increased endoplasmic reticulum (ER) stress, was attenuated by repeated oral administration of TPPU [107]. Additionally, both AUDA [12-(((tricyclo(3.3.1.13,7)dec-1-ylamino)carbonyl)amino)-dodecanoic acid], another sEH inhibitor, and sEH deficiency significantly protected against MPTP-induced toxicity [109,110]. A recent study found that the natural compound kurarinone, an uncompetitive inhibitor of sEH, can alleviate the MPTP-induced behavioral deficits, dopaminergic neurotoxicity, and neuroinflammation via suppressing the activated microglia, including the nuclear factor kappa B (NF-κB) signaling pathway [111]. Furthermore, the inhibition of sEH can suppress the aggregation of α-synuclein, microglia activation, neuroinflammation, and apoptosis, resulting in the neuroprotection activity in the PD model [112,113]. Altogether, increased sEH and the resulting increase in phosphorylation of α-synuclein may play a role in the pathogenesis of PD and that sEH inhibitors can be potential drugs for PD. 5. Stroke Stroke, constituting a loss of blood flow to the brain, is the leading cause of disability and death worldwide. In fact, the global prevalence of stroke in 2019 alone was 101.5 million people [114]. Acute episodes, including ischemic strokes, hemorrhagic strokes, traumatic brain injury, and seizures, can cause neuroinflammation in the brain, resulting in the loss of neurons and activation of resident immune cells [92]. Subsequent activation of immune responses can increase pro-inflammatory cytokines [115] and compromise the brain-blood barrier (BBB), further deteriorating neurodegeneration and exacerbating the injury caused by stroke [116]. Despite extensive efforts to discover better therapies for stroke, treatment options are still limited. The primary current treatment for ischemic stroke is thrombolysis with the tissue plasminogen activator (tPA), which was approved by the Food and Drug Administration. However, it must be administered within a relatively short time window before neuronal loss occurs [117]. While the tPA can be effective for patients with ischemic stroke, it can also aggravate hemorrhagic strokes, which include similar clinical symptoms with ischemia [118]. Thus, developing a therapy that focuses on a single agent that targets multiple mechanisms of ischemic brain injury may prove more effective [119]. As for sEH, it catalyzes the metabolism of EETs, which exhibit potentially beneficial actions in stroke via vasodilation, neuroprotection, promotion of angiogenesis, suppression of platelet aggregation, oxidative stress, and post-ischemic inflammation [120]. In related research, hypertension- and diabetes-associated ischemic stroke risk increased through the EPHX2 gene variants in the Turkish population [121]. Moreover, serum oxylipin changes consistent with higher sEH activity played a key role in vascular cognitive impairment, which is associated with the injury of periventricular subcortical white matter [122]. Collectively, sEH inhibition is a potential intervention for stroke patients based on the beneficial properties of EETs. The inhibition of sEH is also effective for reducing infarct volume, improving memory deficits, and alleviating cognitive impairment and microvasculature augmentation by suppressing neuroinflammation and increasing reparative cytokines and growth factors such as BDNF and doublecortin [123,124,125]. Regarding BDNF, it activates the receptor TrkB to promote the growth and differentiation of nerve cells and to have a neuroprotective effect on the neurons. Thus, a blockade of the BDNF-TrkB signaling pathway by the TrkB inhibitor ANA-12 can abolish the protective effect of sEH gene deletion in the middle cerebral arterial occlusion (MCAO) models of stroke [126], suggesting the role of BDNF-TrkB signaling in the actions of sEH inhibitors. Finally, administration of TPPU or AUDA significantly promoted M2 polarization of microglial cells, which indicates a shift from pro-inflammatory polarized microglia to microglia that primarily release anti-inflammatory cytokines. This can result in the differentiation of oligodendrocytes, protection against white matter integrity, and remyelination against chronic hypoperfusion [127,128,129,130,131]. Interestingly, treatment with t-AUCB (a sEH inhibitor) after ischemic stroke onset has been shown to exert brain protection in a sample of non-diabetic mice, but not in type 2 diabetes mellitus (DM2) mice, while DM2-induced hyperglycemia can abolish t-AUCB-mediated neuroprotection against stroke [132]. SMTP-7 targeting sEH has also been shown to be effective in treating severe embolic stroke in a sample of monkeys under conditions in which tPA treatment causes hemorrhagic infarct-associated premature death [133,134]. Furthermore, related research has shown that treatment with the sEH inhibitor can decrease the activity of matrix metalloproteases (MMP)-2 and MMP-9, increase the expression of tight junction proteins, reduce activation of NF-κB, and suppress the apoptosis to protect the BBB integrity from ischemia [135,136]. Altogether, these studies suggest that sEH inhibition can have multi-target protective effects and alleviate cognitive impairment after a stroke. 6. Conclusions and Future Perspectives As discussed in this review, sEH inhibitors can exert a neuroprotective effect through potent anti-inflammatory actions, including BDNF-TrkB activation for inflammation-related endoplasmic reticulum stress and mitochondrial dysfunction (Figure 3). For example, EETs suppress the MMP-2 and MMP-9, resulting in the prevention of mitochondrial dysfunction. Furthermore, EETs induce the expression of BDNF which binds to its receptor TrkB, resulting in the MEK–ERK–CREB signaling pathway [45,137]. BDNF–TrkB–MEK–ERK–CREB pathway could contribute to neurite outgrowth, neurogenesis, and synaptic plasticity. Furthermore, EETs could reduce oxidative stress and ER stress through suppression of NF-kB and activation of Bcl-2, respectively. In addition, EETs can protect against apoptosis through suppression of the BAX (Bxcl-2-associated X protein) and caspase-3 (see Figure 3). Thus, it is likely that sEH inhibitors can potentially serve as prophylactic or therapeutic drugs for neuropsychiatric disorders such as depression, ASD, schizophrenia, PD, and stroke. Currently, clinical studies using two sEH inhibitors (i.e., GSK2256294 and EC5026) (see Figure 4) in humans are underway [138]. For instance, a recent randomized, double-blind, placebo-controlled study found that treatment with GSK2256294 for seven days in a sample of subjects with obesity and prediabetes (n = 16) effectively inhibited sEH activity in plasma, muscle, and adipose tissue. Although it reduced F2-isoprostanes (a marker of oxidative stress), it did not improve insulin sensitivity or blood pressure [139]. Moreover, a double-blind, randomized placebo-controlled trial indicated that treatment with GSK2256294 for 10 days in a sample of patients (n = 10) with aneurysmal subarachnoid hemorrhage resulted in a considerable increase in the serum EET/DHET ratio at days 7 and 10 but not in the cerebrospinal fluid (CSF) [140]. Conversely, there was decreased CSF inflammatory cytokines following GSK2256294 treatment, but it did not achieve statistical significance [140]. These clinical studies showed that treatment with GSK2256294 can cause an increase in the EET/DHET ratio of blood and tissue through potent sEH inhibition in humans. Furthermore, although a clinical trial of EC5026 in humans is underway by EicOsis Human Health Inc. (Davis, CA, USA) [138], future randomized, double-blind, placebo-control studies using a large sample size are necessary for determining the efficacy of sEH inhibitors in patients with neuropsychiatric disorders. Acknowledgments The authors would like to than our collaborators who are listed as the co-authors of our papers in the reference list. Author Contributions Conceptualization, J.S. and K.H.; writing—original draft preparation, J.S.; writing—review and editing, K.H.; visualization, K.H.; supervision, K.H.; project administration, K.H.; funding acquisition, K.H. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. Figure 1 Metabolism of eicosapentaenoic acid (EPA), docosohexaenoic acid (DHA), and arachidonic acid (ARA). Eicosapentaenoic acid (EPA) is converted into hydroxyeicosapentaenoic acids (HEPEs) and prostaglandin E3 (PGE3) through the lipoxygenase (LOX) and cyclooxygenase (COX) pathways, respectively. Docosohexaenoic acid (DHA) is also converted into neuroprotectin D1 (NPD1) and electrophile oxo-derivatives (EFOXs) through the LOX and COX pathways, respectively. The compounds HEPEs, NPD1, PGE3, and EFOXs act as anti-inflammatory and anti-apoptotic mediators, while EPA and DHA are converted into epoxyeicosatetraenoic acids (EEQs) and epoxydocosapentaenoic acids (EDPs) through the cytochrome P450 (CYP) pathway, respectively. Moreover, these epoxide fatty acids are metabolized into their corresponding diols (DHETEs and DHDPAs) by soluble epoxide hydrolase (sEH), while arachidonic acid (ARA) is converted into leukotrienes and prostaglandins by the LOX and COX pathways, respectively. Finally, ARA is converted into epoxyeicosatrienoic acids (EETs) by the CYP pathway, and these EETs are metabolized into their corresponding diols (DHETs) by sEH. In this case, epoxy fatty acids (EEQs, EDPs, and EETs) have anti-inflammatory effects. Figure 2 Roles of CYP and sEH in the arachidonic acid (ARA) cascade. In addition to the COX and LOX pathways, ARA is metabolized into four epoxyeicosatrienoic acids (EETs) through the cytochrome P450 (CYP) pathway. EETs are then metabolized into the corresponding dihydroxyeicosatetraenoic acids (DHETs) by soluble epoxide hydrolase (sEH). EETs have anti-inflammatory effects, while DHETs are inactive or have fewer active effects. Figure 3 Molecular mechanisms underlying the neuroprotective effect of EETs. EETs suppress the MMP-2 and MMP-9 to prevent mitochondrial dysfunction. EETs induce the expression of BDNF, then stimulate BDNF-TrkB signaling. Subsequently, BDNF-TrkB signaling stimulates the MEK–ERK–CREB signaling pathway, promoting neurite outgrowth, neurogenesis, and synaptic plasticity. Furthermore, EETs reduce oxidative stress and suppress ER stress through suppression of NF-kB and activation of Bcl-2, respectively. Moreover, EETs protect against apoptosis through suppressing the BAX and Caspase-3; MMP-2: matrix metalloproteinase-2; MMP-9: matrix metalloproteinase-9; BDNF: brain derived neurotrophic factor; TrkB: tropomyosin receptor kinase B; MEK: mitogen-activated protein kinase; ERK: extracellular signal-regulated kinase; CREB: cAMP response element binding protein; ER: endoplasmic reticulum; NF-κB: nuclear factor kappa-light-chain-enhancer of activated B cells; Bcl-2: B-cell lymphoma 2; BAX: Bcl-2-associated X protein. Made by Figdraw (www.figdraw.com, accessed on 24 March 2022). Figure 4 Chemical structure of GSK2256294 and EC5026 for human use. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Cheng L. Liu J. Chen Z. The histaminergic system in neuropsychiatric disorders Biomolecules 2021 11 1345 10.3390/biom11091345 34572558 2. Bray N.J. O’Donovan M.C. The genetics of neuropsychiatric disorders Brain Neurosci. Adv. 2018 2 2398212818799271 10.1177/2398212818799271 31179400 3. Kessler R.C. Amminger G.P. Aguilar-Gaxiola S. Alonso J. Lee S. Ustun T.B. Age of onset of mental disorders: A review of recent literature Curr. Opin. Psychiatry 2007 20 359 10.1097/YCO.0b013e32816ebc8c 17551351 4. GBD 2019 Diseases and Injuries Collaborators Global burden of 369 diseases and injuries in 204 countries and territories, 1990–2019: A systematic analysis for the Global Burden of Disease Study 2019 Lancet 2020 396 1204 1222 10.1016/S0140-6736(20)30925-9 33069326 5. GBD 2019 Diseases and Injuries Collaborators Global, regional, and national burden of 12 mental disorders in 204 countries and territories, 1990–2019: A systematic analysis for the Global Burden of Disease Study 2019 Lancet Psychiatry 2022 9 137 150 10.1016/S2215-0366(21)00395-3 35026139 6. COVID-19 Mental Disorders Collaborators Global prevalence and burden of depressive and anxiety disorders in 204 countries and territories in 2020 due to the COVID-19 pandemic Lancet 2021 398 1700 1712 10.1016/S0140-6736(21)02143-7 34634250 7. Gooch C.L. Pracht E. Borenstein A.R. The burden of neurological disease in the United States: A summary report and call to action Ann. Neurol. 2017 81 479 484 10.1002/ana.24897 28198092 8. Hashimoto K. Malchow B. Falkai P. Schmitt A. Glutamate modulators as potential therapeutic drugs in schizophrenia and affective disorders Eur. Arch. Psychiatry Clin. Neurosci. 2013 263 367 377 10.1007/s00406-013-0399-y 23455590 9. Hashimoto K. Targeting of NMDA receptors in new treatments for schizophrenia Expert Opin. Ther. Targets 2014 18 1049 1063 10.1517/14728222.2014.934225 24965576 10. Hashimoto K. Recent advances in the early intervention in schizophrenia: Future direction from preclinical findings Curr. Psychiatry Rep. 2019 21 75 10.1007/s11920-019-1063-7 31278495 11. LeWitt P.A. Chaudhuri K.R. Unmet needs in Parkinson disease: Motor and non-motor Park. Relat. Disord. 2020 80 (Suppl. 1) S7 S12 10.1016/j.parkreldis.2020.09.024 33349582 12. Grosso Jasutkar H. Oh S.E. Mouradian M.M. Therapeutics in the pipeline targeting alpha-synuclein for Parkinson’s disease Pharmacol. Rev. 2022 74 207 237 10.1124/pharmrev.120.000133 35017177 13. Mongan D. Healy C. Jones H.J. Zammit S. Cannon M. Cotter D.R. Plasma polyunsaturated fatty acids and mental disorders in adolescence and early adulthood: Cross-sectional and longitudinal associations in a general population cohort Transl. Psychiatry 2021 11 321 10.1038/s41398-021-01425-4 34059620 14. Healy-Stoffel M. Levant B. N-3 (Omega-3) fatty acids: Effects on brain dopamine systems and potential role in the etiology and treatment of neuropsychiatric disorders CNS Neurol. Disord. Drug Targets 2018 17 216 232 10.2174/1871527317666180412153612 29651972 15. Morgese M.G. Schiavone S. Mhillaj E. Bove M. Tucci P. Trabace L. N-3 PUFA diet enrichment prevents amyloid beta-induced depressive-like phenotype Pharmacol. Res. 2018 129 526 534 10.1016/j.phrs.2017.11.034 29203442 16. Dagnino-Subiabre A. Stress and Western diets increase vulnerability to neuropsychiatric disorders: A common mechanism Nutr. Neurosci. 2021 24 624 634 10.1080/1028415X.2019.1661651 31524571 17. Tesei A. Crippa A. Ceccarelli S.B. Mauri M. Molteni M. Agostoni C. Nobile M. The potential relevance of docosahexaenoic acid and eicosapentaenoic acid to the etiopathogenesis of childhood neuropsychiatric disorders Eur. Child Adolesc. Psychiatry 2017 26 1011 1030 10.1007/s00787-016-0932-4 27988864 18. Daray F.M. Grendas L.N. Rodante D.E. Errasti A.E. Cases G.G. Moix C.F. Uicich R.E. GimÉNez M.I. Puppo S. Fasolino G.H. Polyunsaturated fatty acids as predictors of future suicide attempt Prostaglandins Leukot. Essent. Fat. Acids 2021 165 102247 10.1016/j.plefa.2021.102247 19. Di Miceli M. Bosch-Bouju C. Layé S. PUFA and their derivatives in neurotransmission and synapses: A new hallmark of synaptopathies Proc. Nutr. Soc. 2020 79 388 403 10.1017/S0029665120000129 20. Chen W. Chen Y. Wu R. Guo G. Liu Y. Zeng B. Liao X. Wang Y. Wang X. DHA alleviates diet-induced skeletal muscle fiber remodeling via FTO/m6A/DDIT4/PGC1α signaling BMC Biol. 2022 20 39 10.1186/s12915-022-01239-w 35135551 21. Díaz M. Mesa-Herrera F. Marín R. DHA and its elaborated modulation of antioxidant defenses of the brain: Implications in aging and AD neurodegeneration Antioxidants 2021 10 907 10.3390/antiox10060907 34205196 22. Funk Colin D. Prostaglandins and leukotrienes: Advances in eicosanoid biology Science 2001 294 1871 1875 10.1126/science.294.5548.1871 11729303 23. Orr S.K. Bazinet R.P. The emerging role of docosahexaenoic acid in neuroinflammation Curr. Opin. Investig. Drugs 2008 9 735 743 18600579 24. Borsini A. Stangl D. Jeffries A.R. Pariante C.M. Thuret S. The role of omega-3 fatty acids in preventing glucocorticoid-induced reduction in human hippocampal neurogenesis and increase in apoptosis Transl. Psychiatry 2020 10 219 10.1038/s41398-020-00908-0 32636362 25. Giacobbe J. Benoiton B. Zunszain P. Pariante C.M. Borsini A. The anti-inflammatory role of omega-3 polyunsaturated fatty acids metabolites in pre-clinical models of psychiatric, neurodegenerative, and neurological disorders Front. Psychiatry 2020 11 122 10.3389/fpsyt.2020.00122 32180741 26. Morisseau C. Hammock B.D. Impact of soluble epoxide hydrolase and epoxyeicosanoids on human health Annu. Rev. Pharmacol. Toxicol. 2013 53 37 58 10.1146/annurev-pharmtox-011112-140244 23020295 27. Hashimoto K. Role of soluble epoxide hydrolase in metabolism of PUFAs in psychiatric and neurological disorders Front. Pharmacol. 2019 10 36 10.3389/fphar.2019.00036 30761004 28. Imig J.D. Cervenka L. Neckar J. Epoxylipids and soluble epoxide hydrolase in heart diseases Biochem. Pharmacol. 2022 195 114866 10.1016/j.bcp.2021.114866 34863976 29. Capozzi M.E. Hammer S.S. McCollum G.W. Penn J.S. Epoxygenated fatty acids inhibit retinal vascular inflammation Sci. Rep. 2016 6 39211 10.1038/srep39211 27966642 30. Yanai R. Mulki L. Hasegawa E. Takeuchi K. Sweigard H. Suzuki J. Gaissert P. Vavvas D.G. Sonoda K.H. Rothe M. Cytochrome P450-generated metabolites derived from ω-3 fatty acids attenuate neovascularization Proc. Natl. Acad. Sci. USA 2014 111 9603 9608 10.1073/pnas.1401191111 24979774 31. Wang B. Wu L. Chen J. Dong L. Chen C. Wen Z. Hu J. Fleming I. Wang D.W. Metabolism pathways of arachidonic acids: Mechanisms and potential therapeutic targets Signal Transduct. Target. Ther. 2021 6 94 10.1038/s41392-020-00443-w 33637672 32. Dos Santos L.R. Fleming I. Role of cytochrome P450-derived, polyunsaturated fatty acid mediators in diabetes and the metabolic syndrome Prostaglandins Other Lipid Mediat. 2020 148 106407 10.1016/j.prostaglandins.2019.106407 31899373 33. Hiesinger K. Wagner K.M. Hammock B.D. Proschak E. Hwang S.H. Development of multitarget agents possessing soluble epoxide hydrolase inhibitory activity Prostaglandins Other Lipid Mediat. 2019 140 31 39 10.1016/j.prostaglandins.2018.12.003 30593866 34. Hildreth K. Kodani S.D. Hammock B.D. Zhao L. Cytochrome P450-derived linoleic acid metabolites EpOMEs and DiHOMEs: A review of recent studies J. Nutr. Biochem. 2020 86 108484 10.1016/j.jnutbio.2020.108484 32827665 35. Jones R.D. Liao J. Tong X. Xu D. Sun L. Li H. Yang G.Y. Epoxy-oxylipins and soluble epoxide hydrolase metabolic pathway as targets for NSAID-induced gastroenteropathy and inflammation-associated carcinogenesis Front. Pharmacol. 2019 10 731 10.3389/fphar.2019.00731 31293429 36. Sun C.P. Zhang X.Y. Morisseau C. Hwang S.H. Zhang Z.J. Hammock B.D. Ma X.C. Discovery of soluble epoxide hydrolase inhibitors from chemical synthesis and natural products J. Med. Chem. 2021 64 184 215 10.1021/acs.jmedchem.0c01507 33369424 37. Abdoli A. Taghipour A. Pirestani M. Mofazzal Jahromi M.A. Roustazadeh A. Mir H. Ardakani H.M. Kenarkoohi A. Falahi S. Karimi M. Infections, inflammation, and risk of neuropsychiatric disorders: The neglected role of “co-infection” Heliyon 2020 6 e05645 10.1016/j.heliyon.2020.e05645 33319101 38. Institute of Health Metrics and Evaluation, Global Health Datta Exchange Available online: http://ghdx.healthdata.org/gbd-results-tool?params=gbd-api-2019-permalink/d780dffbe8a381b25e1416884959e88b (accessed on 1 May 2021) 39. Hashimoto K. Soluble epoxide hydrolase: A new therapeutic target for depression Expert Opin. Ther. Targets 2016 20 1149 1151 10.1080/14728222.2016.1226284 27534794 40. Hashimoto K. Molecular mechanisms of the rapid-acting and long-lasting antidepressant actions of (R)-ketamine Biochem. Pharmacol. 2020 177 113935 10.1016/j.bcp.2020.113935 32224141 41. Hashimoto K. Inflammatory biomarkers as differential predictors of antidepressant response Int. J. Mol. Sci. 2015 16 7796 7801 10.3390/ijms16047796 25856677 42. Zhang J.C. Yao W. Hashimoto K. Brain-derived neurotrophic factor (BDNF)-TrkB signaling in inflammation-related depression and potential therapeutic targets Curr. Neuropharmacol. 2016 14 721 731 10.2174/1570159X14666160119094646 26786147 43. Troubat R. Barone P. Leman S. Desmidt T. Cressant A. Atanasova B. Brizard B. El Hage W. Surget A. Belzung C. Neuroinflammation and depression: A review Eur. J. Neurosci. 2021 53 151 171 10.1111/ejn.14720 32150310 44. Carlessi A.S. Borba L.A. Zugno A.I. Quevedo J. Réus G.Z. Gut microbiota–brain axis in depression: The role of neuroinflammation Eur. J. Neurosci. 2021 53 222 235 10.1111/ejn.14631 31785168 45. Wei Y. Chang L. Hashimoto K. Molecular mechanisms underlying the antidepressant actions of arketamine: Beyond the NMDA receptor Mol. Psychiatry 2021 27 559 573 10.1038/s41380-021-01121-1 33963284 46. Chang L. Wei Y. Hashimoto K. Brain-gut-microbiota axis in depression: A historical overview and future directions Brain Res. Bull. 2022 182 44 56 10.1016/j.brainresbull.2022.02.004 35151796 47. Wei Y. Wang T. Liao L. Fan X. Chang L. Hashimoto K. Brain-spleen axis in health and diseases: A review and future perspective Brain Res. Bull. 2022 182 130 140 10.1016/j.brainresbull.2022.02.008 35157987 48. Zhao G. Tu L. Li X. Yang S. Chen C. Xu X. Wang P. Wang D.W. Delivery of AAV2-CYP2J2 protects remnant kidney in the 5/6-nephrectomized rat via inhibition of apoptosis and fibrosis Hum. Gene Ther. 2012 23 688 699 10.1089/hum.2011.135 22260463 49. Morin C. Sirois M. Echave V. Gomes M.M. Rousseau E. EET displays anti-inflammatory effects in TNF-α–stimulated human bronchi Am. J. Respir. Cell Mol. Biol. 2008 38 192 201 10.1165/rcmb.2007-0232OC 17872494 50. Ren Q. Ma M. Ishima T. Morisseau C. Yang J. Wagner K.M. Zhang J.C. Yang C. Yao W. Dong C. Gene deficiency and pharmacological inhibition of soluble epoxide hydrolase confers resilience to repeated social defeat stress Proc. Natl. Acad. Sci. USA 2016 113 E1944 E1952 10.1073/pnas.1601532113 26976569 51. Wang S. Ishima T. Qu Y. Shan J. Chang L. Wei Y. Zhang J. Pu Y. Fujita Y. Tan Y. Ingestion of Faecalibaculum rodentium causes depression-like phenotypes in resilient Ephx2 knock-out mice: A role of brain-gut-microbiota axis via the subdiaphragmatic vagus nerve J. Affect. Disord. 2021 292 565 573 10.1016/j.jad.2021.06.006 34147969 52. Qin X.H. Wu Z. Dong J.H. Zeng Y.N. Xiong W.C. Liu C. Wang M.Y. Zhu M.Z. Chen W.J. Zhang Y. Liver soluble epoxide hydrolase regulates behavioral and cellular effects of chronic stress Cell Rep. 2019 29 3223 3234.e6 10.1016/j.celrep.2019.11.006 31801085 53. Zhang J. Tan Y. Chang L. Hammock B.D. Hashimoto K. Increased expression of soluble epoxide hydrolase in the brain and liver from patients with major psychiatric disorders: A role of brain–liver axis J. Affect. Disord. 2020 270 131 134 10.1016/j.jad.2020.03.070 32339103 54. Yang B. Ren Q. Zhang J.C. Chen Q.X. Hashimoto K. Altered expression of BDNF, BDNF pro-peptide and their precursor proBDNF in brain and liver tissues from psychiatric disorders: Rethinking the brain-liver axis Transl. Psychiatry 2017 7 e1128 10.1038/tp.2017.95 28509900 55. Marowsky A. Burgener J. Falck J.R. Fritschy J.M. Arand M. Distribution of soluble and microsomal epoxide hydrolase in the mouse brain and its contribution to cerebral epoxyeicosatrienoic acid metabolism Neuroscience 2009 163 646 661 10.1016/j.neuroscience.2009.06.033 19540314 56. Xiong W. Cao X. Zeng Y. Qin X. Zhu M. Ren J. Wu Z. Huang Q. Zhang Y. Wang M. Astrocytic epoxyeicosatrienoic acid signaling in the medial prefrontal cortex modulates depressive-like behaviors J. Neurosci. 2019 39 4606 10.1523/JNEUROSCI.3069-18.2019 30902874 57. Ren J. Lu C.L. Huang J. Fan J. Guo F. Mo J.W. Huang W.Y. Kong P.L. Li X.W. Sun L.R. A distinct metabolically defined central nucleus circuit bidirectionally controls anxiety-related behaviors J. Neurosci. 2022 42 2356 2370 10.1523/JNEUROSCI.1578-21.2022 35105676 58. Wu Q. Cai H. Song J. Chang Q. The effects of sEH inhibitor on depression-like behavior and neurogenesis in male mice J. Neurosci. Res. 2017 95 2483 2492 10.1002/jnr.24080 28699310 59. Wu Q. Song J. Meng D. Chang Q. TPPU, a sEH inhibitor, attenuates corticosterone-induced PC12 cell injury by modulation of BDNF-TrkB pathway J. Mol. Neurosci. 2019 67 364 372 10.1007/s12031-018-1230-z 30644034 60. Peng W. Shen Y. Wang Q. Ding J. Wang X. TPPU Pre-treatment rescues dendritic spine loss and alleviates depressive behaviours during the latent period in the lithium chloride-pilocarpine-induced status epilepticus rat model Brain Sci. 2021 11 1465 10.3390/brainsci11111465 34827464 61. Shen Y. Peng W. Chen Q. Hammock B.D. Liu J. Li D. Yang J. Ding J. Wang X. Anti-inflammatory treatment with a soluble epoxide hydrolase inhibitor attenuates seizures and epilepsy-associated depression in the LiCl-pilocarpine post-status epilepticus rat model Brain Behav. Immun. 2019 81 535 544 10.1016/j.bbi.2019.07.014 31306773 62. Borsini A. Nicolaou A. Camacho-Muñoz D. Kendall A.C. Di Benedetto M.G. Giacobbe J. Su K.P. Pariante C.M. Omega-3 polyunsaturated fatty acids protect against inflammation through production of LOX and CYP450 lipid mediators: Relevance for major depression and for human hippocampal neurogenesis Mol. Psychiatry 2021 26 6773 6788 10.1038/s41380-021-01160-8 34131267 63. Anita N.Z. Forkan N. Kamal R. Nguyen M.M. Yu D. Major-Orfao C. Wong S.K. Lanctôt K.L. Herrmann N. Oh P.I. Serum soluble epoxide hydrolase related oxylipins and major depression in patients with type 2 diabetes Psychoneuroendocrinology 2021 126 105149 10.1016/j.psyneuen.2021.105149 33503568 64. Swardfager W. Hennebelle M. Yu D. Hammock B.D. Levitt A.J. Hashimoto K. Taha A.Y. Metabolic/inflammatory/vascular comorbidity in psychiatric disorders; soluble epoxide hydrolase (sEH) as a possible new target Neurosci. Biobehav. Rev. 2018 87 56 66 10.1016/j.neubiorev.2018.01.010 29407524 65. Ren Q. Soluble epoxide hydrolase inhibitor: A novel potential therapeutic or prophylactic drug for psychiatric disorders Front. Pharmacol. 2019 10 420 10.3389/fphar.2019.00420 31105566 66. Atone J. Wagner K. Hashimoto K. Hammock B.D. Cytochrome P450 derived epoxidized fatty acids as a therapeutic tool against neuroinflammatory diseases Prostaglandins Other Lipid Mediat. 2020 147 106385 10.1016/j.prostaglandins.2019.106385 31698143 67. Borsini A. The role of soluble epoxide hydrolase and its inhibitors in depression Brain Behav. Immun. Health 2021 16 100325 10.1016/j.bbih.2021.100325 34514442 68. Bilbo S.D. Block C.L. Bolton J.L. Hanamsagar R. Tran P.K. Beyond infection—Maternal immune activation by environmental factors, microglial development, and relevance for autism spectrum disorders Exp. Neurol. 2018 299 241 251 10.1016/j.expneurol.2017.07.002 28698032 69. Brown A.S. Meyer U. maternal immune activation and neuropsychiatric illness: A translational research perspective Am. J. Psychiatry 2018 175 1073 1083 10.1176/appi.ajp.2018.17121311 30220221 70. Lombardo M.V. Moon H.M. Su J. Palmer T.D. Courchesne E. Pramparo T. Maternal immune activation dysregulation of the fetal brain transcriptome and relevance to the pathophysiology of autism spectrum disorder Mol. Psychiatry 2018 23 1001 1013 10.1038/mp.2017.15 28322282 71. Solek C.M. Farooqi N. Verly M. Lim T.K. Ruthazer E.S. Maternal immune activation in neurodevelopmental disorders Dev. Dyn. 2018 247 588 619 10.1002/dvdy.24612 29226543 72. Conway F. Brown A.S. Maternal immune activation and related factors in the risk of offspring psychiatric disorders Front. Psychiatry 2019 10 430 10.3389/fpsyt.2019.00430 31316403 73. Siniscalco D. Schultz S. Brigida A.L. Antonucci N. Inflammation and neuro-immune dysregulations in autism spectrum disorders Pharmaceuticals 2018 11 56 10.3390/ph11020056 29867038 74. Matta S.M. Hill-Yardin E.L. Crack P.J. The influence of neuroinflammation in autism spectrum disorder Brain Behav. Immun. 2019 79 75 90 10.1016/j.bbi.2019.04.037 31029798 75. Gevezova M. Sarafian V. Anderson G. Maes M. Inflammation and mitochondrial dysfunction in autism spectrum disorder CNS Neurol. Disord. Drug Targets 2020 19 320 333 10.2174/1871527319666200628015039 32600237 76. Jiang H.Y. Xu L.l. Shao L. Xia R.M. Yu Z.H. Ling Z.X. Yang F. Deng M. Ruan B. Maternal infection during pregnancy and risk of autism spectrum disorders: A systematic review and meta-analysis Brain Behav. Immun. 2016 58 165 172 10.1016/j.bbi.2016.06.005 27287966 77. Hashimoto K. Risk of neuropsychiatric disorders in offspring of COVID-19-infected pregnant women and nutritional intervention Eur. Arch. Psychiatry Clin. Neurosci. 2021 271 387 389 10.1007/s00406-020-01148-5 32488522 78. Lins B. Maternal immune activation as a risk factor for psychiatric illness in the context of the SARS-CoV-2 pandemic Brain Behav. Immun. Health 2021 16 100297 10.1016/j.bbih.2021.100297 34308388 79. Hashimoto Y. Suzuki T. Hashimoto K. Mechanisms of action of fluvoxamine for COVID-19: A historical review Mol. Psychiatry 2022 1 10 10.1038/s41380-021-01432-3 33972691 80. Guma E. Plitman E. Chakravarty M.M. The role of maternal immune activation in altering the neurodevelopmental trajectories of offspring: A translational review of neuroimaging studies with implications for autism spectrum disorder and schizophrenia Neurosci. Biobehav. Rev. 2019 104 141 157 10.1016/j.neubiorev.2019.06.020 31265870 81. Haddad F.L. Patel S.V. Schmid S. Maternal immune activation by poly(I:C) as a preclinical model for neurodevelopmental disorders: A focus on autism and schizophrenia Neurosci. Biobehav. Rev. 2020 113 546 567 10.1016/j.neubiorev.2020.04.012 32320814 82. Ma M. Ren Q. Yang J. Zhang K. Xiong Z. Ishima T. Pu Y. Hwang S.H. Toyoshima M. Iwayama Y. Key role of soluble epoxide hydrolase in the neurodevelopmental disorders of offspring after maternal immune activation Proc. Natl. Acad. Sci. USA 2019 116 7083 7088 10.1073/pnas.1819234116 30890645 83. von Ehrenstein O.S. Ling C. Cui X. Cockburn M. Park A.S. Yu F. Wu J. Ritz B. Prenatal and infant exposure to ambient pesticides and autism spectrum disorder in children: Population based case-control study BMJ 2019 364 l962 10.1136/bmj.l962 30894343 84. Pu Y. Yang J. Chang L. Qu Y. Wang S. Zhang K. Xiong Z. Zhang J. Tan Y. Wang X. Maternal glyphosate exposure causes autism-like behaviors in offspring through increased expression of soluble epoxide hydrolase Proc. Natl. Acad. Sci. USA 2020 117 11753 11759 10.1073/pnas.1922287117 32398374 85. Pu Y. Ma L. Shan J. Wan X. Hammock B.D. Hashimoto K. Autism-like behaviors in male juvenile offspring after maternal glyphosate exposure Clin. Psychopharmacol. Neurosci. 2021 19 554 558 10.9758/cpn.2021.19.3.554 34294625 86. Hashimoto K. Hammock B.D. Reply to Reeves and Dunn: Risk for autism in offspring after maternal glyphosate exposure Proc. Natl. Acad. Sci. USA 2021 118 e2016496118 10.1073/pnas.2016496118 33443192 87. Ma M. Ren Q. Fujita Y. Ishima T. Zhang J.C. Hashimoto K. Effects of AS2586114, a soluble epoxide hydrolase inhibitor, on hyperlocomotion and prepulse inhibition deficits in mice after administration of phencyclidine Pharmacol. Biochem. Behav. 2013 110 98 103 10.1016/j.pbb.2013.06.005 23792539 88. Iyer M.R. Kundu B. Wood C.M. Soluble epoxide hydrolase inhibitors: An overview and patent review from the last decade Expert Opin. Ther. Pat. 2022 12 1 19 10.1080/13543776.2022.2054329 89. Hashimoto K. Understanding the link between maternal infections and neurodevelopmental disorders in offspring: The role of abnormalities in metabolism of polyunsaturated fatty acids Brain Behav. Immun. 2019 81 4 5 10.1016/j.bbi.2019.07.010 31394207 90. Kalia L.V. Lang A.E. Parkinson’s disease Lancet 2015 386 896 912 10.1016/S0140-6736(14)61393-3 25904081 91. Ascherio A. Schwarzschild M.A. The epidemiology of Parkinson’s disease: Risk factors and prevention Lancet Neurol. 2016 15 1257 1272 10.1016/S1474-4422(16)30230-7 27751556 92. Kodani S.D. Morisseau C. Role of epoxy-fatty acids and epoxide hydrolases in the pathology of neuro-inflammation Biochimie 2019 159 59 65 10.1016/j.biochi.2019.01.020 30716359 93. Chen J.J. Parkinson’s disease: Health-related quality of life, economic cost, and implications of early treatment Am. J. Manag. Care 2010 16 S87 S93 20297871 94. Zarriello S. Tuazon J.P. Corey S. Schimmel S. Rajani M. Gorsky A. Incontri D. Hammock B.D. Borlongan C.V. Humble beginnings with big goals: Small molecule soluble epoxide hydrolase inhibitors for treating CNS disorders Prog. Neurobiol. 2019 172 23 39 10.1016/j.pneurobio.2018.11.001 30447256 95. Borlongan C.V. Fatty acid chemical mediator provides insights into the pathology and treatment of Parkinson’s disease Proc. Natl. Acad. Sci. USA 2018 115 6322 6324 10.1073/pnas.1807276115 29848628 96. Dehay B. Bourdenx M. Gorry P. Przedborski S. Vila M. Hunot S. Singleton A. Olanow C.W. Merchant K.M. Bezard E. Targeting α-synuclein for treatment of Parkinson’s disease: Mechanistic and therapeutic considerations Lancet Neurol. 2015 14 855 866 10.1016/S1474-4422(15)00006-X 26050140 97. Bloem B.R. Okun M.S. Klein C. Parkinson’s disease Lancet 2021 397 2284 2303 10.1016/S0140-6736(21)00218-X 33848468 98. Hirsch E.C. Hunot S. Neuroinflammation in Parkinson’s disease: A target for neuroprotection? Lancet Neurol. 2009 8 382 397 10.1016/S1474-4422(09)70062-6 19296921 99. Hirsch E.C. Standaert D.G. Ten unresolved questions about neuroinflammation in Parkinson’s disease Mov. Disord. 2021 36 16 24 10.1002/mds.28075 32357266 100. Pallàs M. Vázquez S. Sanfeliu C. Galdeano C. Griñán-Ferré C. Soluble epoxide hydrolase inhibition to face neuroinflammation in Parkinson’s disease: A new therapeutic strategy Biomolecules 2020 10 703 10.3390/biom10050703 101. Wang Q. Liu Y. Zhou J. Neuroinflammation in Parkinson’s disease and its potential as therapeutic target Transl. Neurodegener. 2015 4 19 10.1186/s40035-015-0042-0 26464797 102. Liu L. Chen C. Gong W. Li Y. Edin M.L. Zeldin D.C. Wang D.W. Epoxyeicosatrienoic acids attenuate reactive oxygen species level, mitochondrial dysfunction, caspase activation, and apoptosis in carcinoma cells treated with arsenic trioxide J. Pharmacol. Exp. Ther. 2011 339 451 10.1124/jpet.111.180505 21846841 103. Yuan L. Liu J. Dong R. Zhu J. Tao C. Zheng R. Zhu S. 14,15-epoxyeicosatrienoic acid promotes production of brain derived neurotrophic factor from astrocytes and exerts neuroprotective effects during ischaemic injury Neuropathol. App. Neurobiol. 2016 42 607 620 10.1111/nan.12291 26526810 104. Kuo Y.M. Hsu P.C. Hung C.C. Hu Y.Y. Huang Y.J. Gan Y.L. Lin C.H. Shie F.S. Chang W.K. Kao L.S. Soluble epoxide hydrolase inhibition attenuates excitotoxicity involving 14,15-epoxyeicosatrienoic acid–mediated astrocytic survival and plasticity to preserve glutamate homeostasis Mol. Neurobiol. 2019 56 8451 8474 10.1007/s12035-019-01669-8 31257558 105. Sedelis M. Schwarting R.K.W. Huston J.P. Behavioral phenotyping of the MPTP mouse model of Parkinson’s disease Behav. Brain Res. 2001 125 109 125 10.1016/S0166-4328(01)00309-6 11682102 106. Jackson-Lewis V. Przedborski S. Protocol for the MPTP mouse model of Parkinson’s disease Nat. Protoc. 2007 2 141 151 10.1038/nprot.2006.342 17401348 107. Ren Q. Ma M. Yang J. Nonaka R. Yamaguchi A. Ishikawa K. Kobayashi K. Murayama S. Hwang S.H. Saiki S. Soluble epoxide hydrolase plays a key role in the pathogenesis of Parkinson’s disease Proc. Natl. Acad. Sci. USA 2018 115 E5815 E5823 10.1073/pnas.1802179115 29735655 108. Marsden C.D. Parkinson’s disease Lancet 1990 335 948 949 10.1016/0140-6736(90)91006-V 1691427 109. Qin X. Wu Q. Lin L. Sun A. Liu S. Li X. Cao X. Gao T. Luo P. Zhu X. Soluble epoxide hydrolase deficiency or inhibition attenuates MPTP-induced Parkinsonism Mol. Neurobiol. 2015 52 187 195 10.1007/s12035-014-8833-3 25128026 110. Huang H.J. Wang Y.T. Lin H.C. Lee Y.H. Lin A.M.Y. Soluble epoxide hydrolase inhibition attenuates MPTP-induced neurotoxicity in the nigrostriatal dopaminergic system: Involvement of α-synuclein aggregation and ER stress Mol. Neurobiol. 2018 55 138 144 10.1007/s12035-017-0726-9 28822080 111. Sun C.P. Zhou J.J. Yu Z.L. Huo X.K. Zhang J. Morisseau C. Hammock B.D. Ma X.C. Kurarinone alleviated Parkinson’s disease via stabilization of epoxyeicosatrienoic acids in animal model Proc. Natl. Acad. Sci. USA 2022 119 e2118818119 10.1073/pnas.2118818119 35217618 112. Lakkappa N. Krishnamurthy P.T. Pandareesh M.D. Hammock B.D. Hwang S.H. Soluble epoxide hydrolase inhibitor, APAU, protects dopaminergic neurons against rotenone induced neurotoxicity: Implications for Parkinson’s disease Neurotoxicology 2019 70 135 145 10.1016/j.neuro.2018.11.010 30472438 113. Lakkappa N. Krishnamurthy P.T. Yamjala K. Hwang S.H. Hammock B.D. Babu B. Evaluation of antiparkinson activity of PTUPB by measuring dopamine and its metabolites in Drosophila melanogaster: LC–MS/MS method development J. Pharm. Biomed. Anal. 2018 149 457 464 10.1016/j.jpba.2017.11.043 29169114 114. Virani S.S. Alonso A. Aparicio H.J. Benjamin E.J. Bittencourt M.S. Callaway C.W. Carson A.P. Chamberlain A.M. Cheng S. Delling F.N. Heart Disease and Stroke Statistics-2021 update: A report from the American Heart Association Circulation 2021 143 e254 e743 10.1161/CIR.0000000000000950 33501848 115. Chamorro Á. Meisel A. Planas A.M. Urra X. van de Beek D. Veltkamp R. The immunology of acute stroke Nat. Rev. Neurol. 2012 8 401 410 10.1038/nrneurol.2012.98 22664787 116. Emsley H.C.A. Smith C.J. Gavin C.M. Georgiou R.F. Vail A. Barberan E.M. Hallenbeck J.M. del Zoppo G.J. Rothwell N.J. Tyrrell P.J. An early and sustained peripheral inflammatory response in acute ischaemic stroke: Relationships with infection and atherosclerosis J. Neuroimmunol. 2003 139 93 101 10.1016/S0165-5728(03)00134-6 12799026 117. Adeoye O. Hornung R. Khatri P. Kleindorfer D. Recombinant tissue-type plasminogen activator use for ischemic stroke in the United States Stroke 2011 42 1952 1955 10.1161/STROKEAHA.110.612358 21636813 118. Nor A.M. Ford G.A. Misdiagnosis of stroke Expert Rev. Neurother. 2007 7 989 1001 10.1586/14737175.7.8.989 17678494 119. Iliff J.J. Alkayed N.J. Soluble epoxide hydrolase inhibition: Targeting multiple mechanisms of ischemic brain injury with a single agent Future Neurol. 2009 4 179 199 10.2217/14796708.4.2.179 19779591 120. Zhang W. Iliff J.J. Campbell C.J. Wang R.K. Hurn P.D. Alkayed N.J. Role of soluble epoxide hydrolase in the sex-specific vascular response to cerebral ischemia J. Cereb. Blood Flow Metab. 2009 29 1475 1481 10.1038/jcbfm.2009.65 19471280 121. Demirdöğen B.C. Miçooğulları Y. Özçelik A.T. Adalı O. Missense genetic polymorphisms of microsomal (EPHX1) and soluble epoxide hydrolase (EPHX2) and their relation to the risk of large artery atherosclerotic ischemic stroke in a Turkish population Neuropsychiatr. Dis. Treat. 2020 16 3251 10.2147/NDT.S233992 33994786 122. Yu D. Hennebelle M. Sahlas D.J. Ramirez J. Gao F. Masellis M. Cogo-Moreira H. Swartz R.H. Herrmann N. Chan P.C. Soluble epoxide hydrolase-derived linoleic acid oxylipins in serum are associated with periventricular white matter hyperintensities and vascular cognitive impairment Transl. Stroke Res. 2019 10 522 533 10.1007/s12975-018-0672-5 30443886 123. Matin N. Fisher C. Lansdell T.A. Hammock B.D. Yang J. Jackson W.F. Dorrance A.M. Soluble epoxide hydrolase inhibition improves cognitive function and parenchymal artery dilation in a hypertensive model of chronic cerebral hypoperfusion Microcirculation 2021 28 e12653 10.1111/micc.12653 32767848 124. Hao J. Chen Y. Yao E. Liu X. Soluble epoxide hydrolase inhibition alleviated cognitive impairments via NRG1/ErbB4 signaling after chronic cerebral hypoperfusion induced by bilateral carotid artery stenosis in mice Brain Res. 2018 1699 89 99 10.1016/j.brainres.2018.07.002 30343686 125. Tu R. Armstrong J. Lee K.S.S. Hammock B.D. Sapirstein A. Koehler R.C. Soluble epoxide hydrolase inhibition decreases reperfusion injury after focal cerebral ischemia Sci. Rep. 2018 8 5279 10.1038/s41598-018-23504-1 29588470 126. Chang L.H. Lin H.C. Huang S.S. Chen I.C. Chu K.W. Chih C.L. Liang Y.W. Lee Y.C. Chen Y.Y. Lee Y.H. Blockade of soluble epoxide hydrolase attenuates post-ischemic neuronal hyperexcitation and confers resilience against stroke with TrkB activation Sci. Rep. 2018 8 118 10.1038/s41598-017-18558-6 29311641 127. Taguchi N. Nakayama S. Tanaka M. Single administration of soluble epoxide hydrolase inhibitor suppresses neuroinflammation and improves neuronal damage after cardiac arrest in mice Neurosci. Res. 2016 111 56 63 10.1016/j.neures.2016.05.002 27184295 128. Wang J. Fujiyoshi T. Kosaka Y. Raybuck J.D. Lattal K.M. Ikeda M. Herson P.S. Koerner I.P. Inhibition of soluble epoxide hydrolase after cardiac arrest/cardiopulmonary resuscitation induces a neuroprotective phenotype in activated microglia and improves neuronal survival J. Cereb. Blood Flow Metab. 2013 33 1574 1581 10.1038/jcbfm.2013.111 23820647 129. Chen Y. Tian H. Yao E. Tian Y. Zhang H. Xu L. Yu Z. Fang Y. Wang W. Du P. Soluble epoxide hydrolase inhibition promotes white matter integrity and long-term functional recovery after chronic hypoperfusion in mice Sci. Rep. 2017 7 7758 10.1038/s41598-017-08227-z 28798352 130. Yeh C.F. Chuang T.Y. Hung Y.W. Lan M.Y. Tsai C.H. Huang H.X. Lin Y.Y. Soluble epoxide hydrolase inhibition enhances anti-inflammatory and antioxidative processes, modulates microglia polarization, and promotes recovery after ischemic stroke Neuropsychiatr. Dis. Treat. 2019 15 2927 10.2147/NDT.S210403 31686827 131. Yeh C.F. Chuang T.Y. Hung Y.W. Lan M.Y. Tsai C.H. Huang H.X. Lin Y.Y. Inhibition of soluble epoxide hydrolase regulates monocyte/macrophage polarization and improves neurological outcome in a rat model of ischemic stroke Neuroreport 2019 30 567 572 10.1097/WNR.0000000000001248 30950936 132. Davis C.M. Zhang W.H. Allen E.M. Bah T.M. Shangraw R.E. Alkayed N.J. Soluble epoxide hydrolase blockade after stroke onset protects normal but not diabetic mice Int. J. Mol. Sci. 2021 22 5419 10.3390/ijms22115419 34063817 133. Hasumi K. Suzuki E. Impact of SMTP targeting plasminogen and soluble epoxide hydrolase on thrombolysis, inflammation, and ischemic stroke Int. J. Mol. Sci. 2021 2 954 10.3390/ijms22020954 33477998 134. Suzuki E. Nishimura N. Yoshikawa T. Kunikiyo Y. Hasegawa K. Hasumi K. Efficacy of SMTP-7, a small-molecule anti-inflammatory thrombolytic, in embolic stroke in monkeys Pharmacol. Res. Perspect. 2018 6 e00448 10.1002/prp2.448 30546909 135. Zhang L. Xu S. Wu X. Muse F.M. Chen J. Cao Y. Yan J. Cheng Z. Yi X. Han Z. Protective effects of the soluble epoxide hydrolase inhibitor 1-trifluoromethoxyphenyl-3-(1-propionylpiperidin-4-yl) urea in a rat model of permanent middle cerebral artery occlusion Front. Pharmacol. 2020 11 182 10.3389/fphar.2020.00182 32184732 136. Yi X. Xu C. Huang P. Zhang L. Qing T. Li J. Wang C. Zeng T. Lu J. Han Z. 1-trifluoromethoxyphenyl-3-(1-propionylpiperidin-4-yl) urea protects the blood-brain barrier against ischemic injury by upregulating tight junction protein expression, mitigating apoptosis and inflammation in vivo and in vitro model Front. Pharmacol. 2020 11 1197 10.3389/fphar.2020.01197 32848796 137. Yao W. Cao Q. Luo S. He L. Yang C. Chen J. Qi Q. Hashimoto K. Zhang J.C. Microglial ERK-NRBP1-CREB-BDNF signaling in sustained antidepressant actions of (R)-ketamine Mol. Psychiatry 2021 10.1038/s41380-021-01377-7 138. Hammock B.D. McReynolds C.B. Wagner K. Buckpitt A. Cortes-Puch I. Croston G. Lee K.S.S. Yang J. Schmidt W.K. Hwang S.H. Movement to the clinic of soluble epoxide hydrolase inhibitor EC5026 as an analgesic for neuropathic pain and for use as a nonaddictive opioid alternative J. Med. Chem. 2021 64 1856 1872 10.1021/acs.jmedchem.0c01886 33550801 139. Luther J.M. Ray J. Wei D. Koethe J.R. Hannah L. DeMatteo A. Manning R. Terker A.S. Peng D. Nian H. GSK2256294 decreases sEH (soluble epoxide hydrolase) activity in plasma, muscle, and adipose and reduces F2-isoprostanes but does not alter insulin sensitivity in humans Hypertension 2021 78 1092 1102 10.1161/HYPERTENSIONAHA.121.17659 34455816 140. Martini R.P. Siler D. Cetas J. Alkayed N.J. Allen E. Treggiari M.M. A double-blind, randomized, placebo-controlled trial of soluble epoxide hydrolase inhibition in patients with aneurysmal subarachnoid hemorrhage Neurocrit. Care 2021 10.1007/s12028-021-01398-8

PMC009xxxxxx/PMC9099664.txt

==== Front Int J Mol Sci Int J Mol Sci ijms International Journal of Molecular Sciences 1422-0067 MDPI 10.3390/ijms23095239 ijms-23-05239 Article Salt-Templated Nanoarchitectonics of CoSe2-NC Nanosheets as an Efficient Bifunctional Oxygen Electrocatalyst for Water Splitting Cao Hong 1 https://orcid.org/0000-0001-9026-2175 Li Hailong 1* Liu Linhao 2 Xue Kangning 1 Niu Xinkai 1 Hou Juan 1* Chen Long 2 Ariga Katsuhiko Academic Editor Grasset Fabien Academic Editor Molard Yann Academic Editor 1 Key Laboratory of Ecophysics, Department of Physics, College of Science, Shihezi University, Shihezi 832003, China; c_hong1002@sina.com (H.C.); x-15937874475@sina.com (K.X.); niuxinkai0424@sina.com (X.N.) 2 Key Laboratory for Green Process of Chemical Engineering of Xinjiang Bingtuan, School of Chemistry and Chemical Engineering, Shihezi University, Shihezi 832003, China; thebestleo@sina.cn (L.L.); chenlong2012@sinano.ac.cn (L.C.) * Correspondence: well09131015@126.com (H.L.); hjuan05@sina.com (J.H.) 07 5 2022 5 2022 23 9 523907 4 2022 02 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Recently, the extensive research of efficient bifunctional electrocatalysts (oxygen evolution reaction (OER) and hydrogen evolution reaction (HER)) on water splitting has drawn increasing attention. Herein, a salt-template strategy is prepared to synthesize nitrogen-doped carbon nanosheets encapsulated with dispersed CoSe2 nanoparticles (CoSe2-NC NSs), while the thickness of CoSe2-NC NSs is only about 3.6 nm. Profiting from the ultrathin morphology, large surface area, and promising electrical conductivity, the CoSe2-NC NSs exhibited excellent electrocatalytic of 10 mA·cm−2 current density at small overpotentials of 247 mV for OER and 75 mV for HER. Not only does the nitrogen-doped carbon matrix effectively avoid self-aggregation of CoSe2 nanoparticles, but it also prevents the corrosion of CoSe2 from electrolytes and shows favorable durability after long-term stability tests. Furthermore, an overall water-splitting system delivers a current density of 10 mA·cm−2 at a voltage of 1.54 V with resultants being both the cathode and anode catalyst in alkaline solutions. This work provides a new way to synthesize efficient and nonprecious bifunctional electrocatalysts for water splitting. metal–organic frameworks nitrogen-doped carbon CoSe2 oxygen reduction reaction hydrogen evolution reaction National Natural Science Foundation of China51962032 Autonomous Region Graduate Student Innovation ProgramXJ2021G141 This work was supported by the National Natural Science Foundation of China (No. 51962032) and the Autonomous Region Graduate Student Innovation Program (XJ2021G141). ==== Body pmc1. Introduction Electrochemical water splitting has been an effective approach to generate sustainable and clean H2 energy [1,2,3,4]. To accelerate the production of hydrogen, numerous endeavors have been attempted to explore advanced and stable electrocatalysts that lower the overpotential of OER and HER [5,6,7]. Thus far, precious metal-based catalysts (such as Ir-based for OER and Pt-based for HER) are still the most effective oxygen electrocatalysts, but scarcity and high costs have largely hindered commercial applications. Hence, numerous research endeavors have been devoted to finding non-precious metal alternatives, including transition metal phosphates [8], nitrides [9], dichalcogenides [10,11], and borides [12,13,14], which showed exceptional activity and stability for OER and HER. Among them, cobalt selenide is a promising catalyst candidate. Its intrinsic metallic properties lead it to have higher electrical conductivity and more active edge sites [15,16,17]. For example, Lan et al. fabricated CoSe2 spheres via a facile hydrothermal process and the prepared samples exhibited an excellent electrochemical activity toward OER and HER [18]. In order to achieve higher electrochemical catalytic properties and stability for CoSe2, utilizing a nitrogen-doped carbon material as a supporter to enable CoSe2 particles is a feasible and widely used method. Metal organic frameworks (MOFs) have shown considerable advantages as precursors, and the well-organized framework consists of inorganic metal ions or clusters and N-containing organic ligands that can be converted in situ to nitrogen-doped carbon through pyrolysis [19,20,21,22]. Dong et al. found that pyrolysis and selenization of in situ grown zeolitic imidazolium framework-67 (ZIF-67) can homogeneously anchor CoSe2 nanoparticles (CoSe2/CF) to carbon fiber paper and the obtained CoSe2/CF shows excellent long-term stability and electrocatalytic properties [23]. However, metal sites in the MOF usually induce shrinkage agglomeration during high temperature calcination, which would disrupt the structure of the MOF and prevent exposure of the active site. Jiao et al. developed the SiO2 as templates to inhibit the Fe agglomeration during pyrolysis [24]. Notably, compared with the bulk counterparts, two-dimensional (2D) MOF nanosheets are being increasingly studied in electrocatalysis due to their highly exposed active sites [25], large surface area [26], and enhanced conductivity [27]. To date, synthetic methods of 2D MOF nanosheet preparation usually depends on physical exfoliation and chemical vapor deposition (CVD). Tang et al. prepared a series of ultrathin Ni/Co MOF nanosheets with unsaturated coordination metal active sites by a simple ultrasonic method, and demonstrated that ligand-unsaturated metal atoms are the main active centers of electrocatalytic OER. Nevertheless, how to avoid the aggregation and restacking of exfoliated nanosheets is still a challenge. Wurster et al. engineered heterobimetallic catalysts via CVD and the obtained nanosheets exhibited 300 mV overpotential and high turnover frequencies for OER [28]. Considering the low-yield of traditional methods, Huang et al. developed a salt-template confined method to prepare ultrathin ZIF-67 nanosheets [29]. The Co, N co-doped ultra-thin graphene nanosheets exhibited better electrocatalytic performance than commercial Pt/C catalysts. Thus, there is an urgent demand to develop a simple and cost-effective approach for designing 2D MOF derived efficient electrocatalysts with large surface area and fast mass transfer. Here, CoSe2-NC nanosheet electrocatalysts were prepared using NaCl as a template. Inorganic salt has excellent chemical stability and a smooth surface, which is suitable as a template to build 2D structures. During pyrolysis, the outer layer of ZIF-67 served as the nitrogen source and carbon source for the in situ synthesis of nitrogen-doped carbon. The resultant compounds showed a higher electrochemically active surface area (ECSA) and stability. It also displays excellent OER and HER activity. In addition, the water-splitting cell, based on CoSe2-NC bifunctional catalysts, shows good electrochemical performance, demonstrating that the catalysts with 2D MOF-derived nanosheet structures have great potential for practical applications. 2. Results and Discussion 2.1. Characterization of CoSe2-NC NSs Commercial NaCl powder was selected as template for the one-step synthesis of 2D CoSe2 nanosheets. A schematic diagram of the material fabrication process is shown in Figure 1. First, NaCl powder was mixed with cobalt (II) nitrate and 2-methylimidazole precursors. After vigorous grinding, the organic ligands of imidazole coordinate to Co2+ at room temperature (Figure S1). It should be noted that, in this approach, the excess of NaCl is used to avoid the aggregation of Co during selenization. The EDS images of the NaCl@MOF (Figure S2) show the uniform growth of ZIF-67 on the NaCl surface. SEM image of the 2D MOF which removed the salt template (Figure S3) evidenced that the salt template successfully synthesized ZIF-67 nanosheets. The diffraction peaks of NaCl@MOF and NaCl@CoSe2 (Figure S4) both match the NaCl crystallinity, and no other peaks were displayed, indicating that the formed ZIF-67 layer was relatively thin. Afterwards, the ZIF-67 shell and Se powder were calcined at 750 °C in a N2-protected tube furnace and converted to CoSe2-NC. To further expand the applicability of the synthetic method, amorphous nanosheets with a thickness of about 3.6 nm were effectively corroborated by XRD, Raman, and AFM. As shown in Figure 2, the nanosheets showed no obvious diffraction peaks, indicating that CoSe2-NC NSs are amorphous materials [30]. The graphite (002) peak in CoSe2-NC NSs implied graphitization. Compared with the 3D counterpart, the diffraction peaks of CoSe2-NC NPs were a perfect match with the simulated patterns (JCPDS, No. 09-0234), suggesting high crystallinity. As is known, the activity and number of exposed active sites directly affect the activity of electrocatalysts. Compared to its crystalline counterpart, the non-crystalline structure possesses more unsaturated coordination sites and effective active sites. Raman spectroscopy (Figure 2b) showed two peaks. The peak at 1355 cm−1 was due to the disordered sp3 carbon (D-band) and the peak at 1580 cm−1 indicated the existence of graphite sp2 carbon (G-band) [31]. Typically, the G-band corresponds to the lattice characteristics of graphite, while the D-band corresponds to the vibrational modes of carbon atoms at the edges of graphene [32,33]. The degree of carbon disorder is usually estimated by the value of ID/IG [34,35]. The calculated ratios of ID/IG were 1.63 for CoSe2-NC NSs and 1.26 for CoSe2-NC NPs. The results showed that the salt template prepared nanosheets with a higher ID/IG than the 3D structure, which indicated that the nanosheets had abundant defects and were considered catalytically active sites [36,37]. The AFM images (Figure 2c) were also used to evaluate the thickness of CoSe2-NC NSs. As revealed in the AFM images, Figure 2d suggests that CoSe2-NC NSs exhibited ultrathin nanosheets with a thickness of 3.6 nm. Ultrathin nanosheets can expose abundant catalytic active sites during the OER and HER multiphase reaction interface. To obtain more details of the structure, the morphologies of the CoSe2-NC NSs were observed using TEM. Figure 3a demonstrates that CoSe2-NC was in a sheet-like morphology. According to Figure S6, the dodecahedral shapes of the CoSe2-NC NPs could be well preserved, with dimensions of around 500 nm. Figure 3b shows that the CoSe2 particles were densely interconnected with the graphene layers. HRTEM showed that the interplanar distance of the lattice fringes was 0.258 nm, corresponding to the (111) plane of CoSe2, and 0.35 nm, corresponding to the (002) plane of graphite [38]. Moreover, the elemental mapping images (Figure 3d–h) verified that abundant C and N distributed throughout the entire sample. The results indicated the CoSe2 particles were encapsulated in nitrogen-doped carbon layers, which can induce greater stability during the corrosion of electrolytes. 2.2. Electronic States of CoSe2-NC NSs The elemental compositions of CoSe2-NC NSs were determined by XPS analysis. The XPS spectrum of Co 2p (Figure 4a) can be divided into Co 2p3/2 and Co 2p1/2, which were located at 780.7 and 796.6 eV, and the corresponding satellite peaks were at 786.5 eV and 803.2 eV, respectively. These results indicated the presence of Co2+ [39]. The measured binding energy of 778.1 eV, relative to the reported binding energy of metallic Co, indicated that the Co on the catalyst surface was oxidized by Se elements [40]. The measured binding energy of 778.1 eV, relative to the reported binding energy of metallic Co, indicated that Co on the catalyst surface was oxidized by Se elements [41]. Pyridinic N and pyrrolic N can both coordinate with Co, so the peaks at 781.8 and 797.8 eV could be assigned to Co-N structures. In the Se spectral region (Figure S10), the two main characteristic peaks of CoSe2-NC NSs were located at 54.9 and 55.8 eV, which correspond to the Se 3d5/2 and 3d3/2 orbitals of Se2−, respectively. In addition, the peak located at 60.1 eV indicated the presence of Se-O bonds on the surface. These relative peaks were from Co2+ coordinated to Se ions. The peaks of the C 1s spectrum (Figure 4b) at 284.3 eV, which could be assigned to sp2 hybridized carbons, and the peaks at 285.4 and 286.9 eV, due to the N-sp2 C and N-sp3, demonstrated the successful doping of N into carbon [42]. Notably, the C 1s spectrum of CoSe2-NC NPs (Figure 4e) could be only deconvoluted into sp2 C and N-sp2 C. The presence of the sp3 carbon atoms could disrupt the long-range order of the carbon network and were considered to be defective sites in the sp2 carbon matrix. The analysis results in Figure 4c show that the characteristic peaks were located at 398.4, 399.6, and 400.7 eV, corresponding to pyridine N, pyrrole N, and graphite N, further confirming the formation of nitrogen-doped graphitic carbon [43]. Figure 4f shows that the pyrrole N peak of cobalt selenide is prominent, indicating that the three-dimensional structure of CoSe2-NC NPs has more pyrrole N in the annealing process. However, the pyrrole N species are nitrogen atoms in a five-membered C-N heterocyclic structure, which are unstable due to their special structure. These results indicated that the CoSe2 particles were successfully encapsulated into N-doped carbon (NC) matrix. For increasing active sites, doping the carbon matrix with nitrogen heteroatoms is useful. In addition, the nitrogen formed a strong bond with the internal atoms, which resulted in a high stability of the composite [44]. 2.3. Electrochemical Performance of CoSe2-NC NSs The performance of the OER catalysts were tested in 1 M KOH electrolyte. Typically, the potential of the OER at a current density of 10 mA·cm−2 is defined as Ej=10. The overpotential is subtracted from the Ej=10 value of the catalyst by 1.23 V. Figure 5a shows the LSV curves for the OER electrocatalytic performance of CoSe2-NC NSs, CoSe2-NC NPs, Co-NC, and IrC using a three-electrode configuration. It is a remarkable that the CoSe2-NC NSs catalyst exhibited an advanced electrocatalytic activity with an overpotential of 246.7 mV, which is evidently better than that of other catalysts, and is even better than IrC (overpotential = 322.8 mV. The Tafel slope of CoSe2-NC NSs (72.66 mV·dec−1), CoSe2-NC NPs (91.97 mV·dec−1), Co-NC (78.03 mV·dec−1), and IrC (90.67 mV·dec−1) are displayed in Figure 5b, and suggests that the morphology of the nanosheet directly increases OER activity. Figure 5c shows the overpotential and Tafel slope data of CoSe2-NC NSs, CoSe2-NC NPs, Co-NC, and IrC, indicating that CoSe2-NC NSs have a far greater OER activity than the other samples. The synergistic effect of CoSe2 and nitrogen-doped carbon contributes to the outstanding OER activity of CoSe2-NC NSs. Meanwhile, the stability of the catalyst in electrolyte is a key parameter. As shown in Figure 5d, CoSe2-NC NSs maintained a stable voltage at current density of 10 mA·cm−2 after 132 h of chronopotentiometry testing. The outstanding stability mainly originates from the carbon layers protecting the CoSe2 nanoparticles. To assess the HER electrocatalytic activity, CoSe2-NC NSs, CoSe2-NC NPs, Co-NC, and PtC were explored in 1 M KOH electrolyte using a three-electrode cell. The CoSe2-NC NSs catalyst exhibited favorable electrocatalytic properties (Figure 6a). The overpotential of CoSe2-NC NSs was only 75.6 mV, which is much less than CoSe2-NC NPs (121.1 mV) and Co-NC (133.6 mV). The Tafel slope of 114.4 mV·dec−1 was measured for CoSe2-NC NSs, and was lower than CoSe2-NC NPs and Co-NC. As shown in Figure 6d, the ECSA of CoSe2-NC NSs (13.43 mF·cm−2) was much higher than CoSe2-NC NPs (10.48 mF·cm−2), Co-NC (12.08 mF·cm−2), and IrC (7.25 mF·cm−2). The results indicate that there were more active sites in CoSe2-NC NSs and that these abundant active sites were derived from the low-dimensional nanosheets, which have a large specific surface area (Figure S10). To further explore the HER kinetics, EIS measurements (Figure 6e) were taken in 1.0 M KOH. The smaller semicircle diameter of CoSe2-NC NSs (1.21 Ω) demonstrates a smaller charge-transfer resistance. This reveals that the electrochemical impedance of the CoSe2-NC NSs is much lower, which can effectively accelerate the charge transfer between the electrocatalyst and electrolyte interface. In view of the high performance of the prepared CoSe2-NC NS electrodes for both OER and HER, a two-electrode cell was set up in 1.0 M KOH using CoSe2-NC NSs as both the cathode and anode. Figure 7a shows the LSV curves of the CoSe2-NC NSs, which exhibited excellent overall water splitting activity. For the LSV measurement, the cell voltage of the CoSe2-NC NSs-based water splitting cell at 10 mA·cm−2 was only 1.54 V, even below that of the PtC||IrC (1.65 V). After 12 h of constant current testing, the catalyst voltage showed no obvious change, which indicated that the catalyst has good electrochemical activity and stability. The salt template promotes the formation of ultrathin nanosheets. The strong bonding of CoSe2 and the carbon layer ensures the immobilization of the active component, which is beneficial for improving the durability of the electrochemical process for overall water splitting. 3. Materials and Methods 3.1. Materials and Reagents Co(NO3)2·6H2O (99%), 2-methylimidazole (2-MeIm, 99%), and selenium powder (99.999%) were supplied by Sigma Aldrich (Missouri, USA). Methanol (99%), ethanol (99%), NaCl and KOH were bought from Chemical Reagent (Guangzhou, China). Nafion solution (5%) was purchased from Hesen (Shanghai, China). The ultrapure water (18 MΩ) used in the experiments was prepared using Hhitech equipment (Shanghai, China). Commercial catalysts (Pt/C, 20 wt%, Ir/C, 5 wt%) for comparison were bought from Macklin (Shanghai, China). All chemicals in the experiment were used directly without further purification. 3.2. Synthesis of Co-NC A total of 1.97 g 2-methylimidazole was dissolved in a mixed solvent with 20 mL of methanol and 20 mL of ethanol. Meanwhile, 0.87 g Co(NO3)2·6H2O was dissolved in another mixed solvent with 20 mL of methanol and 20 mL of ethanol. Then, the above two solutions were mixed under continuous stirring for 1 min and the final solution was kept at room temperature for 24 h. The resultant purple ZIF-67 precipitate was collected using centrifugation and washed several times with ethanol and ultrapure water and then dried in an oven at 60 °C for 12 h. 3.3. Synthesis of CoSe2-NC NPs The ZIF-67 particles and 0.1 g Se powder were dispersed in ceramic boats and the temperature in the furnace was raised to 750 °C at a rate of 2 °C·min−1. After that, the furnace was naturally cooled to room temperature. During the pyrolysis, the furnace was under a N2 atmosphere. To remove the free metal ions, the prepared black powder product was stirred in 0.5 M hydrochloric acid for 12 h. The samples were collected by centrifugation and washed repeatedly with deionized water and then dried at 80 °C. 3.4. Synthesis of CoSe2-NC NSs First, 0.363 g Co(NO3)2·6H2O, 0.411 g 2-methylimidazole and 4 g NaCl salt were mixed and ground in a mortar. After that, ZIF-67 coated on NaCl nanocrystals surface (denoted as NaCl@ZIF-67) was obtained. Then, the as-obtained product and 0.1 g Se powder were selenized under a N2 atmosphere at 750 °C for 2 h, 2 °C·min−1. After being cooled, the powders were washed with 0.5 M HCl solution and deionized water to remove the NaCl templates and impurities. 3.5. Material Characterization X-ray diffraction (XRD) was carried out using an X’ Pert PRO with a Cu Ka radiation diffractometer (k = 1.5418 Å). Raman spectra were detected using a Horiba JobinYvon, LabRAM HR800. Atomic force microscopy (AFM) images were determined using a Bruker Dimension ICON. The morphology and structures of the catalysts were measured by scanning electron microscopy (SEM, ZEISS Sigma 300) and an energy dispersive spectrometer (EDS). Transmission electron microscopy (TEM) was performed on a Tecnai G2 F30. The Brunauer–Emmet–Teller (BET) surface areas were measured at 77 K using a NOVA 2000 (Quantachrome, Boynton Beach, FL, USA). X-ray photoelectron spectroscopy (XPS) analyses were performed on a Thermo Scientific K-Alpha with Al Ka radiation. 3.6. Electrochemical Performance Prior to catalyst loading, nickel foam was acid washed to remove oxide impurities from the surface. The 1 cm × 1 cm nickel foam was immersed in 1 M HCl solution and sonicated for 15 min, and then sonicated in deionized water and anhydrous ethanol solution for 15 min. A homogeneous suspension of the catalyst was formed by dispersing 2 mg of catalyst in a mixture of 250 µL of deionized water, 700 µL of anhydrous ethanol, and 50 µL of Nafion with sonication for 1 h. Then, the suspension (100 μL) was dripped onto the pre-polished nickel foam and dried in a vacuum at 60 ℃. The mass loading of the active materials in this paper was, on average, 0.2 mg·cm−2. Electrocatalytic OER and HER measurements were tested at room temperature in a standard three-electrode setup, which was carried out on a CHI 760E electrochemical workstation (Chenhua, Shanghai, China). Nickel foam with electrocatalyst, graphite rod, and Ag/AgCl electrode filled with saturated KCl were selected as the working electrode, counter electrode, and reference electrode, respectively. The electrolyte used was KOH solution at 1.0 M (pH 13.6). To ensure the O2/H2O equilibrium at 1.23 V vs. RHE, all electrochemical experiments were performed in the O2 saturated condition. Linear sweep voltammetry (LSV) and cyclic voltammetry (CV) curves were measured at a scan rate of 5 mV·s−1. Tafel slopes were calculated according to Tafel equation: η = a + b log(j). CV was measured in the potential window of a non-Faraday process at different scan rates from 10 to 100 mV·s−1. The slope Cdl was obtained by fitting the current density versus scan rate as a linear relationship To further investigate electrocatalytic kinetics, electrochemical impedance spectroscopy (EIS) measurements were carried out in the frequency range 10 kHz to 10 mHz. The stability of the catalyst was determined by chronopotentiometry measurements at j = 10 mA·cm−2. 4. Conclusions In summary, CoSe2 nanoparticles embedded into nitrogen-doped carbon nanosheets were successfully synthesized using a salt-template strategy. The ultrathin nanosheets formed by the salt could effectively avoid self-aggregation of the CoSe2 particles, while the 2D structure could promote more efficient electron transfer between reactants and catalysts. In comparison with CoSe2-NC NPs and bare Co-NC NPs, the CoSe2-NC NSs exhibited remarkable OER and HER properties. In addition, the robust structure can maintain excellent stability during the reaction. As a result, this work provides a new strategy for the design of CoSe2-based bifunctional electrocatalysts with excellent catalytic performance and long-term stability. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ijms23095239/s1. Click here for additional data file. Author Contributions Writing, review and editing, H.C., methodology, H.L., formal analysis, L.L. and X.N., data curation, K.X., visualization and supervision, J.H., funding acquisition, L.C. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Schematic illustration of the overall synthetic procedure of CoSe2-NC NSs. Figure 2 (a) XRD pattern of CoSe2-NC NSs and CoSe2-NC NPs, (b) Raman spectra of CoSe2-NC NSs and CoSe2-NC NPs, (c,d) AFM image of CoSe2-NC NSs. Figure 3 (a,b) TEM images of the CoSe2-NC NSs, (c) HRTEM image of the CoSe2-NC NSs, (d–h) EDS elemental mapping images of CoSe2-NC NSs. Figure 4 (a–c) High-resolution XPS spectra of Co 2p, C 1s and N 1s, for the CoSe2-NC NSs, (d–f) High-resolution XPS spectra of Co 2p, C 1s, and N 1s for the CoSe2-NC NPs. Figure 5 (a) LSV curves of CoSe2-NC NSs, CoSe2-NC NPs, Co-NC, IrC and NF, (b) Tafel slopes of CoSe2-NC NSs, CoSe2-NC NPs, Co-NC and IrC, (c) Tafel slopes and overpotential for OER of CoSe2-NC NSs, CoSe2-NC NPs, Co-NC and IrC, (d) stability of CoSe2-NC NSs in chronopotentiometry curve. Figure 6 (a) LSV curves of CoSe2-NC NSs, CoSe2-NC NPs, Co-NC, PtC and NF. (b) Tafel slopes of CoSe2-NC NSs, CoSe2-NC NPs, Co-NC and PtC. (c) Tafel slopes and overpotential for HER of CoSe2-NC NSs, CoSe2-NC NPs, Co-NC and PtC. (d) Cdl curves of CoSe2-NC NSs, CoSe2-NC NPs, Co-NC, PtC and IrC. (e) EIS test of CoSe2-NC NSs, CoSe2-NC NPs, Co-NC, PtC and IrC. (f) LSV curves initial and after 24 h of continuous CV. Figure 7 (a) LSV curves of the CoSe2-NC NSs and PtC||IrC with a two-electrode system in alkaline electrolyte. (b) Stability of CoSe2-NC NSs in 1 M KOH at a current density of 10 mA·cm−2 over 12 h. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Wang Z. Li C. Domen K. Recent developments in heterogeneous photocatalysts for solar-driven overall water splitting Chem. Soc. Rev. 2019 48 2109 2125 10.1039/C8CS00542G 30328438 2. Li H. Chen L. Jin P. Li Y. Pang J. Hou J. Peng S. Wang G. Shi Y. NiCo2S4 microspheres grown on N, S co-doped reduced graphene oxide as an efficient bifunctional electrocatalyst for overall water splitting in alkaline and neutral pH Nano Res. 2022 15 950 958 10.1007/s12274-021-3580-z 3. Pang J.X. Fu H.H. Kong W.W. Jiang R. Ye J.H. Zhao Z.Y. Hou J. Sun K.S. Zheng Y. Chen L. Design of NiCo2O4 nanoparticles decorated N, S co-doped reduced graphene oxide composites for electrochemical simultaneous detection of trace multiple heavy metal ions and hydrogen evolution reaction Chem. Eng. J. 2022 433 133854 10.1016/j.cej.2021.133854 4. Fan C. Zhai X. Chen L. Peng S. Jiang R. Yu J. Li Y. Zhang Y. Kong W. Ge G. Synthesis and electrocatalytic mechanism of ultrafine MFe2O4 (M: Co, Ni, and Zn) nanocrystallites: M/Fe synergistic effects on the electrochemical detection of Cu(ii) and hydrogen evolution reaction performances J. Mater. Chem. A 2021 9 22277 22290 10.1039/D1TA05646H 5. Ye J.H. Zhai X.W. Chen L. Guo W. Gu T.T. Shi Y.L. Hou J. Han F. Liu Y. Fan C.C. Oxygen vacancies enriched nickel cobalt based nanoflower cathodes: Mechanism and application of the enhanced energy storage J. Energy Chem. 2021 62 252 261 10.1016/j.jechem.2021.03.030 6. You B. Sun Y. Innovative Strategies for Electrocatalytic Water Splitting Acc. Chem. Res. 2018 51 1571 1580 10.1021/acs.accounts.8b00002 29537825 7. Wang D. Chang Y.-X. Li Y.-R. Zhang S.L. Xu S.L. Well-dispersed NiCoS2 nanoparticles/rGO composite with a large specific surface area as an oxygen evolution reaction electrocatalyst Rare Metals 2021 40 3156 3165 10.1007/s12598-021-01733-0 8. He P. Yu X.Y. Lou X.W. Carbon-Incorporated Nickel-Cobalt Mixed Metal Phosphide Nanoboxes with Enhanced Electrocatalytic Activity for Oxygen Evolution Angew. Chem. 2017 56 3897 3900 10.1002/anie.201612635 28244625 9. Huang T. Chen Y. Lee J.-M. Two-Dimensional Cobalt/N-Doped Carbon Hybrid Structure Derived from Metal–Organic Frameworks as Efficient Electrocatalysts for Hydrogen Evolution ACS Sustain. Chem. Eng. 2017 5 5646 5650 10.1021/acssuschemeng.7b00598 10. Sheng H. Janes A.N. Ross R.D. Kaiman D. Huang J. Song B. Schmidt J.R. Jin S. Stable and selective electrosynthesis of hydrogen peroxide and the electro-Fenton process on CoSe2 polymorph catalysts Energy Environ. Sci. 2020 13 4189 4203 10.1039/D0EE01925A 11. Kong D. Wang H. Lu Z. Cui Y. CoSe2 nanoparticles grown on carbon fiber paper: An efficient and stable electrocatalyst for hydrogen evolution reaction J. Am. Chem. Soc. 2014 136 4897 4900 10.1021/ja501497n 24628572 12. Gupta S. Patel M.K. Miotello A. Nainesh P. Metal Boride-Based Catalysts for Electrochemical Water-Splitting: A Review Adv. Funct. Mater. 2019 30 1906481 1906509 10.1002/adfm.201906481 13. Guo F. Wu Y. Chen H. Liu Y. Yang L. Ai X. Zou X. High-performance oxygen evolution electrocatalysis by boronized metal sheets with self-functionalized surfaces Energy Environ. Sci. 2019 12 684 692 10.1039/C8EE03405B 14. Masa J. Weide P. Peeters D. Sinev L. Xia W. Sun Z. Somsen C. Muhler M. Schuhmann W. Amorphous Cobalt Boride (Co2B) as a Highly Efficient Nonprecious Catalyst for Electrochemical Water Splitting: Oxygen and Hydrogen Evolution Adv. Energy Mater. 2016 6 1502313 1502323 10.1002/aenm.201502313 15. Zhang L. Lu C. Ye F. Pang R. Liu Y. Wu Z. Shao Z. Sun Z. Hu L. Selenic Acid Etching Assisted Vacancy Engineering for Designing Highly Active Electrocatalysts toward the Oxygen Evolution Reaction Adv. Mater. 2021 33 e2007523 10.1002/adma.202007523 33656778 16. Li K. Zhang J. Wu R. Yu Y. Zhang B. Anchoring CoO Domains on CoSe2 Nanobelts as Bifunctional Electrocatalysts for Overall Water Splitting in Neutral Media Adv. Sci. 2016 3 1500426 10.1002/advs.201500426 17. Zhong W. Wang Z. Gao N. Huang L. Lin Z. Liu Y. Meng F. Deng J. Jin S. Zhang Q. Coupled Vacancy Pairs in Ni-Doped CoSe for Improved Electrocatalytic Hydrogen Production Through Topochemical Deintercalation Angew. Chem. 2020 59 22743 22748 10.1002/anie.202011378 32896011 18. Lan K. Li J. Zhu Y. Gong L. Li F. Jiang P. Niu F. Li R. Morphology engineering of CoSe2 as efficient electrocatalyst for water splitting J. Colloid Interface Sci. 2019 539 646 653 10.1016/j.jcis.2018.12.044 30616195 19. Duan J. Chen S. Zhao C. Ultrathin metal-organic framework array for efficient electrocatalytic water splitting Nat. Commun. 2017 8 15341 10.1038/ncomms15341 28580963 20. Xu Y. Li B. Zheng S. Wu P. Zhan J. Xue H. Xu Q. Pang H. Ultrathin two-dimensional cobalt–organic framework nanosheets for high-performance electrocatalytic oxygen evolution J. Mater. Chem. A 2018 6 22070 22076 10.1039/C8TA03128B 21. Li C. Zhao D.-H. Long H.L. Li M. Recent advances in carbonized non-noble metal–organic frameworks for electrochemical catalyst of oxygen reduction reaction Rare Met. 2021 40 2657 2689 10.1007/s12598-020-01694-w 22. Zhao S. Wang Y. Dong J. He C.T. Yin H. An P. Zhao K. Zhang X. Gao C. Zhang L. Ultrathin metal–organic framework nanosheets for electrocatalytic oxygen evolution Nat. Energy 2016 1 16184 10.1038/nenergy.2016.184 23. Sun C. Dong Q. Yang J. Dai Z. Lin J. Chen P. Huang W. Dong X. Metal–organic framework derived CoSe2 nanoparticles anchored on carbon fibers as bifunctional electrocatalysts for efficient overall water splitting Nano Res. 2016 9 2234 2243 10.1007/s12274-016-1110-1 24. Jiao L. Zhang R. Wan G. Yang W. Wan X. Zhou H. Shui J. Yu S.H. Jiang H.L. Nanocasting SiO2 into metal-organic frameworks imparts dual protection to high-loading Fe single-atom electrocatalysts Nat. Commun. 2020 11 2831 10.1038/s41467-020-16715-6 32504040 25. Zheng Y.R. Gao M.R. Gao Q. Li H.H. Xu J. Wu Z.Y. Yu S.H. An efficient CeO2/CoSe2 Nanobelt composite for electrochemical water oxidation Small 2015 11 182 188 10.1002/smll.201401423 25115699 26. Liu X. Liu Y. Fan L.Z. MOF-derived CoSe2 microspheres with hollow interiors as high-performance electrocatalysts for the enhanced oxygen evolution reaction J. Mater. Chem. A 2017 5 15310 15314 10.1039/C7TA04662F 27. Zhang Y. Zhang C. Guo Y. Liu D. Yu Y. Zhang B. Selenium vacancy-rich CoSe2 ultrathin nanomeshes with abundant active sites for electrocatalytic oxygen evolution J. Mater. Chem. A 2019 7 2536 2540 10.1039/C8TA11407B 28. Wurster B. Grumelli D. Hotger D. Gutzler R. Kern K. Driving the Oxygen Evolution Reaction by Nonlinear Cooperativity in Bimetallic Coordination Catalysts J. Am. Chem. Soc. 2016 138 3623 3626 10.1021/jacs.5b10484 26937997 29. Huang L. Zhang X. Han Y. Wang Q. Fang Y. Dong S. In situ synthesis of ultrathin metal–organic framework nanosheets: A new method for 2D metal-based nanoporous carbon electrocatalysts J. Mater. Chem. A 2017 5 18610 18617 10.1039/C7TA05821G 30. Liu C. Wang J. Wan J. Cheng Y. Huang R. Zhang C. Hu W. Wei G. Yu C. Amorphous Metal-Organic Framework-Dominated Nanocomposites with Both Compositional and Structural Heterogeneity for Oxygen Evolution Angew. Chem. 2020 59 3630 3637 10.1002/anie.201914587 31788950 31. Chang S.H. Danilovic N. Chang K.C. Subbaraman R. Paulikas A.P. Fong D.D. Highland M.J. Baldo P.M. Stamenkovic V.R. Freeland J.W. Functional links between stability and reactivity of strontium ruthenate single crystals during oxygen evolution Nat. Commun. 2014 5 4191 10.1038/ncomms5191 24939393 32. Ding J. Bu L. Guo S. Zhao Z. Zhu E. Huang Y. Huang X. Morphology and Phase Controlled Construction of Pt–Ni Nanostructures for Efficient Electrocatalysis Nano Lett. 2016 16 2762 2767 10.1021/acs.nanolett.6b00471 26950511 33. Wang T. Yang C. Liu Y. Yang M. Li X. He Y. Li H. Chen H. Lin Z. Dual-Shelled Multidoped Hollow Carbon Nanocages with Hierarchical Porosity for High-Performance Oxygen Reduction Reaction in Both Alkaline and Acidic Media Nano Lett. 2020 20 5639 5645 10.1021/acs.nanolett.0c00081 32639740 34. Gui Y. Liu X. Dou Y. Zhang L. Al-Mamun M. Jiang L. Yin H. He C.-T. Zhao H. Manipulating the assembled structure of atomically thin CoSe2 nanomaterials for enhanced water oxidation catalysis Nano Energy 2019 57 371 378 10.1016/j.nanoen.2018.12.063 35. Aboagye A. Elbohy H. Kelkar A.D. Qiao Q. Zai J. Qian X. Zhang L. Electrospun carbon nanofibers with surface-attached platinum nanoparticles as cost-effective and efficient counter electrode for dye-sensitized solar cells Nano Energy 2015 11 550 556 10.1016/j.nanoen.2014.10.033 36. Zhang H. Liu X. He G. Zhang X. Bao S. Hu W. Bioinspired synthesis of nitrogen/sulfur co-doped graphene as an efficient electrocatalyst for oxygen reduction reaction J. Power Sources 2015 279 252 258 10.1016/j.jpowsour.2015.01.016 37. Zhao D. Zhang S. Yin G. Du C. Wang Z. Wei J. Effect of Se in Co-based selenides towards oxygen reduction electrocatalytic activity J. Power Sources 2012 206 103 107 10.1016/j.jpowsour.2012.01.018 38. Wu X. Han S. He D. Yu C. Lei C. Liu W. Zheng G. Zhang X. Lei L. Metal Organic Framework Derived Fe-Doped CoSe2 Incorporated in Nitrogen-Doped Carbon Hybrid for Efficient Hydrogen Evolution ACS Sustain. Chem. Eng. 2018 6 8672 8678 10.1021/acssuschemeng.8b00968 39. Cai T. Zhao L. Hu H. Li T. Li X. Guo S. Li Y. Xue Q. Xing W. Yan Z. Stable CoSe2/carbon nanodice@reduced graphene oxide composites for high-performance rechargeable aluminum-ion batteries Energy Environ. Sci. 2018 11 2341 2347 10.1039/C8EE00822A 40. Gupta S. Patel N. Miotello A. Kothari D. Cobalt-Boride: An efficient and robust electrocatalyst for Hydrogen Evolution Reaction J. Power Sources 2015 279 620 625 10.1016/j.jpowsour.2015.01.009 41. Kim J. Kim H. Kim S.-K. Ahn S.-H. Electrodeposited amorphous Co-P-B ternary catalyst for hydrogen evolution reaction J. Mater. Chem. A 2018 6 6282 6288 10.1039/C7TA11033B 42. Zhou W. Lu J. Zhou K. Yang L. Ke Y. Tang Z. Chen S. CoSe2 nanoparticles embedded defective carbon nanotubes derived from MOFs as efficient electrocatalyst for hydrogen evolution reaction Nano Energy 2016 28 143 150 10.1016/j.nanoen.2016.08.040 43. Liu H. Jin M. Zhan D. Wang J. Cai X. Qiu Y. Lai L. Stacking faults triggered strain engineering of ZIF-67 derived Ni-Co bimetal phosphide for enhanced overall water splitting Appl. Catal. B Environ. 2020 272 118951 10.1016/j.apcatb.2020.118951 44. Xia B.Y. Yan Y. Li N. Wu H.B. Lou X.W. Wang X. A metal–organic framework-derived bifunctional oxygen electrocatalyst Nat. Energy 2016 1 15006 10.1038/nenergy.2015.6

PMC009xxxxxx/PMC9099665.txt

==== Front Int J Environ Res Public Health Int J Environ Res Public Health ijerph International Journal of Environmental Research and Public Health 1661-7827 1660-4601 MDPI 10.3390/ijerph19095019 ijerph-19-05019 Article Analysis of Factors Influencing Polish Students’ Opinions on Euthanasia https://orcid.org/0000-0002-7283-5491 Stokłosa Iga 1*† Stokłosa Maciej 1† https://orcid.org/0000-0003-2404-393X Więckiewicz Gniewko 1 Porwolik Mateusz 2 Bugajski Maciej 3 https://orcid.org/0000-0001-9516-0709 Masarczyk Wilhelm 1 Męcik-Kronenberg Tomasz 4 Piegza Magdalena 1 https://orcid.org/0000-0002-5748-0063 Pudlo Robert 1 Gorczyca Piotr 1 Bert Fabrizio Academic Editor 1 Department and Clinic of Psychiatry, Medical University of Silesia, 42-612 Tarnowskie Gory, Poland; maciej.piotr.stoklosa@gmail.com (M.S.); gniewkowieckiewicz@gmail.com (G.W.); wil.m@poczta.fm (W.M.); mpiegza@sum.edu.pl (M.P.); rpudlo@sum.edu.pl (R.P.); pgorczyca@sum.edu.pl (P.G.) 2 Department of Ophtalmology, Medical University of Silesia, 40-514 Katowice, Poland; pporwolik@gmail.com 3 National Research Institute of Oncology, State Research Institute, 31-115 Krakow, Poland; maciek.bugajski@gmail.com 4 Department of Pathomorphology, Medical Univeristy of Silesia, 41-800 Zabrze, Poland; patolog@interia.pl * Correspondence: iga.florczyk@gmail.com † These authors contributed equally to this work. 20 4 2022 5 2022 19 9 501913 3 2022 19 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Due to the continuous development of palliative care and methods of pain relief in the last moments of patients’ lives, we are faced with the question of how long we should take measures to delay inevitable death for, without providing prolonged therapy against the patient’s will. For terminally ill and people experiencing unimaginable suffering, euthanasia is considered as one of the possible options for a dignified farewell. The aim of the study was to determine the views on euthanasia in a group of students from Polish universities. Questionnaire responses were analyzed for 9686 students (79.9% of women and 20.1% of men) aged 18–35 years. Respondents were asked to complete their own questionnaire on demographic data and attitudes toward euthanasia in the case of severe terminal illness or paralysis preventing independent living. Euthanasia was significantly more acceptable among women (85% vs. 75%, p < 0.0001; 69% vs. 62%, p < 0.0001) and non-believers [98% vs. 97% (denominations other than Christian), 84% (other forms of Christianity), 69% (Roman Catholic); p < 0.0001] in every case considered. Religious affiliation was the factor that most influenced attitudes toward euthanasia. Among the other elements influencing attitudes toward euthanasia, residence and field of study were distinguished. Considering the large sample size and specific conclusions, the article should be an important argument in the discussion on euthanasia. death ethics euthanasia public health students ==== Body pmc1. Introduction One of the most ethically controversial issues discussed by the public in many countries, not only from a medical but also from a moral point of view, is euthanasia. By definition, euthanasia is a procedure that aims to end the life of a severely chronically ill person in a painless way when the person experiences unimaginable suffering [1]. This process is divided into three forms: active euthanasia, passive euthanasia, and assisted suicide. Passive euthanasia is the failure to continue to support the patient’s life, while active euthanasia is the taking of measures aimed at ending a person’s life under the influence of compassion for them and at their expressed request [2]. Assisted suicide is a process in which a physician facilitates a patient’s death by providing the patient with the necessary resources or information to enable the patient to commit the death act [3]. Euthanasia means providing the patient with means to commit suicide, while all activities aimed at taking one’s own life are carried out by the patient [4]. Currently (2022), euthanasia is legalized in Europe in Belgium, the Netherlands, and Luxembourg, and, according to recent reports, this procedure is recognized as legal in Spain [5,6]. The remaining countries in the world where euthanasia is available are Canada and Colombia [7]. Depending on the country, the reasons for euthanasia vary, but most often this procedure is possible in the case of a serious and incurable disease, both physical and psychological, which causes unimaginable suffering that cannot be alleviated [8]. People who support the possibility of euthanasia state as their main argument that they want to ensure that suffering and terminally ill people have the right to die with dignity. In addition, many patients in a vegetative state do not want to be a burden to their family members, and demand the “right to die”. Another argument of the proponents of this procedure is the possibility of using healthy organs from patients who choose euthanasia, and transferring them to people waiting for transplantation [8]. Opponents of euthanasia believe that this practice can eliminate terminally ill people from society, while the current development of palliative care allows suffering patients to relieve their pain and improve the comfort of their functioning. In addition, opponents of euthanasia mention as counterarguments the fear of possible abuse, as well as the violation and constant shifting of the ethical boundaries of the current criteria for approval of euthanasia [9,10]. Euthanasia remains an issue that, unlike other moral dilemmas such as abortion and in vitro fertilization (IVF), is forbidden in all major religious cultures of the world [11]. In recent years, there have been frequent discussions about the indications for the possibility of euthanasia in children suffering from terminal illnesses [12]. Now, 20 years after the legalization of euthanasia in the Benelux countries, there is also a growing percentage of people who choose to receive help to end their lives, and are struggling with severe mental disorders such as dementia, major depression, bipolar disorder, and conversion disorder [13,14]. Many mental disorders, including depression, remain controversial as a criterion for assisted suicide. In these cases, the ability to give informed consent by the patient remains a topic of discussion, especially when it is accompanied by the deterioration of cognitive function [15]. The same can be said about cognitive disorders and dementia in general. From the available data, it can be concluded that, in the Netherlands, euthanasia was reported as the cause of death in 2.9% of deaths (2010), while in Belgium, this figure was 4.6% (2013), but these figures are constantly increasing—4.4% in the Netherlands (2017), while in Belgium, the number of reported cases of euthanasia in 2021 increased by 10.39%, as compared to 2020 [16,17,18]. According to the Annual Report, the main cause of euthanasia in countries that have legalized this procedure is cancer, but also many comorbidities and neurological disorders [19]. In addition, it should be noted that reporting euthanasia as a cause of death is a complex and often problematic issue. Therefore, it has been suggested that the number of assisted deaths is sometimes underestimated and that there are actually more of them than the statistical data indicates [20]. The aim of this study was to determine attitudes toward euthanasia among students at Polish universities—people who are also studying in the field of medicine and may in the future come into direct contact with moral dilemmas related to the issue of patient life and death. In our study, we aimed to identify the factors that determine young adults’ attitudes toward euthanasia. In addition, we wanted to draw attention to a topic that has remained a frequently discussed aspect of ending human life in recent years in an ethical and moral context, around which many controversies and questions arise that often remain unanswered. It is important from the perspective of public health to understand the needs and expectations of Polish society regarding this difficult moral and medical dilemma. 2. Materials and Methods 2.1. Sample Characteristics A total of 9824 students aged 18–35 years from 40 Polish universities participated in the study. The criteria for inclusion in the study were student status and correct completion of the questionnaire. After a thorough analysis of the collected responses, the inclusion criteria were met by 9686 respondents, whose questionnaires were then subjected to statistical analysis. Among the students, two groups were distinguished—respondents studying in the medical field (medicine and paramedical studies), and students of other scientific fields. 2.2. Questionnaire The questionnaire included questions on sociodemographic data and attitudes toward the aspect of euthanasia (Appendix A). In terms of population data, students indicated their gender, age, field and year of study, and place of residence. A corresponding euthanasia questionnaire allowed respondents to select the “for” or “against” option in two cases involving the circumstances of euthanasia: (1) a serious, incurable illness that causes unimaginable suffering; (2) permanent paralysis of the body that substantially limits the ability to function independently. Respondents were asked whether the method of euthanasia should be legal in a particular case, whether they would choose euthanasia themselves, and also whether they would choose euthanasia for their relatives if they asked for it. 2.3. Distribution of the Questionnaire Students were asked to complete their own questionnaire in Polish, which was prepared by the authors of this paper, and then carefully analyzed by two independent experts in the field of public health and ethics. In order to ensure the greatest possible anonymity and convenience in filling out the questionnaire, and to reach the most diverse and large group of students, it was decided to distribute the survey via the Internet, using Google Forms, Google’s proprietary platform that enables the creation of anonymous questionnaires that are convenient for both creators and participants. A total of 40 Polish universities were randomly selected, and then, using social media, the researchers managed to reach online groups in which students from each university had joined, and in which, with the administrator’s consent, the form was distributed to be filled out and participants informed about the purpose of the study, and the appropriate consents to participate in the study were obtained. In each survey we sent, participants received information about the purpose of the study and were asked for consent to participate in it by selecting the appropriate option in the consent form available to them immediately before going to the appropriate questionnaire. The survey was available to participants from 10 October to 10 November 2018. Initially, the questionnaire was tested on a group of 30 students from the Medical University of Silesia in Katowice. Since the project was conducted online in a manner that prevented the identification of respondents, the study did not require approval from the Bioethics Committee. The study was exploratory in nature, and intended to determine students’ views on euthanasia. The currently described topic of euthanasia is one of the parts of the project on attitudes of Polish university students towards controversial ethical and health issues, for example euthanasia, abortion, and IVF. The first part of the study on attitudes towards abortion was published in the International Journal of Environmental Research and Public Health (IJERPH) in December 2021 [21]. 2.4. Statistical Analysis The collected data were analyzed using the STATISTICA 13.3 program (StatSoft, Krakow, Poland). The significance level was set at p < 0.01. For comparison of qualitative variables, the chi-square test was used, and for quantitative variables that did not conform to the normal distribution, the Mann–Whitney U test was used to compare two groups. For additional groups, the Kruskal–Wallis test was used. To evaluate the strength of correlation, the Spearman rank test was used. 3. Results The characteristics of the group and the distribution of responses by gender are shown in Table 1. A total of 9686 respondents (79.9% women and 20.1% men) participated in the study. The age of women in the study was significantly higher than that of men. Women are statistically significantly more supportive of the legality of euthanasia, more often explaining the possibility of euthanasia or the acceptance of euthanasia among their family members compared to male respondents. The distribution of responses as a function of medical school or other studies is shown in Table 2. Due to the possibility of direct contact with the topic of euthanasia during education and work, it was decided in the following study to divide the respondents into groups studying medicine and those studying outside the medical field. Medical students were significantly less likely than other majors to support the legality of euthanasia, and would be less likely to provide euthanasia in certain cases queried by the researchers. The distribution of responses as a function of reported religion is shown in Table 3. The study also analyzed the influence of religion on attitudes toward euthanasia. Faith is a statistically significant and strong factor determining respondents’ views on euthanasia. Opposition to euthanasia increases according to group affiliation in the following order: (1) non-believers, (2) believers in religions other than Christian, (3) believers in non-Roman Catholic faiths, (4) Roman Catholic. The distribution of results according to the size of the respondent’s city of origin is shown in Table 4. The acceptance of euthanasia increases with the size of the respondent’s city of origin. The distribution of responses as a function of the respondent’s field of study is shown in Table 5. Significant differences were found between fields of study in attitudes toward euthanasia; in every case analyzed, the differences were statistically significant. The most conducive of euthanasia were social sciences, arts, and natural sciences, while the opposite views were noticed in the field of education, law, and religious subjects. There was a discrepancy between the medical field and paramedical fields (such as emergency services, nursing, medical caregivers), where responses differed by an average of 10%, placing them at two poles of the list. In every case studied, age was an important factor influencing respondents’ views. Age was significantly higher in the groups that were in favor of each analyzed case of euthanasia than in the groups that had a different opinion. The average age of the respondents depending on the given answer in each analyzed case is shown in Table 6. 4. Discussion Euthanasia continues to be a complex ethical and legal issue. A wide-ranging discussion has consistently argued that legalizing the practice could be a dangerous entry into a downward spiral and the possibility of significant abuse [22]. A review of the scientific literature suggests that in countries that have allowed the possibility of euthanasia, as well as in countries that have decided to legalize so-called assisted suicide, many terminally ill people choose to avail themselves of the opportunity to end their lives with the participation of others [19]. To date, studies have been conducted in many countries to determine the attitudes of certain social groups toward assisted suicide [23,24]. Furthermore, in Poland, an attempt was made to characterize the views on the phenomenon of euthanasia among medical students from two Polish universities (588 respondents), but so far no similar study has been conducted in Poland among such a large group of people who are also not associated with the medical profession, which should be emphasized [25]. The factor determining the views on euthanasia among our respondents is gender. Women are significantly more likely to support the possibility of legalizing euthanasia, both in the case of a terminal, serious illness and in the case of paralysis that prevents independent living. Women are also more liberal about undertaking this procedure for themselves and their loved ones. This finding contrasts with data from the Kuwait study, in which men showed greater tolerance of euthanasia [26]. In a study conducted among health science students in Papua New Guinea, no significant differences in attitudes toward this phenomenon were found between the genders [27]. Age was found to be a factor significantly influencing attitudes toward the possibility of assisted termination of existence. As in other studies conducted worldwide, the tendency was observed that older students are more open to the phenomenon of euthanasia than younger people, who are more rigorous about legalization and possible implementation in the case of themselves or their relatives [28]. Respondents’ place of residence also has a significant influence on their respective attitudes toward euthanasia-both in terms of legalization and whether euthanasia is permitted among relatives or in their own case. Residents of larger cities show greater tolerance on these issues, which is consistent with the views of respondents over 50 years of age in Austria, where greater approval of euthanasia is observed among residents of urban areas [29]. Euthanasia is considered in all major religious cultures to interfere with the life of another human being, thereby compromising the sanctity of human life, and is a prohibited practice [30]. Among the respondents of this study, it has also been shown that students who belong to a particular religion denomination do not agree with the legalization of this practice. Students professing the Roman Catholic faith were most strongly opposed to euthanasia. Tolerance of the phenomenon of euthanasia grows accordingly among the group of students who report belonging to other groups of Christianity, and then to other religions. The greatest liberalism in the field of euthanasia is shown by people who identify themselves as non-believers. In a 2015 study, religion was also identified among medical students as one of the factors that strongly influence opposition to activities that may contribute to a faster death of patients, and Muslims in particular were opposed to euthanasia [31]. Similar trends were found in a study conducted in England and Wales, where Muslims and Catholics were strong opponents of euthanasia, while Protestants and non-believers showed more liberal attitudes toward issues related to death [32]. Looking at attitudes toward euthanasia among students in specific fields of study, medical school students were characterized by considerable reluctance to legalize euthanasia and implement it. Similar results were obtained in a study conducted in Germany, which demonstrated that both in 2004, and 12 years later in a 2016 study, only a small proportion of students supported the legalization of euthanasia and assisted suicide, possibly indicating that as future physicians they would not be able to perform such an operation on a patient [33]. Another study that examined the views of future physicians also found that medical students do not support euthanasia because it is contrary to the principles of medicine to care for patients by using the available options of medicine and palliative care and to provide them with a dignified death [34]. Our study concluded that the most cautious groups in terms of attitudes toward euthanasia were students from faculties related to cultural studies, religious studies, education, and law. Researchers in India came to different conclusions. There, it was found that among the specific professional groups in which attitudes toward euthanasia were studied, the proponents of this phenomenon were mainly judges [35]. This discrepancy can be explained by the influence of professional experience and age of judges on the views on euthanasia in the study cited above, considering that in our study we asked students, for example, young adults who had not yet practiced their profession, about their views. On the other hand, students in the humanities, natural sciences, and social sciences were significantly more in favor of euthanasia. This might be related to the fact that people who consider themselves artistically talented are characterized by a greater openness to new experiences and boundary crossing, and their attitudes are nonconformist to a greater extent [36]. 5. Limitations This study is exploratory in nature; therefore we could not raise any particular research questions, which could be considered a limit in itself. It was conducted to determine the views of young adults in Poland on the issue of euthanasia, which has become increasingly popular in global forums in recent years, but is still a controversial topic treated as an ethical dilemma. In our study, we included only undergraduate students in the questionnaire, therefore there is no complete evaluation of this phenomenon among non-educated people; through evaluation, this particular group could provide more interesting insights. In addition, the study used an original questionnaire which has not been validated, therefore it lacks psychometric evaluation; however, we found no specific questionnaire on euthanasia, and it seemed important to create a questionnaire regarding the practice that could be used in international research to learn about the views of people from different countries on this issue, and then to unify them and indicate the factors that determine them. However, the large number of respondents, which in itself is of great value, encouraged the authors to publish this manuscript. 6. Conclusions From the presented demographical data, it can be concluded that in every case, women show greater tolerance of euthanasia than men. Age influences respondents’ views on euthanasia—the older respondents are, the greater acceptance of euthanasia. The larger the city in which they live, the greater the favor of euthanasia respondents’ views are. The factor that strongly determines attitudes toward euthanasia is religion—the greatest opposition to euthanasia is expressed by the respondents declaring their affiliation to the Roman Catholic Church, and the smallest is observed among non-believers. Considering education—students of religious studies, pedagogy and law are characterized by more cautious views, unlike students of arts, science, and social sciences. Medical students showed greater opposition to euthanasia than paramedics. Author Contributions Conceptualization, M.S. and I.S.; methodology, M.S.; software M.S.; validation M.P. (Magdalena Piegza), T.M.-K. and G.W.; formal analysis M.S.; investigation M.P. (Mateusz Porwolik), M.B. and W.M; resources, M.S.; data curation, M.S.; writing—original draft I.S., preparation, I.S. and W.M.; writing—review and editing, G.W., M.P. (Magdalena Piegza) and R.P.; visualization M.S. and I.S.; supervision R.P. and P.G.; project administration, I.S. and M.S.; funding acquisition, G.W. and P.G. All authors have read and agreed to the published version of the manuscript. Funding This research received no external funding. Institutional Review Board Statement Ethical review and approval were waived for this study, due to the research being a web-based, anonymous study, and thus this study did not require one. Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement Data supporting reported results are available on request from the study team. Conflicts of Interest The authors declare no conflict of interest. Appendix A Gender (a) Male (b) Female Year of birth Branch of study (a) Agriculture (b) Another (c) Artistic (d) Economics and management (e) Education (f) Humanistic (g) Medical (h) Medical, non-medical (i) Military (j) Mining and metallurgy (k) Natural (l) Social Sciences (m) Sport (n) Strict sciences—non-technical (o) Technology University Year of study Place of residence (a) village (b) town of <20k inhabitants (c) town of 20–50k inhabitants (d) town of 50–100k inhabitants (e) town of 100–200k inhabitants (f) town of 200–500k inhabitants (g) town of >500k inhabitants What is your religion? Are you religiousy active? (a) Yes (b) No Do you think that euthanasia should be allowed in the following cases? (A) serious, incurable illness that causes unimaginable suffering; YES/NO (B) permanent paralysis of the body that substantially limits the ability to function independently; YES/NO Would you undergo euthanasia in the following cases? (A) serious, incurable illness that causes unimaginable suffering; YES/NO (B) permanent paralysis of the body that substantially limits the ability to function independently; YES/NO Would you allow a family member to be euthanized in the following cases? (A) serious, incurable illness that causes unimaginable suffering; YES/NO (B) permanent paralysis of the body that substantially limits the ability to function independently; YES/NO ijerph-19-05019-t001_Table 1 Table 1 Distribution of answers according to the gender of the respondents. All of the data are given in %. All Male Female χ 2/Z ** p Value Amount 9686 1947 7739 % 100 20 80 Mean Age 23.7 ± 3.6 23.0 ± 2.9 23.9 ± 3.8 8.1 ** <0.0001 ** Do you think that euthanasia should be allowed in the following cases? Yes: Severe disease 83 75 85 92.9 <0.0001 * Paralysis 67 62 69 28.9 <0.0001 * Would you undergo euthanasia in the following cases? Yes: Severe disease 73 62 76 118.9 <0.0001 * Paralysis 59 49 62 72.1 <0.0001 * Would you allow a family member to be euthanized in the following cases? Yes: Severe disease 72 62 74 71.2 <0.0001 * Paralysis 52 44 54 36.5 <0.0001 * * p Value for Chi-square test; ** p Value or Z Value for Mann–Whitney U test. ijerph-19-05019-t002_Table 2 Table 2 Distribution of responses depending on the study of medicine or other studies. All of the data are in %. All Non-Med Medical χ2 p Value Amount 9686 8582 1104 % 100 89 11 Do you think that euthanasia should be allowed in the following cases? Yes: Severe disease 83 84 75 41.7 <0.0001 * Paralysis 67 69 57 52.6 <0.0001 * Would you undergo euthanasia in the following cases? Yes: Severe disease 73 74 64 37.3 <0.0001 * Paralysis 59 60 49 36.1 <0.0001 * Would you allow a family member to be euthanized in the following cases? Yes Severe disease 72 72 Paralysis 52 52 * p Value for Chi-square test; Non-med—non medical. ijerph-19-05019-t003_Table 3 Table 3 Distribution of answers depending on the declared religion. Believers All Non-Believers Non-Christians Other Christians Roman Catholics χ2 p Value Amount 9686 3647 452 253 5334 % 100 37.6 4.7 2.6 55.1 Do you think that euthanasia should be allowed in the following cases? Yes: Severe disease 83 98 97 84 69 1217.9 <0.0001 * Paralysis 67 91 89 67 48 1607.9 <0.0001 * Would you undergo euthanasia in the following cases? Yes: Severe disease 73 95 92 68 55 1349.5 <0.0001 * Paralysis 59 86 81 59 40 1299.6 <0.0001 * Would you allow a family member to be euthanized in the following cases? Yes: Severe disease 72 96 91 69 52 1288.2 <0.0001 * Paralysis 52 84 77 53 30 1351.3 <0.0001 * * p Value for Chi-square test. ijerph-19-05019-t004_Table 4 Table 4 Distribution of answers depending on the size of the respondent’s city of origin. All Village <20k 20–50k 50–100k 100–200k 200–500k >500k χ2 p Value Amount 9686 1785 924 929 859 827 1347 3015 % 100 18 10 10 9 9 14 31 Do you think that euthanasia should be allowed in the following cases? Yes: Severe disease 83 75 81 82 81 85 84 87 97.9 <0.0001 * Paralysis 67 53 64 63 63 72 71 76 238.1 <0.0001 * Would you undergo euthanasia in the following cases? Yes: Severe disease 73 63 72 73 70 76 73 79 106.1 <0.0001 * Paralysis 59 48 55 59 56 61 61 67 121 <0.0001 * Would you allow a family member to be euthanized in the following cases? Yes: Severe disease 72 58 68 73 70 76 73 79 146.5 <0.0001 * Paralysis 52 35 44 50 48 57 58 63 216.8 <0.0001 * * p Value for Chi-square test. ijerph-19-05019-t005_Table 5 Table 5 Distribution of answers depending on the field of study of the respondent. Do You Think That Euthanasia Should Be Allowed in the Following Cases? Yes: Would You Undergo Euthanasia in the Following Cases? Yes: Would You Allow a Family Member to Be Euthanized in the Following Cases? Yes: Severe Disease Paralysis Severe Disease Paralysis Severe Disease Paralysis Social Sciences 94% 82% 87% 76% 84% 70% Artistic 89% 78% 79% 71% 78% 62% Natural 87% 72% 79% 66% 78% 60% Medical, non-MD 85% 65% 75% 59% 74% 56% Sport 85% 71% 80% 67% 76% 44% Humanistic 85% 74% 77% 65% 75% 58% Another 84% 69% 76% 62% 72% 53% Agriculture 84% 62% 65% 57% 73% 37% Military 84% 70% 82% 70% 80% 58% All 83% 67% 73% 59% 72% 52% Economics and management 82% 64% 72% 56% 68% 44% Mining and metallurgy 82% 62% 68% 56% 69% 41% University of Technology 80% 64% 67% 53% 67% 46% Strict sciences—non-technical 80% 62% 68% 53% 67% 45% Medical 75% 57% 64% 49% Education 72% 53% 61% 49% 55% 33% Law 67% 63% 62% 43% 56% 29% Religious 64% 63% 67% 65% 63% 61% χ2 180.2 208.6 162.2 159.4 132.8 179.6 p Value <0.0001 * <0.0001 * <0.0001 * <0.0001 * <0.0001 * <0.0001 * * p Value for Chi-square test; non-MD—non-medical doctor. ijerph-19-05019-t006_Table 6 Table 6 The average age of the respondents depending on the given answer in relation to the given cases. Mean Age With p Value Answers for Each Question Yes No Do you think that euthanasia should be allowed in the following cases? Yes: Severe disease 23.8 23.4 3.2 0.0012 ** Paralysis 24 23.2 8.6 <0.0001 ** Would you undergo euthanasia in the following cases? Yes: Severe disease 23.9 23.3 5.7 <0.0001 ** Paralysis 24 23.2 8.1 <0.0001 ** Would you allow a family member to be euthanized in the following cases? Yes: Severe disease 24.2 23.4 7.1 <0.0001 ** Paralysis 24.4 23.2 10.7 <0.0001 ** ** p Value for Mann–Whitney U Test. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Varkey B. Defining Euthanasia and the Need to be Circumspect in the Usage of the Term Med. Princ. Pract. 2020 29 499 500 10.1159/000508791 32645706 2. Śmigacz K. Concepts and types of euthanasia Proceedings of the 3rd Student Scientific Conference of the University of Law and Administration in Rzeszów Rzeszów, Poland 3 July 2015 3. American Medical Association Opinion 5.7 Physician-assisted suicide Code of Medical Ethics: American Medical Association Available online: https://www.ama-assn.org/delivering-care/physician-assisted-suicide (accessed on 9 April 2022) 4. Bliźniewska-Kowalska K.M. Talarowska M. Szewczyk A. Gałecki P. Assisted suicide: Ethical and legal problem Psychiatria 2019 16 44 49 5. Cohen J. Van Landeghem P. Carpentier N. Deliens L. Public acceptance of euthanasia in Europe: A survey study in 47 countries Int. J. Public Health 2013 59 143 156 10.1007/s00038-013-0461-6 23558505 6. Rada A.G. Spain will become the sixth country worldwide to allow euthanasia and assisted suicide BMJ 2021 372 n147 10.1136/bmj.n147 33452018 7. These People Want to Die. Will Their Countries Allow Euthanasia? Available online: https://www.nbcnews.com/news/latino/people-want-die-will-countries-allow-euthanasia-rcna3307 (accessed on 5 March 2022) 8. Euthanasia & Physician-Assisted Suicide (PAS) around the World Available online: https://euthanasia.procon.org/euthanasia-physician-assisted-suicide-pas-around-the-world/ (accessed on 5 March 2022) 9. Math S.B. Chaturvedi S.K. Euthanasia: Right to life vs right to die Indian J. Med. Res. 2012 136 899 902 23391785 10. Mushtaq A. The right to live. Or die? A Perspective on Voluntary Euthanasia Pak. J. Med. Sci. 1969 30 1159 1160 10.12669/pjms.305.5777 25225548 11. Skura-Madziała A. Religions of the World and their Positions on Euthanasia. Annales Ethics Econ. Life 2009 12 25 34 12. Brouwer M. Kaczor C. Battin M.P. Maeckelberghe E. Lantos J.D. Verhagen E. Should Pediatric Euthanasia be Legalized? Pediatrics 2018 141 e20171343 10.1542/peds.2017-1343 29317518 13. Appelbaum P.S. Should Mental Disorders Be a Basis for Physician-Assisted Death? Psychiatr. Serv. 2017 68 315 317 10.1176/appi.ps.201700013 28196462 14. Dierickx S. Deliens L. Cohen J. Chambaere K. Euthanasia for people with psychiatric disorders or dementia in Belgium: Analysis of officially reported cases BMC Psychiatry 2017 17 203 10.1186/s12888-017-1369-0 28641576 15. Pudlo R. Life not worth living Psychiatra 2016 2 14 15 16. Dom G. Stoop H. Haekens A. Sterckx S. Euthanasia and assisted suicide in the context of psychiatric disorders: Sharing experiences from the Low Countries Psychiatr. Polska 2020 54 661 672 10.12740/PP/124078 33386719 17. Groenewoud A.S. Atsma F. Arvin M. Westert G.P. Boer T.A. Euthanasia in the Netherlands: A claims data cross-sectional study of geographical variation BMJ Support. Palliat. Care 2021 10.1136/bmjspcare-2020-002573 33446488 18. Belgian Euthanasia Numbers of 2021 Released Available online: https://wfrtds.org/belgian-euthanasia-numbers-of-2021-released/ (accessed on 9 April 2022) 19. Annual Reports Available online: https://english.euthanasiecommissie.nl/the-committees/documents/publications/annual-reports/2002/annual-reports/annual-reports (accessed on 5 March 2022) 20. Cohen J. Dierickx S. Penders Y.W.H. Deliens L. Chambaere K. How accurately is euthanasia reported on death certificates in a country with legal euthanasia: A population-based study Eur. J. Epidemiol. 2018 33 689 693 10.1007/s10654-018-0397-5 29680997 21. Stokłosa M. Stokłosa I. Porwolik M. Bugajski M. Więckiewicz G. Męcik-Kronenberg T. Piegza M. Pudlo R. Gorczyca P. Mercy Procedure or a Nightmare? Attitude towards Pregnancy Termination among Polish University Students Int. J. Environ. Res. Public Health 2021 19 213 10.3390/ijerph19010213 35010470 22. Lewis P. The Empirical Slippery Slope from Voluntary to Non-Voluntary Euthanasia J. Law Med. Ethics 2007 35 197 210 10.1111/j.1748-720X.2007.00124.x 17341228 23. Crusat-Abelló E. Fernández-Ortega P. Conocimientos y actitudes de las enfermeras en torno a la eutanasia a nivel internacional y nacional: Revisión de la literatura Enfermería Clínica 2021 31 268 282 10.1016/j.enfcli.2021.01.004 24. Gutierrez-Castillo A. Gutierrez-Castillo J. Guadarrama-Conzuelo F. Jimenez-Ruiz A. Ruiz-Sandoval J.L. Euthanasia and physician-assisted suicide: A systematic review of medical students’ attitudes in the last 10 years J. Med. Ethics Hist. Med. 2020 13 22 10.18502/jmehm.v13i22.4864 33552455 25. Leppert W. Gottwald L. Majkowicz M. Kazmierczak-Lukaszewicz S. Forycka M. Ciałkowska-Rysz A. Kotlińska-Lemieszek A. A Comparison of Attitudes Toward Euthanasia Among Medical Students at Two Polish Universities J. Cancer Educ. 2012 28 384 391 10.1007/s13187-012-0414-4 26. Abohaimed S. Matar B. Al-Shimali H. Al-Thalji K. Al-Othman O. Zurba Y. Shah N. Attitudes of Physicians towards Different Types of Euthanasia in Kuwait Med. Princ. Pract. 2019 28 199 207 10.1159/000497377 30703772 27. Kolodziejczyk I. Kuzma J. Knowledge and Attitudes Towards Abortion and Euthanasia Among Health Students in Papua New Guinea Adv. Med. Educ. Pract. 2020 11 977 987 10.2147/AMEP.S281199 33363426 28. Karaahmetoglu G.U. Kutahyalioglu N.S. Attitudes Toward Euthanasia Among Turkish University Students Omega—J. Death Dying 2017 79 174 190 10.1177/0030222817729616 28914147 29. Stolz E. Mayerl H. Gasser-Steiner P. Freidl W. Attitudes towards assisted suicide and euthanasia among care-dependent older adults (50+) in Austria: The role of socio-demographics, religiosity, physical illness, psychological distress, and social isolation BMC Med. Ethics 2017 18 71 10.1186/s12910-017-0233-6 29212490 30. Sabriseilabi S. Williams J. Dimensions of religion and attitudes toward euthanasia Death Stud. 2020 46 1149 1156 10.1080/07481187.2020.1800863 32755292 31. Pomfret S. Mufti S. Seale C. Medical students and end-of-life decisions: The influence of religion Future Health J. 2018 5 25 29 10.7861/futurehosp.5-1-25 32. Francis L.J. McKenna U. Sahin A. Religion, Human Rights and Matters of Life and Death: Exploring Attitude Toward Abortion and Euthanasia Among Adolescents in England and Wales Religion, Education and Human Rights Springer Nature Cham, Switzerland 2018 139 159 10.1007/978-3-319-98773-6_6 33. Anneser J. Jox R.J. Thurn T. Borasio G.D. Physician-assisted suicide, euthanasia and palliative sedation: Attitudes and knowledge of medical students GMS J. Med. Educ. 2016 15 33 10.3205/zma001010 34. Rhee J.Y. Callaghan K.A. Allen P. Stahl A. Brown M.T. Tsoi A. McInerney G. Dumitru A.-M.G. A Medical Student Perspective on Physician-Assisted Suicide Chest 2017 152 475 477 10.1016/j.chest.2017.06.007 28629916 35. Singh S. Gupta S. Singh I. Madaan N. Awareness and attitude of select professionals toward euthanasia in Delhi, India Indian J. Palliat. Care 2016 22 485 490 10.4103/0973-1075.191856 27803572 36. Feist G.J. A Meta-Analysis of Personality in Scientific and Artistic Creativity Pers. Soc. Psychol. Rev. 1998 2 290 309 10.1207/s15327957pspr0204_5 15647135

PMC009xxxxxx/PMC9099666.txt

==== Front Animals (Basel) Animals (Basel) animals Animals : an Open Access Journal from MDPI 2076-2615 MDPI 10.3390/ani12091055 animals-12-01055 Article Morphology and Chemical Coding of Rat Duodenal Enteric Neurons following Prenatal Exposure to Fumonisins https://orcid.org/0000-0002-7790-3780 Kras Katarzyna 1 https://orcid.org/0000-0002-4152-560X Rudyk Halyna 2 https://orcid.org/0000-0002-4118-6302 Muszyński Siemowit 3 https://orcid.org/0000-0002-7574-2653 Tomaszewska Ewa 4 https://orcid.org/0000-0001-5873-9666 Dobrowolski Piotr 5 https://orcid.org/0000-0002-9374-2454 Kushnir Volodymyr 2 Muzyka Viktor 2 Brezvyn Oksana 2 https://orcid.org/0000-0002-3643-3524 Arciszewski Marcin B. 1* Kotsyumbas Ihor 2 Li Robert Academic Editor 1 Department of Animal Anatomy and Histology, University of Life Sciences in Lublin, 12 Akademicka St., 20-950 Lublin, Poland; katarzyna.kras@up.lublin.pl 2 State Scientific Research Control Institute of Veterinary Medicinal Products and Feed Additives, Donetska St. 11, 79000 Lviv, Ukraine; galusik.77@gmail.com (H.R.); wolodjak@gmail.com (V.K.); muzyka@scivp.lviv.ua (V.M.); brezvun@gmail.com (O.B.); dir@scivp.lviv.ua (I.K.) 3 Department of Biophysics, Faculty of Environmental Biology, University of Life Sciences in Lublin, 13 Akademicka St., 20-950 Lublin, Poland; siemowit.muszynski@up.lublin.pl 4 Department of Animal Physiology, Faculty of Veterinary Medicine, University of Life Sciences in Lublin, 12 Akademicka St., 20-950 Lublin, Poland; ewaRST@interia.pl 5 Department of Functional Anatomy and Cytobiology, Faculty of Biology and Biotechnology, Maria Curie-Sklodowska University, Akademicka St. 19, 20-033 Lublin, Poland; piotr.dobrowolski@mail.umcs.pl * Correspondence: mb.arciszewski@wp.pl 19 4 2022 5 2022 12 9 105501 4 2022 16 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Simple Summary Food contamination with toxins produced by various species of fungi (hereafter: mycotoxins) occurs frequently, and, for this reason, their activity in the bodies of animals is the subject of wide-ranging research. Although many aspects of the activity of fumonisins (mycotoxins produced by Fusarium) have been described, an open question is whether prenatal exposure to fumonisins may result in morphological changes, including in the enteric nervous system (ENS). The present study revealed that fumonisin administered via the gastrointestinal tract to pregnant rats did not substantially change the structure of the intestine/ENS of the offspring, but altered the neurochemical profile of its enteric neurons. Abstract Fumonisins (FBs), including fumonisin B1 and B2 produced by the fungus Fusarium verticillioides, are widespread mycotoxins contaminating crop plants as well as processed food. The aim of the experiment was to determine whether the exposure of 5-week-old pregnant rats to FBs at 60 mg/kg b.w. (group FB60) or 90 mg/kg b.w. (group FB90) results in morphological changes in the duodenum of weaned offspring, particularly the enteric nervous system (ENS). In addition, the levels of expression of galanin and vasoactive intestinal polypeptide (VIP) in the ENS were analysed by immunofluorescence in the control and experimental groups of animals. No significant morphological changes in the thickness of the muscle layer or submucosa of the duodenum were noted in group FB60 or FB90. In group FB90 (but not FB60), there was a significant increase in the width of the villi and in the density of the intestinal crypts. Immunofluorescence analysis using neuronal marker Hu C/D showed no significant changes in group FB60 or FB90 in the morphology of the duodenal ENS, i.e., the myenteric plexus (MP) and submucosal plexus (SP), in terms of the density of enteric ganglia in the MP and SP, surface area of MP and SP ganglia, length and width of MP and SP ganglia, surface area of myenteric and submucosal neurons, diameter of myenteric and submucosal neurons, density of myenteric and submucosal neurons, and number of myenteric and submucosal neurons per ganglion. In both groups, there was an increase (relative to the control) in the percentage of Hu C/D-IR/VIP-IR (IR-immunoreactive) and Hu C/D-IR/galanin-IR myenteric and submucosal neurons in the ganglia of both the MP and SP of the duodenum. In addition, in groups FB60 and FB90, there was an increase in the number of nerve fibres showing expression of VIP and galanin in the mucosa, submucosa and circular muscle layer of the duodenum. The results indicate that prenatal exposure to FBs does not significantly alter the histological structure of the duodenum (including the ENS) in the weaned offspring. The changes observed in the chemical code of the myenteric and submucosal neurons in both experimental groups suggest harmful activity of FBs, which may translate into activation of repair mechanisms via overexpression of neuroprotective neuropeptides (VIP and galanin). mycotoxins enteric nervous system neuropeptides small intestine rat ==== Body pmc1. Introduction In the 1970s, increased incidence of equine leukoencephalomalacia was observed in South Africa, which was eventually linked to the consumption of maize contaminated with fungi. The main fungus species isolated from contaminated maize was Fusarium verticillioides, which thereafter attracted the attention of researchers [1,2]. In subsequent years, a correlation was observed between the consumption of contaminated maize and increased occurrence of oesophageal cancer in humans [3], which prompted researchers to search for and isolate new substances with mycotoxin activity. In 1988, Gelderblom et al. extracted, purified and chemically tested two substances isolated from Fusarium verticillioides strain MRC 826 with carcinogenic potential, which were called fumonisin B1 (FB1) and fumonisin B2 (FB2) [4]. Subsequently, mass spectrometry and high-resolution nuclear magnetic resonance were used to describe the exact structure of the isolated FBs [5]; for review, see also [6]. FBs have been shown to be similar in their structure and activity to sphingolipids, and both substances act as ceramide synthase inhibitors and cause accumulation of bioactive intermediates of sphingolipid metabolism [7,8]. Given that sphingolipids are ubiquitous cell membrane components involved in a variety of cellular processes, such as signal transduction and cell survival [9,10], the potentially harmful effects of FBs in mammals are of particular importance. Moreover, it has recently been found that intoxication with FBs may result in induction of oxidative stress [11,12], autophagy activation [13], or even induction of epigenetic changes [14,15]. The numerous mechanisms of action of FBs may explain why, despite their very poor absorption following oral administration and thus limited accumulation in tissues [16], these mycotoxins cause substantial changes in many mammalian systems, particularly the digestive organs [8,17]. It should be noted that, due to its function, it is the gastrointestinal tract (especially the initial parts of the small intestine) that is most exposed to direct contact with mycotoxins. A number of functional changes have been observed in the small intestine of pigs and rats experimentally intoxicated with FBs, including villi flattening, atrophy and fusion, changes in the size of the enteric ganglia, segment-dependent reduction of goblet cell numbers, lymphatic infiltrates, and disturbance of the permeability of the intestinal barrier [18,19,20,21,22]. It has also been postulated that FB intoxication changes the proportions of gut microbiota [23]. While that study primarily focused on the effects of FB administration on postnatal development and in adults, it must be emphasized that FBs are classified as a potentially teratogenic factor [24]. In newborns which were potentially exposed to FB as foetuses via administration of the mycotoxin to the mother, various changes in the structure of the gastrointestinal tract were observed [25], as well as neural tube defects (NTDs) [26]. However, it is not entirely clear to what extent FB administered orally crosses the placental barrier of pregnant females and to what degree it is neurotoxic to developing foetuses. Therefore, the aim of the present study was to use immunohistochemical staining to investigate the potentially toxic effects of FBs on the structure of the small intestine (duodenum) and the corresponding enteric nervous system (ENS) of weaned rats whose mothers were treated orally with various doses of mycotoxin during pregnancy. Additionally, taking into account the presence of mechanisms of neuroplasticity in the ENS, we investigated whether the administration of FBs influenced expression levels of neuroprotective neuropeptides (galanin and vasoactive intestinal polypeptide, VIP) in the enteric neurons of the duodenum of weaned rats. 2. Materials and Methods 2.1. Fumonisins FB1 and FB2 were produced as previously described [27] from a culture of Fusarium verticillioides in a maize grain medium. Briefly, autoclaved coarsely cracked maize seeds were inoculated with F. verticillioides cultured in Petri dishes with tryptone glucose yeast extract broth (from the biobank at the Laboratory of Mycotoxicology, Institute of Veterinary Medicine of the National Academy of Agrarian Sciences of Ukraine, Kiev, Ukraine) and kept in a dark place (24 °C) for the next 4 weeks. Next, the incubation was stopped and the maize seeds were collected and autoclaved at 121 °C for 15 min. Then, the grains were dried at 80–90 °C for 120 min, ground, and stored at −20 °C. Liquid chromatography analysis showed that the FB1:FB2 ratio was 3:1 (73% to 27%). FBs were extracted from dried seeds with ethyl alcohol and analysed by ELISA. Finally, after determining the concentration, the extract was evaporated to obtain a concentration of 100 mg/mL. 2.2. Animals The study was conducted in compliance with the guidelines of the Declaration of Helsinki and approved by the Institutional Ethics Committee of the State Scientific Research Control Institute of Veterinary Medicinal Products and Feed Additives in Lviv, Ukraine (#132 676-Adm/08/2020, 28 February 2020). The experiment was conducted on pregnant 5-week-old Wistar rats (n = 18) receiving a mixture of FB1 and FB2 from the 7th day of pregnancy (to omit the period of organogenesis). Rats were housed individually in polypropylene cages (380 × 200 × 590 mm) and underwent a one-week acclimatization period to become accustomed to the laboratory conditions. They were kept at a temperature of 21 ± 3 °C, humidity of 55 ± 5%, with a 12 h/12 h day/night cycle, and had free access to water. After the acclimatization period, the pregnant rats were randomly divided into three groups: one control (n = 6) and two experimental (FB60 and FB90, n = 6 each). All animals were fed ad libitum with mycotoxin free standard diet for laboratory rodents. Pregnant rats from group FB60 were daily intoxicated by intragastric administration of FB1 + FB2 (3:1 ratio) at a dose of 60 mg/kg b.w., whereas animals from group FB90 (n = 6) received 90 mg/kg b.w. of FB1 + FB2 (3:1 ratio) in the same way. The FB mixture was administered in 0.5 mL of 0.9% saline solution until parturition. The administered dose of 90 mg FB1 + FB2/kg b.w. was equal to 1/10 of the established LD50 value, which was sufficient to induce subclinical intoxication [22], while the dose of 60 mg FB1 + FB2/kg b.w. was equal to 1/15 of the LD50 value. Rats from the control group (n = 6) received saline solution in a corresponding amount and manner. After birth, the offspring were culled to the same number of offspring in each litter and kept with their mothers until weaning at 28 days. One randomly selected weaned rat (28 days old) from each litter in every group fasted for 24 h and was euthanized by CO2 inhalation. 2.3. Tissue Processing In each control and experimental rat, the abdomen was opened with a midline incision and the duodenum was located. Starting 1.0 cm from the gastric pylorus, a duodenal fragment 2.0 cm in length was collected from each animal. This material was carefully washed in saline solution (37 °C) and transferred to a cold buffered formalin solution (4 °C, pH = 7.4) for 24 h. Then, the fixed duodenum was cut into two 1 cm parts (one for morphometric analysis and one for immunofluorescence analysis). The material for immunofluorescent (IF) staining was washed twice in 0.01 M phosphate buffered saline (PBS) for 10 min and then transferred to a transparent container filled with 16% sucrose solution (4 °C) supplemented with 0.01% bacteriostatic sodium azide (Avantor Performance Materials Poland S.A., Gliwice, Poland). The sucrose solution was replaced with fresh solution every day until the material fell to the bottom of the container. Subsequently, immediately after the final washing in PBS, the material was embedded in Tissue-Tek® O.C.T.™ Compound (Sakura Finetek USA, Inc., Torrance, CA, USA) and frozen in dry ice. Then, the frozen sections were cut to a thickness of 8 µm with a cryostat (HM 525 NX, Thermo Scientific, Waltham, MA, USA). Every fourth section was placed on adhesive glass slides (Superfrost Plus, Thermo Scientific, Waltham, MA, USA), and finally the slides were stored at −20 °C until further IF analysis. The remaining duodenal fragments, intended for morphometric analyses, were first dehydrated and cleared in Ottix Plus and Ottix Sharper (DiaPath, Martinengo, Italy) and then saturated with paraffin using a tissue processor (STP 120, Thermo Scientific, Waltham, MA, USA). Then, the processed material was embedded in paraffin blocks using the EC 350 automatic sample preparation system (Especialidades Médicas Myr S.L., Tarragona, Spain). Finally, the paraffin blocks were cut into 4 μm thick sections with a microtome (Microm HM 360, Microm, Walldorf, Germany). Every fourth section was placed on adhesive glass slides (Superfrost Plus, Thermo Scientific) and kept in an incubator (CG Wamed, Warsaw, Poland) at 37 °C for 12 h. 2.4. Immunofluorescence Selected slides were first warmed at room temperature (RT) for 30 min, and then the sections were outlined with a hydrophobic marker (ImmEdge™ Hydrophobic Barrier Pen, Vector Laboratories, Burlingame, CA, USA). Next, the sections were washed in 0.01 M PBS with 0.25% Triton X-100 (Sigma-Aldrich, Saint Louis, MO, USA) (3 × 10 min), covered with bovine serum albumin to block non-specific protein binding sites, and washed again. Then, solutions of primary antibodies (details of antisera are listed in Table 1) were dropped onto the slides and left for overnight incubation in a dark humid chamber (RT). Mouse anti Hu C/D antibodies (used as a pan-neuronal marker for visualization of enteric neurons) were combined with either rabbit antisera raised against VIP or against galanin. The next day, the excess antibody solution was removed, and the slides were again washed in PBS and incubated for 1 h with a combination of species-specific and fluorochrome-conjugated secondary antibodies. Following incubation, the slides were washed for the last time, mounted in phosphate-buffered glycerol (pH = 8.2), and finally coverslipped. Two different procedures were used to test the specificity of the antibodies. In the first step, considered a negative control, sections not exposed to primary antibodies (omitted or replaced with non-immunoreactive sera) were stained. The second procedure was a pre-absorption experiment in which primary antibodies were mixed with an excess of target synthetic protein before incubation. No positive immunoreaction was detected in any of the control sections. 2.5. Cell Counting, Imaging and Statistical Analysis Immunofluorescence-labelled sections were viewed with an epifluorescence microscope (BX61 Olympus, Nagano, Japan) equipped with interference filter cubes optimized for detection of Alexa Fluor 595 (MWIY2, excitation/emission wavelength 545–580 nm; Olympus, Tokyo, Japan) and Alexa Fluor 488/FITC (MNIBA2, excitation/emission wavelength 470–490 nm; Olympus). Images were captured using Cell^M 2.3 software (Olympus cellSens Standard) and a digital camera (C11440-36U, Hamamatsu Photonics, Shizuoka, Japan) connected to a standard PC. Images were taken under a 20× objective at a resolution of 1024 × 1024 pixels. Each stained section was examined in its entirety, and expression of biologically active substances was assessed by analysing all the ganglia seen on the cross-section of the duodenum of each animal (but no fewer than 100 myenteric and submucosal neurons per animal). Further morphometric analysis of the duodenal ENS structure was performed using ImageJ 1.52 software [28]. The following parameters were measured for myenteric and submucosal ganglia: neuron area, neuron diameter, ganglion area, ganglion length, ganglion width, number of ganglia per mm, number of neurons per mm, and number of neurons per ganglion. Next, the percentages of galanin-immunoreactive (galanin-IR) and VIP-immunoreactive (VIP-IR) perikarya were calculated and expressed as a percentage of the total numbers of Hu C/D-IR myenteric/submucosal neurons. The densities of galanin-IR and VIP-IR nerve fibres were also counted using ImageJ and a grid mask (each square of 1000 μm2). The grid mask was superimposed on the original images, and the number of nerve fibres per square was counted. The densities of galanin-IR and VIP-IR fibres were also estimated on a semi-quantitative scale: very numerous (>2.51 IR nerve fibres per 1000 μm2), numerous (2.1–2.5), moderate (1.51–2), few (0.01–1.5) and absent (0) [29]. 2.6. Morphometry Slides for morphometric analysis were dewaxed with xylene and rehydrated with descending grades of ethyl alcohol. Goldner’s trichrome staining was applied to differentiate the layers of duodenal wall [30]. The analyses were performed using ImageJ software [28]. The thickness of the muscular layer, mucosa and submucosa was measured using a straight line tool (30 measurements per parameter on each duodenum slice). The length and width of undamaged intestinal villi and the length and width of the crypts were measured using a segmented line tool (10 measurements per parameter on each duodenum slice). Additionally, the numbers of villi and crypts per mm were calculated. An experimental unit was a single rat. Measurements were averaged per animal prior statistical analysis 2.7. Statistical Analysis The data were analysed using Statistica 13.3 software (TIBCO Software Inc., 2017; Palo Alto, CA, USA). Parameters were compared between the control group and the experimental groups. One-way analysis of variance (ANOVA) was performed, followed by Tukey’s honestly significant difference (HSD) test. The normality of data distribution was tested using the Shapiro–Wilk test, and homogeneity of variance was tested with Levene’s test. Statistically significant differences were assumed for p < 0.05. Results are expressed as means ± standard deviation. 3. Results 3.1. ENS Morphology Immunostaining with Hu C/D antibodies enabled precise identification of enteric neurons (the presence of Hu C/D was confined to the neuronal soma, neuroplasm and nuclei), as neuronal perikarya stained using this neuronal marker were bright and clearly distinguishable against the dark background (see Figure 1 and Figure 2). Administration of FBs at 60 and 90 mg/kg b.w. during pregnancy did not significantly change either the mean numbers of myenteric ganglia per mm or the mean numbers of myenteric neurons per ganglion (see Table 2). No statistical changes (vs. control) in the geometry (mean area, length, and width) of the single myenteric ganglion were observed in either group FB60 or group FB90. The mean area and diameter of a single Hu C/D-IR myenteric neuron did not change after treatment during pregnancy with either 60 or 90 mg/kg b.w. Neither the mean numbers of submucosal ganglia nor the numbers of Hu C/D-IR submucosal neurons per ganglion changed significantly (vs. control) in group FB60 or FB90 (see Table 3). The administration of FBs during pregnancy at 60 mg/b.w. did not affect the area, length or width of the submucosal ganglia. Similarly, FBs given during pregnancy at 90 mg/b.w. did not cause any significant differences (vs. control) in the dimensions of submucosal ganglia. In animals from group FB60 and group FB90, there were no significant differences in the mean area or diameter of Hu C/D-IR submucosal neurons in comparison to the controls. 3.2. Morphometric Measurements of the Intestine The use of Goldner’s trichrome staining method enabled visualization of individual histological layers in the duodenal wall of the control and experimental animals (Figure 3). In animals from FB60 and FB90 groups, the mean thickness of the duodenal longitudinal smooth muscle layer and the circular muscle layer were not significantly changed when compared to controls (see Table 4). The mean thickness of the submucosal layer in animals from group FB60 and FB90 was not significantly different to that of the control animals. Similarly, administration of different doses of FBs during pregnancy had no effect on the mean thickness of the mucosal layer. In group FB60, the mean density of the mucosal villi and the mean villus height and width were not changed vs. controls. However, although in group FB90 the mean numbers of villi per mm of intestine remained unchanged to controls, the mean villus width was significantly greater. No statistical differences in the mean height of the villi of animals from control and FB90 group were found. Animals from group FB90 had a significantly higher density of duodenal crypts than the controls, but this effect was not observed in group FB60. There were no statistically significant differences between the crypt dimensions (mean depth and width) in the control, FB60 and FB90 groups. 3.3. Immunofluorescence In the control animals, immunofluorescent staining revealed a small subpopulation of Hu C/D-IR myenteric duodenal neurons expressing VIP (see Table 5). FBs administered during pregnancy at 60 and 90 mg/kg b.w. resulted in a significant increase in VIP in Hu C/D-IR myenteric neurons. The proportions of Hu C/D-IR/VIP-IR submucosal neurons were higher in both experimental groups and these values were statistically significant when compared to controls. VIP was also detected in nerve fibres supplying the intestinal layers of animals in the control and experimental groups. Administration of FBs caused a statistically significant increase in the numbers of VIP-IR nerve fibres in the circular muscle layer, submucosal layer and mucous layer in both groups (see Figure 1). In the control animals, the presence of galanin was noted in both myenteric and submucosal neurons. Prenatal administration of FBs at both doses significantly changed the subpopulation of Hu C/D-IR/galanin-IR myenteric neurons in both groups. Similarly, an increase in numbers of Hu C/D-IR/galanin-IR submucosal neurons was observed in groups FB60 and FB90. The administration of FBs at 60 and 90 mg/kg b.w. during pregnancy significantly changed the mean density of galanin-IR nerve fibres supplying the circular muscle layer, the submucosa and the mucosa (see Figure 2). 4. Discussion The results of the study clearly showed that FBs administered to pregnant rats (in two different doses) did not cause significant morphological or measurable changes in the layer structure of the duodenum of their offspring, but they did induce certain changes in the duodenal ENS. Previously published research reports suggest factors that may potentially influence the effect of FBs in the prenatal period. First, it is clear that previous studies have used various doses of FBs. Due to the ubiquity of FBs in food for people (cereals including rye, wheat, barley and rice) and animal feed (maize and maize-based products), many research centres have attempted to establish safe levels of FB intake that would not negatively affect animal health, but the results are often conflicting. As insufficiently decontaminated feed may contain many similarly acting mycotoxins at the same time, their synergistic and at times additive effects must be taken into account, which makes determination of maximum levels even more difficult [31]. According to FDA recommendations from 2001, safe levels of FBs (FB1 + FB2 + FB3) for animals depend on the species and purpose of the animal, as livestock animals, like horses, sheep or pigs, or laboratory animals like rodents are more sensitive to FB when compared to, for example, poultry, and range from 5 to 100 mg/kg of feed, with a median of 25 and an average of 37.5 mg/kg feed [32]. The levels of FBs used in our experiment (60 and 90 mg/kg b.w.) amount to 1/15 and 1/10 of the LD50 value for FBs, determined in a preliminary study as 900 mg/kg b.w. for rats. Importantly, both levels were higher than those inducing NTDs in the offspring of pregnant mice receiving FBs intraperitoneally from 7.5 and 8.5 days of gestation [33]. The existence of various mechanisms of action of FBs on developing foetuses has been postulated. One of these assumes inhibition of sphingolipid synthesis, which leads to the accumulation of intermediate metabolites, and this in turn can disturb signalling cascades involved in embryonic development, including proliferation, differentiation and migration of cells [34,35]. Therefore, it can also be presumed that, after FBs pass through the placenta of the pregnant female, the most likely effect should be disturbance of embryonic or foetal development. Inhibition of sphingolipid synthesis may also indirectly lead to disturbances in the supply of folic acid to the developing foetus [24], as well as to depletion of glycosphingolipid reserves, which in turn can disturb the expression and function of folate receptor 1, FOLR1 [36]. Folic acid is commonly known to play a key role in nervous system development, and thus blockage of its transfer to the foetus generally results in NTDs. This relationship between administration of FB and the development of NTDs has been confirmed in many studies, e.g., using rat embryos exposed to the toxin at levels from 0.2 to 217 ppm [37] and mouse embryos in a range from 0.71 to 71.2 ppm [38], but also in studies in which FB was administered to pregnant mice at levels from 2.5 to 100 mg/kg b.w. [8,33,39]. Increased incidence of NTDs has also been observed in human foetuses whose mothers’ diet consisted mainly of maize products contaminated with F. verticillioides [36]. Another factor considered by many researchers to be the main cause of disturbances in the development of foetuses potentially exposed to FBs in utero is maternal toxicity. Under that assumption, penetration of FBs through the placental barrier is not obvious. Several studies using pregnant rats, hamsters, mice and rabbits did not show NTDs, but other developmental disturbances were observed, involving a reduction in body weight [38,40,41,42], deformations [43] or increased foetal mortality [42,43], which is largely explained by the secondary effects of FBs via maternal toxicity. These studies also showed that sensitivity to various amounts of FB is species-dependent. In most cases, as in the present study, FB was administered beginning at 7 to 9 days of gestation, omitting the organogenesis period, but—what is important in light of the present study—the start of administration coincided with the migration of nerve stem cells colonizing the digestive tract and ultimately forming the ENS. This is because ENS development in rodents begins relatively late; cells derived from the neural crest do not begin to migrate towards the gastrointestinal tract until day 9.5 of embryonic development [44]. Normal ENS development depends on numerous factors and is regulated at several levels. The large number of processes regulating ENS development and thus the large number of signalling molecules enable potential interactions between these molecules and exogenous compounds to which foetuses may be exposed if the compounds penetrate the placental barrier [45]. In addition, nervous system cells largely owe their special structure and function to sphingolipids, which help to preserve the normal physiology of the neuron or oligodendrocyte, including their differentiation, polarization, and in the case of neurons, synapse formation and synaptic conduction [46]. To our surprise, however, in the present study, we did not observe significant changes in ENS structure in either the MP or the SP. Previous experiments have shown that the addition of FBs to the diet can significantly affect the morphology of ENS neurons. Specifically, administration of feed with the addition of FBs to rats for 42 days reduced the surface area of neurons in both the SP and the MP of the jejunum [47]. In another experiment, 21-day intragastric intoxication of 5-week-old rats with FBs significantly reduced the surface area of ENS ganglia in the duodenum (in this case only in the SP) and the jejunum (in the SP and MP) [22]. Other mycotoxins are known to have similar effects on ENS morphology as in the case of FBs. For example, deoxynivalenol (DON) administered orally to adult rats for 42 days reduced the area of muscle neurons and the area of MP ganglia in the jejunum [48]. While there are few reports of ENS damage caused by mycotoxins, there are numerous reports concerning the effects of these toxins on the structure of gastrointestinal tract organs and their consequences. A number of mycotoxins produced by Fusarium spp. (besides FBs, these include DON, T2 toxin and zearalenone (ZEN)), as well as others such as patulin, ochratoxin or aflatoxin, are usually absorbed directly through the mucosa of individual parts of the digestive tract and induce classical symptoms of food poisoning such as diarrhoea and vomiting [49,50]. Direct contact of the mucosa of the small intestine with the toxin may explain the changes observed in its histological structure, i.e., decreased villus length, disruption of the epithelial barrier, changes in the number of goblet cells, and the presence of local inflammation combined with intensified apoptosis [49,51]. In our study, however, the small intestinal wall exhibited none of the structural changes typically associated with mycotoxins. This may be explained in part by the route of administration of the toxin, as the animals were exposed to FBs via umbilical cord blood, bypassing the gastrointestinal tract. The changes we observed in the histological structure of the duodenum included an increase in the width of the villi and the density of crypts in group FB90, which suggests that the harmful effect of FBs administered in the prenatal period was partially dose-dependent. Although no morphological changes were observed in the ENS, there seems to have been a reaction to the harmful activity of FBs via neuroplasticity. Harmful factors (e.g., inflammation, axotomy or toxins) are known to cause neurons to produce and release certain neuroprotective substances, such as VIP and galanin [52,53,54,55]. It should be borne in mind that these neuropeptides also take part physiologically in the regulation of the motility and secretory activity of the intestines [56,57]. In addition, in pathological states, apart from the neuroprotective function mentioned above, VIP is involved in immunomodulation of the intestines, while galanin modulates inflammatory processes within the intestines [49,52,53]. Moreover, it has been suggested that increased expression of galanin may result from its role in tissue regeneration following inflammation or organ damage [52]. These observations are consistent with the results of the present study, as FB intoxication caused an increase in the percentages of both VIP-positive and galanin-positive neurons in both the MP and the SP of the duodenum, as well as an increase in the density of VIP-positive and galanin-positive nerve fibres in the circular muscle layer, submucosa, and mucosa of the duodenum. Similar results have been obtained in studies determining the effects of other mycotoxins on chemical coding of ENS neurons. Oral administration of T2 toxin and ZEN to immature pigs for 42 days resulted in an increase in the percentage of VIP-positive neurons in both the SP and MP and an increase in the density of VIP-positive nerve fibres in the circular muscle layer and mucosa of the small intestine [49]. Interestingly, immature pigs receiving ZEN orally for 42 days showed a significant decrease in the density of galanin-positive nerve fibres in the circular muscle layer of the ileum [58], which suggests that this toxin may primarily have impaired the peristalsis of the intestine. These last results are in contrast with those obtained in our study, but it should be noted that the route of administration of toxins was different in the two experiments (via the placenta vs. via the digestive tract). 5. Conclusions The results of the study indicated that prenatal exposure to FBs via administration to pregnant rats in the amount of 60 mg/kg b.w. and 90 mg/kg b.w. does not significantly affect the histological structure of the duodenum of weaned rats, although certain changes, i.e., a decrease in villus width and crypt density, were seen in the case of the higher dose. Prenatal exposure to FBs (at both doses) also had no significant effect on the morphology of the duodenal ENS; however, changes were noted in the chemical code of ENS neurons, which can be linked to the neuroprotective activity of galanin and VIP in response to the harmful activity of FB. Author Contributions Conceptualization, H.R., M.B.A. and E.T.; methodology, P.D. and K.K.; formal analysis, M.B.A. and K.K.; investigation, E.T., H.R., S.M., M.B.A., P.D., V.K., V.M., O.B., I.K. and K.K.; resources, I.K. and M.B.A.; data curation, E.T., H.R., S.M., M.B.A., P.D., V.K., V.M., O.B., I.K. and K.K.; writing—original draft preparation M.B.A. and K.K.; writing—review and editing, M.B.A. and K.K.; visualization, K.K.; supervision, M.B.A.; project administration, K.K.; funding acquisition, M.B.A. All authors have read and agreed to the published version of the manuscript. Funding This research was co-financed by the Polish National Agency for Academic Exchange (https://nawa.gov.pl/en/ (accessed on 14 April 2022)), Grant No. PPN/BUA/2019/1/00024/U/00001, and the Ministry of Education and Science of Ukraine through the project 0120U104053 (M/89-2020). Institutional Review Board Statement The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Ethics Committee of State Scientific Research Control Institute of Veterinary Medicinal Products and Feed Additives in Lviv, Ukraine (#132 676-Adm/08/2020, 28 February 2020). Informed Consent Statement The study was conducted in compliance with the guidelines of the Declaration of Helsinki and approved by the Institutional Ethics Committee of the State Scientific Research Control Institute of Veterinary Medicinal Products and Feed Additives in Lviv, Ukraine (#132 676-Adm/08/2020, 28 February 2020). Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 The immunoreactivity of Hu C/D (green) and VIP (red) in myenteric (MG) and submucosal (SG) ganglia of the rat duodenum: (A) control group; (B) FB60 group; (C) FB90 group; empty arrowheads indicate Hu C/D-IR/VIP-IR neurons, arrows indicate VIP-IR fibres. Figure 2 The immunoreactivity of Hu C/D (green) and galanin (red) in myenteric (MG) and submucosal (SG) ganglia of the rat duodenum: (A) control group; (B) FB60 group; (C) FB90 group; empty arrowheads indicate Hu C/D-IR/galanin-IR neurons, arrows indicate galanin-IR fibres. Figure 3 Representative pictures of villus (A–C); crypts (D–F) muscular and submucosal layer (G–I) in Goldner’s trichrome stained sections of rat duodenum from the control group (A,D,F), FB60 group (B,E,H) and FB90 group (C,F,I). Please note measurement schemes of the length (a long dashed yellow line) and the thickness (a short dashed yellow line) of the villus, the depth (a long green dashed line) and the width (a short green dashed line) of the crypts, the thickness of the submucosal layer (a continuous red line) and the thickness of the the circular muscle layer (a continuous blue line) and the thickness of the longitudinal muscle layer (a continuous black line). animals-12-01055-t001_Table 1 Table 1 Characteristics of primary and secondary antisera used for immunofluorescence experiments. Primary Antibody Host Dilution Code Source Anti-Human Neuronal Protein Hu C/Hu D (anti-Hu C/D) Mouse 1:400 A-21271 Molecular Probes, Eugene, OR, USA Anti-galanin Rabbit 1:1400 4600–5004 Biogenesis, Paterson, NJ, USA Anti-vasoactive intestinal peptide (VIP) Rabbit 1:1200 ab22736 Abcam, Cambridge, UK Anti-mouse Alexa Fluor 488 Donkey 1:800 A11029 Thermo Fisher Scientific, Waltham, MA, USA Anti-rabbit Alexa Fluor 594 Donkey 1:800 A21207 Thermo Fisher Scientific, Waltham, MA, USA animals-12-01055-t002_Table 2 Table 2 Effect of the fumonisins intoxication on the morphology of the rat duodenal myenteric plexus (mean ± SD). Parameter Group p-Value Control FB60 FB90 Neuron area (µm2) 122.83 ± 15.98 114.88 ± 12.93 96.17 ± 11.18 n.s. Neuron diameter (µm) 11.18 ± 1.16 11.21 ± 1.03 10.56 ± 0.98 n.s. Ganglion area (µm2) 864.21 ± 297.79 786.33 ± 169.95 726.76 ± 281.17 n.s. Ganglion length (µm) 92.70 ± 32.15 73.22 ± 17.80 76.68 ± 24.36 n.s. Ganglion width (µm) 13.18 ± 2.11 12.71 ± 1.26 12.03 ± 2.32 n.s. Ganglia (per 1 mm) 1.84 ± 0.84 1.59 ± 0.50 1.37 ± 0.64 n.s. Neurons per ganglion 5.10 ± 2.66 5.31 ± 2.46 4.97 ± 1.09 n.s. Statistical significance: n.s.—not significant. animals-12-01055-t003_Table 3 Table 3 Effect of the fumonisins intoxication on the morphology of the rat duodenal submucosal plexus (mean ± SD). Parameter Group p-Value Control FB60 FB90 Neuron area (µm2) 123.15 ± 44.50 117.86 ± 20.87 110.19 ± 15.53 n.s. Neuron diameter (µm) 11.56 ± 1.56 11.21 ± 1.03 11.41 ± 1.07 n.s. Ganglion area (µm2) 412.58 ± 78.73 370.70 ± 84.75 336.72 ± 57.80 n.s. Ganglion length (µm) 44.60 ± 12.86 39.72 ± 9.49 32.97 ± 7.17 n.s. Ganglion width (µm) 13.07 ± 3.47 11.02 ± 2.71 12.28 ± 1.81 n.s. Ganglia (per 1 mm) 1.41 ± 0.64 1.02 ± 0.68 0.71 ± 0.17 n.s. Neurons per ganglion 2.10 ± 0.62 1.89 ± 0.27 1.76 ± 0.45 n.s. Statistical significance: n.s.—not significant. animals-12-01055-t004_Table 4 Table 4 Effect of the fumonisin intoxication on the rat duodenum morphology (mean ± SD). Parameter Group p-Value Control FB60 FB90 Longitudinal muscle layer thickness (µm) 23.49 ± 4.41 21.06 ± 4.71 19.43 ± 4.46 n.s. Circular muscle layer thickness (µm) 52.18 ± 8.22 42.56 ± 6.38 40.77 ± 8.34 n.s. Submucosa thickness (µm) 27.04 ± 5.77 30.42 ± 11.88 27.59 ± 6.05 n.s. Mucosa thickness (µm) 819.32 ± 94.53 753.71 ± 41.71 825.51 ± 53.65 n.s. Villus height (µm) 589.70 ± 71.70 557.55 ± 57.70 623.26 ± 37.37 n.s. Villus width (µm) 75.90 ± 6.28 a 76.59 ± 3.56 a 100.85 ± 15.76 b <0.001 Density of villi (mm−1) 8.24 ± 1.41 6.39 ± 1.12 8.82 ± 1.38 n.s. Crypt depth (µm) 229.63 ± 38.71 196.16 ± 30.07 202.25 ± 31.13 n.s. Crypt width (µm) 36.23 ± 3.89 39.18 ± 3.28 31.82 ± 3.40 n.s. Density of crypts (mm−1) 22.99 ± 2.80 a 21.16 ± 2.35 a 26.79 ± 1.54 b 0.002 Statistical significance: n.s.—not significant; Different letters indicate significant differences between groups. animals-12-01055-t005_Table 5 Table 5 Effects of fumonisins intoxication on the chemical coding of the rat duodenal enteric neurons (mean ± SD). Parameter Group p-Value Control FB60 FB90 VIP-IR myenteric neurons 11.61 ± 1.61 a 15.14 ± 2.35 b 17.65 ± 4.10 b 0.008 VIP-IR submucosal neurons 24.84 ± 3.35 a 27.46 ± 2.55 b 30.23 ± 3.08 b 0.024 VIP-IR fibres Muscular layer 0.48 ± 0.07 a 0.58 ± 0.07 b 0.66 ± 0.06 b <0.001 Submucosa 0.23 ± 0.04 a 0.30 ± 0.06 b 0.32 ± 0.06 b 0.041 Mucosa 0.43 ± 0.04 a 0.77 ± 0.07 b 0.81 ± 0.12 b <0.001 galanin-IR myenteric neurons 21.05 ± 2.84 a 27.47 ± 4.85 b 30.02 ± 5.90 b 0.014 galanin-IR submucosal neurons 19.21 ± 3.35 a 25.97 ± 4.76 b 27.06 ± 4.74 b 0.014 galanin-IR fibres Muscular layer 0.83 ± 0.12 a 0.94 ± 0.12 b 1.11 ± 0.07 b <0.001 Submucosa 0.13 ± 0.04 a 0.23 ± 0.07 b 0.25 ± 0.05 b 0.004 Mucosa 0.12 ± 0.04 a 0.17 ± 0.05 b 0.19 ± 0.01 b 0.017 Statistical significance: n.s.—not significant; Different letters indicate significant differences between groups. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Kellerman T.S. Marasas W.F. Pienaar J.G. Naudé T.W. A mycotoxicosis of equidae caused by Fusarium moniliforme sheldon. A preliminary communication Onderstepoort J. Vet. Res. 1972 39 205 208 4664322 2. Marasas W.F. Kellerman T.S. Pienaar J.G. Naudé T.W. Leukoencephalomalacia: A mycotoxicosis of Equidae caused by Fusarium moniliforme Sheldon Onderstepoort J. Vet. Res. 1976 43 113 122 1012650 3. Marasas W.F.O. Wehner F.C. Van Rensburg S.J. Van Schalkwyk D.J. Mycoflora of corn produced in human esophageal cancer areas in Transkei, southern Africa Phytopathology 1981 8 792 796 10.1094/Phyto-71-792 4. Gelderblom W.C. Jaskiewicz K. Marasas W.F. Thiel P.G. Horak R.M. Vleggaar R. Kriek N.P. Fumonisins—Novel mycotoxins with cancer-promoting activity produced by Fusarium moniliforme Appl. Environ. Microbiol. 1988 54 1806 1811 10.1128/aem.54.7.1806-1811.1988 2901247 5. Bezuidenhout S.C. Gelderblom W.C. Gorst-Allman C.P. Horak R.M. Marasas W.F. Spiteller G. Vleggaar R. Structure elucidation of the fumonisins, mycotoxins from Fusarium moniliforme J. Chem. Soc. Chem. Commun. 1988 11 743 745 10.1039/c39880000743 6. ApSimon J.W. Structure, synthesis, and biosynthesis of fumonisin B1 and related compounds Environ. Health Perspect. 2001 109 245 249 10.1289/ehp.01109s2245 7. Marasas W.F.O. Riley R.T. Hendricks K.A. Stevens V.L. Sadler T.W. van Waes J.G. Missmer S.A. Cabrera J. Torres O. Gelderblom W.C.A. Fumonisins disrupt sphingolipid metabolism, folate transport, and neural tube development in embryo culture and in vivo: A potential risk factor for human neural tube defects among populations consuming fumonisin-contaminated maize J. Nutr. 2004 134 711 716 10.1093/jn/134.4.711 15051815 8. Voss K.A. Smith G.W. Haschek W.M. Fumonisins: Toxicokinetics, mechanism of action and toxicity Anim. Feed Sci. Technol. 2007 137 299 325 10.1016/j.anifeedsci.2007.06.007 9. Merrill A.H. Sullards M.C. Jr. Wang E. Voss K.A. Riley R.T. Sphingolipid metabolism: Roles in signal transduction and disruption by fumonisins Environ. Health Perspect. 2001 109 (Suppl. 2) 283 289 10. Hannun Y.A. Obeid L.M. Sphingolipids and their metabolism in physiology and disease Nat. Rev. Mol. Cell Biol. 2017 19 175 191 10.1038/nrm.2017.107 29165427 11. Wang X. Wu Q. Wan D. Liu Q. Chen D. Liu Z. Martínez-Larrañaga M.R. Martínez M.A. Anadon A. Yuan Z. Fumonisins: Oxidative stress-mediated toxicity and metabolism in vivo and in vitro Arch. Toxicol. 2016 90 81 101 10.1007/s00204-015-1604-8 26419546 12. Zhang W. Zhang S. Zhang M. Yang L. Cheng B. Li J. Shan A. Individual and combined effects of Fusarium toxins on apoptosis in PK15 cells and the protective role of N-acetylcysteine Food Chem. Toxicol. 2018 111 27 43 10.1016/j.fct.2017.10.057 29109043 13. Yin S. Guo X. Li J. Fan L. Hu H. Fumonisin B1 induces autophagic cell death via activation of ERN1-MAPK8/9/10 pathway in monkey kidney MARC-145 cells Arch. Toxicol. 2016 90 985 996 10.1007/s00204-015-1514-9 25925693 14. Demirel G. Alpertunga B. Ozden S. Role of fumonisin B1 on DNA methylation changes in rat kidney and liver cells Pharm. Biol. 2015 53 1302 1310 10.3109/13880209.2014.976714 25858139 15. Huang D. Cui L. Sajid A. Zainab F. Wu Q. Wang X. Yuan Z. The epigenetic mechanisms in Fusarium mycotoxins induced toxicities Food Chem. Toxicol. 2019 123 595 601 10.1016/j.fct.2018.10.059 30599843 16. Norred W.P. Plattner R.D. Chamberlain W.J. Distribution and excretion of [14C]fumonisin B1 in male Sprague-Dawley rats Nat. Toxins. 1993 1 341 346 10.1002/nt.2620010604 8167955 17. Chen J. Wei Z. Wang Y. Long M. Wu W. Kuca K. Fumonisin B1: Mechanisms of toxicity and biological detoxification progress in animals Food Chem. Toxicol. 2021 149 111977 10.1016/j.fct.2021.111977 33428988 18. Theumer M.G. López A.G. Masih D.T. Chulze S.N. Rubinstein H.R. Immunobiological effects of fumonisin B1 in experimental subchronic mycotoxicoses in rats Clin. Diagn. Lab. Immunol. 2002 9 149 155 10.1128/CDLI.9.1.149-155.2002 11777845 19. Bracarense A.P. Lucioli J. Grenier B. Drociunas Pacheco G. Moll W.D. Schatzmayr G. Oswald I.P. Chronic ingestion of deoxynivalenol and fumonisin, alone or in interaction, induces morphological and immunological changes in the intestine of piglets Br. J. Nutr. 2011 107 1776 1786 10.1017/S0007114511004946 21936967 20. Grenier B. Bracarense A.P. Schwartz H.E. Lucioli J. Cossalter A.M. Moll W.D. Schatzmayr G. Oswald I.P. Biotransformation approaches to alleviate the effects induced by Fusarium mycotoxins in swine J. Agric. Food Chem. 2013 61 6711 6719 10.1021/jf400213q 23758213 21. Basso K. Gomes F. Bracarense A.P. Deoxynivanelol and fumonisin, alone or in combination, induce changes on intestinal junction complexes and in E-cadherin expression Toxins 2013 5 2341 2352 10.3390/toxins5122341 24287571 22. Rudyk H. Tomaszewska E. Arciszewski M.B. Muszyński S. Tomczyk-Warunek A. Dobrowolski P. Donaldson J. Brezvyn O. Kotsyumbas I. Histomorphometrical changes in intestine structure and innervation following experimental fumonisins intoxication in male Wistar rats Pol. J. Vet. Sci. 2020 23 77 88 32233304 23. Oswald I.P. Desautels C. Laffitte J. Fournout S. Peres S.Y. Odin M. Le Bars P. Le Bars J. Fairbrother J.M. Mycotoxin fumonisin B1 increases intestinal colonization by pathogenic Escherichia coli in pigs Appl. Environ. Microbiol. 2003 69 5870 5874 10.1128/AEM.69.10.5870-5874.2003 14532038 24. Lumsangkul C. Chiang H.I. Lo N.W. Fan Y.K. Ju J.C. Developmental toxicity of mycotoxin fumonisin B₁ in animal embryogenesis: An overview Toxins 2019 11 114 10.3390/toxins11020114 30781891 25. Collins T.F. Shackelford M.E. Sprando R.L. Black T.N. Láborde J.B. Hansen D.K. Eppley R.M. Trucksess M.W. Howard P.C. Bryant M.A. Effects of fumonisin B1 in pregnant rats Food Chem. Toxicol. 1998 36 397 408 10.1016/S0278-6915(97)00170-1 9662415 26. Gelineau-van Waes J. Starr L. Maddox J. Aleman F. Voss K.A. Wilberding J. Riley R.T. Maternal fumonisin exposure and risk for neural tube defects: Mechanisms in an in vivo mouse model Birth Defects Res. A Clin. Mol. Teratol. 2005 73 487 497 10.1002/bdra.20148 15959874 27. Tomaszewska E. Rudyk H. Dobrowolski P. Donaldson J. Świetlicka I. Puzio I. Kamiński D. Wiącek D. Kushnir V. Brezvyn O. Changes in the Intestinal Histomorphometry, the Expression of Intestinal Tight Junction Proteins, and the Bone Structure and Liver of Pre-Laying Hens Following Oral Administration of Fumonisins for 21 Days Toxins 2021 13 375 10.3390/toxins13060375 34070555 28. Schneider C.A. Rasband W.S. Eliceiri K.W. NIH Image to ImageJ: 25 years of image analysis Nat. Methods 2012 9 671 675 10.1038/nmeth.2089 22930834 29. Księżarczyk M.M. Arciszewski M.B. Immunohistochemical study on the expressions of neuropeptides in the superficial and deep gland of the third eyelid of pigs Anat. Histol. Embryol. 2021 50 579 587 10.1111/ahe.12663 33559911 30. Bancroft J.D. Layton C. Connective and other mesenchymal tissues with their stains Bancroft s Theory and Practice of Histological Techniques 8th ed. Suvarna S.K. Layton C. Bancroft J.D. Elsevier Amsterdam, The Netherlands 2019 153 175 31. Alassane-Kpembi I. Kolf-Clauw M. Gauthier T. Abrami R. Abiola F.A. Oswald I.P. Puel O. New insights into mycotoxin mixtures: The toxicity of low doses of Type B trichothecenes on intestinal epithelial cells is synergistic Toxicol. Appl. Pharmacol. 2013 272 191 198 10.1016/j.taap.2013.05.023 23735874 32. US Food and Drug Administration Guidance for Industry: Fumonisin Levels in Human Foods and Animal Feeds United States Food and Drug Administration Washington, DC, USA 2001 33. Voss K.A. Gelineau-van Waes J.B. Riley R.T. Fumonisins: Current research trends in developmental toxicology Mycotoxin Res. 2006 22 61 69 10.1007/BF02954559 23605503 34. Riley R.T. Merrill A.H. Jr. Ceramide synthase inhibition by fumonisins: A perfect storm of perturbed sphingolipid metabolism, signaling, and disease J. Lipid Res. 2019 60 1183 1189 10.1194/jlr.S093815 31048407 35. Chen Y. Liu Y. Sullards M.C. Merrill A.H. Jr. An introduction to sphingolipid metabolism and analysis by new technologies Neuromolecular Med. 2010 12 306 319 10.1007/s12017-010-8132-8 20680704 36. Gelineau-van Waes J. Voss K.A. Stevens V.L. Speer M.C. Riley R.T. Maternal fumonisin exposure as a risk factor for neural tube defects Adv. Food Nutr 2009 56 145 181 37. Flynn T.J. Stack M.E. Troy A.L. Chirtel S.J. Assessment of the embryotoxic potential of the total hydrolysis product of fumonisin BI using organogenesis-staged at embryos Food Chem. Toxicol. 1997 35 1135 1141 10.1016/S0278-6915(97)85466-X 9449218 38. Sadler T.W. Merrill A.H. Stevens V.L. Sullards M.C. Wang E. Wang P. Prevention of fumonisin B1-induced neural tube defects by folic acid Teratology 2002 66 169 176 10.1002/tera.10089 12353213 39. Liao Y.J. Yang J.R. Chen S.E. Wu S.J. Huang S.Y. Lin J.J. Chen L.R. Tang P.C. Inhibition of fumonisin B1 cytotoxicity by nanosilicate platelets during mouse embryo development PLoS ONE 2014 9 e112290 10.1371/journal.pone.0112290 25383881 40. Penner J.D. Casteel S.W. Pittman L. Jr. Rottinghaus G.E. Wyatt R.D. Developmental toxicity of purified fumonisin B1 in pregnant Syrian hamsters J. Appl. Toxicol. 1998 18 197 203 10.1002/(SICI)1099-1263(199805/06)18:3<197::AID-JAT497>3.0.CO;2-R 9685049 41. LaBorde J.B. Terry K.K. Howard P.C. Chen J.J. Collins T.F. Shackelford M.E. Hansen D.K. Lack of embryotoxicity of fumonisin B1 in New Zealand white rabbits Fundam. Appl. Toxicol. 1997 40 120 128 10.1006/faat.1997.2380 9398494 42. Reddy R.V. Johnson G. Rottinghaus G.E. Casteel S.W. Reddy C.S. Developmental effects of Fumonisin B1 in mice Mycopathologia 1996 134 161 166 10.1007/BF00436724 8981781 43. Floss J.L. Casteel S.W. Johnson G.C. Rottinghaus G.E. Krause G.F. Developmental toxicity of Fumonisinin Syrian hamsters Mycopathologia 1994 128 33 38 10.1007/BF01104276 7708090 44. Lake J.I. Heuckeroth R.O. Enteric nervous system development: Migration, differentiation, and disease Am. J. Physiol.—Gastrointest. Liver Physiol. 2013 305 G1 G24 10.1152/ajpgi.00452.2012 23639815 45. Nagy N. Goldstein A.M. Enteric nervous system development: A crest cell s journey from neural tube to colon Semin. Cell Dev. Biol. 2017 66 94 106 10.1016/j.semcdb.2017.01.006 28087321 46. Olsen A.S.B. Færgeman N.J. Sphingolipids: Membrane microdomains in brain development, function and neurological diseases Open Biol. 2017 7 170069 10.1098/rsob.170069 28566300 47. Sousa F.C. Schamber C.R. Amorin S.S. Natali M.R. Effect of fumonisin-containing diet on the myenteric plexus of the jejunum in rats Auton. Neurosci. 2014 185 93 99 10.1016/j.autneu.2014.08.001 25183308 48. Rissato D.F. de Santi Rampazzo A.P. Borges S.C. Sousa F.C. Busso C. Buttow N.C. Natali M. Chronic ingestion of deoxynivalenol-contaminated diet dose-dependently decreases the area of myenteric neurons and gliocytes of rats Neurogastroenterol. Motil. 2020 32 e13770 10.1111/nmo.13770 31793155 49. Gonkowski S. Gajęcka M. Makowska K. Mycotoxins and the Enteric Nervous System Toxins 2020 12 461 10.3390/toxins12070461 50. Payros D. Alassane-Kpembi I. Pierron A. Loiseau N. Pinton P. Oswald I.P. Toxicology of deoxynivalenol and its acetylated and modified forms Arch. Toxicol. 2016 90 2931 2957 10.1007/s00204-016-1826-4 27663890 51. Broom L. Mycotoxins and the intestine Anim. Nutr. 2015 1 262 265 10.1016/j.aninu.2015.11.001 29766989 52. Brzozowska M. Całka J. Review: Occurrence and Distribution of Galanin in the Physiological and Inflammatory States in the Mammalian Gastrointestinal Tract Front. Immunol. 2021 11 602070 10.3389/fimmu.2020.602070 33552060 53. Arciszewski M.B. Ekblad E. Effects of vasoactive intestinal peptide and galanin on survival of cultured porcine myenteric neurons Regul. Pept. 2005 125 185 192 10.1016/j.regpep.2004.08.036 15582731 54. Rytel L. Lozano M.A. Gonkowski S. Neuregulin 1 (NRG-1) as a neuronal active substance in the porcine intrahepatic nerve fibers in physiological conditions and under the influence of bisphenol A (BPA) Ann. Anim. Sci. 2020 20 1339 1350 10.2478/aoas-2020-0038 55. Bulc M. Całka J. Zielonka Ł. Dąbrowski M. Palus K. Effect of chemically-induced diabetes mellitus on phenotypic variability of the enteric neurons in the descending colon in the pig Ann. Anim. Sci. 2021 21 1403 1422 10.2478/aoas-2020-0121 56. Iwasaki M. Akiba Y. Kaunitz J.D. Recent advances in vasoactive intestinal peptide physiology and pathophysiology: Focus on the gastrointestinal system F1000Research 2019 8 18039 10.12688/f1000research.18039.1 31559013 57. Botella A. Delvaux M. Bueno L. Frexinos J. Intracellular pathways triggered by galanin to induce contraction of pig ileum smooth muscle cells J. Physiol. 1992 458 475 486 10.1113/jphysiol.1992.sp019428 1284568 58. Gonkowski S. Obremski K. Calka J. The Influence of Low Doses of Zearalenone on Distribution of Selected Active Substances in Nerve Fibers Within the Circular Muscle Layer of Porcine Ileum J. Mol. Neurosci. 2015 56 878 886 10.1007/s12031-015-0537-2 25772391

PMC009xxxxxx/PMC9099667.txt

==== Front Sensors (Basel) Sensors (Basel) sensors Sensors (Basel, Switzerland) 1424-8220 MDPI 10.3390/s22093600 sensors-22-03600 Article A 12-b Subranging SAR ADC Using Detect-and-Skip Switching and Mismatch Calibration for Biopotential Sensing Applications Nguyen Cong Luong Phan Huu Nhan https://orcid.org/0000-0002-9160-2183 Lee Jong-Wook * Diéguez Angel Academic Editor Information and Communication System-on-Chip (SoC) Research Center, School of Electronics and Information, Kyung Hee University, Yongin 17104, Korea; luongnguyen@savarti.com (C.L.N.); phnhan2310@gmail.com (H.N.P.) * Correspondence: jwlee@khu.ac.kr; Tel.: +82-312-013730 09 5 2022 5 2022 22 9 360011 4 2022 08 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). This paper presents a 12-b successive approximation register (SAR) analog-to-digital converter (ADC) for biopotential sensing applications. To reduce the digital-to-analog converter (DAC) switching energy of the high-resolution ADC, we combine merged-capacitor-switching (MCS) and detect-and-skip (DAS) methods, successfully embedded in the subranging structure. The proposed method saves 96.7% of switching energy compared to the conventional method. Without an extra burden on the realization of the calibration circuit, we achieve mismatch calibration by reusing the on-chip DAC. The mismatch data are processed in the digital domain to compensate for the nonlinearity caused by the DAC mismatch. The ADC is realized using a 0.18 μm CMOS process with a core area of 0.7 mm2. At the sampling rate fS = 9 kS/s, the ADC achieves a signal-to-noise ratio and distortion (SINAD) of 67.4 dB. The proposed calibration technique improves the spurious-free dynamic range (SFDR) by 7.2 dB, resulting in 73.5 dB. At an increased fS = 200 kS/s, the ADC achieves a SINAD of 65.9 dB and an SFDR of 68.8 dB with a figure-of-merit (FoM) of 13.2 fJ/conversion-step. analog-to-digital converter successive approximation register signal-to-noise merged capacitor switching capacitor mismatch National Research Foundation of Korea2021R1A2B5B01001475 Ministry of Science and ICT2020M3H2A1076786 This work was supported in part by the Basic Science Research Program through the National Research Foundation of Korea (No. 2021R1A2B5B01001475) and in part by National R&D Program through the National Research Foundation of Korea funded by the Ministry of Science and ICT (No. 2020M3H2A1076786). ==== Body pmc1. Introduction Portable biomedical sensing applications demand low-power consumption for long battery operation. The human biopotentials have low-frequency bandwidth, up to a few kHz [1]. The amplitude of an electrocardiogram (ECG) is around 1 mV. An electroencephalogram (EEG) has an amplitude from 10 to 100 µV over a frequency band from 0.5 Hz to 150 Hz. The local field potential (LFP) has a typical amplitude of 1 mV over 1 Hz to 200 Hz. The biopotentials are low-amplitude signals, which must be amplified before signal processing. The next important block for signal processing will be the analog-to-digital converter (ADC). Thus, the performance of the amplifier and ADC determines the quality of the measured biopotentials. For digitizing the amplified signal, successive approximation register (SAR) ADC is suitable, with its energy-efficient structure for medium resolution. Moreover, the scaling-friendly structure of the SAR ADC has drawn continued research interest [2]. The basic building blocks of the SAR ADC include a comparator, a digital-to-analog converter (DAC), and SAR logic. The power consumption of the SAR logic, which is mostly digital, can be reduced by lowering the supply voltage. The comparator power can be reduced using a dynamic structure. Thus, researchers have investigated various energy-efficient DAC switching methods—for example, split-DAC [3], monotonic switching [4], set and down [5], and energy saving [6]. The work in [7] introduces a merged-capacitor-switching (MCS) method. In this approach, DAC capacitors are switched from the common-mode (CM) voltage VCM to ground or reference voltage VREF. This method not only saves switching energy but also effectively handles the issues related to CM variations. Behavioral simulation of a 10-bit SAR ADC shows 93.4% less switching energy than the conventional method. A high-resolution SAR ADC demands a relatively large DAC area and switching energy. Switching energy is usually dominant in the capacitors of the most significant bit (MSB) segment rather than those in the least significant bit (LSB) segment. The work in [8] achieves efficient DAC switching using detect-and-skip (DAS). This approach allows skipping capacitor switching in the MSB segment, still generating the correct residue for the DAC. Implemented in 40 nm CMOS, the 10-bit SAR ADC achieves a signal-to-noise ratio and distortion (SINAD) of 55.6 dB and a spurious-free dynamic range (SFDR) of 76.2 dB; however, this work uses a split capacitor for the DAC, which is suboptimal in terms of energy efficiency compared to the MCS. If we can combine the merits of DAS and MCS, this approach can further reduce the DAC switching energy. Moreover, the work in [8] does not support DAC calibration; therefore, it cannot handle the mismatch caused by the parasitics and process variations. Several approaches investigate techniques for DAC mismatch calibration. The work in [9] presents an on-chip dual calibration method for comparator offset and DAC mismatch. The work in [10] presents an energy-efficient ADC using digital domain calibration without additional analog circuits. The work in [11] presents a 13-bit SAR ADC with on-chip calibration for capacitor error compensation. Implemented in 130 nm CMOS, the ADC achieves a SINAD of 66.3 dB and an SFDR of 71 dB by consuming 1.47 μW; this approach suffers from a relatively large area of 0.9 mm2. A digital calibration method is presented using a sub-radix-2 redundant architecture [12], which can handle dynamic errors in the conversion process. The work in [13] analyzes the characteristics of the nonbinary-weighted capacitive DAC and a bottom-up weight calibration technique; however, these works have the drawback of altering the full-scale weight, which is different from the ideal value. In this work, we present a 12-bit subranging SAR ADC suitable for low-power biopotential sensing applications. We reduce the DAC switching energy by combining MCS and DAS methods, successfully embedded in the subranging structure. Behavioral simulation of a 12-bit SAR ADC shows that the proposed method reduces 96.7% of switching energy compared to the conventional method, which is up to 9.2% lower than the previous state-of-the-art [7]. Without an extra burden on the realization of the on-chip calibration circuit, we implement digital domain calibration to compensate for the nonlinearity caused by the DAC mismatch. To address the drawback of the previous approach altering the weight of the full scale, we adopt a normalized full-scale weight for the subranging ADC. Using the proposed calibration technique, the ADC fabricated in 0.18 μm CMOS demonstrates successful operation and performance improvement. At a sampling rate of 200 kS/s, the ADC achieves a SINAD of 65.9 dB with a figure-of-merit (FoM) of 13.2 fJ/conversion-step. An SFDR of 68.8 dB is achieved near the Nyquist frequency. The novelty of this work is efficiently combining MCS and DAS for a high-resolution ADC and implementing a digital domain calibration using a normalized full-scale weight for the subranging ADC. 2. Design 2.1. Subranging SAR ADC Figure 1a shows the proposed subranging ADC. It includes a 7-bit coarse SAR ADC, a 12-bit fine SAR ADC, a DAS controller, a calibration (CAL) logic, and an output buffer. The coarse ADC includes the DAC consisting of seven binary-weighted capacitors CCk (k = 1 to 7). The fine ADC includes the DAC designed with twelve binary-weighted capacitors. For mismatch calibration, we divide the DAC into a 7-bit MSB segment of capacitors Ci (i = 6 to 12) and a 5-bit LSB segment of capacitor Cj (j = 1 to 5). MCS is used for coarse and fine ADCs to save DAC switching energy. The analog input is sampled into the two ADCs at the same time. Top-plate sampling is performed using a bootstrapped switch operating with 1.8 V [9]. After sampling the input, the coarse ADC sequentially generates 7-bit output DOUT,C[12:6]. Then, the DAS controller and fine ADC are enabled by the signal CDONE (coarse done). The DAS controller decodes DOUT,C[12:6], and sets the switches for Ci (i = 6 to 12) of the fine ADC. This operation generates the residue in the fine DAC. Then, the SAR logic of the fine ADC sequentially determines the switch states of the remaining Cj (j = 1 to 5) to generate DOUT,F[5:1]. The DOUT,C[12:6] and DOUT,F[5:1] are combined in the output buffer to generate the ADC output DOUT[12:1] with the end-of-conversion (EOC) signal. Figure 1b shows the timing sequence for the subranging ADC, which consists of calibration and conversion modes. The calibration mode includes three steps: reset, mismatch measurement, and data loading. When the reset signal becomes high, calibration mode starts with the calibration-enabled signal CAL. In this mode, the DAC inputs are disconnected from the analog input. During this time, the calibration code DCAL[6:1] for Ci (i = 6 to 12) is generated and loaded two times (positive and negative DAC). Two bootstrap switches are used to set the bottom plate of the DAC capacitor to VCM. These switches are controlled by the output CALp,n of the CAL logic at the beginning of each calibration cycle. The data loading occurs at the falling edge of EOC_CAL (end of calibration), which captures DCAL[6:1]. After finishing the calibration, the ADC enters conversion mode. Figure 2 shows the timing sequence of the ADC in conversion mode. It shows the internal DAC control signals, VC[k] (k = 1 to 7) for the coarse ADC, VF[i] (i = 6 to 12) for the MSB segment of the fine ADC, and VF[j] (j = 1 to 5, 6ex) for the LSB segment of the fine ADC. VF[6ex] is the control signal for C6ex, which is an additional capacitor for mismatch calibration. The input signal is sampled into the coarse and fine ADCs by the sampling clock CLKS. All VC[k], VF[i], and VF[j] are connected to VCM. One cycle after CLKS, VC[k] is switched to either VREF or ground, depending on the comparator output. After VC[1] is determined, the signal CDONE becomes high, indicating that the coarse ADC has finished quantization. Then, the DAS controller is enabled, which decodes DOUT,C[12:6] from the coarse ADC to determine VF[i] using the DAS operation. The controller decides which VF[i] is skipped or switched (either VREF or ground) to generate the residue for VDAC,p and VDAC,n of the fine DAC. Here, VDAC,p and VDAC,n are the top-plate voltage of the positive and negative DAC, respectively. The fine ADC waits for one cycle after VF[i] switching so that the values of VDAC,p and VDAC,n are stabilized before it starts quantizing the remaining VF[j]. The input CM voltage is constant during MCS. In the previous work [4,9], the comparator is implemented with a PMOS differential pair because this comparator is designed for monotonic switching. When the previous comparator is used for MCS, it can result in a relatively large offset at the input of the comparator. In this work, we use a comparator having complementary input stages, which allows rail-to-rail range and reduces the kickback noise [14]. 2.2. Merged Capacitor Switching with Detect and Skip Figure 3a shows an example waveform of the DAC when DOUT[9:6] = 0101 is generated using MCS. Figure 3b shows the waveform when MCS and DAS are combined. The two methods generate the same residue for VDAC,p and VDAC,n; however, MCS can waste energy by performing unnecessary switching. By combining MCS and DAS, unnecessary switching can be avoided. The DAS controller decides which capacitor can be skipped for switching. Using DOUT,C[12:6] from the coarse ADC, the DAS operation can be summarized as follows:(1) DOUT,C[MSB-1] = DOUT,C[MSB] → switch CMSB|DOUT,C[MSB-1] ≠ DOUT,C[MSB] → skip CMSB, (2) DOUT,C[MSB-2] = DOUT,C[MSB] → switch CMSB-1|DOUT,C[MSB-2] ≠ DOUT,C[MSB] → skip CMSB-1, …, (3) DOUT,C[MSB-k + 1] = DOUT,C[MSB] → switch CMSB-N+2|DOUT,C[MSB-k + 1] ≠ DOUT,C[MSB] → skip CMSB-N+2, (4) Switch CMSB-N+1, Where MSB = 12 and k is the binary capacitor index of the coarse ADC. We note that the MCS and DAS method is more effective for a relatively smaller input since most switching can be skipped. Because the mismatch effect of the skipped capacitors is also removed, DAS can provide the additional benefit of improved linearity. Figure 4 shows the schematic of the DAS controller. When CDONE is enabled, the output DOUT,C[12:6] is input to the DAS control switch through the logic gates. In the beginning, VC[i] is connected to VCM. Depending on the logic value, VC[i] is connected to the ground if DOUT,C[i] is high, or VC[i] is connected to VREF. To evaluate the effectiveness of various switching methods, we compare the switching energy of a 12-bit ADC. The switching energy EMono(i) of the ith capacitor in the monotonic switching can be expressed as (1) EMono(i)=CN−i+1CTVREF2CT−CN−i+1−∑m=N−i+2NCm(bN−i+1⊕bm¯) where index i is from 1 to N = 12, CT is the total capacitance of each DAC branch, and bm is the binary bit value. The switching energy EMCS(i) of the ith capacitor in the MCS can be expressed as [15] (2) EMCS(i)=12−CN−i+12CTVREF2CN−i+1+12CN−i+1CTVREF2∑m=N−i+2N(−1)bN−i+1⊕bm¯Cm A detailed derivation of Equations (1) and (2) can be found in the Appendix A and Appendix B, respectively. In the subranging ADC, switching energy can be divided into MSB and LSB segments of the DAC. The switching energy of the MSB segment can be expressed as (3) EDAS(MSB)=VREF22CSW1−CSWCT where CSW is the sum of switched capacitors. The switching energy of the LSB segment is calculated using 6-bit MCS. The total switching energy is obtained using (4) Etotal=EDAS(MSB)+∑i=16EMCS(i) Figure 5 compares the switching energy of a 12-bit ADC normalized using VREF and the unit capacitor C1. The split capacitor scheme saves 37.5% of energy on average compared with the conventional method [1]. The monotonic switching saves up to 81%. The energy is further reduced using the MCS to 87.5%. Finally, the average switching energy saved is up to 96.7% when combining MCS and DAS in the subranging ADC, which is 9.2% lower than the previous state-of-the-art [7]. This result neglects the energy of the 7-bit coarse ADC, which is relatively small compared to the energy of the 12-bit fine ADC. We note that the switching energy is a normalized value using VREF and C1, independent of the technology node. Relatively low power can still be achieved using the conventional method—for example, 0.084 μW for a 10-bit ADC [8] and 0.38 μW for a 12-bit ADC [16]. Because the SAR ADC is realized using mostly digital logic, except for the comparator, low power can be achieved using scaled-down CMOS technology; the works [8] and [16] are realized using 40 nm and 65 nm CMOS processes, respectively. 3. Mismatch Calibration 3.1. DAC Capacitor Mismatch Calibration Figure 6a shows one example of a DAC configuration for reading out the mismatch of Ci (i = 6 to 12), one of the 7-bit MSB segments of the DAC. The proposed calibration method reuses the 6-bit DAC to measure the weight error of Ci. The 6-bit DAC consists of 5-bit LSB capacitors (C1 to C5) and one extra capacitor C6ex. Assuming that the 6-bit DAC has sufficient intrinsic linearity, the mismatch of each Ci is sequentially measured. The digital representation DCAL[6:1] of the mismatch is generated from the CAL logic. The positive DAC branch is evaluated first, and the negative DAC branch is calibrated next. During the positive DAC calibration, VDAC,n (negative input of the comparator) is connected to VCM. Figure 6b shows the waveform of VC[i] during calibration Ci in the positive branch. Here, VC[i] is the control signal connected to the bottom plate of the DAC capacitor. In the sampling phase, VC[i] of all capacitors are connected to VCM. In the next cycle, the bottom plate of the upper group capacitors (C12 to Ci + 1) is connected to VCM, while the bottom plate of the lower group capacitors (Ci−1 to C6) is connected to the ground. The switching results in VDAC,p are (5) VDAC,p=VCM+(wi*−∑j=6i−1wj*)VCM where wi* is the weight of Ci with mismatch error. Without mismatch, VDAC,p will be equal to VCM. The mismatch causes VDAC,p to deviate from VCM, which is measured by the 6-bit DAC. The bit weight difference between Ci and the sum of lower group capacitors (Ci-1 to C6), which is quantized by the 6-bit DAC, can be expressed as (6) wi*−∑j=6i−1wj*=∑j=16wjbj+qj where wj is the ideal weight, bj is the binary value, and qj is the quantization error. The values of the LSB segment capacitors (C6ex, C5, …, C1) are assumed to be linear with wj*=wj (j = 1, …, 6). The C6ex is added to provide sufficient coverage for weight extraction. The value of C6ex is 16CU, which is small compared to the total capacitance CT = 2048CU of each DAC, where CU = C1 is the unit capacitor of the DAC. To simplify the SAR logic, C6ex can be activated only during the calibration mode while connected to VCM in the conversion mode; however, the addition of C6ex causes the actual weight of each capacitor to deviate from the ideal binary weight. The DAC mismatch calibration is based on the idea that the addition of C6ex does not significantly change the weight of each capacitor. To preserve the correct weight of each capacitor, we handle the issue using an alternative approach: (1) C6ex is used in both calibration and conversion mode; in the conversion mode, C6ex serves as a redundant capacitor to improve the ADC linearity; (2) mismatch calibration is designed by including the weight of C6ex; then, the total weight of the 12-bit DAC is increased from 2048 to 2064 (see Table 1). 3.2. Mismatch Error of DAC Capacitor The VDAC,p and VDAC,n at the inputs of the comparator can be expressed as (7) VDAC,p=VIN,p+∑i∈Ω(1−2bi)(wi−Δwpi)VCM+∑j=16(1−2bj)wjVCM (8) VDAC,n=VIN,n−∑i∈Ω(1−2bi)(wi−Δwni)VCM−∑j=16(1−2bj)wjVCM where VIN,p and VIN,n are the sampled input voltages at the positive and negative DAC, respectively. The bi is the binary value of the DAC capacitor in the MSB segment (C12, …, C6), and bj is the value of capacitors in the LSB segment (C6ex, C5, …, C1). The Ω is the group of switched capacitors by the DAS controller. The wi is the ideal weight of ith capacitor in the MSB segment of the DAC, which is the ratio between Ci and CT. The wj is the ideal weight of the jth capacitor in the LSB segment. The Δwpi and Δwni are the weight errors of the ith capacitor in the positive and negative branches of the fine DAC, respectively. Table 1 shows the ideal weight of each capacitor. The second term of (7) and (8) is the amount of change caused by the mismatch of the MSB capacitors. The third term represents the change caused by the mismatch of the LSB capacitors. At the end of conversion, both VDAC,p and VDAC,n approach VCM as (9) VIN,p−VIN,n+∑i∈Ω1−2bi2wi−Δwpi−ΔwniVCM+2∑j=161−2bjwjVCM≅0 Noting VREF = (VIN,p + VIN,n), we can rearrange (9) as (10) 2VIN,pVREF=∑i∈Ω2bi−1wi−Δwpi+Δwni2+2∑j=16bjwj+−∑j=16wj+1 By multiplying 211 on both sides of (10), we obtain (11) 212VIN,pVREF=12∑i∈Ω2bi−1Wi−ΔWi+∑j=16bjWj+W0 where Wi = 212wi, ΔWi = (ΔWpi + ΔWni)/2 is the average error of the positive and negative branch, ΔWpi = 212(Δwpi), Wni = 212(Δwni), and W0 = 211(127/129) = 2016.248. At this moment, the weight error ΔWi is unknown, and the method of calculating ΔWi is presented in the next subsection. 3.3. Weight Error Extraction We assume that the overall mismatch of the DAC is averaged out and normalize the full scale to one [17]. Then, the sum of weight for C12 can be expressed as (12) w12*+∑i=611wi*+∑j=16wj=1 The calibration code d12 for C12 can be expressed as (13) d12=w12*−∑i=611wi* When we substitute (13) into (12), we obtain (14) 12−w12*=12∑j=16wj−d12 where wj is the weight of the capacitors in the 6-bit DAC. We note that Δ12=(w12−w12*)Δ12 = (w12 – w*12) is the weight error of C12 and w12 = (64/129) is the ideal weight of C12. Then, we obtain (15) Δ12=12∑j=16wj−d12−1258 Similarly, the sum of weight for C11 can be expressed as (16) w11*+∑i=610wi*+∑j=16wj=1−w12* Using the calibration code d11 for C11, we obtain (17) d11=w11*−∑i=611wi* Noting that Δ11 = (w11 – w*11) Δ11=(w11−w11*) is the weight error of C11, where w11 = (32/129) is the ideal weight, we obtain (18) Δ11=12∑j=16wj−d11−1129−Δ12 Similarly, we obtain the remaining weights. For example, the weight error Δ6 of C6 can be expressed as (19) Δ6=12∑j=16wj−d6−1129−Δ12−Δ11−Δ10−Δ9−Δ8−Δ7 The digital representation of sampled input VIN,p can be expressed as DIN,p = 212(VIN,p/VREF). Then, the result (10) can be rearranged as (20) DIN,p=12∑i∈Ω2bi−1Wi−12∑i∈Ω2bi−1ΔWi+∑j=16bjWj+W0 The first two terms of (20) represent the contribution of the MSB segment of the DAC, which can be positive or negative. The third term is the contribution of the LSB segment. The last term is the average output value. A similar definition can be proposed for DIN,n = 212(VIN,n/VREF) for the sampled input VIN,n. Figure 7 shows the block diagram implementing Equation (20) to calculate DIN,p. Using the logic value of the skipped MSB group, it calculates the contribution of the MSB and LSB segment capacitors. The calculation is performed off-chip using Matlab. Figure 8 shows the floor plan of the coarse DAC. We use a common-centroid layout to reduce the capacitor mismatch. Because capacitors need to be connected to the outside of the DAC, the metal route increases the coupling with neighboring capacitors. The effect of additional coupling is usually more sensitive to small capacitors. We reduce the effect by placing the capacitors of the LSB segment close to the edge of the DAC. Dummy capacitors are added around the DAC periphery to reduce the mismatch caused by the edge effect. A similar technique is used for the fine DAC. We use a behavioral model to investigate the ADC performance depending on the mismatch. Monte Carlo simulations with 1000 samples are performed using the DAC capacitor mismatch rate of 1.0%, 1.5%, 2.0%, and 2.5%. Figure 9 compares the effective number of bits (ENOB) probability distribution before and after calibration. Before calibration, the average ENOB decreases from 11.1 bits to 10.2 bits when the mismatch increases from 0.5% to 2%. In the case of a 1% mismatch, the average ENOB increases from 10.8 bits to 11.2 bits after calibration. The standard deviation is reduced from 0.44 bit to 0.15 bit. In the case of a 1.5% mismatch, the average ENOB improves from 10.5 bits to 11.3 bits. The result shows that calibration effectively handles ENOB degradation with the mismatch rate. The minimum capacitor value allowed by the process is 21.2 fF (4 × 4 μm2). Based on the process datasheet, the unit capacitor in the coarse DAC is designed to be larger than the minimum value to achieve a 1% mismatch rate, which is 54 fF (6.72 × 6.72 μm2). Figure 10 shows the power breakdown of the ADC. Overall power including output buffer is 5.08 μW at fS = 200 kS/s. The breakdown shows that the SAR logic of fine ADC, the DAS controller, and the SAR logic of coarse ADC consume 39.5%, 18.9%, and 16.7% of the overall power, respectively. 4. Measured Results Figure 11 shows a microphotograph of the ADC fabricated in a 0.18 μm CMOS process. The core area is 0.7 mm2. The coarse ADC occupies 8.5% of the overall area. The IC is mounted on a test board using the chip-on-board (COB) technique. Biopotentials typically exhibit signal frequencies less than 1 kHz. In this measurement, we choose an input frequency fIN = 1.12k kHz. Figure 12 shows the comparison of the measured output spectra of the ADC before and after calibration. A differential sinusoidal signal with 0.9 V amplitude is applied for dynamic performance testing. The measured data are obtained from the fast Fourier transform (FFT) spectrum with 32768 points. After calibration, SINAD and SFDR are improved by 5.04 dB and 7.21 dB, respectively, resulting in an ENOB of 10.9 bits. The third harmonic located at 3fIN, which is related to the nonlinearity of the ADC, is reduced from −66.3 dB to −77.2 dB. Figure 13 shows the output spectrum using a near-Nyquist input frequency and the sampling rate fS = 9 kS/s. The SINAD and SFDR are improved by 5.44 dB and 2.94 dB, respectively, resulting in an ENOB of 10.5 bits. Additionally, we characterize the dynamic performance at increased fIN. Figure 14 shows the measured spectra of the ADC at fIN = 24.981 kHz and 97.857 kHz (near the Nyquist frequency) after the calibration. The ADC achieves a SINAD of 65.9 dB and an SFDR of 68.8 dB for fIN = 24.981 kHz. The level of the third harmonic located at 3fIN is −68.7 dB, indicating that further improvement of the ADC nonlinearity is needed. Figure 15 shows the measured SFDR and SINAD as a function of fIN. The effective resolution bandwidth (ERBW) is the input frequency where the SINAD drops by 3 dB (1/2 LSB or 0.5 bit) from its value for low-frequency input. The result shows that ERBW is around 100 kHz, approximately half of the sampling frequency (Nyquist frequency). Figure 16 compares the differential nonlinearity (DNL) and integral nonlinearity (INL) of the ADC before and after calibration. A total of 32786 codes are collected to build a histogram. Both INL and DNL are improved, and the calibration has a more desirable effect on INL than DNL, as expected. The peak INL decreases from +3.4/−3.48 LSB to +2.05/−2.24 LSB. The peak DNL is +2.19/−1 LSB before calibration, improving to +1.31/−1 LSB. Table 2 shows the comparison with the previous works. The work in [8] presents a subranging SAR ADC using the DAS method. They use split capacitor switching, which consumes more energy than MCS. A similar observation can be made for the work in [16], which uses the swap-to-reset DAC switching method. The low power consumption can be attributed to the scaled-down technology, 65 nm CMOS [16] and 40 nm CMOS [8,18]. When we compare the ADC realized using a similar CMOS process [2,19], our work achieves a better Walden’s figure-of-merit (FOMW) of 13.2 fJ/conv.-step and Schreier’s figure-of-merit (FOMS) of 170.4 dB. The work in [11] presents a 13-bit SAR ADC with on-chip calibration realized in a relatively large area (0.9 mm2) using a 0.13 μm CMOS process. Our work realizes the subranging ADC, consisting of coarse and fine ADC, in 0.7 mm2 using a 0.18 μm CMOS process. The work in [20] presents good dynamic performance; however, it consumes relatively high power, leading to an FoMS of 114.5 dB. The footnote shows the relationship between ENOB and SINAD. This equation does not explicitly consider the process gain related to the FFT. We can estimate the process gain using GFFT = 10∙log(NF/2), where NF is the number of points processed in the FFT. Each nth FFT bin can be considered as the output from a narrow bandpass filter with a center frequency at (nfS/NF). A large number of samples improves the frequency resolution and decreases the amount of noise in the bin’s passband. For an NF-point FFT, the average value of the noise contained in each frequency bin is reduced by GFFT below the root-mean-square (rms) value of the quantization noise. 5. Conclusions We investigate a 12-bit subranging SAR ADC for low-power biopotential sensing applications. A new DAC switching method is proposed by combining the MCS and DAS methods, successfully embedded in the subranging structure. Analysis of the DAC switching energy shows that the proposed method saves 96.7% of switching energy compared to the conventional method. To handle the DAC mismatch, we implement digital domain calibration without the extra burden of an on-chip calibration circuit. A simple method of extracting the weight error is presented by reusing the 6-bit DAC. The mismatch data are successfully processed in the digital domain to compensate for the nonlinearity caused by the DAC mismatch. The proposed ADC fabricated in 0.18 μm CMOS demonstrates successful operation and performance improvement using the proposed calibration technique. At a sampling rate of 200 kS/s, the ADC achieves SINAD of 65.9 dB and SFDR of 68.8 dB, with an FoM of 13.2 fJ/conversion- step. The contributions of this paper can be summarized as follows: (1) this work proposes an energy-efficient DAC switching method by combining MCS and DAS, (2) the proposed switching method is successfully implemented in a 12-bit subranging ADC, and (3) this work proposes a digital domain calibration using a normalized full-scale weight method. The result will be useful for realizing a low-power ADC for battery-powered, portable biomedical sensing applications. Acknowledgments The chip fabrication and CAD tools were supported by the IDEC (IC Design Education Center). Author Contributions C.L.N. designed the ADC and setup, performed the experimental work, and wrote the manuscript. H.N.P. designed the mismatch calibration algorithm and wrote the manuscript. J.-W.L. conceived the project, organized the paper’s content, and edited the manuscript. Corresponding author: J.-W.L. All authors have read and agreed to the published version of the manuscript. Conflicts of Interest The authors declare no conflict of interest. Appendix A. Monotonic Switching Figure A1 shows the monotonic switching diagram for a 3-bit example (N = 3). After the sampling switches are turned off, the comparator directly performs the first comparison without switching any capacitor. Figure A1 Monotonic switching diagram for a 3-bit ADC. (1) After the first comparison, the MSB bit bN is determined. The bottom plate of capacitor CN (positive DAC branch if bN = 1 or negative DAC branch if bN = 0) is switched to the ground. Then, the comparator input (VDAC,p if bN = 1 or VDAC,n if bN = 0) is reduced by an amount of (CN/CT)VREF. The total capacitance connected to VREF is (CT − CN) in the corresponding DAC branch. The switching energy that VREF supplies to the DAC at the first cycle (ϕ1) can be expressed as (A1) EMono(1)=CNCTVREF2CT−CN=12VREF2CN=CVREF2 where CN = 2C, CT = 4C is the total capacitance of each DAC branch, and C is the unit capacitance. (2) After the second comparison, bN−1 is determined. The bottom plate of capacitor CN−1 (positive DAC branch if bN−1 = 1 or negative DAC branch if bN−1 = 0) is switched to the ground. Then, the comparator input is reduced by (CN−1/CT)VREF. We consider the four cases as follows: - If bN = 0, bN−1 = 0: total capacitance connected to VREF of the negative DAC branch is CT − CN−1 − CN; - If bN = 0, bN−1 = 1: total capacitance connected to VREF of the positive DAC branch is CT − CN−1; - If bN = 1, bN−1 = 0: total capacitance connected to VREF of the negative DAC branch is CT − CN−1; - If bN = 1, bN−1 = 1: total capacitance connected to VREF of the positive DAC branch is CT − CN−1 − CN. We can express the four cases using a single equation that describes the total capacitance connected to VREF as (A2) CT−CN−1−(bN−1⊕bN¯)CN where (bN−1 ⊕ bN) is the XOR of the current bit (bN) and the previous bit (bN−1) value. The switching energy that VREF supplies to the DAC at the second cycle (ϕ2) can be expressed as (A3) EMono(2)=CN−1CTVREF2CT−CN−1−(bN−1⊕bN¯)CN (3) After the third comparison, bN−2 is determined. The bottom plate of capacitor CN−2 is switched to the ground. The comparator input is reduced by an amount of (CN−2/CT)VREF. There are eight switching cases, and we can express the cases using a single equation that describes the total capacitance connected to VREF as (A4) CT−CN−2−(bN−2⊕bN¯)CN−(bN−2⊕bN−1¯)CN−1 The switching energy that VREF supplies to the DAC at the third cycle (ϕ3) can be expressed as (A5) EMono(3)=CN−2CTVREF2CT−CN−3+1−(bN−2⊕bN¯)CN−(bN−2⊕bN−1¯)CN−1    =CN−3+1CTVREF2CT−CN−3+1−∑m=N−3+2NCm(bN−3+1⊕bm¯) By generalizing the above result for an N-bit ADC, we obtain Equation (1) of the main text. Appendix B. Merged Capacitor Switching Figure A2 shows the merged capacitor switching diagram for a 3-bit example. After the sampling switches are turned off, the comparator directly performs the first comparison without switching any capacitor. Switching energy calculation is similar to monotonic switching, except that switching is performed from VCM rather than VREF. (1) After the first comparison, the MSB bit bN is determined. The bottom plate of capacitor CN (positive DAC branch if bN = 1 or negative DAC branch if bN = 0) is switched from VCM to VREF. Then, one comparator input (VDAC,p if bN = 1 or VDAC,n if bN = 0) is increased by the amount of (CN/CT)(VREF/2). The opposite comparator input is reduced by the amount of (CN/CT)(VREF/2). The total capacitance connected to VREF is CN. The switching energy that VREF supplies to the DAC at the first cycle (ϕ1) can be expressed as (A6) EMCS(1)=12−CN2CTVREF2CN=12−14VREF22C=CVREF22 (2) After the second comparison, bN−1 is determined. The bottom plate of capacitor CN−1 (positive DAC branch if bN−1 = 1 or negative DAC branch if bN−1 = 0) is switched from VCM to VREF. Then, one comparator input is increased by the amount of (CN−1/CT)(VREF/2). The opposite comparator input is reduced by the amount of (CN−1/CT)(VREF/2). We consider one of the four cases (bN = 0, bN−1 = 0). There are two capacitors (CN and CN−1) in the positive branch connected to VREF. The energy that VREF supplies to the DAC at this step can be expressed as (A7) EMCS(2)=VREF−VCM−CN−1CTVREF2CN−1VREF+VREF−VREF−CN−1CTVREF2CNVREF    =12−CN−12CTCN−1VREF2−12CN−1CTVREF2CN=18CVREF2 Similar equations can be derived for the remaining three cases. Then, we can express the four cases using a single equation, and the switching energy at ϕ2 can be expressed as (A8) EMCS(2)=12−CN−12CTVREF2CN−1±12CN−1CTVREF2CN    =12−CN−12CTVREF2CN−1+12CN−1CTVREF2(−1)bN−1⊕bN¯CN=38VREF2C±28VREF2C where the ± sign is replaced with (−1)bN−1⊕bN¯. Figure A2 Merged capacitor switching diagram for a 3-bit ADC. (3) After the third comparison, bN−2 is determined. There are eight cases, and we can express the cases using a single equation. Then, the switching energy at ϕ3 can be expressed as (A9) EMCS(3)=12−CN−22CTVREF2CN−2+12CN−2CTVREF2±CN±CN−1    =12−CN−22CTVREF2CN−2+12CN−2CTVREF2(−1)bN−2⊕bN¯CN+(−1)bN−2⊕bN−1¯CN−1 where the first ± sign in the second term is replaced with (−1)bN−2⊕bN¯, and the second ± sign is replaced with (−1)bN−2⊕bN−1¯. By generalizing the above result, we obtain Equation (2) of the main text. Figure 1 (a) Block diagram of the subranging ADC. (b) Timing sequence of the ADC. Figure 2 Timing sequence of the ADC in the conversion mode. Figure 3 Example waveform of the DAC switching using (a) MCS only, (b) MCS and DAS method. Figure 4 Schematic of the DAS controller. Figure 5 Comparison of the DAC switching energy. Figure 6 (a) Example DAC configuration for reading out the mismatch of the capacitor Ci. (b) Waveforms of the control signal VC[i] at the bottom plate of the DAC capacitor during calibration Ci. Figure 7 Block diagram of processing of the calibrated digital output. Figure 8 Floor plan of the coarse DAC. Figure 9 Comparison of the ENOB probability distribution before and after calibration. Mismatch rate is (a) 0.5%, (b) 1.0%, (c) 1.5%, (d) 2.0%. Figure 10 Power breakdown of the ADC. Figure 11 Microphotograph of the fabricated ADC. Figure 12 Measured output spectra of the ADC (a) before calibration, (b) after calibration. fIN = 1.124 kHz, fS = 9 kS/s. Figure 13 Measured output spectra of the ADC (a) before calibration, (b) after calibration. fIN = 4.403 kHz, fS = 9 kS/s. Figure 14 (a) Measured output of the ADC. (b) Measured output near the Nyquist frequency. fS = 200 kS/s. Figure 15 Measured SINAD and SFDR at different input frequencies. fS = 200 kS/s. Figure 16 Measured static nonlinearity of the ADC. sensors-22-03600-t001_Table 1 Table 1 Ideal weight of DAC. DAC Capacitor Capacitance (CU) Ideal Weight C 12 1024 64/129 C 11 512 32/129 C 10 256 16/129 C 9 128 8/129 C 8 64 4/129 C 7 32 2/129 C 6 16 1/129 C 6ex 16 1/129 C 5 8 1/258 C 4 4 1/516 C 3 2 1/1032 C 2 1 1/2064 C 1 1 1/2064 Total 2064 1 sensors-22-03600-t002_Table 2 Table 2 Performance comparison. [2] [8] [11] [16] [18] [19] [20] This Work Tech. (nm) 180 40 130 65 40 180 65 180 Supply (V) 0.75 0.45 0.5 0.8 1.0 1.0 1.2 1.8/1.0 Resolution (bit) 11 10 13 12 13 11 13 12 Rate (kS/s) 10 200 40 40 6400 1000 50,000 200 SINAD (dB) 60.5 55.6 66.3 64.2 64.1 63.4 70.9 67.4 SFDR (dB) 72.0 76.2 71.0 88.2 68.8 76.6 84.6 73.5 ENOB † (bit) 9.8 8.95 10.7 10.4 10.4 10.3 11.5 10.9 Calibration No No Yes No Yes No No Yes Power (μW) 0.25 0.084 1.47 0.38 46 24 1000 5.08 Area (mm2) 0.13 0.007 0.9 0.11 0.07 0.1 0.05 0.7 FoMW * (fJ/conv.-step) 28.8 0.85 21.8 7.1 2.2 19.9 6.9 13.2 FoMS ** (dB) 163.5 176.4 167.6 171.5 172.5 166.6 114.9 170.4 † ENOB = (SINAD − 1.76)/6.02. * FoMW=PowerfS×2ENOB,** FoMS=SINAD+10logfS2×Power. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Saberi M. Lotfi R. Mafinezhad K. Serdijn W.A. Analysis of power consumption and linearity in capacitive digital-to-analog converters used in successive approximation ADCs IEEE Trans. Circuits Syst. I Reg. Pap. 2011 58 1736 1748 10.1109/TCSI.2011.2107214 2. Sadollahi M. Temes G. An 11-bit 250-nW 10-kS/s SAR ADC with doubled input range for biomedical application Proceedings of the IEEE International Midwest Symposium on Circuits and Systems Boston, MA, USA 6–9 August 2017 385 388 3. Yoshioka M. Ishikawa K. Takayama T. Tsukamoto S. A 10-b 50-MS/s 820-μW SAR ADC with on-chip digital calibration IEEE Trans. Biomed. Circuits Syst. 2010 4 410 416 10.1109/TBCAS.2010.2081362 23850758 4. Liu C. Chang S. Huang G. Lin Y. A 10-bit 50-MS/s SAR ADC with a monotonic capacitor switching procedure IEEE J. Solid-State Circuits 2010 45 731 740 10.1109/JSSC.2010.2042254 5. Liu C. Chang S. Huang G. Lin Y. A 0.92 mW 10-bit 50-MS/s SAR ADC in 0.13 μm CMOS process Proceedings of the Symposium on VLSI Circuits, Digest of Technical Papers Kyoto, Japan 16–18 June 2009 236 237 6. Chang Y. Wang C. Wang C. A 8-bit 500-KS/s low power SAR ADC for biomedical applications Proceedings of the IEEE Asian Solid-State Circuits Conference Jeju, Korea 12–14 November 2007 228 231 7. Hariprasath V. Guerber J. Lee S.-H. Moon U. Merged capacitor switching based SAR ADC with highest switching energy-efficiency Electron. Lett. 2010 46 620 621 10.1049/el.2010.0706 8. Tai H.-Y. Hu Y.-S. Chen H.-W. Chen H.-S. A 0.85fJ/conversion-step 10b 200 kS/s subranging SAR ADC in 40nm CMOS Proceedings of the IEEE International Solid-State Circuits Conference Digest of Technical Papers San Francisco, CA, USA 9–13 February 2014 196 197 9. Lee J.H. Park D. Cho W. Phan H.N. Nguyen C.L. Lee J.-W. A 1.15 μW, 200 kS/s 10-b monotonic SAR ADC using dual on-chip calibrations and accuracy enhancement techniques Sensors 2018 18 3486 10.3390/s18103486 30332815 10. Um J.Y. Kim Y.-J. Song E.-W. Sim J.-Y. Park H.-J. A digital-domain calibration of split-capacitor DAC for a differential SAR ADC without additional analog circuits IEEE Trans. Circuits Syst. I Reg. Pap. 2013 60 2845 2856 10.1109/TCSI.2013.2252475 11. Ha H.S. Lee S.-K. Kim B.S. Park H.-J. Sim J.-Y. A 0.5-V, 1.47-μW 40-kS/s 13-bit SAR ADC with capacitor error compensation IEEE Trans. Circuits Syst. II Exp. Briefs 2014 61 840 844 10.1109/TCSII.2014.2350378 12. Liu W. Huang P. Chiu Y. A 12-bit 45-MS/s, 3-mW redundant successive-approximation-register analog-to-digital converter with digital calibration IEEE J. Solid-State Circuits 2011 46 2661 2672 10.1109/JSSC.2011.2163556 13. Zhang D. Alvandpour A. Analysis and calibration of nonbinary-weighted capacitive DAC for high-resolution SAR ADCs IEEE Trans. Circuits Syst. II Exp. Briefs 2014 61 666 670 10.1109/TCSII.2014.2331111 14. Choi C. Lee J.-W. An 11.8-fJ/conversion-step noise shaping SAR ADC with embedded passive gain for energy-efficient IoT sensors Sensors 2022 22 869 10.3390/s22030869 35161615 15. Guerber J. Venkatram H. Oh T. Moon U. Enhanced SAR ADC energy efficiency from the early reset merged capacitor switching algorithm Proceedings of the IEEE International Symposium on Circuits and Systems Seoul, Korea 20–23 May 2012 2361 2364 16. Liu M. Roermund A. Harpe P. A 7.1fJ/conv.-step 88 dB-SFDR 12b SAR ADC with energy-efficient swap-to-reset Proceedings of the IEEE European Solid-State Circuits Conference Lausanne, Switzerland 12–15 September 2016 409 412 17. Chang D.-J. Kim W. Seo M.-J. Hong H.-K. Ryu S.-T. Normalized-full-scale-referencing digital-domain linearity calibration for SAR ADC IEEE Trans. Circuits Syst. I Reg. Pap. 2017 64 322 332 10.1109/TCSI.2016.2612692 18. Lim Y. Flynn M.P. A 1 mW 71.5 dB SNDR 50 MS/s 13 bit fully differential ring amplifier based SAR-assisted pipeline ADC IEEE J. Solid-State Circuits 2015 50 2901 2911 10.1109/JSSC.2015.2463094 19. Ding M. Harpe P. Liu Y.-H. Busze B. Philips K. Groot H. A 46 μW 13 b 6.4 MS/s SAR ADC with background mismatch and offset calibration IEEE J. Solid-State Circuits 2016 52 423 432 10.1109/JSSC.2016.2609849 20. Chen L. Sanyal A. Ma J. Sun N. A 24-μW 11-bit 1-MS/s SAR ADC with a bidirectional single-side switching technique Proceedings of the IEEE European Solid State Circuits Conference Venice, Italy 22–26 September 2014 219 222

PMC009xxxxxx/PMC9099668.txt

==== Front Int J Environ Res Public Health Int J Environ Res Public Health ijerph International Journal of Environmental Research and Public Health 1661-7827 1660-4601 MDPI 10.3390/ijerph19094947 ijerph-19-04947 Article Exploring Risk Factors of Recall-Associated Foodborne Disease Outbreaks in the United States, 2009–2019 https://orcid.org/0000-0001-8999-5822 Sanchez Emily 12* https://orcid.org/0000-0002-1375-5387 Simpson Ryan B. 1 Zhang Yutong 1 Sallade Lauren E. 1 https://orcid.org/0000-0002-9562-4734 Naumova Elena N. 1* Amalaradjou Mary Anne Academic Editor 1 Friedman School of Nutrition Science and Policy, Tufts University, Boston, MA 02111, USA; ryan.simpson@tufts.edu (R.B.S.); yutong.zhang@tufts.edu (Y.Z.); lauren.sallade@tufts.edu (L.E.S.) 2 Army Medical Department Student Detachment, U.S. Army Medical Center of Excellence, Fort Sam Houston, San Antonio, TX 78234, USA * Correspondence: emily.sanchez@tufts.edu (E.S.); elena.naumova@tufts.edu (E.N.N.); Tel.: +1-(608)-449-3194 (E.S.); +1-617-636-2927 (E.N.N.) 19 4 2022 5 2022 19 9 494731 1 2022 16 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Earlier identification and removal of contaminated food products is crucial in reducing economic burdens of foodborne outbreaks. Recalls are a safety measure that is deployed to prevent foodborne illnesses. However, few studies have examined temporal trends in recalls or compared risk factors between non-recall and recall outbreaks in the United States, due to disparate and often incomplete surveillance records in publicly reported data. We demonstrated the usability of the electronic Foodborne Outbreak Reporting System (eFORS) and National Outbreak Reporting System (NORS) for describing temporal trends and outbreak risk factors of food recalls in 1998–2019. We examined monthly trends between surveillance systems by using segmented time-series analyses. We compared the risk factors (e.g., multistate outbreak, contamination supply chain stage, pathogen etiology, and food products) of recalls and non-recalls by using logistic regression models. Out of 22,972 outbreaks, 305 (1.3%) resulted in recalls and 9378 (41%) had missing recall information. However, outbreaks with missing recall information decreased at an accelerating rate of ~25%/month in 2004–2009 and at a decelerating rate of ~13%/month after the transition from eFORS to NORS in 2009–2019. Irrespective of the contaminant etiology, multistate outbreaks according to the residence of ill persons had odds 11.00–13.50 times (7.00, 21.60) that of single-state outbreaks resulting in a recall (p < 0.001) when controlling for all risk factors. Electronic reporting has improved the availability of food recall data, yet retrospective investigations of historical records are needed. The investigation of recalls enhances public health professionals’ understanding of their annual financial burden and improves outbreak prediction analytics to reduce the likelihood and severity of recalls. data credibility foodborne outbreaks food recalls National Outbreak Reporting System (NORS) ==== Body pmc1. Introduction Every year, ~48 million Americans get sick, 128,000 are hospitalized, and 3000 die from foodborne diseases [1]. From 2013 to 2018, the US Department of Agriculture’s (USDA) Economic Research Service estimated that the value of preventing foodborne illnesses, a measure of demand for reduction in mortality risk, increased by 12%, from $12.8 to $14.4 billion (USD), respectively [1]. These estimates account for inpatient and outpatient hospital costs and costs of prescription drugs and medical supplies used to treat infected persons [1]. One safety measure deployed to prevent foodborne illnesses is food recalls, or when a manufacturer or distributor voluntarily removes food products from commerce due to their expected risk to human health [2]. In 2011, the Food Marketing Institute and Grocery Manufacturers Association reported that food recalls cost ~$10 M/recall in direct costs to food companies [3]. The study further noted that ~23% of annual recalls exceed ~$30 M/recall in direct costs, which accounts for product retrieval, storage destruction, and regulatory notifications throughout the supply chain [3,4]. However, these costs underestimate recalls’ total financial burden, as direct costs exclude government fines, food-safety hygiene compliance penalties, lawsuits, lost sales, and damaged brand reputation [5,6,7]. Indirect costs have long-lasting impacts; for example, a 2010 Harris Interactive Poll found that 55% of consumers would switch brands if consuming a recalled product and 36% of consumers would never purchase the product or brand again if they learned of a recalled product [3]. The early detection of outbreaks and identification of contaminated food products is crucial in reducing the health burdens of foodborne outbreaks. While research has explored temporal trends and risk factors associated with foodborne outbreaks, less is known regarding food recalls [8,9,10]. The USDA’s Food Safety and Inspection Service (FSIS) and Department of Health and Human Services’ Food and Drug Administration (FDA) are charged with monitoring and investigating recalls [11]. The FSIS routinely inspects and regulates meat, poultry, and processed egg recalls, while the FDA regulates all other products [12]. When conducting traceback investigations, both agencies publicly report the product manufacturer and ingredients of contaminated food products [12,13]. The FSIS also reports annual tallies of recalls by class (i.e., public risk level), suspected contaminant (i.e., pathogen, allergen, substance, or chemical/toxin), product type (i.e., beef, pork, poultry, etc.), and volume of food recalled [13]. The FDA provides event-based descriptive summaries of implicated food company name, a product description, product type, and recall termination date [12]. While the current literature on recalls largely explores their impact on consumer trust and behaviors, few publications provide in-depth descriptions of recall risk factors or any information on non-recall outbreaks [14,15]. In one study exploring the direct contribution of meat and poultry recalls to the food waste stream, the FSIS meat and poultry recalls from 1994 to 2015 were found to be predominantly attributed to Listeria, undeclared allergens, and Shiga toxin-producing Escherichia coli (STEC) [16]. Additional reports summarize outbreak investigations with brief mention of the outbreak resulting in a food recall or not [17,18,19]. Furthermore, resource allocation to investigate and report foodborne outbreaks may be determined by a public health agency’s mission and subsequently effect the frequency of outbreak-related recalls reported. That is, since the FSIS regulates meat, poultry and processed eggs, microbial sampling and testing programs are more routinely conducted for Salmonella, Campylobacter, Listeria monocytogenes, and STEC [20]. Lastly, despite analyzing implicated product types and ingredients, neither the USDA nor the FDA reports the supply chain stage of contaminant introduction or locations of food preparation and consumption by ill persons. These data are critical for improving outbreak prediction analytics to enhance food traceability and expedite public health responses to foodborne outbreaks and recalls in accordance with the FDA’s 2020 New Era of Smarter Food Safety Blueprint [21,22]. Comparisons between the risk factors of recall and non-recall outbreaks can improve food safety technology and interagency collaborative efforts for rapid outbreak response detection and mitigation [22]. In 2012/2013, the Government Accountability Office (GAO) reported that the USDA’s and the FDA’s recall surveillance systems were unreliably, inconsistently, and incompletely reporting recall record data, preventing its usage for analytical purposes [23]. For example, annual tallies and descriptive reports lack the comprehensiveness to extensively study recall record temporal trends and risk factors. Though agencies report recalls’ dates, reporting formats are not conducive to performing time-series analyses. This impedes data users from investigating the longitudinal impact of food safety policies and justifying strategies for continued regulatory oversight and enforcement [24]. Additionally, the inability to conduct time-series analyses prevents the inspection of recall seasonality imperative for public health professionals to prepare for and mitigate the intensity of seasonal foodborne outbreaks and illnesses [8,25,26,27,28]. While data-quality-related issues have improved per a 2018 GAO report, USDA agency officials still noted that data are poorly utilized within analytical workflows to inform food safety and consumption guidelines [29]. In fact, the most recent 2020/2021 GAO report emphasized that gaps in the USDA/FDA foodborne illness and recall surveillance stress the need for improved surveillance capacity on these topics by the Centers for Disease Control and Prevention (CDC) [30]. Though not responsible for investigating recalls, the CDC has conducted thorough event-based surveillance of waterborne and foodborne outbreaks since 1971 and 1973, respectively [31]. In 1998, CDC surveillance transitioned from the paper-based Foodborne Outbreak Reporting System (pFORS) to electronic reporting (eFORS) [32]. In contrast to the FSIS and FDA, the CDC’s records include data on where, when, how many persons, what food sources, and which pathogens are associated with outbreaks [32]. In November 2004, the CDC began identifying if outbreaks resulted in recalls and reporting recall-related traceback information [33]. In January 2009, the CDC integrated all outbreak surveillance data streams into the National Outbreak Report System (NORS), which monitors, tracks, and reports on 45 person-to-person, zoonotic, environmental, and unknown/indeterminate sources of outbreaks [32]. Despite the NORS’s comprehensiveness, public health professionals have only recently begun exploring the usability of NORS for describing recall records [34]. These studies have largely explored the likelihood of increased morbidity and mortality of recalls compared to non-recalls [34]. To the best of our knowledge, no studies have utilized recall record data to examine temporal trends, particularly before and after November 2004 or January 2009. Additionally, no studies have compared risk factors between recalls and non-recalls, perhaps due to the limited completeness of surveillance records [35]. Thus, further investigation of recall records reported by the CDC is warranted. In this study, we demonstrated the use of recall record data from electronic Foodborne Outbreak Reporting System and National Outbreak Reporting System for investigating temporal trends and risk factors of food recalls in 1998–2019. First, we described monthly trends and seasonality of recalls, non-recalls, and outbreaks with missing recall information, using segmented negative binomial regression models with respect to three delineated periods: prior to November 2004 (surveillance reporting begins under eFORS), prior to January 2009 (foodborne outbreak reporting revised to include if any product was recalled from an outbreak), and after January 2009 (surveillance reporting begins under NORS). Next, we examined risk factors (e.g., multistate exposure outbreak, contamination supply chain stage, pathogen etiology, and food products) associated with recalls, using logistic regression models. Our findings highlight improvements in CDC surveillance reporting over time; however, they still note extensive incomplete records on food recalls. 2. Materials and Methods 2.1. Data Source On 4 March 2021, we requested CDC surveillance records for foodborne and waterborne outbreaks from 1 January 1998 to 31 December 2019. The NORS Foodborne and Animal Contact Team investigated the accuracy and quality of data prior to distribution. Due to the volume of surveillance records requested, data were separated into 65 data tables, each corresponding to characteristics of the time, location, pathogen, preparation/consumption location, food ingredients, and food inspection methods of an outbreak. NORS used unique outbreak identifiers to harmonize records across tables and provided a comprehensive dictionary to describe variables and their units of measurement per table [36]. We created etiology-, state-, and county-specific binary indicator variables (1 = present, and 0 = absent). Etiology-specific variables included 29 pathogens (adenovirus, Anisakis, astrovirus, Bacillus, Brucella, Campylobacter, Ciguatoxin, Clostridium, Cryptosporidium, Cyclospora, Entamoeba, Enterobacter, Enterococcus, E. coli, Giardia, Hepatitis, Listeria, norovirus, Proteus, rotavirus, Salmonella, sapovirus, Shigella, Staphylococcus, Streptococcus, Toxoplasma, Trichinella, Vibrio, and Yersinia) and 5 toxins/poisons (shellfish poison, ciguatoxin, plant/herbal toxins, puffer fish tetrodotoxin, and scombroid toxin). States included the 50 US states; Washington, D.C.; and three US territories (Guam, Puerto Rico, and Republic of Palau). We created multi-etiology, multistate exposure, and multi-county variables for outbreaks associated with numerous etiologies or where an outbreak was caused by exposures in multiple states, respectively [37]. Reported recall record data included the type of food product, a description of the recalled product, and the product’s brand or lot number, a distinct combination of letters, numbers, or symbols that correspond to the complete history of the manufacturer, processing, packing, holding, and distribution of a product [38]. We examined the recall status according to three categories: outbreaks resulting in recalls (recalls), outbreaks not resulting in recalls (non-recalls), and outbreaks with missing recall information (missing). Under NORS, recall-related traceback information was incorporated into the recall record. NORS classified 3 supply chain contamination points, namely before preparation (i.e., production, harvesting, packaging, and transporting), preparation (i.e., cooking, retail, consumption), and unknown. The before preparation category was further disaggregated into 3 subcategories: pre-harvest (e.g., traceback to producer farms and fields), preprocessing (e.g., traceback to leaking produce cleaning and storage facility), and unknown preparation. NORS reports food products by using the Interagency Food Safety Analytics Collaboration (IFSAC) classification scheme. Developed in 2011 by the CDC, FSIS, and FDA, IFSAC is a 5-level food-categorization hierarchy that specifies 234 food categories [39]. Level 1 refers to the coarsest categorization of food groups (e.g., aquatic animals, land animals, plants, and other foods), while Level 5 disaggregates groups by processing, preparation, and consumption type (e.g., fermented, cured, salt-cured, etc.). NORS reported 23 locations where ill persons prepared and consumed contaminated foods implicated in causing an outbreak, including restaurants (e.g., fast food, sit-down, and other), function halls (e.g., private home, banquet facility, and caterer), community gathering areas (e.g., daycare, school, prison/jail, religious location, camp, picnic, and fair), and other locations (e.g., grocery store, workplace, nursing home, assisted living facility, hospital, and home) [40]. These categories were not mutually exclusive; we created dichotomous variables for each location, as well as multi-location food preparation or consumption. 2.2. Investigating Temporal Trends We conducted a segmented time-series regression analysis to investigate monthly count trends and seasonality in outbreaks and recalls over our 22-year study period. We divided our study period into 3 critical periods according to 2 critical points: November 2004, or when the CDC began food recall reporting; and January 2009, or when surveillance transitioned from eFORS to NORS (Table 1). To estimate the mean monthly counts of outbreaks, recalls, non-recalls, and outbreaks missing recall information for the entire study period, we applied a generalized linear model with a negative binomial distribution and logarithmic link function:Model 1: ln[E(Yi)]=β0 Model 2: ln[E(Yi)]=β0+β1t∗zi where Yi is the estimated mean of i-outcome (e.g., outbreaks, recalls, non-recalls, and missing); β0 is the estimated mean for i-outcome for the study period of 264 months; t is the consecutive time, in months, ranging from 1 to 264 sequentially; and z is a binary indicator variable for i-critical period ranging from 1 to 3, where the outcome of interest occurred (z = 1). By exponentiating the model’s intercept, we calculated the estimated mean, exp{β0}, and their 95% confidence interval estimates, exp{β0±1.96se}. The study period of 264 months divided into three critical periods was marked with knots, or critical points where a represents the start of critical period 2 at study month 82 and b represents the start of critical period 3 at study month 132. Using the selected periods, we developed a segmented negative binomial regression model to examine the temporal trends across the three critical periods for all outcomes:Model 3: ln[E(Yt,i)]=β0+β1t+β2(t−a)+β3(t−b) Model 4: ln[E(Yt,i)]=Model 3+β4t2+β5(t−a)2+β6(t−b)2 Model 5: ln[E(Yt,i)]=Model 4+β7sin(2πωt)+β8cos(2πωt)+β9sin(2πω(t−a))+β10cos(2πω(t−a))+β11sin(2πω(t−b))+β12cos(2πω(t−b)) where Yt,i represents the monthly counts of i-outcome (e.g., outbreaks, recalls, non-recalls, and missing) in t-month; t is the consecutive time in months, ranging from 1 to 264, sequentially; and a and b are the locations of the critical points at 82 and 132 months, respectively. Moreover, t, (t−a), and (t−b); and t2, (t−a)2, and (t−b) 2 are the linear and quadratic trends of continuous time-series variables in months, respectively. In addition, sin(2πωt), sin(2πω(t−a), and  sin(2πω(t−b) ; and cos(2πωt), cos(2πω(t−a)), and cos(2πω(t−b)) are the sinusoidal and co-sinusoidal harmonic terms, respectively, with a frequency of ω=1/M, where M=12 represents the length of the annual cycle in months. We assessed the contribution of linear and quadratic trend terms in Model 4. The linear term indicated overall increases (β1t > 0, β2(t−a) > 0, β3(t−b) > 0) or decreases (β1t < 0, β2(t−a) < 0, β3(t−b) < 0), while the quadratic term indicated acceleration (β1t2 > 0, β2(t−a)2 > 0, β3(t−b)2 > 0) or deceleration (β1t2 < 0, β2(t−a)2 < 0, β3(t−b)2 < 0) within each critical period. We calculated the trend contribution by multiplying each coefficient by the trend-associated time unit to recover the corresponding predicted rates:TCi,j,m,k=|βm(t−tk)j||β1t|+|β2(t−a)|+|β3(t−b)|+|β4t2|+|β5(t−a)2|+|β6(t−b)2| where TCi,j,m,k is the contribution of the i-outcome for j-trend (j = 1 for linear term, j = 2 for quadratic term) in the βm coefficient, with m ranging from 1 to 6 for k-continuous time series variable (e.g., a and b; for summary of model coefficients and diagnostics, see Supplementary Table S1). The trend terms across all critical periods were summed to 1.00 per outcome regression model. We determined seasonality by the significance of either harmonic term in Model 5. 2.3. Assessing Risk Factors Associated with Recalls Based on the trend analyses, we found that very few non-recalls occurred as monthly counts of outbreaks missing recall information increased in Period 1. However, while monthly counts of non-recalls began to rise, outbreaks missing recall information decreased in Period 2. To better understand the risk factors associated with an outbreak resulting in a recall, we chose to conduct risk-factor analyses amongst recalls and non-recalls aggregated with outbreaks missing recall information, subsequently referred to as non-recalls. In these analyses, we considered the following risk factors: multistate exposure outbreaks, supply chain contamination stage, pathogen etiology, and IFSAC Level 1 category food products. We analyzed multistate exposure outbreaks by using a binary variable where single-state exposure outbreaks were the reference. We analyzed supply chain contamination stage by using a 3-level categorical variable (i.e., before preparation, preparation, or unknown) where before preparation was the reference. We analyzed IFSAC Level 1 category food products by using a 4-level categorical variable (i.e., land animals, aquatic animals, plants, or other) where land animals were the reference. We restricted our analyses and independently evaluated 5 etiologies (i.e., Salmonella, E. coli, Listeria, norovirus, and scombroid toxin), as they attributed to 46.8% of all outbreaks and 78.7% of all recalls. We analyzed pathogen etiology by using a binary variable indicating whether the specific pathogen was associated with the outbreak or not. First, we continued to explore patterns of missingness among risk factors, using frequency tables. Second, we compared differences in frequencies of recalls and non-recalls. Third, we examined the likelihood of a recall with each factor, using univariate logistic regression models. Lastly, we performed multivariate models in a stepwise order, where parameters were specified in accordance with univariate findings:Model 6: ln(Pr[Ri])=β0+β1(Si)+β2(Ci)+β3(Fi)+β4(Di) where Ri is a recall for i-outbreak (reference: non-recall and missing combined); Si is a binary variable indicating multistate exposure of illness for i-outbreak; Ci is a categorical variable indicating the supply chain contamination stage of i-outbreak; Fi is the IFSAC Level 1 category for i-outbreak; and Di is a binary variable indicating specific etiology associated with i-outbreak. In a sub-analysis, we examined the likelihood of identifying recalls (n = 305) during the before-preparation supply chain stage, using logistic regression models and the following risk factors: IFSACL Level 1 category, pathogen etiology, and preparation and consumption locations. We created a dichotomous variable for contamination stage (i.e., before preparation or preparation) by setting outbreaks of unknown preparation stage to missing and using preparation stage as the reference. We restricted our analyses to the 3 most common locations for preparation and consumption (i.e., home, diner, restaurant and other), which accounted for 38.7% and 54.1% of all outbreaks and all recalls, respectively. We analyzed preparation and consumption locations by using a binary variable indicating whether the specific location was associated with the outbreak resulting in a recall or not. We explored associations between supply chain stage and outbreak etiology, food product, and location of preparation or consumption:Model 7: ln(Pr[Cr])=β0+β1(Fr)+β2(Dr)+β3(Lr) where Cr is the r-recall identified in before-preparation supply chain stage; Fr is the IFSAC Level 1 category for i-outbreak; Dr is a binary variable indicating specific etiology associated with i-outbreak; and Lr is a binary variable indicating specific locations where persons prepared or consumed contaminated foods associated with r-recall. We defined statistical significance as α < 0.05. We evaluated model goodness-of-fit for all models by using the Akaike’s Information Criterion (AIC). We performed data extraction, alignment, management, and cleaning by using Excel 2016 Version 16.59 and Stata SE/16.1 software. We conducted statistical analyses and created data visualizations by using Stata SE/16.1 and RStudio Version 1.2.5042 software. 3. Results 3.1. Investigating Temporal Trends NORS reported 22,792 outbreaks from 1 January 1998 to 31 December 2019, of which 305 (1.3%) resulted in food recalls, 13,109 (57.5%) stated no recall, and 9378 (41.1%) had information missing. The initiation of recall reporting resulted in an increase of reported recalls from 0.06 (0.01, 0.23) to 1.23 (0.42, 3.48) in Period 2 and to 1.79 (<0.01, 1566.96) in Period 3 (Table 2). The estimated monthly mean of outbreaks declined from 106.68 (84.62, 134.52) in Period 1 to 92.53 (78.15, 109.68) in Period 2 (p < 0.001) and to 71.05 (34.47, 146.55) in Period 3 (p < 0.001). The estimated monthly means of non-recalls were higher in Period 2 (79.12 (54.33, and 115.73)) compared to Periods 1 and 3 (7.49 (5.39, 10.40), p < 0.001, and 64.03 (12.75, 103.11), p < 0.001, respectively), whereas outbreaks missing recall information were lower in Period 2 (2.18 (0.01, 312.28)) compared to Periods 1 and 3 (99.03 (44.96, 219.89), p < 0.001) and 5.12 (0.09, 284.86), p = 0.16, respectively) (Table 2 and Figure 1). Across the entire study period, outbreaks, non-recalls, and outbreaks missing recall information increased by 0.06%/month, 0.24%/month, and 1.90%/month, respectively, whereas recalls decreased by 0.01%/month (Supplementary Table S1; Figure 1). We found no significant linear or quadratic trends in monthly outbreaks in Period 1, though outbreaks with missing recall information steadily decreased by 1.03%/month. Though recall information was not formally collected until November 2004, NORS does report non-recalls consistently from January 1998 to November 2004. Non-recalls in Period 1 decreased at an accelerating rate of 1.78%/month (−3.20, −0.34); p = 0.014). Unexpectedly, before the official collection of recall status data, in Period 1, NORS reported one recall in April 1998, June 2002, and April 2004; and two recalls in June 2004. In Period 2, outbreaks increased at a decelerating rate, by 2.13%/month (0.64, 3.65), which continued to increase during Period 3, though at an accelerating rate (2.70%/month (1.46, 3.94)). Similarly, non-recalls increased at a decelerating rate by 10.36%/month (7.76, 13.03) in Period 2, followed by increases at an accelerating rate in Period 3 (9.30%/month (7.64, 10.98)). In contrast, outbreaks with missing recall information decreased across both periods, first at an accelerating rate in Period 2 (25.73%/month (−28.25, −23.20) and then at a decelerating rate in Period 3 (12.65%/month (−15.67, −9.53)). Though we found no significant trends in Period 2, recalls steadily decreased by 3.02%/month (−4.35, −1.76) in Period 3. Across critical periods, we found that outbreaks decreased by 0.77%/month from Periods 1 to 2 and increased by 0.16%/month and 0.72%/month from Periods 2 to 3 and Periods 1 to 3, respectively. Similarly, outbreaks missing recall information decreased by 6.87%/month from Periods 1 to 2 but increased by 2.97%/month from Periods 2 to 3 and 7.77%/month from Periods 1 to 3. In contrast, both non-recalls and recalls increased by 2.14%/month and 3.02%/month, respectively, between Periods 1 and 2, followed by decreases of similar magnitudes from Periods 2 to 3 (2.14%/month and 2.88%/month, respectively). Both non-recalls and recalls increased slightly between Periods 1 and 3 (0.27%/month and 1.15%/month, respectively). When examining trend contributions and modeling diagnostics, we found that linear trends contributed to 96.6–98.5% of the overall trend for all models compared to just 1.5–3.4% for quadratic terms (Supplementary Table S2). The model fit improved in Model 4, as indicated by a ~0.52–12.4% reduction in AIC for all outcomes. These findings suggested the need for inclusion of quadratic terms when examining seasonal patterns of outcomes. Outbreaks, non-recalls, and outbreaks with missing recall information demonstrated significant seasonality in at least one critical period (Supplementary Table S1; Figure 2). While seasonal patterns of outbreaks appeared visually in all periods, harmonic terms were only significant in Period 1. Non-recalls had significant seasonal patterns in both Periods 1 and 2, whereas outbreaks with missing recall information had significant seasonal patterns in Period 2 only. Though insignificant, outbreaks with missing recalls appeared to have a seasonal pattern in Period 1, also with maximum counts reported in both May and December. All outcomes shared similar patterns, such that maximum counts occurred in April/May, while minimum counts occurred in September/October. 3.2. Comparing Risk Factors—Food Recalls The temporal analyses showed that the reporting of recalls and non-recalls began in Period 2 and continued through Period 3. In comparison, outbreaks missing recall information largely occurred in Period 1, with minimal reporting during Periods 2 and 3. Due to the opposite trends seen in non-recalls and outbreaks missing recall information over the study period, we continued to explore missingness amongst risk factors for outbreaks resulting in a recall with those resulting in non-recalls combined with outbreaks missing recall information. We found extensive missing data among outbreak risk factors (Table 3). Only 7.6% of outbreaks (51.5% of recalls and 7.0% of non-recalls) had non-missing records for all factors (Figure 3). The location of outbreak exposure had no missing data in our study period. In contrast, 75.9% of outbreaks (n = 17,292) had missing supply chain contamination–stage data, including 41.3% of recalls (n = 126) and 76.3% of non-recalls (n = 17,166). While only 3.28% of recalls had missing etiology information (n = 10), nearly one-third of non-recalls failed to report this risk factor (n = 7383; 32.4%). Similarly, 12.8% of recalls (n = 39) failed to report IFSAC Level 1 information compared to 68.6% of non-recalls (n = 15,459). In a sub-analysis of IFSAC, reporting proved even scarcer in further disaggregated subcategories, with 3.08% of outbreaks missing IFSAC Level 2 (n = 125 of 4058 outbreaks) and 20.92% of outbreaks missing IFSAC Level 3 (n = 823 of 3933 outbreaks) (Figure 4). Both recalls and non-recalls had limited missing records for the preparation (9.51% and 5.01%, respectively) and consumption location (10.16% and 5.23%, respectively) of contaminated foods. We found that 58.3% of recalls (n = 164) were the result of multistate exposure outbreaks compared to only 1.70% of non-recalls (n = 382; Table 4). The univariate analyses demonstrated that the odds of multistate exposure outbreaks resulting in a recall were 24.75 times (18.87, 32.55; p < 0.001) that of single-state exposure outbreaks—the single-most influential risk factor found. Similarly, 30.16% of recalls (n = 92) were associated with plant food products, compared to 5.36% of non-recalls (n = 1205); the odds of plant foods resulting in recall were 74% higher (OR = 1.74, 1.31, 2.31; p < 0.001) than outbreaks associated with land animals or their byproducts. In contrast, only 2.62% of recalls (n = 8) occurred within the preparation supply chain stage, whereas 47.21% of recalls (n = 144) occurred within the before-preparation stage. We found that the odds of recall following a preparation-stage outbreak were 95% lower (OR = 0.05, 0.01, 0.11; p < 0.001) than a recall following a before-preparation-stage outbreak. Norovirus and Salmonella accounted for 28.74% (n = 6550) and 13.04% (n = 2972) of all outbreaks. Salmonella, E. coli, and Listeria outbreaks accounted for 33.44% (n = 102), 22.62% (n = 69), and 9.84% (n = 30) of recalls compared to only 12.76% (n = 2870), 2.68% (n = 603), and 0.30% (n = 67) of non-recalls, respectively. The odds of Salmonella-, E. coli-, and Listeria-associated outbreaks resulting in a recall were 1.91 times (1.46, 2.49), 5.27 times (3.86, 7.12), and 16.83 times (9.79, 28.79) that of non-Salmonella, non–E. coli, and non-Listeria outbreaks, respectively (p < 0.001). In contrast, norovirus outbreaks accounted for only 8.20% of recalls (n = 25) with 57% lower odds (OR = 0.43, 0.26, 0.67) of resulting in a recall compared to non-norovirus outbreaks (p < 0.001). The odds of scombroid-poisoning-associated outbreaks resulting in a recall was 46% lower (OR = 0.54, 0.29, 0.90) compared to non-scombroid-poisoning-associated outbreaks. Locations where contaminated foods were prepared had nearly identical patterns with respect to recall status as with consumption locations. We found that 17.70% (n = 54) and 18.69% (n = 57) of recalls had contaminated foods prepared or consumed, respectively, in multiple locations compared to only 7.16% (n = 1611) and 4.41% (n = 991) of non-recalls. Outbreaks with multiple locations for preparation and consumption had odds of 3.59 times (2.57, 4.93) and 4.75 times (3.39, 6.56), respectively, that of single preparation or consumption location outbreaks to result in a recall. Similarly, we found that 22.30% (n = 68) and 40.33% (n = 123) of recalls were either prepared or consumed at the home, respectively. Outbreaks with at-home preparation and consumption had odds of 1.36 times (1.00, 1.83) and 2.11 times (1.62, 2.73), respectively, that of outbreaks with away-from-home preparation or consumption to result in a recall. In contrast, outbreaks with food preparation or consumption at restaurants had 80–82% lower odds (0.11, 0.32) of resulting in a recall compared to non-restaurant outbreaks. We found similar patterns when examining the combined effect of all risk factors, with fully adjusted multivariate models having the lowest reported AIC values (Table 5; Supplementary Table S3). Irrespective of contaminant etiology, multistate exposure outbreaks had odds that were 11.00–13.50 times (7.00, 21.60) that of single-state exposure outbreaks to result in recall (p < 0.001). In contrast, outbreaks where supply chain contamination occurred in the preparation and unknown stages had 93–97% and 53–62% lower odds, respectively, of resulting in a recall (p < 0.05) compared to the before-preparation stage. Though outbreaks associated with other foods had significantly greater odds to result in recall compared to land animals, we assumed that the results were spurious due to small sample size within this category. Across contaminant etiologies, we found that Listeria- and norovirus-associated outbreaks had odds of 5.81 times (2.20, 16.40) and 4.93 times (2.39, 9.82) that of non-Listeria and non-norovirus outbreaks of resulting in recall, respectively (p < 0.001). Though of a lesser magnitude, E. coli–associated outbreaks had similarly higher odds of 1.86 (1.08, 3.18) resulting in a recall compared to non–E. coli outbreaks. We found no significant findings for either Salmonella- or scombroid-poisoning-associated outbreaks. 3.3. Comparing Risk Factors—Supply Chain Contamination Stage After comparing risk factors by recall status, we aimed to examine the likelihood of supply chain contamination in the preparation stage compared to the before-preparation stage among recalls. This analysis would have provided critical information on where within the supply chain recalls commonly occur to inform guidelines for improving outbreak analytics to enhance food traceability in accordance with the 2020 New Era of Smarter Food Safety Blueprint [21,22]. However, due to an insufficient sample size, we were unable to perform these logistic regression analyses. Of the 305 recalls identified in our study period, 144 recalls (47.21%) were identified in the before-preparation supply chain stage, while only 8 and 27 recalls (5.56% and 8.85%, respectively) were identified in the preparation or unknown stages. Among the before-preparation stage recalls, we found that 20.83% and 28.47% (n = 30 and n = 41, respectively) occurred within the pre-harvest and pre-processing stages, respectively, compared to 36.73% (n = 396) and 9.46% (n = 102) of the 1078 before preparation stage non-recalls. 4. Discussion Our study demonstrated the usability of CDC foodborne national surveillance records for investigating food recalls. In doing so, we described temporal trends of recalls for the past two decades and identified risk factors most likely to drive recall occurrence. We found that, while improving since the transition from eFORS to NORS, recall records and information on recall-related risk factors were largely incomplete. Approximately 41.1% (n = 9378) of the 22,792 outbreaks reported from 1 January 1998 to 31 December 2019 had a missing recall status. However, our findings suggest that outbreaks missing recall information occurred most frequently before November 2004, with substantial improvements after November 2004 and January 2009, following changes in data-collection methods and reporting standards. Furthermore, only 7.6% of outbreaks (51.5% of recalls and 7.0% of non-recalls) had non-missing records for all factors. These findings alone suggest that current publicly available surveillance records may be insufficient to adequately investigate the financial and human-health burdens of food recalls and foodborne/waterborne outbreaks more broadly. The New Era of Smarter Food Safety Blueprint aims to enhance interagency communications, design interoperable tools, and improve the timeliness of foodborne outbreak responses [21,22]. While acknowledging the importance of data quality, the Blueprint fails to promote interagency harmonization of existing recall surveillance systems between the FDA, USDA, and CDC. Both the FDA and USDA report traceback information on expenses, manufacturers, and volume of recalled foods not currently traced by the CDC [2,12,13], whereas the CDC traceback investigations identify supply chain contamination locations and where ill persons prepared and consumed contaminated foods. Harmonizing recall record data across these agencies could lead to more comprehensive estimates of healthcare and economic burdens, a better understanding of the impact food recalls has on food waste, and predictive analytics of foodborne outbreaks. Furthermore, reporting standards impede the ease of temporally or spatially aligning data across agencies or other environmental datasets. In contrast, eFORS and NORS provide comprehensive information on all foodborne outbreaks, thus enabling both descriptions of temporal trends and comparisons of recall-associated risk factors. However, 41.3% and 12.8% of recall-associated records lack information on supply chain contamination stage and IFSAC Level 1 grouping. Other food- and waterborne disease research has explored the supplementation of food-safety surveillance systems with hospitalization records for more precise and complete reporting of notifiable diseases [41,42]. By refocusing collaborative efforts toward interdepartmental data harmonization and considering triangulation of additional public-health-system data, these agencies can create more comprehensive and complete outbreak and recall surveillance vital to understanding food traceability at refined spatiotemporal scales. Though the volume and velocity of newly reported data increase annually, the CDC must continue to allocate fiscal and personnel resources to check the quality and accuracy of reported data. From January 1998 to November 2004, we found a consistent decrease in the reporting of outbreaks with missing recall information. However, we also found consistent reporting of non-recalls and five reported recalls, despite these records preceding the formal mandate to conduct recall traceback investigations. These findings may reflect the CDC’s attempt to modify historic records, as this will greatly improve the precision and accuracy of temporal trend and risk factor analyses on recalls in future studies. However, these findings may also reflect reporting anomalies requiring further investigation by CDC data-quality and accuracy personnel. Overall, our finding further illustrates the need for interagency collaboration on and greater attention to improving the quality of existing data amidst plans for strengthening surveillance capacity [21,22]. In fact, temporal trends from November 2004 to December 2019 already demonstrated the advantages of regulatory oversight and enforcement of improved data-reporting protocols. After the standardization of reporting of food recalls, found that outbreaks with missing recall information decreased at an accelerating rate, by ~25%/month, while non-recalls decreased at a decelerating rate, by ~10%/month. The expansion of surveillance capacity from eFORS to NORS brought further reductions in outbreaks with missing recall information at a decelerating rate, by ~13%/month. Such extensive reductions in missing recall information over time illustrate the importance of standardized outbreak surveillance reporting and improved usability of CDC surveillance data for investigating food recalls over time. In our study period, recalls increased by ~3%/month after the beginning of standard traceback investigations and by ~3%/month, again, after the transition to NORS. However, our trend analysis demonstrated that recalls steadily decreased by ~3%/month from January 2009 to December 2019. These trends reflect the improvements in food traceability in the supply chain and, thereby, the mitigating of food recalls after the enactment of the Food Safety Modernization Act in 2009 [43]. Signed into law in 2011, this legislation enabled the FDA to impose mandatory produce safety standards, controls, and inspections for potential hazards in food production, distribution, transport, and retail facilities [44]. Subsequent appendices to the law have mandated increased frequency of food safety inspections, distribution of supply chain records, and testing of food company products to improve early detection and warning of potential outbreaks [45]. Continued support for regulatory oversight and technological advancement on food traceability throughout the supply chain is critical for the continued reduction and prevention of recall events. Investigations of seasonality can also improve emergency and incident response coordination and enhance early warnings of foodborne outbreaks. In our prior work, we demonstrated stable seasonal patterns of foodborne illnesses in the United States and how these patterns can be examined and understood visually [8,27,46,47]. Across all studies, outbreak peak timing ranged from early July to late August for most enteric infections. More recently, we found that foodborne outbreak severity, measured using an 11-metric index score, similarly peaked in June–September (Simpson et al. (Personal Communication)). In this study, we found that foodborne outbreaks slightly preceded illness and outbreak severity peaks, as the maximum count of outbreaks occurred in April/May, with minimal counts in September/October. These findings suggest that outbreaks and illness may have synchronized seasonal patterns requiring further investigation to determine the exact lags between seasonal peaks of illnesses, outbreaks, outbreak severity, and recalls. The early onset of outbreaks further emphasizes the need for increased product testing, safety inspections, and toxicological-hazard screenings in April–June annually. However, we cannot discount that these temporal patterns may also reflect changes in annual resources and make the efficient identification of contaminated products through the harmonization of trace-back data even more crucial to food safety [48]. As 47.21% of recalls were associated with the before-preparation supply chain stage, our results suggest that food traceability operations and data reporting must more closely target pre-harvest and preprocessing techniques among producers [21,22]. This will improve data completeness and allow for a closer examination of food traceability, using CDC surveillance data, which we could not perform, due to sample size limitations. Improved monitoring of food safety earlier in the supply chain may reduce both the volume and severity of seasonal outbreaks and illnesses. Among risk factors, we found that multistate exposure outbreaks consistently had odds of ~10–15 times that of single-state exposure outbreaks to result in a recall. This underscores one of the Blueprint’s main directives of enhanced outbreak responsiveness, rapid traceback deployment, and strengthened root-cause analyses to identify the location of outbreaks and recalls [21,22]. In doing so, multistate exposures outbreaks can be more thoroughly contained to minimize the volume and severity of ill persons per recall. Furthermore, improved traceback investigations will better identify food distribution and retail pathways to mitigate the expansiveness of outbreaks within the supply chain. These efforts must more readily target outbreaks associated with E. coli, Listeria, and norovirus, as these thee etiologies had odds of ~1.5–6 times that of non–E. coli, non-Listeria, and non-norovirus outbreaks to result in food recalls. Our study was subject to several limitations. First, recall records were vulnerable to reporting bias of high-priority pathogens and food products that are most burdensome because they cause pathogen-related deaths. In 2015, just five pathogens (i.e., Salmonella, Toxoplasma gondii, Listeria monocytogenes, Campylobacter, and norovirus) caused 90% of the economic burden imposed by foodborne outbreaks [49]. Similarly, in this study, we found that Salmonella, Listeria and norovirus outbreaks predominantly resulted in a food recall. Second, further risk factor analysis by IFSAC Levels 2–5 and the sub-analysis of recalls in the preparation stage compared to the before-preparation stage were not possible, due to insufficient sample size. These analyses would have determined major food types or subtypes and supply chain contamination locations with higher probabilities of contamination resulting in a recall, thus informing the prioritization of traceback food products and locations by regulatory agencies. Next, the pathogen etiology and the location of preparation and consumption of contaminated foods were not originally mutually exclusive variables. By creating multi-pathogen and multi-location food preparation or consumption variables, we might have introduced potential multiplicity when comparing specific pathogen etiology, or location of preparation or consumption to their respective reference groups. Lastly, we paid sufficient attention to missing data and the structure of the missing data [50]. On the surface, we could handle missing data by using imputation; however, due to structural missingness, this could create bias. To our knowledge, this study is the first to utilize NORS recall record data to examine temporal trends, particularly before and after November 2004 and January 2009. Additionally, our comparison of risk factors between recalls and non-recalls highlighted existing biases in reporting influenced by available resources or outbreak healthcare and economic burden. Future directions should include more granular analyses of contaminant etiology and preparation and consumption locations and explore the relationship of these risk factors on foodborne outbreak severity. With the FDA taking a new approach to food safety via the New Era of Smarter Food Safety Blueprint, we urge food-safety and public-health agencies to collaborate more closely and standardize data-reporting protocols, thereby improving the spatiotemporal alignment and harmonization of publicly reported national surveillance databases on food recalls. 5. Conclusions Food recalls impose an extensive fiscal burden on the food economy in the United States, in addition to recall- and outbreak-associated foodborne illnesses. However, current national surveillance systems lack sufficient data quality and completeness for establishing precise and accurate early outbreak and recall detection and warnings. While data quality has improved over time, as the result of federal food-safety policies, further regulatory oversight is still needed. Future policy regulations must standardize timely and thorough data reporting of food recall and outbreak events to improve the traceability of food throughout the supply chain and responsiveness to multistate exposure outbreak events. Acknowledgments We acknowledge Ziming Dou for her support in data analyses. Supplementary Materials The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/ijerph19094947/s1. Table S1: Summary of model coefficients and diagnostics estimated by using segmented negative binomial regression analyses for linear, quadratic, and harmonic trends; Table S2: Estimated trend contribution of linear and quadratic trend, using Model 6 for outbreaks, recalls, non-recalls, and outbreaks missing recall information; and Table S3: Logistic regression results examining the likelihood of foodborne and waterborne outbreaks resulting in food recalls. Click here for additional data file. Author Contributions E.S., R.B.S. and L.E.S. contributed to the data extraction and statistical analyses performed to complete this manuscript; E.S., R.B.S. and Y.Z. contributed to the construction of visual aids and validation of both extracted data and statistical methods; E.S. contributed to the original drafting of the manuscript, while E.S. and R.B.S. contributed to the editing of manuscript text; E.N.N. contributed to the review and editing of the manuscript, as well as project supervision, administration, and funding acquisition. All authors have read and agreed to the published version of the manuscript. Funding The views represented in this article are solely those of the authors and do not represent the views of the United States Government, the Department of Defense, or the US Army. Additionally, this article does not represent the endorsement of any organization or association by the authors or any United States Government agency. The US Army Medical Center of Excellence supported Emily Sanchez via the Long Term Health and Education Training Program. This research is based upon work supported in part by the Office of the Director of National Intelligence (ODNI), Intelligence Advanced Research Projects Activity (IARPA), via 2017-17072100002 (Naumova–PI). The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies, either expressed or implied, of ODNI, IARPA, or the US Government. The US Government is authorized to reproduce and distribute reprints for governmental purposes, notwithstanding any copyright annotation therein. The United States Department of Agriculture (USDA) National Institute of Food and Agriculture (NIFA) Cooperative State Research, Education, and Extension Service Fellowship supported Ryan B. Simpson via grant award number 2020-38420-30724 (Naumova-PI). This manuscript was completed for a course enrolled in the National Science Foundation Innovations in Graduate Education Program grant (Award #1855886, Naumova-PI) entitled SOLution-oriented, STudent-Initiated, Computationally Enhanced (SOLSTICE) Training. This work, in part, was supported by the STOP Spillover project through the United States Agency for International Development (USAID). The contents are the responsibility of STOP Spillover and do not necessarily reflect the views of USAID or the United States Government. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement The Centers for Disease Control and Prevention (CDC) publicly report records for the electronic Foodborne Outbreak Reporting System (eFORS) and National Outbreak Reporting System (NORS) on their data dashboard [36]. We received more detailed records than provided publicly through a formal data request with the NORS Foodborne and Animal Contact Team. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Stacked time-series plots of monthly outbreaks, recalls, non-recalls, and outbreaks with missing recall information (top to bottom rows, respectively), as reported by the electronic Foodborne Outbreak Reporting System (eFORS) and National Outbreak Reporting System (NORS) in 1998–2019. We present segmented time-series model results, adjusting for linear trends only (Model 3; Panel (A)); linear and quadratic trends (Model 4; Panel (B)); and linear, quadratic, and harmonic trends (Model 5; Panel (C)). Within each plot, we report observed counts (gray bars) with fitted model results (red lines) and indicate critical periods by using different background colors. Figure 2 Boxplots of monthly counts of outbreaks (A), recalls (B), non-recalls (C), and outbreaks missing recall information (D), as reported by the electronic Foodborne Outbreak Reporting System (eFORS) and National Outbreak Reporting System (NORS) in 1998–2019. Each of the 3 critical periods is indicated by using a different box color. Grey dots represent outliers. Figure 3 A Sankey Diagram of the distribution of recalls and non-recalls with and without missing information across all risk factors examined by using Model 6. We report frequencies and percentages of foodborne and waterborne outbreaks associated with each risk factor for all 22,792 outbreaks reported by the electronic Foodborne Outbreak Reporting System (eFORS) and National Outbreak Reporting System (NORS) in 1998–2019. Blue and orange colors define recalls and non-recalls, respectively. Outbreaks with missing recall status or risk-factor information are defined with orange terminal nodes. We calculated percentages according to the frequency of observations available for each risk factor, which include recall status, single- or multistate exposure outbreak, supply chain contamination stage, Interagency Food Safety Analytics Collaboration (IFSAC) Level 1 food categorization, and etiology of contaminant. Other IFSAC 1 includes outbreaks associated with Other (n = 36), Unclassifiable (n = 33), Undetermined (n = 229), and Invalid (n = 3) food products. For contaminant etiology, we list the 5 etiologies of interest in our study (Salmonella, E. coli, norovirus, Listeria, and scombroid poisoning), as well as Other Etiology to account for contaminants not considered in our analyses. Figure 4 A Sankey Diagram of a sub-analysis examining the distribution of recalls with and without missing information across the Interagency Food Safety Analytics Collaboration (IFSAC) food categorization for Levels 1–3. We report frequencies and percentages of foodborne and waterborne outbreaks associated with each category for all 22,792 outbreaks reported by the electronic Foodborne Outbreak Reporting System (eFORS) and National Outbreak Reporting System (NORS) in 1998–2019. Blue and orange colors define outbreaks with non-missing and missing recall status information, respectively. We calculated percentages according to the frequency of observations in each level. Level 1 describes overarching food groups, including aquatic animals, land animals, plants, and other foodstuffs. Level 2 further categorizes groups into fish/shellfish, other aquatic animals, dairy, game, meat/poultry, eggs, oils/sugars, produce, grains/beans, and seeds/nuts. Level 3 provides more refined categories by specific food subtypes. ijerph-19-04947-t001_Table 1 Table 1 Critical periods used to examine trends and seasonality of CDC foodborne and waterborne outbreak surveillance records with segmented regression analyses. Period Start Date Duration Description 1 January 1998 82 Months Surveillance reporting begins under eFORS. 2 November 2004 50 Months The CDC’s Investigation of a Foodborne Outbreak revised to include if any food product was recalled from an outbreak. 3 January 2009 132 Months The CDC transitions reporting foodborne outbreak data from eFORS to NORS. ijerph-19-04947-t002_Table 2 Table 2 Summary statistics for monthly outbreaks, recalls, non-recalls, and outbreaks missing recall information, as reported by the electronic Foodborne Outbreak Reporting System (eFORS) and National Outbreak Reporting System (NORS) in 1998–2019. Statistic Full Study Period January 1998–December 2019 Period 1 January 1998–October 2004 Period 2 November 2004–December 2008 Period 3 January 2009–December 2019 Outbreaks Mean (95% CI) 86.33 (83.45, 89.33) 106.68 (84.62, 134.52) 92.53 (78.15, 109.68) 71.05 (34.47, 146.55) Median (LQR, UQR) 82.00 (68.00, 103.25) 109.00 (87.00, 121.00) 87.00 (79.00, 103.00) 69.00 (60.00, 81.00) Min, Max 23, 165 71, 156 61, 165 23, 116 L-Skew, L-Kurt 0.12, 0.11 0.02, 0.04 0.22, 0.19 0.09, 0.13 Recalls Mean (95% CI) 1.15 (1.00, 1.34) 0.06 (0.01, 0.23) 1.23 (0.42, 3.48) 1.79 (<0.01, 1566.96) Median (LQR, UQR) 1.00 (0.00, 2.00) 0.00 (0.00, 0.00) 1.00 (0.00, 2.00) 2.00 (1.00, 3.00) Min, Max 0, 6 0, 2 0, 4 0, 6 L-Skew, L-Kurt 0.30, 0.02 0.94, 0.85 0.20, <0.01 0.13, 0.11 Non-Recalls Mean (95% CI) 49.65 (44.69, 55.18) 7.49 (5.39, 10.40) 79.12 (54.33, 115.73) 64.03 (12.75, 103.11) Median (LQR, UQR) 57.00 (11.00, 73.00) 6.50 (5.00, 10.00) 78.50 (66.25, 97.25) 63.00 (54.00, 74.00) Min, Max 0, 162 0, 22 14, 162 20, 103 L-Skew, L-Kurt −0.01, 0.01 0.14, 0.09 -0.009, 0.20 0.07, 0.13 Outbreaks Missing Recall Information Mean (95% CI) 35.52 (29.62, 42.60) 99.03 (44.96, 219.89) 2.18 (0.01, 312.28) 5.12 (0.09, 284.86) Median (LQR, UQR) 7.00 (3.00, 81.00) 101.00 (83.25, 112.50) 1.00 (0.00, 3.00) 4.00 (3.00, 18.00) Min, Max 0, 146 57, 146 0, 146 0, 18 L-Skew, L-Kurt 0.39, 0.01 0.03, 0.07 0.75, 0.52 0.20, 0.13 95% CI, 95% confidence interval; LQR, lower quartile range; UQR, upper quartile range; min, minimum; max, maximum; L-Skew, L-skewness; L-Kurt, L-kurtosis. ijerph-19-04947-t003_Table 3 Table 3 Frequency and percentage of outbreaks with missing information by recall status (total, recall, and non-recall) and risk factor. We extracted data from the electronic Foodborne Outbreak Reporting System (eFORS) and National Outbreak Reporting System (NORS) in 1998–2019. Risk factors include location of residence for ill persons, etiology of contaminant, location of preparation and consumption of contaminated foods, Interagency Food Safety Analytics Collaboration (IFSAC) Level 1 food categorization, and supply chain contamination stage. We list risk factors in ascending order by percentage of outbreaks with missing information. Outbreak Recall (n = 305) Non-Recall (n = 22,487) Total (n = 22,792) Risk Factors n % n % n % Location of Outbreak Exposure 0 0.00 0 0.00 0 0.00 Etiology 10 3.28 7373 32.79 7383 32.39 Preparation Location 29 9.51 1127 5.01 1156 5.07 Consumption Location 31 10.16 1175 5.23 1206 5.29 IFSAC Level 1 39 12.79 15,420 68.57 15,459 67.83 Supply Chain Contamination Stage 126 41.31 17,166 76.34 17,292 75.87 ijerph-19-04947-t004_Table 4 Table 4 Frequency and percentage of foodborne and waterborne outbreaks, overall, by recall status (recall vs. non-recall) and by outbreak risk factors. We extracted data from the electronic Foodborne Outbreak Reporting System (eFORS) and National Outbreak Reporting System (NORS) in 1998–2019. Risk factors include location of outbreak exposure, supply chain contamination stage, Interagency Food Safety Analytics Collaboration (IFSAC) Level 1 food categorization, contaminant etiology, and locations of preparation and consumption of contaminated foods. We supplement descriptive statistics with odds ratio estimates (and 95% confidence intervals) from univariate logistic regressions. Outbreak Recall (n = 305) Non-Recall (n = 22,487) Total (n = 22,792) Univariate Odds Ratio Risk Factors n % n % n % 95% Conf. Int. Location of Outbreak Exposure Single-State 1 Multistate 164 53.77 382 1.70 546 2.40 24.75 (18.87, 32.55) a Supply Chain Contamination Stage Before Preparation 144 47.21 1078 4.79 1222 5.36 1 Preparation 8 2.62 3022 13.44 3030 13.29 0.05 (0.01, 0.11) a Unknown 27 8.85 1221 5.43 1248 5.48 0.88 (0.54, 1.39) IFSAC Level 1 Land Animals 111 36.39 2534 11.27 2645 11.60 1 Aquatic Animals 56 18.36 1506 6.70 1562 6.85 0.85 (0.61, 1.17) Plants 92 30.16 1205 5.36 1297 5.69 1.74 (1.31, 2.31) a Other Foods 6 1.97 181 0.80 187 0.82 0.76 (0.29, 1.60) Contaminant Etiology Non-Salmonella 193 63.28 12,244 54.45 12,437 54.57 1 Salmonella 102 33.44 2870 12.76 2972 13.04 1.91 (1.46, 2.49) a Non–E. coli 226 74.10 14,511 64.53 14,737 64.66 1 E. coli 69 22.62 603 2.68 672 2.95 5.27 (3.86, 7.12) a Non-Listeria 265 86.89 15,047 66.91 15,312 67.18 1 Listeria 30 9.84 67 0.30 97 0.43 16.83 (9.79, 28.79) a Non-Norovirus 270 88.52 8589 38.20 8859 38.87 1 Norovirus 25 8.20 6525 29.02 6550 28.74 0.43 (0.26, 0.67) a Non-Scombroid Poisoning 281 92.13 14,668 65.23 14,949 65.59 1 Scombroid Poisoning 14 4.59 446 1.98 460 2.02 0.54 (0.29, 0.90) b Preparation Location Non-Home 208 68.20 19,460 86.54 19,668 86.29 1 Home 68 22.30 1900 8.45 1968 8.63 1.36 (1.00, 1.83) b Non-Diner 247 80.98 17,597 78.25 17,844 78.29 1 Diner 29 9.51 3763 16.73 3792 16.64 0.78 (0.51, 1.17) Non-Restaurant 255 83.61 13,128 58.38 13,383 58.72 1 Restaurant 21 6.89 8232 36.61 8253 36.21 0.18 (0.11, 0.27) a Single Location 222 72.79 19,749 87.82 19,971 87.62 1 Multiple Locations 54 17.70 1611 7.16 1665 7.31 3.59 (2.57, 4.93) Consumption Location Non-Home 151 49.51 17,876 79.49 18,027 79.09 1 Home 123 40.33 3436 15.28 3559 15.62 2.11 (1.62, 2.73) a Non-Diner 249 81.64 17,872 79.48 18,121 79.51 1 Diner 25 8.20 3440 15.30 3465 15.20 0.77 (0.48, 1.16) Non-Restaurant 257 84.26 14,863 66.10 15,120 66.34 1 Restaurant 17 5.57 6449 28.68 6466 28.37 0.20 (0.11, 0.32) a Single Location 217 71.15 20,321 90.37 20,538 90.11 1 Multiple Locations 57 18.69 991 4.41 1048 4.60 4.75 (3.39, 6.56) a Superscripts indicate statistical significance at p < 0.001 (a), and p < 0.05 (b). ijerph-19-04947-t005_Table 5 Table 5 Logistic regression results examining the likelihood of foodborne and waterborne outbreaks resulting in food recalls, as reported by the electronic Foodborne Outbreak Reporting System (eFORS) and National Outbreak Reporting System (NORS) in 1998–2019. We selected risk factors according to univariate logistic regression results and added factors in a stepwise order. Risk factors include multistate exposure outbreaks (reference: single-state exposure outbreaks), supply chain contamination stage (reference: before-preparation stage), IFSAC Level 1 food categories (reference: land animals), and presence of a contaminant etiology (reference: absence of or unknown etiology). We report fully specified models for 5 contaminant etiologies, namely Salmonella, E. coli, Listeria, norovirus, and scombroid poisoning associated outbreaks. We report the odds ratio estimates (and 95% confidence intervals), Akaike’s Information Criterion (AIC), and the number of observations per model. Risk Factors Salmonella E. coli Listeria Norovirus Scombroid Poisoning Location of Outbreak Exposure Multistate 13.00 (8.17, 21.00) a 11.60 (7.39, 18.50) a 11.00 (7.00, 17.60) a 13.50 (8.59, 21.60) a 12.60 (8.05, 20.10) a Supply Chain Contamination Stage Preparation 0.06 (0.02, 0.17) a 0.07 (0.02, 0.19) a 0.06 (0.02, 0.16) a 0.03 (0.01, 0.10) a 0.06 (0.02, 0.17) a Unknown 0.44 (0.25, 0.77) b 0.47 (0.26, 0.82) b 0.38 (0.20, 0.68) b 0.41 (0.23, 0.72) b 0.43 (0.24, 0.75) b IFSAC Level 1 Aquatic Animals 0.58 (0.35, 0.95) b 0.69 (0.42, 1.13) 0.64 (0.40, 1.03) 0.46 (0.27, 0.77) b 0.54 (0.32, 0.90) b Plants 0.71 (0.43, 1.14) 0.72 (0.44, 1.16) 0.72 (0.43, 1.17) 0.67 (0.41, 1.08) 0.71 (0.43, 1.15) Other Foods 6.45 (1.17, 29.2) b 6.94 (1.26, 31.50) b 7.16 (1.31, 32.60) b 4.91 (0.90, 22.30) b 6.53 (1.18, 29.70) b Etiology Etiology Present 0.84 (0.53, 1.31) 1.86 (1.08, 3.18) b 5.81 (2.20, 16.40) a 4.93 (2.39, 9.82) a 1.76 (0.77, 3.73) Modeling Diagnostics AIC 782.78 778.41 770.48 766.33 781.52 Observations 1596 1596 1596 1596 1596 Superscripts indicate statistical significance at p < 0.001 (a), and p < 0.05 (b). Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Hoffmann S. Ahn J.-W. Economic Cost of Major Foodborne Illnesses Increased $2 Billion from 2013 to 2018 Available online: https://www.ers.usda.gov/amber-waves/2021/april/economic-cost-of-major-foodborne-illnesses-increased-2-billion-from-2013-to-2018/ (accessed on 1 April 2022) 2. FSIS Food Recalls Available online: https://www.fsis.usda.gov/wps/portal/fsis/topics/food-safety-education/get-answers/food-safety-fact-sheets/production-and-inspection/fsis-food-recalls/fsis-food-recalls (accessed on 11 November 2020) 3. Recall: The Food Industry’s Biggest Threat to Profitability. Food Safety Magazine 11 October 2012 Available online: https://www.food-safety.com/articles/2542-recall-the-food-industrys-biggest-threat-to-profitability (accessed on 2 January 2022) 4. Product Recalls: How Recalls Damage Your Brand and Bottom Line Available online: https://marel.com/en/news/food-recalls-true-cost (accessed on 2 January 2022) 5. SmartSense The True Costs of Food Recalls 2019 2022 Available online: https://blog.smartsense.co/costs-of-food-recalls (accessed on 2 January 2022) 6. Heneghan C. More than Money: What a Recall Truly Costs Available online: https://www.fooddive.com/news/more-than-money-what-a-recall-truly-costs/426855/ (accessed on 2 January 2022) 7. Koeris M. The True Costs You Endure during a Food Recall Available online: https://foodsafetytech.com/column/the-true-costs-you-endure-during-a-food-recall/ (accessed on 2 January 2022) 8. Simpson R.B. Zhou B. Naumova E.N. Seasonal synchronization of foodborne outbreaks in the United States, 1996–2017 Sci. Rep. 2020 10 17500 10.1038/s41598-020-74435-9 33060743 9. Sivapalasingam S. Friedman C.R. Cohen L. Tauxe R.V. Fresh produce: A growing cause of outbreaks of foodborne illness in the United States, 1973 through 1997 J. Food Prot. 2004 67 2342 2353 10.4315/0362-028X-67.10.2342 15508656 10. White A. Cronquist A. Bedrick E.J. Scallan E. Food source prediction of shiga toxin-producing Escherichia coli outbreaks using demographic and outbreak characteristics, United States, 1998–2014 Foodborne Pathog. Dis. 2016 13 527 534 10.1089/fpd.2016.2140 27526280 11. Food Safety: USDA and FDA Need to Better Ensure Prompt and Complete Recalls of Potentially Unsafe Food United States Government Accountability Office Washington, DC, USA 2004 12. Recalls, Market Withdrawals, & Safety Alerts Available online: https://www.fda.gov/safety/recalls-market-withdrawals-safety-alerts (accessed on 17 October 2021) 13. Recalls & Public Health Alerts Available online: https://www.fsis.usda.gov/recalls (accessed on 8 October 2020) 14. Magno F. Managing product recalls: The effects of time, responsible vs. opportunistic recall management and blame on consumers’ attitudes Procedia-Soc. Behav. Sci. 2012 58 1309 1315 10.1016/j.sbspro.2012.09.1114 15. Liao C. Luo Y. Zhu W. Food safety trust, risk perception, and consumers’ response to company trust repair actions in food recall crises Int. J. Environ. Res. Public Health 2020 17 1270 10.3390/ijerph17041270 32079171 16. Gorton A. Stasiewicz M.J. Twenty-two years of U.S. meat and poultry product recalls: Implications for food safety and food waste J. Food Prot. 2017 80 674 684 10.4315/0362-028X.JFP-16-388 28294687 17. Routh J.A. Pringle J. Mohr M. Bidol S. Arends K. Adams-Cameron M. Hancock W.T. Kissler B. Rickert R. Folster J. Nationwide outbreak of multidrug-resistant Salmonella Heidelberg infections associated with ground turkey: United States, 2011 Epidemiol. Infect. 2015 143 3227 3234 10.1017/S0950268815000497 25865382 18. Centers for Disease Control and Prevention (CDC) Multistate outbreak of Salmonella typhimurium infections associated with eating ground beef—United States, 2004 MMWR Morb. Mortal. Wkly. Rep. 2006 55 180 182 16498383 19. Outbreak of invasive listeriosis associated with the consumption of hog head cheese—Louisiana, 2010 MMWR Morb. Mortal. Wkly. Rep. 2011 60 401 405 21471946 20. Summary of FSIS Government Micorbiological Sampling Programs Frequencies Available online: https://www.fsis.usda.gov/sites/default/files/import/Microbiological-Frequencies.pdf (accessed on 3 April 2022) 21. Tracking and Tracing of Food Available online: https://www.fda.gov/food/new-era-smarter-food-safety/tracking-and-tracing-food (accessed on 8 October 2021) 22. New Era of Smarter Food Safety Blueprint Available online: https://www.fda.gov/food/new-era-smarter-food-safety/new-era-smarter-food-safety-blueprint (accessed on 17 October 2021) 23. Food Safety: FDA’s Food Advisory and Recall Process Needs Strengthening United States Government Accountability Office Washington, DC, USA 2012 24. Page E.T. Trends in Food Recalls: 2004–2013 Economic Information Bulletin Number 191 U.S. Department of Agriculture, Economic Research Service Washington, DC, USA 2018 10.22004/ag.econ.276244 25. O’Brien E. Xagoraraki I. Understanding temporal and spatial variations of viral disease in the US: The need for a one-health-based data collection and analysis approach One Health 2019 8 100105 10.1016/j.onehlt.2019.100105 31709295 26. Sher A.A. Ashraf M.A. Mustafa B.E. Raza M.M. Epidemiological trends of foodborne Campylobacter outbreaks in the United States of America, 1998–2016 Food Microbiol. 2021 97 103751 10.1016/j.fm.2021.103751 33653524 27. Naumova E.N. Jagai J.S. Matyas B. DeMaria A. MacNeill I.B. Griffiths J.K. Seasonality in six enterically transmitted diseases and ambient temperature Epidemiol. Infect. 2007 135 281 292 10.1017/S0950268806006698 17291363 28. Koluman A. KiKiCi A. Kahraman T. iNCiLi G.K. Food safety and climate change: Seasonality and emerging food borne pathogens J. Gastroenterol. Res. 2017 1 24 29 29. GAO-18-272 Food Safety: USDA Should Take Further Action to Reduce Pathogens in Meat and Poultry Products U.S. Government Accountability Office Washington, DC, USA 2018 30. GAO-21-23 Food Safety: CDC Could Further Strengthen Its Efforts to Identify and Respond to Foodborne Illnesses U.S. Government Accountability Office Washington, DC, USA 2020 31. Centers for Disease Control and Prevention (CDC) National Center for Immunization and Respiratory Diseases, D.o.V.D. about NORS Available online: https://www.cdc.gov/nors/about.html (accessed on 30 September 2021) 32. Using NORS Data Available online: https://www.cdc.gov/nors/data/using-nors.html (accessed on 31 March 2022) 33. Investigation of a Foodborne Outbreak CDC 52.13 Revised November, 2004 CDC Washington, DC, USA 2004 34. Qiu Q. Dewey-Mattia D. Subramhanya S. Cui Z. Griffin P.M. Lance S. Lanier W. Wise M.E. Crowe S.J. Food recalls associated with foodborne disease outbreaks, United States, 2006–2016 Epidemiol. Infect. 2021 149 e190 10.1017/S0950268821001722 34275497 35. Marus J.R. Magee M.J. Manikonda K. Nichols M.C. Outbreaks of Salmonella enterica infections linked to animal contact: Demographic and outbreak characteristics and comparison to foodborne outbreaks-United States, 2009–2014 Zoonoses Public Health 2019 66 370 376 10.1111/zph.12569 30821071 36. Forms & Guidance Available online: https://www.cdc.gov/nors/forms.html (accessed on 8 December 2021) 37. National Outbreak Reporting System Guidance Appendix B: Reporting Multistate Exposure and Residency Outbreaks Centers for Disease Control and Prevention Atlanta, GA, USA 2017 38. Code for Federal Regulations Title 21—Food and Drugs U.S. Food & Drug Administration Silver Spring, MD, USA 2018 39. IFSAC Food Categorization Scheme Available online: https://www.cdc.gov/foodsafety/ifsac/projects/food-categorization-scheme.html (accessed on 8 December 2021) 40. National Outbreak Reporting System (NORS) Guidance Centers for Disease Control and Prevention Atlanta, GA, USA 2017 41. Mor S.M. DeMaria A. Naumova E.N. Hospitalization records as a tool for evaluating performance of food- and water-borne disease surveillance systems: A Massachusetts case study PLoS ONE 2014 9 e93744 10.1371/journal.pone.0093744 24740304 42. Alarcon Falconi T.M. Cruz M.S. Naumova E.N. The shift in seasonality of legionellosis in the USA Epidemiol. Infect. 2018 146 1824 1833 10.1017/S0950268818002182 30099976 43. H.R.875—Food Safety Modernization Act of 2009 U.S. Government Publishing Office Washington, DC, USA 2009 44. Hoffmann S.U.S. Food Safety Policy Enters a New Era Available online: https://www.ers.usda.gov/amber-waves/2011/december/us-food-safety-policy/ (accessed on 2 January 2022) 45. Full Text of the Food Safety Modernization Act U.S. Government Publishing Office Washington, DC, USA 2017 46. Chui K.K. Webb P. Russell R.M. Naumova E.N. Geographic variations and temporal trends of Salmonella-associated hospitalization in the U.S. elderly, 1991–2004: A time series analysis of the impact of HACCP regulation BMC Public Health 2009 9 447 10.1186/1471-2458-9-447 19958556 47. Castronovo D.A. Chui K.K. Naumova E.N. Dynamic maps: A visual-analytic methodology for exploring spatio-temporal disease patterns Environ. Health 2009 8 61 10.1186/1476-069X-8-61 20042115 48. Powell M.R. Crim S.M. Hoekstra R.M. Williams M.S. Gu W. Temporal patterns in principal Salmonella serotypes in the USA; 1996–2014 Epidemiol. Infect. 2018 146 437 441 10.1017/S0950268818000195 29436316 49. Hoffmann S. Maculloch B. Batz M. Economic Burden of Major Foodborne Illnesses Acquired in the United States United States Department of Agriculture Economic Research Service Washington, DC, USA 2015 50. Zhang Y. Simpson R.B. Sallade L.E. Sanchez E. Monahan K.M. Naumova E.N. Evaluating completeness of foodborne outbreak reporting in the United States, 1998–2019 Int. J. Environ. Res. Public Health 2022 19 2898 10.3390/ijerph19052898 35270590

PMC009xxxxxx/PMC9099669.txt

==== Front Int J Mol Sci Int J Mol Sci ijms International Journal of Molecular Sciences 1422-0067 MDPI 10.3390/ijms23094565 ijms-23-04565 Review Demystifying the Neuroprotective Role of Neuropeptides in Parkinson’s Disease: A Newfangled and Eloquent Therapeutic Perspective Behl Tapan 1* https://orcid.org/0000-0002-6001-125X Madaan Piyush 1 Sehgal Aayush 1 Singh Sukhbir 1 https://orcid.org/0000-0002-1078-2478 Makeen Hafiz A. 2 Albratty Mohammed 3 Alhazmi Hassan A. 34 https://orcid.org/0000-0002-6876-7270 Meraya Abdulkarim M. 2 https://orcid.org/0000-0003-3236-1292 Bungau Simona 56* Lee Bae Hwan Academic Editor 1 Chitkara College of Pharmacy, Chitkara University, Rajpura 140401, India; piyushmadaan4811@gmail.com (P.M.); aayushsehgal00@gmail.com (A.S.); sukhbir.singh@chitkara.edu.in (S.S.) 2 Pharmacy Practice Research Unit, Department of Clinical Pharmacy, College of Pharmacy, Jazan University, Jazan 45142, Saudi Arabia; hafiz@jazanu.edu.sa (H.A.M.); ameraya@jazanu.edu.sa (A.M.M.) 3 Department of Pharmaceutical Chemistry, College of Pharmacy, Jazan University, Jazan 45142, Saudi Arabia; malbratty@jazan.edu.sa (M.A.); hasalhazmi@gmail.com (H.A.A.) 4 Substance Abuse and Toxicology Research Center, Jazan University, Jazan 45142, Saudi Arabia 5 Department of Pharmacy, Faculty of Medicine and Pharmacy, University of Oradea, 410028 Oradea, Romania 6 Doctoral School of Biomedical Sciences, University of Oradea, 410028 Oradea, Romania * Correspondence: tapanbehl31@gmail.com (T.B.); simonabungau@gmail.com (S.B.) 20 4 2022 5 2022 23 9 456522 3 2022 18 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Parkinson’s disease (PD) refers to one of the eminently grievous, preponderant, tortuous nerve-cell-devastating ailments that markedly impacts the dopaminergic (DArgic) nerve cells of the midbrain region, namely the substantia nigra pars compacta (SN-PC). Even though the exact etiopathology of the ailment is yet indefinite, the existing corroborations have suggested that aging, genetic predisposition, and environmental toxins tremendously influence the PD advancement. Additionally, pathophysiological mechanisms entailed in PD advancement encompass the clumping of α-synuclein inside the lewy bodies (LBs) and lewy neurites, oxidative stress, apoptosis, neuronal-inflammation, and abnormalities in the operation of mitochondria, autophagy lysosomal pathway (ALP), and ubiquitin–proteasome system (UPS). The ongoing therapeutic approaches can merely mitigate the PD-associated manifestations, but until now, no therapeutic candidate has been depicted to fully arrest the disease advancement. Neuropeptides (NPs) are little, protein-comprehending additional messenger substances that are typically produced and liberated by nerve cells within the entire nervous system. Numerous NPs, for instance, substance P (SP), ghrelin, neuropeptide Y (NPY), neurotensin, pituitary adenylate cyclase-activating polypeptide (PACAP), nesfatin-1, and somatostatin, have been displayed to exhibit consequential neuroprotection in both in vivo and in vitro PD models via suppressing apoptosis, cytotoxicity, oxidative stress, inflammation, autophagy, neuronal toxicity, microglia stimulation, attenuating disease-associated manifestations, and stimulating chondriosomal bioenergetics. The current scrutiny is an effort to illuminate the neuroprotective action of NPs in various PD-experiencing models. The authors carried out a methodical inspection of the published work procured through reputable online portals like PubMed, MEDLINE, EMBASE, and Frontier, by employing specific keywords in the subject of our article. Additionally, the manuscript concentrates on representing the pathways concerned in bringing neuroprotective action of NPs in PD. In sum, NPs exert substantial neuroprotection through regulating paramount pathways indulged in PD advancement, and consequently, might be a newfangled and eloquent perspective in PD therapy. Parkinson’s disease neuropeptides substance P ghrelin neuropeptide Y neurotensin pituitary adenylate cyclase-activating polypeptide neuroprotective action ==== Body pmc1. Introduction Parkinson’s disease (PD), a clinical state portrayed around 20.5 decades ago by an English surgeon named James Parkinson as paralysis agitans, is currently acknowledged as the second leading, mystifying, and incapacitating neurodegenerative disease in older individuals [1]. The condition is represented by a tetrad of cardinal manifestations, viz., rigor, tremor, postural deformities, and bradykinesia [2,3]. The aforenamed pivotal manifestations are ascribable to the deterioration of dopaminergic (DArgic) nerve cells in the substantia nigra pars compacta (SN-PC), a region pinpointed in the midbrain [4]. With a tremendous upsurge in the prevalence and incidence rates across the different regions of the nation, PD is emerging as a grievous age-associated and intricate malady [5,6]. Owing to the profound nerve cell protection exhibited by estrogen in females, the females display a de-escalated possibility of encountering PD comparably to males [7]. Although the exact etiopathogenesis of the malady remains perplexing, multifaceted, and vague, extensive data robustly propounds that aging, genetic predisposition, and subjection to environmental toxins unitedly partake in the progression of the malady [8,9,10,11,12,13,14]. The pathophysiological processes embroiled in PD comprehends the clumping of α-synuclein inside the lewy bodies (LBs) and lewy neurites, oxidative stress, apoptosis, neuronal-inflammation, and abnormalities in the operation of mitochondria, autophagy lysosomal pathway (ALP), and ubiquitin–proteasome system (UPS). However, the interrelationship among these processes is still inexplicit [8,15]. Up to the present time, the treatment with the assistance of dopamine (DA) precursor (levodopa), DA agonists, catechol-O-methyltransferase (COMT) inhibitors, and monoamine oxidase B (MAO-B) inhibitors principally focuses on the mitigation of PD-related manifestations, but hitherto no therapeutic candidate has been indicated to totally abolish the progression of the ailment [16,17,18]. Neuropeptides (NPs) are designated as tiny, protein-comprising additional messenger substances that are fundamentally generated and liberated by nerve cells inside the entire nervous system, comprehending the central nervous system (CNS) and the peripheral nervous system (PNS) [19,20]. NPs are synthesized in the cell body from their large protein precursors denominated as prepropeptides (which are synthesized on palade granules at the endoplasmic reticulum and processed by means of Golgi complex). The NPs are principally transcripted and translated from the prepropeptides genes. Further, the activation of prepropeptides carried by proteases/proteinases (peptide bonds hydrolyzing biocatalysts) culminates in the conversion of prepropeptides into propeptides, and at last, following the stimulation of converting biocatalysts, NPs are procured [21,22,23]. Proteolytic processing has been reported to significantly partake in the activation, partial inactivation, or inactivation of the modulatory peptides, for instance, NPs. Proteases, otherwise denominated as proteinases, carry out the breakdown, and as a consequence, might activate, inactivate, or liberate other proteins/peptides [24,25]. These modulatory proteases are fundamentally pinpointed on the surface of the cell or are liberated by the cells. Although the duo forms of proteases tend to perform identical operations, the existing literature has elucidated that cell-surface proteases possess significantly greater regulatory and specialized operations in comparison to those proteases which are liberated by the cells [25]. The cell-surface proteases exert their action by carrying out deterioration of the duo, i.e., the bioactive peptides, and the cellular operations. Apart from the proteolytic deterioration, and conjugation and oxidation reactions, the eradication via filtration/diffusion, is of paramount importance [25]. The majority of the cell-surface proteases have been originally marked as the clearing biocatalysts despite the fact that they possess the aptitude to break the peptides having not more than eighty residues. Currently, they are considered as the regulatory proteases (for instance, angiotensin-converting enzyme (ACE), endothelin-converting enzyme (ECE), neutral endopeptidase (NEP), and dipeptidyl peptidase IV (DPP IV)) that hold the enormous aptitude to modulate the activation/inactivation of NPs [23]. Following their synthesis, NPs are stored/packaged in the large and dense vesicles, and finally their liberation is facilitated by means of an expulsion process termed exocytosis (following depolarization of the cell) [21,22,23]. Thereafter, NPs undergo interaction with receptors, namely G-protein coupled receptors (GPCRs), in order to instigate their physiological actions and regulate nerve cell operation [19,26]. The NPs and their seven transmembrane domain (7TM) receptors/GPCRs are pinpointed ubiquitously in the body, and they usually exist in amalgamation with classic neurotransmitters [18,20,21,27,28]. NPs consequentially partake in the modulation of the immune system, biological equilibrium (for instance, biotransformation of blood sugar, blood pressure, equilibrium in the water content, feeding behavior, stress reaction, and pain), and neuronal protection [29]. Presently, numerous NPs have been elucidated to exhibit substantial neuronal protection in both in vivo and in vitro models of PD, for instance, substance P (SP) [30], ghrelin [31,32], neuropeptide Y (NPY) [33], neurotensin [34], pituitary adenylate cyclase-activating polypeptide (PACAP) [35], nesfatin-1 [36], and somatostatin (SST) [37] via suppressing apoptosis, cytotoxicity, oxidative stress, autophagy, inflammation, nerve cell toxicity, microglia stimulation, attenuating disease-associated manifestations, and stimulating chondriosomal bioenergetics. In the current scrutiny, the authors attempt to enlighten the linkage between the aforenamed NPs and PD, and elucidate the pathways by means of which these NPs contribute to significant nerve cell protection in PD. A comprehensive examination has been carried out by means of existing literature, i.e., both review and research articles, which were searched through esteemed and well-renowned medical databases, for instance, PubMed, MEDLINE, EMBASE, Frontier, etc., by employing particular keywords in the theme of our paper. The outcome is an explanatory work that would be a tremendously valuable resource for upcoming papers in this respective discipline. 2. Understanding the Etiopathogenic Pathways Underlying Parkinson’s Disease PD, a mystified, multifaceted, and debilitating malady, is depicted by the devastation of DArgic nerve cells inside the SN-PC (pinpointed in the midbrain region), which eventually contributes to DA scantiness in the striatal region. It has been elucidated that deposition of a protein termed α-synuclein within the LBs and lewy neurites is thought to be the characteristic neuropathogenic hallmark of PD [38]. Generally, PD is marked as a motor system ailment exhibiting a quadriad of imperative manifestations, comprehending rigor (stiffness), tremor (shaking in the hands, feets, and legs), postural deformities (abnormal balance and body posture), and bradykinesia (slowed/difficult movement). Nonmotor manifestations, such as urinary and sexual abnormalities, sleep disturbances, psychosis, dementia, anxiety, apathy, depression, constipation, and erectile dysfunction, are also encountered by individuals suffering from PD; however, they are somewhat less apparent in comparison to motor manifestations (Figure 1) [39]. 2.1. Understanding the Etiological Processes Underlying Parkinson’s Disease Analogous to other frequently emerging age-associated nerve cell deteriorating ailments, PD also commences owing to the amalgamation of the trio, namely aging, genetic predisposition, and subjection to environmental toxins, and is recklessly impacting the global economy [40,41]. According to a meta-analysis, the incidence of PD is expanding at an alarming rate in advanced and well-established parts of the nation, and it has been revealed to escalate abruptly with aging [42]. Another investigation has demonstrated that people falling under the age range of 30–40 years rarely experience PD, while it hits nearly 2% of people going above 60–70 years of age grade, and nearly 5% of people falling above the age range of 80–90 years nationwide (Figure 1) [43]. Even the gender differences enormously impact the PD emergence, i.e., males are comparably more prone to PD than females because of the nerve cell safeguarding action of estrogen in females in the initial stages of the disease [44,45]. There are no further gender differences in PD following its progression to the disastrous or later/severe stages, although estrogen partakes in bringing considerable neuronal protection in PD, and it renders no safeguarding effects following the commencement of clinical manifestations [45]. Apart from aging, the duo, namely genetic profile and subjection to environmental toxins, are recognized as fundamental participants indulged in the commencement and advancement of different forms of PD, by triggering DArgic nerve cell demise, as illustrated in Figure 2. Over the last few years, a profusion of exploration in the domain of PD has culminated in the elucidation that nearly 5–10% of the delayed commencement sorts of the PD are immensely related with genetic mutations [46]. To date, numerous genes have been reported to be implicated in the PD evolution, encompassing α-synuclein (SNCA) [46,47,48] Parkin RBR E3 ubiquitin–protein ligase (Parkin) [49], ubiquitin carboxy (C)-terminal hydrolase L1 (UCHL1) [50], PTEN-induced kinase 1 (PINK1) [51], protein deglycase (DJ-1) [52], leucine-rich repeat kinase 2 (LRRK2) [53], glucocerebrosidase (GBA) [54], vacuolar protein sorting 35 (VPS35) [55], neuronal P-type adenosine triphosphate (ATP)ase gene (ATP13A2) [56], high temperature requirement A2 (HTRA2) [57], and synaptojanin 1 (SYNJ1) [58]. In addition, a plethora of corroborations profoundly indicate that exposure to nerve cell toxic agents (1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) [38,51,52], and 6-hydroxy DA (6-OHDA) [38,51]), pesticides (paraquat) [46,59,60], rotenone [46,59,60], dieldrin [61], zineb [46,62], ziram [46,63] and thiram [64], fungicides (nabam [65], and maneb [65]), and solvents (methanol (CH3OH) [66], perchloroethylene (PERC) [66,67], trichloroethylene (TCE) [66,67], and carbon tetrachloride (CCl4) [66,67], considerably escalate the expansion of PD (via triggering the demise of DArgic nerve cells), and the concomitant range of motor as well as nonmotor manifestations. In addition, a recent study has revealed that transition metals, in particular iron (Fe) and copper (Cu), via elevating the oxidative stress, lipid peroxidation, and clumping of α-synuclein inside the LBs, substantially contribute to the progression of PD [68]. 2.2. Understanding the Pathogenic Processes Underlying Parkinson’s Disease Even though the pathogenesis underlying PD remains mystified and equivocal, a myriad of scrutinizations in the foregone years led to the unfolding of several pathways of paramount importance that markedly participate in the PD progression. These comprehend oxidative stress, dysfunction of the ALP, abnormality in the UPS, mitochondrial devastation, nerve cell inflammation, clumping of α-synuclein, and programmed cell death/apoptosis (portrayed in Figure 2) [8,15,41,46,59,68]. 2.2.1. Oxidative Stress and Parkinson’s Disease Mounting corroborations elucidated that, amidst the multiple processes deeply entangled in the pathogenesis of PD, oxidative stress has reaped a noteworthy prominence. Pursuant to one of the widely acknowledged theories, namely the free radical theory, or otherwise denominated as the oxidative stress theory (which was propounded by a renowned biochemist named Denham Harman during the mid-20th century), the chondriosome/power plants of the cell (mitochondria) are reckoned to be a “hotspot” for degenerative events [69]. The investigators set forth that an anomalous complex-I operation inside the chondriosome has been detected in the case of PD, that significantly intercedes with the ATP formation within the cells, and in turn culminates in cellular demise [70]. Further, the nitrogen-comprising low molar mass compounds denominated as biogenic amines (BAs) pinpointed inside the brain, for instance, DA and 5-hydroxytryptamine (5-HT)/serotonin, have been expounded to exhibit consequential antioxidant/free-radical scavenging abilities [71]. Howbeit, the DA fragmentation precipitated by MAO-B, in amalgamation with oxygen (O2) existent in the stable ground state, markedly contributes to the generation of eminently reactive, pernicious, and unstable substances denominated as oxygen radicals/reactive oxygen species (ROS) [72]. Furthermore, an inspection of human brain autopsies has displayed a consequential plummet in the quantities of an imperative tripeptide antioxidant designated as glutathione (GSH), an upsurge in the Fe and malondialdehyde (MDA) quantities, and oxidative harm to macromolecules (lipids, and polypeptides) [73,74,75]. Another study has spotted that those individuals with PD exhibit plummeted functioning of a biocatalyst with antioxidant abilities, termed catalase (CAT), and an upsurge in the lipid hydroperoxides (LOOH), MDA, and the functioning of a biocatalyst possessing antioxidant abilities, designated as superoxide dismutase (SOD) [76]. In accordance with these examinations, MDA is speculated to be the significant biomarker of the malady, while SOD and LOOH are profoundly related to delayed manifestations of the malady. Moreover, several investigations on human beings and animal models with PD have elucidated that a signaling molecule, namely nitrogen monoxide/nitric oxide (NO), significantly participates in multiple pathogenic mechanisms, viz., inflammatory processes, oxidative damage, deoxyribonucleic acid (DNA) devastation, excitotoxicity, S-nitrosylation of numerous proteins, and mitochondrial impairment, and eventually culminates into nerve cell deterioration [76,77,78]. The aforestated explorations markedly highlight the participation of oxidative stress in the PD progression. 2.2.2. Autophagy Lysosomal Pathway Dysfunction and Parkinson’s Disease Existing data has promulgated that, in the case of PD, numerous ALP-related components have been found to be considerably plummeted or undermined, which displays an immense resemblance to the UPS pathway results. Published literature has elucidated that fundamental and eminent protein constituent of the single phospholipid bilayer (lysosomal membrane), encompassing lysosome-associated membrane protein 1 (LAMP1) and lysosome-associated membrane protein 2A (LAMP2A), and heat shock proteins (HSPs), otherwise denominated as molecular chaperones, encompassing heat shock cognate protein 70 (HSC70) and hereditary spastic paraplegia type 35 (HSP35), were substantially plummeted during autopsy of substantia nigra (SN) of individuals experiencing PD [79,80]. In addition, myriad genes have been expounded to partake in the ALP, for instance, SNCA, DJ-1, GBA, LRRK2, PINK1, transmembrane protein 175 (TMEM175), cathepsin B (CTSB), cathepsin D (CTSD), sphingomyelin phosphodiesterase 1 (SMPD1); however, mutations in these genes can culminate in the dysfunctioning of the ALP, and finally contributes to PD evolution [79,81,82,83]. Pursuant to another study, the clumping of α-synuclein and tau is triggered by abnormal autophagic lysosomal breakdown [81]. These aforementioned corroborations profoundly imply that ALP dysfunction significantly contributes to the pathogenesis of PD. 2.2.3. Ubiquitin–Proteasome System Dysfunction and Parkinson’s Disease The available literature has revealed that abnormal operation of the UPS is a prominent feature of multiple nerve cell deteriorating ailments, which are usually marked by impaired protein clumping. Autopsy analysis of the SN of PD individuals has displayed a noteworthy decline in the UPS biocatalyst operation compared to the brains of the healthy individuals, providing robustly solid corroboration for such aberrations in the case of PD [84]. Current publications have delineated the active engagement of mutations or alterations in several genes, viz., UCHL1, DJ-1, Parkin, SNCA, and PINK1, in prompting proteasomal irregularities, and consequently PD advancement [85,86]. In addition, it has been elucidated that an imperative ubiquitin E3 ligase, namely tumor necrosis factor receptor-associated factor 6 (TRAF6), is over expressed in the brains of individuals experiencing PD, and TRAF6 facilitates the trio, i.e., Lys6-, Lys27-, and Lys29-associated ubiquitination of α-synuclein and DJ-1, and in turn might precipitate the insoluble and polyubiquitinated mutant DJ-1 protein clumping. Further, autopsy analysis of the human brain with PD has displayed that the duo, namely α-synuclein and TRAF6, interact in identical spatial compartments, i.e., colocalize (inside the LBs) [85,87]. These examinations highlight the newfangled aptitude for TRAF6 and for aberrant ubiquitination in the pathology of PD. Another investigation has reported significant forfeiture of only α-subunits of 26 or 20S proteasomes inside DArgic nerve cells, disruption in the 20S proteasomal biocatalyst operations inside the SN-PC, and de-escalation in the quantities of proteasome activator 700 (PA700) and proteasome activator 28 (PA28) within the SN-PC of individuals experiencing PD [88]. These studies strongly suggest the active engagement of UPS dysfunction in the pathology of PD. 2.2.4. Mitochondrial Devastation and Parkinson’s Disease Existing work has disclosed that mitochondrial devastation is actively indulged in the PD advancement. Recently, it has been elucidated that abnormal expansion and fragmentation of the existing power plants of the cell (mitochondrial biogenesis), flawed breakdown of disrupted and redundant mitochondria (mitophagy), impaired electron transport chain (ETC) operation, escalated formation of ROS, abnormal trafficking, disrupted calcium equilibrium, alterations in the incessant processes of mitochondrial merging (mitochondrial dynamics), and, presumably, additional concomitant processes that significantly impact on the operation of mitochondria can all partake in the PD-related mitochondrial devastation [89,90]. Aside from producing an organic, energy-rendering molecule termed ATP, the power plants of cells also share their significant involvement in the modulation of calcium equilibrium, cellular demise via programmed cell death, generation and conveyance of Fe-sulphur(S) clusters, haem generation, and cellular expansion and fragmentation, which have all been demonstrated to be drastically altered in different forms of PD [89,91]. Moreover, environmentally-precipitated PD might emerge following the subjection to deleterious constituents, for instance, MPTP, rotenone, and paraquat, that consequentially suppress ETC (principally via suppressing the mitochondrial complex-I operation) [90,92]. Present-day exploration has demonstrated that alterations in numerous genes, viz., SNCA, PINK1, LRRK2, Parkin, DJ-1, and HTRA2, might elicit mitochondrial devastation and ultimately culminate in PD emergence [93]. These investigations highly indicate the critical participation of mitochondrial devastation in PD pathology. 2.2.5. Apoptosis, Nerve Cell Inflammation, and Parkinson’s Disease Apoptosis and nerve cell inflammation are regarded as the cardinal players in the PD pathology. Autopsy examination of the brain of individuals with PD have displayed the duo, namely autophagy and apoptosis [94]. In addition, it has been elucidated that a tiny class of inducible transcription factors, designated as Nuclear Factor kappa-light-chain-enhancer of activated B cells (NF-κB), has escalated inside the DArgic cells of PD patients [95]. Moreover, apoptosis and inflammation-associated events within the encephalon of PD-experiencing individuals are corroborated by numerous episodes, encompassing elevated quantities of the p53 gene (tumor suppressor), an inflammatory mediator denominated as interferon-gamma (IFN-γ), NF-ΚB, caspase activation within the SN, and alteration in the proapoptotic gene operation [96,97,98,99]. In accordance with another investigation, microglia stimulation was noticed inside the SN of PD-experiencing individuals, inevitably culminating in the liberation of proinflammatory molecules, encompassing interferons (IFNs), tumor necrosis factor-alpha (TNF-α), and interleukins (ILs), which, as a consequence, contributes to apoptosis in PD [59,100]. Likewise, α-synuclein clumping as well stimulates microglia, culminating in the protracted and insidious neuronal deterioration within the SN of PD-experiencing individuals [101]. Despite the fact that the processes eliciting PD-related microgliosis are nebulous, a dark brown colored pigment pinpointed within the cells, denominated as neuromelanin-embracing DA-producing neurons, was demonstrated to be profoundly vulnerable to the inflammation-related episodes in the malady. Still, it is elusive whether inflammatory events or episodes taking place in the vicinity of nerve cells are the fundamental culprit of PD or just a repercussion of the ailment. 2.2.6. α-Synuclein Clumping and Parkinson’s Disease Pursuant to published data, the two, namely erroneous folding and clumping of proteins, are integral pathogenic hallmarks of nerve cell deteriorating maladies [102]. Another investigation has revealed that the procurement of α-synuclein’s baneful operation, instead of forfeiture of its usual operation, is regarded as the causal factor of PD [103]. The α-synuclein accumulation, which contributes to its clumping, is profoundly associated with the procurement of the pernicious operation of α-synuclein. Moreover, erroneously folded α-synuclein accumulation is greatly exacerbated by alterations in the expression of a gene termed SNCA, which are probably precipitated by the trio, i.e., gene mutations, gene duplication, and gene triplication. This occurrence tremendously encouraged the scientists to perform an exhaustive investigation into the in vitro mechanism of α-synuclein clumping. Therefore, these findings probably bolster the speculation that the early-commencement form of PD is provoked by the expeditious fibril formation (fibrillation) of α-synuclein [104,105]. In addition, recombinant synthetic α-synuclein has the capability to effectuate clumping in vitro so as to bring about the fibrillation process, correspondingly to those pinpointed in vivo [106]. According to another investigation, the preformed fibrils (PFF) might disseminate in in vitro nerve cell culture in a “prion-like” fashion, and when administered directly inside the brain of a mouse by means of injection, they might disseminate in vivo in a similar fashion, culminating in the generation of pSer129-α-synuclein-positive LBs-like clumps, and eventually contribute to the emergence of PD [107]. Aside from clumping, α-synuclein enormously impacts tyrosine 3-monooxygenase/tyrosine hydroxylase (TH), autophagosome protein, HSP, and ubiquitin [102]. Additionally, it has been displayed that the different positions for ubiquitination may induce distinct consequences on α-synuclein clumping [108]. Moreover, it has been revealed that LBs-like α-synuclein clumps precipitate disruption in the entire macroautophagy via de-escalating the clearance of autophagosomes, which, as a consequence, culminates in escalated cellular demise [109]. Despite the fact that an escalating amount of corroborations strongly indicate that α-synuclein clumping is disrupted in PD, the definitive implication in the pathology of the malady is still elusive and mystified, and more investigation is required to clearly examine its contribution in the PD pathogenesis. 3. Deciphering the Neuroprotective Role of Neuropeptides in Parkinson’s Disease Existing investigations have promulgated that NPs exert significant neuroprotective action in various PD models via inhibiting the pivotal pathways/mechanisms/processes involved in the evolution of the malady, and which are expounded in detail in the following sub-sections. Figure 3 describes the NPs that exhibit neuroprotective action in PD. 3.1. Neuroprotective Role of Substance P in Parkinson’s Disease Substance P, an NP pertaining to one of the elephantine and eminently recognized peptide families, termed tachykinin (TK)/neurokinin (NK), comprehends eleven amino acid units and was first originated nearly nine decades ago by a Swedish pharmacologist named Ulf Svante von Euler and an English pharmacologist named Sir John Henry Gaddum [110]. The amino acid sequence of SP is as follows: H-Arg-Pro-Lys-Pro-Gln-Gln-Phe-Ple-Gly-Leu-Met-NH2 (Table 1) [111,112]. SP is enciphered by a protein coding gene pinpointed on human chromosome 7, namely the tachykinin precursor 1 gene (TAC1), and is known to provoke intestinal smooth muscle contraction [103]. It has been promulgated that SP is distributed in numerous regions of the human body, encompassing the encephalon, spinal column, intrinsic nervous system (INS)/enteric nervous system (ENS), largest body organ (skin), blood circulating vessels, and peripheral sensory nerves. In addition to SP, two more allied NPs comprising ten amino acid units, namely neurokinin A (NKA)/substance K and neurokinin B (NKB)/neuromedin K, have been spotted, and in conjunction, these NPs represent the TK family [113]. The trio, namely SP, NKA, and NKB, arise via the splitting of a precursor protein termed preprotachykinin, and each one of them behaves as a chemical messenger (neurotransmitter)/neuromodulator inside the CNS and the periphery. Even though the trio possesses identical functions, SP is still regarded as an ascendant member of the family. Most notably, SP is recognized to exhibit varied bodily functions by means of its interaction with three distinct categories of GPCRs/heptahelical receptors, namely neurokinin 1 (NK1), neurokinin 2 (NK2), and neurokinin 3 (NK3) [114,115]. Nearly four decades ago, a conducted investigation reported a consequential plummet in the SP-like immunoreactivity inside the nigral area and the outer compartment of the paleostriatum in individuals experiencing PD [116], which was subsequently corroborated by another investigation [117]. Another study elucidated individuals experiencing trouble in the pharynx while deglutition (pharyngeal dysphagia) demonstrated considerably de-escalated SP quantities in a body fluid termed spit/saliva, which is generated to a large extent in the salivary glands, in comparison to individuals experiencing PD with usual pharyngeal deglutition performance [118]. In addition, several investigations have detected alterations in the SP levels in 6-OHDA-instigated PD experimental models of rats. An investigation conducted around 33 years ago has revealed that the denervation of DA substantially de-escalated the SP levels inside the duo, namely the SN and the striate nucleus, following 21–28 days of 6-OHDA lesion [119]. In contrast, another investigation, conducted nearly a decade ago, has indicated that following 3–21 days of 6-OHDA therapy, quantities of SP were found to be considerably escalated inside the SN region of the brain [120]. This indicated that the 6-OHDA lesion precipitated a significant upsurge in SP quantities at first, and afterwards declined [121]. Up to the present time, the outcomes of SP therapy in PD are still considered tendentious. A recent investigation utilized a 1-methyl-4-phenylpyridinium ion(MPP+)-subjected DArgic nerve cell line (MES23.5 cells) to highlight the nerve cell safeguarding abilities of SP (when introduced at a 0.1 µM concentration) in cellular models of PD [30]. This investigation has displayed that SP by means of the NK1 receptor significantly safeguarded the MES23.5 cells against MPP+-precipitated apoptosis and cytotoxicity via de-escalating the entry of calcium ions, caspase-3 stimulation, ROS formation, and modulating the mitochondrial membrane potential (MtMP) [30]. Contrariwise, several investigations have elucidated that SP introduction consequentially elicits the DArgic nerve cell demise in PD models. To illustrate, the further introduction of SP in the 6-OHDA-prompted PD model escalated the evolution of the malady, with animals exhibiting intense motor abnormalities and worsened DArgic cell demise [120]. Another investigation has elucidated that therapy with the aid of 6-OHDA in mesostriatal organotypic coculture significantly escalated the levels of the duo, namely SP, and a biocatalyst and cellular demise indicator, named lactate dehydrogenase (LDH), thereby worsened the cellular demise [122]. Moreover, this investigation has reported that the generation of LDH was further skyrocketed following the integrated administration of SP and 6-OHDA but was plummeted following the integrated administration of 6-OHDA and an NK1 receptor antagonist named N-acetyl-L-tryptophan (NAT) [122]. In the same way, the duo, namely agonists of the SP receptor and antagonists of the SP receptor, exhibited significant nerve cell protective actions in PD. A selective agonist of the NK1 receptor or an analog of SP, namely septide, has shown considerable nerve cell protective action against 6-OHDA-prompted pernicious repercussions following twenty-four hours of pretherapy at a 2 μM concentration via the suppression of the programmed cell death pathways and the stimulation of the protein kinase B (PKB/Akt) pathways (signal transduction pathways) [123]. Most notably, the antiprogrammed cell death action of septide was not reliant on caspase, which is congruent with another published paper highlighting the calpain-1-reliant nerve cell protective action of SP inside the cerebellar granule cells [124]. In addition, another exploration has elucidated that a selective agonist of NK3 receptor, namely senktide, following its administration in a dose of 0.2 mg/kg, reinstated the temporal order memory in the 6-OHDA-lesioned hemiparkinsonian rat model [125]. Howbeit, the intracerebroventricular introduction of the two neoteric, potent, and selective antagonists of the NK1 receptor, namely L-733060 and NAT, as well de-escalated the cellular demise provoked by 6-OHDA exposure, and finally contributed to a noteworthy upgradation in the motor operations [120]. In addition, another investigation has promulgated that NAT and another selective, potent, and neoteric antagonist of NK1 receptor, namely lanepitant (LY303870), markedly de-escalated the levodopa-precipitated anomalous, not controllable, and involuntary movements of muscles (dyskinesia), without influencing the promising medicinal outcomes of levodopa in animal rat models experiencing PD [126,127]. Furthermore, it has been revealed that the immune cells of the CNS, termed microglia, imitate the operation of professional phagocytes, termed macrophages, inside the encephalon. Additionally, it has been reported that the density of microglia inside the SN-PC region of the encephalon is markedly greater in comparison to the circumjacent regions of the encephalon [128]. Published literature has displayed that SP may be in part blameworthy for the greater microglia density. The microglia density in the SN region was markedly de-escalated in animal mice models lacking endogenous SP (TAC1−/−) or NK1 receptor (NK1R−/−) [118]. Moreover, this investigation has highlighted that SP captivated the microglia by means of a trio, namely NK1 receptor, protein kinase C delta (PKCδ), and reduced nicotinamide adenine dinucleotide phosphate (NADPH) oxidase in a pathway-reliant way. Figure 4 portrays the neuroprotective role of SP in PD. 3.2. Neuroprotective Role of Ghrelin in Parkinson’s Disease Ghrelin, an inimitable gastric peptide/hunger hormone, comprehending twenty-eight amino acid units, was originally uncovered nearly 23 years ago by a Japanese researcher named Dr. Masayasu Kojima and fellow workers, and is chiefly liberated from the unfed stomach, but as well pinpointed in the tissues of the peripheral region, for instance, the ovary (a female gonad), testicle (a male reproductive gland), kidney, lymphocytes, pancreas (a mixed gland), placenta (a nonpermanent huge pan-shaped fetal organ that evolves in the course of pregnancy), pituitary (a master gland), and small bowel [129,130]. The amino acid sequence of ghrelin is as follows: NH2-Gly-Ser-[Ser(n-octanoyl)]-Phe-Leu-Ser-Pro-Glu-His-Gln-Arg-Val-Gln-Gln-Arg-Lys-Glu-Ser-Lys-Lys-Pro-Pro-Ala-Lys-Leu-Gln-Pro-Arg-COOH (Table 1) [131]. Ghrelin, initially spotted in the rat stomach, exists as an endogenous ligand of the rhodopsin-like or category A GPCRs, namely growth hormone (GH) secretagogue receptor 1a (GHS-R1a)/ghrelin receptor (ghrelinR), which is principally pinpointed in the encephalon and tissues of the peripheral region [130,132]. Ghrelin, upon interaction with its receptor, i.e., GHS-R1a, is able to stimulate the liberation of GH from adenohypophysis in order to elevate the concentration of calcium ions within the cells by means of the inositol 1,4,5-trisphosphate (InsP3) signaling pathway [123]. Ghrelin arises through the peptidal bond cleavage (proteolytic breakdown) of the two, namely proghrelin and preproghrelin [132]. Two significant types of ghrelin have been recognized in the liquid connective tissue (blood), namely acyl-ghrelin and non-acyl-ghrelin, and amongst the two, acyl-ghrelin has the aptitude to interact with GHS-R1a in order to exhibit physiological consequences [130]. In accordance with published literature, the acylated type of ghrelin, i.e., acyl-ghrelin, markedly renders nerve cell protection in PD [133,134,135]. It has been revealed that a consequential plummet in the concentrations of the duo, namely ghrelin and ghrelinR, are cognized to partake in the PD pathogenesis. Pursuant to a recent study, individuals experiencing PD display a consequential de-escalation in the fasting concentrations of the two, namely acyl-ghrelin (active type) and the entire ghrelin, which is further accompanied by a remarkable decline in the active type, in comparison to the salubrious control individuals [136]. Another investigation has reported that in experimental animal models with PD, the genetic deletion of GHSR significantly elevated the forfeiture of DA-forming nerve cells of the SN region of the encephalon, and consequently plummeted the DA concentrations in the striatal region, which may be turned back following the selective restimulation of GHSRinside the catecholaminergic (CArgic) nerve cells [137]. In addition, it has been promulgated that the introduction of a greatly employed ghrelinR antagonist, namely [D-Lys3]-GHRP6, through the microinjection/intracerebroventricular route within the SN region of a healthy experimental mice model might provoke PD-analogous motor coordination impairment [138]. Furthermore, an investigation conducted around 14 years ago initially demonstrated ghrelin’s nerve cell protective action in the MPTP-instigated experimental mouse model with PD [139], which was thereafter corroborated by numerous investigations [137,140,141]. It has been elucidated that ghrelin acted against or restrained cellular deprivation precipitated by exposure to a pesticide, i.e., rotenone [142,143], upgraded the abnormal rotarod motor performance in the experimental PD mouse model precipitated by subjection to MPTP [133], as well as mediated the nerve cell protective abilities of a nutritional strategy that markedly de-escalates the ingestion of calories without malnourishment, i.e., caloric restriction (CR) [144]. In addition, it has also been demonstrated that ghrelin has the aptitude to carry out the electrical stimulation of DA-forming nerve cells by suppressing the potassium (K+) channels (KCNQ/Kv7) and A-type voltage-gated K+ channels, as well as pursue the up-regulation of pacemaker channels/hyperpolarization-activated cyclic nucleotide-gated (HCN) channels so as to upgrade MPP+ suppression on DA nerve cell excitation [145,146]. In addition, it has been recently discovered that therapy with the aid of ghrelin contributed to a substantial escalation in the amount/number of neural stem cells (NSCs) of the midbrain, facilitated in vitro as well as ex vivo DArgic nerve cell differentiation via the canonical Wnt signaling pathway, and thereupon renders the prospect that subjection to ghrelin may be a newfangled, effective, and propitious strategy for PD therapy [147]. Another investigation has revealed that persistent exposure to a neoteric, and brain-permeable agonist of ghrelinR, namely HM01, was recognized to mitigate the nonmotor manifestations provoked by the 6-OHDA lesion in an experimental rat model experiencing PD, comprehending changes in the ingestion of water (H2O), consumption of food, weight of excrement, and body weight [148]. Moreover, Dpr3ghr, an analog of ghrelin, has been reported to safeguard human neuroblastoma (SH-SY5Y) cells (by escalating the expression of B-cell lymphoma-2 (Bcl-2), and de-escalating the Bcl-2-associated X protein (Bax) expression and Bax/Bcl-2 ratio) from the enormously perilous two carbonyl (C=O) groups comprising the organic compound, possessing the molecular formula CH3C(O)CHO, namely methylglyoxal (MGO)-precipitated programmed cell death and nerve cell toxicity [149]. The pathways by means of which ghrelin exerts its nerve cell protective actions are tortuous [150]. An early investigation, which employed a subacute MPTP-precipitated experimental model of mouse with PD, has indicated that the nerve cell protective outcomes of ghrelin may be strongly associated with a substantial plummet in the caspase-3-elicited programmed cell death through the modulation of the expression of the two, namely Bax and Bcl-2, within the DA-forming nerve cells of the nigral region of the encephalon [130]. Additionally, it has been revealed that ghrelin acted against MPP+ and rotenone-provoked nerve cell toxicity inside the MES23.5 cells and the principal retinal output cells, named retinal ganglion cells (RGCs), via regulating the MtMP, eradicating the formation of ROS, and suppressing the mitochondrial complex-I operation and stimulation of caspase-3 [142,143,151]. The nerve cell protective action rendered by ghrelin is markedly reliant upon the duo, namely chondriosomal biogenesis and chondriosome-associated oxidative damage [121]. It has been propounded that a chondriosomal protein, namely uncoupling protein 2 (UCP2)-reliant changes in two, i.e., chondriosomal respiration and bio-energetic status purveying, prompted by subjection to ghrelin, may construct the DA-forming nerve cells immensely invulnerable to cell destruction [137]. Moreover, it has been expounded that ghrelin has exhibited its actions against oxidative stress by consequentially escalating the operation of two, namely CAT and Cu-zinc (Zn) SOD, suppressing the translocation of NF-κB, and plummeting the MDA levels [152]. In addition, it has been revealed that the nerve cell protective action provoked by ghrelin exposure was enormously reliant upon an imperative energy sensor, i.e., 5′ adenosine monophosphate-activated protein kinase (AMPK), and substantially escalated the devastation of the deteriorated chondriosome (mitophagy) inside the DA-generating nerve cells [153]. Another study has elucidated that in an MPTP-subjected experimental PD mouse model, ghrelin might possess its nerve cell protective outcomes via suppressing the microglia stimulation, which, as a consequence, suppresses the liberation of a couple of inflammatory mediators, viz., TNF-α and interleukin-1β (IL-1β) [141]. Figure 5 describes the neuroprotective role of ghrelin in PD. 3.3. Neuroprotective Role of Neuropeptide Y in Parkinson’s Disease NPY, a prolific and foremost NP, initially originated around four decades ago from the pig encephalon by a well-renowned researcher named Kazuhiko Tatemoto and fellow workers, pertains to a category of three thirty-six amino acid units comprising peptides, and the remaining two, i.e., peptide YY (PYY) and pancreatic polypeptide (PP), exist as gastrointestinal hormones [154,155]. The amino acid sequence of NPY is as follows: Tyr-Pro-Ser-Lys-Pro-Asp-Asn-Pro-Gly-Glu-Asp-Ala-Pro-Ala-Glu-Asp-Leu-Ala-Arg-Tyr-Tyr-Ser-Ala-Leu-Arg-His-Tyr-Ile-Asn-Leu-Ile-Thr-Arg-Gln-Arg-Tyr-NH2 (Table 1) [156]. NPY is profusely pinpointed in the peripheral region and the varied regions of the encephalon, such as the nucleus tractus solitarius, hypothalamus, hippocampus, cerebral mantle, accumbens nucleus, amygdala (the fear center of the encephalon), and locus coeruleus (LC) [155,157]. NPY has been elucidated to interact with NPY receptors that pertain to the rhodopsin-like or category A GPCRs [158]. Up to the present day, five receptors of NPY, namely Y1 receptor (Y1R), Y2 receptor (Y2R), Y4 receptor (Y4R), Y5 receptor (Y5R), and Y6 receptor (Y6R), displaying nonidentical operations, have been cloned and recognized from mammals, and multiple other receptors, for instance, Y3 receptor (Y3R), and have been speculated (relying upon their therapeutic aptitude) to employ numerous animal and human tissues, but were not cloned or clearly recognized until now [155,158,159]. Amongst the five, i.e., Y1R, Y2R, Y4R, Y5R, and Y6R, only Y6R is operational in animals, for instance, in mice and rabbits, whereas the rest of the receptors are operational in human beings [158,160]. In addition, NPY receptors fundamentally bind with G inhibitory (Gi)/G0 proteins and culminate in the suppression of adenyl cyclase (AC), and eventually contribute to the suppression of the build-up of cyclic adenosine monophosphate (cAMP) and the regulation of the two, i.e., K+ channels and calcium channels. Moreover, a pair of NPY receptors, namely Y2R and Y4R, as well binds to a protein termed Gq protein, culminating in the escalated formation of InsP3 through phospholipase C-beta (PLC-β) stimulation inside the smooth muscle cells of the rabbit [158]. Published work has expounded that NPY significantly participates in the regulation of several operations, for instance, the consumption of food, learning (procurement of novel knowledge and skills, or flourishing the one known previously), mood (a conscious mind state), memory (procurement, storage, and retrieval of information) [161], and nerve cell protection against nerve-cell-deteriorating ailments [162]. Numerous investigations entailing animal models and human beings have demonstrated substantial alterations in the levels of NPY. Around 31 years ago, an investigation involving individuals experiencing PD displayed markedly de-escalated levels of NPY in their tissues of the adrenal medulla [163]. After about a year, another research team evaluated the levels of NPY immunoreactivity within the fluid encircling the encephalon and the spinal column, termed cerebrospinal fluid (CSF), of 10 individuals suffering from PD, and concluded that the NPY levels were consequentially plummeted, comparably to healthy subjects, reflecting a marked decrease in the liberation of NPY or a marked elevation in the turnover of NPY [164]. Furthermore, an autopsy examination of the encephalon samples of the PD individuals has displayed an upsurge in the amount of NPY messenger ribonucleic acid (mRNA)-positive cells within their trio, namely accumbens nucleus, caudate nucleus, and putamen [165]. Another investigation has demonstrated a considerable decline in the NPY-positive cells, as well as the deprivation of axons within the two, namely caudate nucleus and putamen, of individuals experiencing lubag syndrome/X-linked dystonia of panay [166]. Moreover, individuals with lubag syndrome/X-linked dystonia of panay have indicated the paucity of the labeling of NPY within the subventricular zone (SVZ), together with a consequential forfeiture of proliferating cell nuclear antigen (PCNA)-representing progenitor cells (PGCs) [166]. Another group of researchers have described that those individuals experiencing the duo, i.e., PD and depression at the same time, exhibited a marked upsurge in the NPY levels within their CSF in comparison to individuals experiencing depression alone [167]. In addition, the deterioration of the DArgic pathway, named the nigrostriatal pathway, provoked a prodigious upsurge in the NPY-like immunoreactivity within the striatal region of the MPTP-subjected experimental C57 black 6 (C57BL/6) mice model [168]. It has been shown that NPY renders substantial nerve cell protection in PD through multifarious processes associated with the ailment. Pursuant to an investigation, the Y2R emanates to be an eminent and cardinal receptor of NPY, which is reported to be enormously liable for the mediation of nerve cell protective action of NPY, since the nerve cell protection rendered by NPY therapy was ceased in mice subjected to the Y2R antagonist and in mice lacking Y2R [169]. The nerve cell protective action of NPY was initially discovered in the duo, i.e., in vivo and in vitro 6-OHDA-precipitated experimental PD models [169]. Further, this investigation has strongly suggested that the nerve cell protective action of NPY is attained to a certain extent via the stimulation of two signaling pathways, namely Akt and mitogen-activated protein kinase (MAPK), which might culminate into robust enhancement in the viability of DA-forming nerve cells of the nigral area of the encephalon [169]. Recently, it has also been elucidated that in experimental rat models exposed to 6-OHDA, NPY treatment culminated in the suppression of microglia in two regions, i.e., striatal and SN, which, as a consequence, facilitated the anti-inflammatory action of NPY in the PD [170]. Furthermore, it has been reported that the subjection to NPY resulted in the suppression of an imperative constituent of the gram-negative bacteria, termed lipopolysaccharide (LPS)-triggered NO formation, and the liberation of IL-1β inside the microglia [171]. In accordance with another exploration, NPY treatment, via pursuing the stimulation of the phosphoinositide 3-kinase (PI3K)-spliced form of X-box binding protein 1 (XBP1s)-precipitated binding immunoglobulin protein (BiP)/78-kDa glucose-regulated protein (GRP78) pathway, markedly exhibited a safeguarding action against an endoplasmic reticulum stress (ER stress)-prompted nerve cell demise [172]. Additionally, the prior introduction of NPY significantly de-escalated the stimulation or operation of the duo, namely caspase-3 and caspase-4, in the duration of the ER stress reaction [172]. Figure 6 depicts the neuroprotective role of NPY in PD. In addition, a protein termed abrineurin/brain-derived neurotrophic factor (BDNF), which is implicated in the growth, differentiation, maintenance, and promotion of the viability of nerve cells, was presumed to be deeply entangled in bringing about the nerve cell protective action of NPY [173]. The forfeiture of DA-forming nerve cells of the SN in PD is speculated to be provoked by the plummeted BDNF expression [173,174]. To date, no literature exists that could clearly expound the putative association amongst the duo, namely NPY and BDNF expression, in PD. Therefore, more research is tremendously required in this particular domain to scrutinize the potential consequences of NPY on the expression of BDNF in PD. 3.4. Neurprotective Role of Neurotensin in Parkinson’s Disease Neurotensin, an NP comprehending thirteen amino acid units, was initially originated nearly 49 years ago from the calf hypothalamus by two prestigious researchers, namely Robert Carraway and Susan E. Leeman, and denominated owing to its localization in the nerve cells and hypotensive action (blood pressure-decreasing aptitude) [175,176]. The amino acid sequence of neurotensin is as follows: pyr-Glu-Leu-Tyr-Glu-Asn-Lys-Pro-Arg-Arg-Pro-Tyr-Ile-Leu-OH (Table 1) [177]. This modulatory endogenous peptide is pinpointed in two regions, i.e., the CNS (principally the pituitary and hypothalamus) and the peripheral region (principally the gastrointestinal tract), where it behaves as a modulator/transmitter for nerve cells and locally as a hormone, respectively [178,179]. Currently, three receptors for neurotensin have been spotted, namely neurotensin receptor 1 (NTR1), neurotensin receptor 2 (NTR2), and neurotensin receptor 3 (NTR3), and neurotensin, upon interaction with these receptors, exhibits its bioactivities. The duo, namely NTR1 and NTR2, pertains to the rhodopsin-like or category A 7TM domain GPCRs, whereas the third one, namely NTR3, a category I single TM domain sorting receptor, otherwise designated as sortilin, pertains to the category of vacuolar protein sorting 10 protein (VPS10P) domain receptors [179,180]. Consequently, the actions of neurotensin enormously rely upon the two, i.e., the type of receptor and the dispersal inside the regions of the body. Pursuant to the existing data, neurotensin is consequentially implicated in the DArgic pathway. In accordance with histological examinations of rat encephalon, ample neurotensin-comprising fibers have been detected in regions possessing plenteous DA content, for instance, SN and ventral tegmentum [181]. Further, two other investigations have displayed a bifold escalation in the neurotensin content inside the SN area of encephalon of PD-experiencing individuals [182,183]. In addition, it has been reported that individuals experiencing PD exhibited markedly escalated neurotensin content within their plasma in comparison to normal healthy individuals, and four nontreated individuals also exhibited significantly elevated neurotensin content within their plasma in comparison to individuals subjected to levodopa therapy [184]. In addition, the neurotensin receptor mRNA quantities were discovered to be markedly elevated within the rat encephalon DA-forming nerve cells of the duo, namely the SN and ventral tegmentum [185]. Howbeit, in PD-experiencing individuals, considerably plummeted quantities or zero/nil mRNA expression for neurotensin receptors were detected inside the ventral tier of the SN region of the encephalon [185]. Moreover, a consequential decline in the concentrations of neurotensin receptors was reported within the two regions, namely the paleostriatum and putamen of PD-experiencing individuals [186,187,188]. Published literature has proven that neurotensin and its analogs have the aptitude to render significant nerve cell protective action in PD-experiencing experimental animal models. It has been described that the introduction of the duo, i.e., neurotensin 8–13 and [D-Tyr11]-neurotensin, through the intracerebroventricular route consequentially de-escalated the 6-OHDA-precipitated PD manifestations (tremor, and rigor) [189]. Another investigation has displayed two neoteric analogs of neurotensin, namely neurotensin2 and neurotensin4, might amplify the liberation of DA within the striatal region, de-escalate the rotations provoked by exposure to an immensely potent agonist of DA, namely apomorphine, and upgrade memory and learning [34]. Howbeit, it has been elucidated that, via escalating the fundamental excitatory neurotransmitter (glutamate)-provoked nerve cell toxicity by elevating the calcium levels within the cells or/and N-methyl-D-aspartate (NMDA)-triggered signaling of glutamate, neurotensin elevated the deterioration of the two, viz., cortical nerve cells and DArgic mesencephalic nerve cells [190]. Therefore, more exploration is enormously needed in this respective discipline to elucidate the potential role of neurotensin in PD. 3.5. Neuroprotective Role of Pituitary Adenylate Cyclase-Activating Polypeptide in Parkinson’s Disease PACAP, an enormously preserved and multitalented NP, comprehending twenty-seven amino acid units, i.e., PACAP27/thirty-eight amino acid units, i.e., PACAP38, was initially originated around 33 years ago from the extracts of sheep hypothalamus by a well-known endocrinologist named Akira Arimura and fellow workers, and pertains to the class of secretin-glucagon-vasoactive intestinal polypeptide (VIP) [191,192]. The amino acid sequence of PACAP27 is as follows: H-His-Ser-Asp-Gly-Ile-Phe-Thr-Asp-Ser-Tyr-Ser-Arg-Tyr-Arg-Lys-Gln-Met-Ala-Val-Lys-Lys-Tyr-Leu-Ala-Ala-Val-Leu-NH2, whereas the amino acid sequence of PACAP38 is as follows: H-His-Ser-Asp-Gly-Ile-Phe-Thr-Asp-Ser-Tyr-Ser-Arg-Tyr-Arg-Lys-Gln-Met-Ala-Val-Lys-Lys-Tyr-Leu-Ala-Ala-Val-Leu-Gly-Lys-Arg-Tyr-Lys-Gln-Arg-Val-Lys-Asn-Lys-NH2 (Table 1) [193,194]. PACAP is robustly engaged in the regulation of multiple biological operations within the CNS and peripheral region upon interaction with three category B GPCRs, otherwise designated as secretin class of GPCRs, namely PACAP type I receptor (PAC1), VIP receptor 1 (VPAC1), and VIP receptor 2 (VPAC2) [191,195]. Amongst the aforementioned three receptors, PACAP exhibits 103-fold greater affinity towards PAC1 in comparison to the duo, namely VPAC1 and VPAC2 [185]. PACAP is reported to be enciphered by a gene named adenylate cyclase activating polypeptide 1 (ADCYAP1) [196,197]. At elevated concentrations, PACAP is principally pinpointed in the accumbens nucleus, SN, hippocampus, cerebellum, hypothalamus, and the bed nucleus of the stria terminalis (BNST) [121,198]. Within the CNS, PACAP behaves as a neurotrophic factor, neurohormone, neuroregulator, and neurotransmitter [199]. In addition, it has been revealed that mRNA that is enormously implicated in enciphering PACAP receptors has been detected within the SN [199]. Another investigation has elucidated that in MPTP-instigated experimental macaque monkey PD models, a consequential plummet in the immunological signal of PAC1 receptor was recognized within various regions of the basal nuclei, encompassing the caudate nucleus, inner and outer regions of the paleostriatum, and putamen [199]. Numerous investigations have displayed the nerve cell protective abilities of PACAP in various experimental models experiencing PD [200,201]. It has been depicted that PACAP markedly declines the DArgic nerve cell deprivation provoked by 6-OHDA exposure, upgrades behavioral impairment [202], de-escalates the fundamental manifestation of PD (hypokinesia) [203], and slows the decline in DA content. PACAP might have the aptitude to forestall the impaired polypeptide chain generation precipitated by the subjection to MPTP and reduce cognitive deterioration [204]. Another investigation has reported that PACAP was significantly capable of de-escalating the programmed cell death and easing the conversion of cellular demise from the late to early phase in a rotenone-prompted cellular PD model [205]. Additionally, PACAP therapy has safeguarded the SH-SY5Y cells from a nerve cell active candidate, named 1-methyl-6,7-dihydroxy-1,2,3,4-tetrahydroisochinolin, precipitated deleterious consequences via consequentially decreasing the programmed cell death and the related chemical alterations [206]. In addition, it has been elucidated that PACAP27 as well markedly de-escalated the DArgic nerve cell deprivation and motor abnormalities in a prostaglandin J2-provoked experimental PD model [207]. Existing publications have demonstrated that PACAP renders substantial nerve cell protection in PD through multifaceted processes. Initially, the nerve cell protective action of PACAP was immensely related to its inflammation-reducing abilities. It has been reported that prior therapy of SH-SY5Y cells with a tremendously potent agonist of the PACAP receptor, namely PACAP (1–38), significantly contributed to a dosage-reliant de-escalation of deleterious repercussions provoked by the mediators of inflammation [208]. Moreover, PACAP has been depicted to possess considerable antiautophagic abilities. A recent exploration has demonstrated that PACAP therapy considerably plummeted the autophagic operation in MPTP-provoked experimental PD models via regulating the concentrations of a protein termed sequestosome-1/p62, and by carrying out the formation of microtubule-associated protein light chain 3-phosphatidylethanolamine conjugate (LC3-II) [195]. Another research group has revealed that the safeguarding action of PACAP against cellular demise precipitated by rotenone exposure was markedly suppressed through the introduction of suppressors of the trio, namely p38 MAPK, protein kinase A (PKA), and extracellular signal-regulated kinase (ERK) [205]. Consequently, the nerve cell protective actions of PACAP were attained via the stimulation of PKA signaling process, and also the two downstream signals, i.e., p38 MAPK and ERK [205]. This nerve cell protective action was cognized to be immensely linked to a balance among DA-acetylcholine (ACh) processes within the nerve cell pathway of the basal nuclei. Moreover, it has been elucidated that the introduction of PACAP27 through the intravenous route in an MPTP-precipitated experimental mouse model with PD rendered significant nerve cell protective action via altering the DArgic, as well as cholinergic synaptic conveyance, by means of escalating the operation of DA 2 receptors (D2R) and the expression of ATP-sensitive potassium channel (KATP) subunits within the striatal region of the basal nuclei [209]. Furthermore, therapy with the aid of PACAP/agonists of PACAP receptors culminated in the significant de-escalation in the SH-SY5Y cells toxicity provoked by 1-methyl-6,7-dihydroxy-1,2,3,4-tetrahydroisochinolin exposure via suppressing caspase-3 expression and elevating the expression of the duo, namely BDNF and phosphorylated cAMP-response element binding protein (p-CREB) [210]. In addition, the nerve cell protective action rendered by PACAP therapy is thought to be immensely linked to microglia. PACAP might have the aptitude to significantly reduce LPS-provoked microglia stimulation and the concomitant formation, as well as the liberation of NO and TNF-α, respectively [210]. Howbeit, another investigation has revealed that PACAP27 was incapable of restraining the stimulation of microglia in the prostaglandin J2-precipitated experimental PD mouse model [207]. As a result, further scrutiny is greatly required in this respective discipline to ascertain the clear participation of PACAP in the stimulation of microglia in PD. VIP, an NP identical to PACAP, comprises twenty-eight amino acid units and was initially derived from porcine duodenum nearly 52 years ago by two renowned researchers, namely Sami I Said and Viktor Mutt [211]. Recently, a research group has elucidated that VIP exhibits the identical nerve cell protective actions as that of PACAP [212]. Another investigation has reported that VIP therapy restrained the MPTP-elicited deprivation of two, i.e., DArgic nerve cells and nerve fibers, within the nigrostriatal pathway via suppressing the liberation of inflammatory mediators, for instance, NO, ROS, IL-1β, and TNF-α [213]. In accordance with another recent investigation, PACAP and VIP cotherapy markedly restrained the liberation of NO, interleukin-6 (IL-6), matrix metallopeptidase 9 (MMP-9), and cluster of differentiation molecule 11b (CD11b) in the rotenone-subjected microglial cells of mice (BV-2 cells), and therefore might contribute to considerable nerve cell protection in PD [214]. Figure 7 represents the neuroprotective role of PACAP in PD. 3.6. Neuroprotective Role of Nesfatin-1 in Parkinson’s Disease Nesfatin-1 is an inimitable and enormously effective appetite-suppressing NP, comprising eighty-two amino acid units procured from the three hundred and ninety-six amino acid units comprising a parent protein termed nucleobindin-2 (NUCB2), and was initially originated around 16 years ago from the hypothalamus of the rat by a Japanese researcher named Shinsuke Oh-I and fellow workers, and is extensively pinpointed in the duo, i.e., the CNS and the peripheral areas [215,216,217]. The amino acid sequence of nesfatin-1 is as follows: Val-Pro-Ile-Asp-Ile-Asp-Lys-Thr-Lys-Val-Gln-Asn-Ile-His-Pro-Val-Glu-Ser-Ala-Lys-Ile-Glu-Pro-Pro-Asp-Thr-Gly-Leu-Tyr-Tyr-Asp-Glu-Tyr-Leu-Lys-Gln-Val-Ile-Asp-Val-Leu-Glu-Thr-Asp-Lys-His-Phe-Arg-Glu-Lys-Leu-Gln-Lys-Ala-Asp-Ile-Glu-Glu-Ile-Lys-Ser-Gly-Arg-Leu-Ser-Lys-Glu-Leu-Asp-Leu-Val-Ser-His-His-Val-Arg-Thr-Lys-Leu-Asp-Glu-Leu (Table 1) [218]. It has been reported that within the encephalon, nesfatin-1 is primarily pinpointed in the nucleus tractus solitarius, master gland, infundibular nucleus, dorsal vagal nucleus, supraoptic nucleus, and hypothalamic paraventricular nucleus [219]. Moreover, nesfatin-1 might have the aptitude to permeate across the blood–brain barrier through nonsaturable pathways in two ways, i.e., blood to encephalon and encephalon to blood [220]. In addition, nesfatin-1 partakes in the modulation of the blood sugar equilibrium and the expenditure of energy, and also exhibits anti-inflammatory as well as antiprogrammed cell death abilities [208]. Despite emerging corroboration for the plethora of NP’s abilities, the GPCR that is implicated in the mediation of these actions remains obscure. Pursuant to a recently published investigation, the blood concentration of nesfatin-1 in individuals experiencing PD was considerably plummeted in comparison to control individuals [221]. The nerve cell protective action of nesfatin-1 in PD might be owing to its aptitude to suppress inflammation, oxidative stress, and programmed cell death. Several investigations have depicted the protective action of nesfatin-1 against inflammation in the encephalon. It has been reported that nesfatin-1 therapy markedly de-escalated the NF-κB expression and the concentrations of inflammatory mediators like IL-1β, TNF-α, and IL-6 in experimental rat models with intracranial injury, strongly indicating that nesfatin-1 may inhibit the NF-κB-reliant inflammatory reactions [222]. Furthermore, therapy with the aid of nesfatin-1 significantly suppressed the acute encephalon damage following the subarachnoid hemorrhage-prompted diffusion and build-up of neutrocytes and the escalated quantities of inflammatory mediators [223]. Another investigation has revealed that in MES23.5 DA nerve cells, nesfatin-1 therapy potentially recovered the suddenly falled MtMP provoked by subjection to rotenone, as well as reinstated the mitochondrial complex-I operation [224]. According to another recent exploration, nesfatin-1 introduction in the brain ischemia tremendously suppresses lipid peroxidation and escalates the operation of biocatalysts exhibiting antioxidant abilities, namely GSH and SOD [225]. Apart from this, nesfatin-1′s ability against oxidative destruction has also been depicted in subarachnoid hemorrhage models [223]. Further, prior therapy with nesfatin-1 markedly safeguards MES23.5 DA nerve cells from nerve cell toxicity precipitated by rotenone exposure via suppressing programmed cell death and improving abnormal mitochondrial operation [224]. Another research group has reported that the antiprogrammed cell death action of nesfatin-1 inside the DA-forming nerve cells was fundamentally attained via the C-Raf/ERK1/2-reliant antiprogrammed cell death mechanism, by means of which nesfatin-1 consequentially inhibits the caspase-3 operation, and finally contributes to the suppression of programmed cell death [226] In the same investigation, the PKA suppressor did not suppress the nesfatin-1 action, tremendously propounding the absence of engagement of the PKA pathway in rendering antiprogrammed cell death action [226]. By and large, the actual and circumstantial molecular pathways engaged in rendering the nerve cell protective action of nesfatin-1 yet remain unexplored. Despite the fact that the pathways implicated in bringing about nesfatin-1′s nerve cell protective action have been depicted in numerous experimental models experiencing varied sort of neurological maladies, it is still greatly inexplicit whether these nerve cell protective pathways are pertinent to PD. Consequently, an in-depth exploration is immensely required so as to attain corroborations for the employment of nesfatin-1 in therapeutic settings. 3.7. Neuroprotective Role of Somatostatin in Parkinson’s Disease SST, otherwise designated as growth hormone-inhibiting hormone (GHIH), is a renowned cyclic ring structure containing modulatory NP, occurring in two active types, the first one comprises of fourteen amino acid units, while the second one comprises of twenty-eight amino acid units, and was initially originated nearly 48 years ago from the extracts of sheep hypothalamus by Paul Brazeau and fellow workers, and is fundamentally pinpointed in the duo, i.e., the CNS and the peripheral areas [227,228]. The amino acid sequence of fourteen amino acid units comprising SST is as follows: Ala-Gly-Cys-Lys-Asn-Phe-Phe-Trp-Lys-Thr-Phe-Thr-Ser-Cys, whereas the amino acid sequence of twenty-eight amino acid units comprising SST is as follows: Ser-Ala-Asn-Ser-Asn-Pro-Ala-Met-Ala-Pro-Arg-Glu-Arg-Lys-Ala-Gly-Cys-Lys-Asn-Phe-Phe-Trp-Lys-Thr-Phe-Thr-Ser-Cys (Table 1) [229]. SST is generated in numerous regions within the body, encompassing the CNS, digestive tract, hypothalamic region, and pancreas [230]. SST is reported to behave as a neurotransmitter, nerve cell regulator, and highly active suppressor of dysfunctional cellular multiplication and several secretory pathways, upon interaction with five 7TM domain GPCRs, viz., SST receptor 1 (SSTR1), SST receptor 2 (SSTR2), SST receptor 3 (SSTR3), SST receptor 4 (SSTR4), and SST receptor 5 (SSTR5) [231,232]. SST markedly exerts wide range of suppressing actions on the duo, namely exocrine secretion (such as pancreatic biocatalysts, stomach acid, and intestinal fluid), and endocrine secretion (such as VIP, GH, PP, cholecystokinin, glucagon, secretin, insulin, and gastrin) [228,231,233,234]. Up to the present time, only a single research group has explored the nerve cell protective abilities of SST in nerve cell deteriorating diseases, for instance, PD [37]. In this investigation, researchers have employed an LPS-subjected experimental PD rat model so as to examine the significant outcomes of SST therapy on DArgic nerve cell deterioration provoked by LPS exposure. Pursuant to this investigation, LPS exposure markedly contributed to DArgic nerve cell deprivation, whereas a consequential plummet in the nerve cell demise via SST prior therapy was strongly corroborated by means of immunohistochemical staining of the duo, i.e., TH- and Nissl-positive cells [37]. In addition, this study has reported that SST substantially suppressed the formation of the ROS and LPS-provoked microglia operation [37]. Moreover, the widely employed analytical assay, namely the enzyme-linked immunosorbent assay (ELISA), has shown that therapy with SST before subjection to LPS markedly de-escalated the formation of inflammatory mediators, for instance, prostaglandin E2, TNF-α, and IL-1β. In addition, immunoblotting inspection has demonstrated that the introduction of SST before LPS exposure culminated in the significant reduction in the LPS-prompted expression of the trio, namely NF-κB p-p65, inducible NO synthase, and cyclooxygenase-2 [37]. These outcomes depicted that SST has the aptitude to inhibit the stimulation of microglia and the NF-κB mechanism, and, consequently, declines the nerve cell inflammation and oxidative destruction, and finally suppresses the DArgic nerve cell deprivation elicited by LPS exposure in nerve cell deteriorating diseases such as PD [29]. Howbeit, more investigation is greatly needed in this particular domain to decipher the restorative aptitude of SST in PD. ijms-23-04565-t001_Table 1 Table 1 The amino acid units and amino acid sequence of fundamental neuropeptides that contribute to significant neuroprotection in Parkinson’s disease. Neuropeptide Amino Acid Units Amino Acid Sequence Ref. Substance P 11 H-Arg-Pro-Lys-Pro-Gln-Gln-Phe-Ple-Gly-Leu-Met-NH2 [111,112] Ghrelin 28 NH2-Gly-Ser-[Ser(n-octanoyl)]-Phe-Leu-Ser-Pro-Glu-His-Gln-Arg-Val-Gln-Gln-Arg-Lys-Glu-Ser-Lys-Lys-Pro-Pro-Ala-Lys-Leu-Gln-Pro-Arg-COOH [131] Neuropeptide Y 36 Tyr-Pro-Ser-Lys-Pro-Asp-Asn-Pro-Gly-Glu-Asp-Ala-Pro-Ala-Glu-Asp-Leu-Ala-Arg-Tyr-Tyr-Ser-Ala-Leu-Arg-His-Tyr-Ile-Asn-Leu-Ile-Thr-Arg-Gln-Arg-Tyr-NH2 [156] Neurotensin 13 pyr-Glu-Leu-Tyr-Glu-Asn-Lys-Pro-Arg-Arg-Pro-Tyr-Ile-Leu-OH [177] Pituitary adenylate cyclase-activating polypeptide 27 H-His-Ser-Asp-Gly-Ile-Phe-Thr-Asp-Ser-Tyr-Ser-Arg-Tyr-Arg-Lys-Gln-Met-Ala-Val-Lys-Lys-Tyr-Leu-Ala-Ala-Val-Leu-NH2 [193,194] 38 H-His-Ser-Asp-Gly-Ile-Phe-Thr-Asp-Ser-Tyr-Ser-Arg-Tyr-Arg-Lys-Gln-Met-Ala-Val-Lys-Lys-Tyr-Leu-Ala-Ala-Val-Leu-Gly-Lys-Arg-Tyr-Lys-Gln-Arg-Val-Lys-Asn-Lys-NH2 Nesfatin-1 82 Val-Pro-Ile-Asp-Ile-Asp-Lys-Thr-Lys-Val-Gln-Asn-Ile-His-Pro-Val-Glu-Ser-Ala-Lys-Ile-Glu-Pro-Pro-Asp-Thr-Gly-Leu-Tyr-Tyr-Asp-Glu-Tyr-Leu-Lys-Gln-Val-Ile-Asp-Val-Leu-Glu-Thr-Asp-Lys-His-Phe-Arg-Glu-Lys-Leu-Gln-Lys-Ala-Asp-Ile-Glu-Glu-Ile-Lys-Ser-Gly-Arg-Leu-Ser-Lys-Glu-Leu-Asp-Leu-Val-Ser-His-His-Val-Arg-Thr-Lys-Leu-Asp-Glu-Leu [218] Somatostatin 14 Ala-Gly-Cys-Lys-Asn-Phe-Phe-Trp-Lys-Thr-Phe-Thr-Ser-Cys [229] 28 Ser-Ala-Asn-Ser-Asn-Pro-Ala-Met-Ala-Pro-Arg-Glu-Arg-Lys-Ala-Gly-Cys-Lys-Asn-Phe-Phe-Trp-Lys-Thr-Phe-Thr-Ser-Cys 4. Conclusions PD, a multifaceted and incapacitating condition, is depicted by the devastation of DA-forming nerve cells inside the SN-PC, which ultimately culminates into DA scantiness in the striatal area. The malady is represented by four key manifestations, viz., rigor, postural deformities, tremor, and bradykinesia. Although the exact etiopathology of the condition is perplexing, multifactorial, and equivocal, the procured data greatly suggested that aging, genetic predisposition, and subjection to environmental toxins unitedly participate in the emergence of PD. The pathological pathways indulged in PD encompasses the clumping of α-synuclein within the LBs and lewy neurites, oxidative stress, apoptosis, neuronal-inflammation, and abnormalities in the operation of mitochondria, ALP, and UPS. Presently, the therapy with the help of DA precursor (levodopa), DA agonists, COMT inhibitors, and MAO-B inhibitors fundamentally concentrates on the mitigation of disease-concerned manifestations, but hitherto no therapeutic approach has been signified to halt the advancement of the disease. NPs are termed as tiny, protein-comprehending additional messenger substances that are fundamentally generated and liberated by nerve cells inside the entire nervous system. NPs are synthesized in the cell body from their large protein precursors, designated as prepropeptides (which are synthesized on palade granules at the endoplasmic reticulum and processed by means of the Golgi apparatus). The NPs are mainly transcripted and translated from the prepropeptides genes. Additionally, the activation of prepropeptides carried by proteases results in the conversion of prepropeptides into propeptides, and lastly, NPs are derived, following the stimulation of converting biocatalysts. Proteolytic processing consequentially partakes in the activation, partial inactivation, or inactivation of the modulatory peptides, such as NPs. Proteases bring about the breakdown, and as a result, may activate, inactivate, or liberate other proteins/peptides. Amongst the two, i.e., cell-surface proteases and proteases, which are liberated by the cells, the cell-surface proteases exhibit enormously greater regulatory and specialized functions. The cell-surface proteases display their action by deteriorating the two, i.e., the bioactive peptides and the cellular functions. Aside from the proteolytic deterioration and synthetic alterations, the removal by means of filtration/diffusion is of utmost and remarkable importance. The preponderance of the cell-surface proteases has been originally depicted as the clearing of biocatalysts, even though they exhibit the potential to cleave the peptides possessing not more than 80 residues. To date, they are presumed as the modulatory proteases (such as ACE, ECE, NEP, and DPP IV) that possess the tremendous tendency to modulate the activation or inactivation of NPs. After the completion of the synthesis, NPs are stored within the large and dense vesicles, and finally they are released via exocytosis. Subsequently, NPs undergo interaction with GPCRs so as to instigate their biological actions and regulate nerve cell operation. NPs consequentially participate in the regulation of the immune system, biological equilibrium (such as the biotransformation of blood sugar, blood pressure, equilibrium in the water content, stress reaction, feeding behavior, and pain), and nerve cell protection. In the current methodical review, the authors spotlighted the emerging status and nerve cell protective role of various NPs, encompassing SP, ghrelin, NPY, neurotensin, PACAP, nesfatin-1, and SST in PD. Pursuant to present-day literature, changes in the expressions of above-stated NPs, as well as their respective GPCRs, were significantly observed within the PD-associated areas, predominantly the SN-striatal area. Numerous pathways, comprehending the suppression of microglia stimulation, cytotoxicity, programmed cell death, oxidative stress, inflammation, autophagy, nerve cell toxicity, the stimulation of chondriosomal bioenergetics, and the de-escalation of disease-related manifestations, emerge to be entailed in NPs-prompted nerve cell protective action in PD. In addition, the trio, namely analogs (septide, Dpr3ghr, HM01, neurotensin2, and neurotensin4), agonists (septide, senktide, HM01, and PACAP (1–38)), and antagonists (NAT, L-733060, LY303870, and [D-Lys3]-GHRP6), of these NPs were as well employed in order to safeguard against nerve cell toxin-precipitated DArgic nerve cell devastation, as well as motor and nonmotor deficiencies, thereby furnishing a newfangled and propitious therapeutic perspective for the management of PD. At present, there are multifarious concerns associated with the peptide therapy, such as the natural peptides exhibiting deprived absorption, distribution, metabolism, and excretion (ADME) profiles, together with little half-life, expeditious clearance, and little solubility/permeability. These issues might be addressed by the continual exploration in this respective domain so as to discover whether these alterations in the content of NPs within the plasma/CSF can be employed as biomarkers in PD. Numerous strategies have emerged with the potential to enormously upgrade the peptide stability by means of structural alteration, escalating permeability and half-life, and de-escalating the clearance and proteolysis. Multiple strategies not only upgrade the stability, but as well upgrade the additional ADME properties; for instance, the conjugation to larger molecules might upgrade the stability and plummet the renal clearance, and cyclic ring formation might elevate the stability as well as the permeability. Further explorations of pharmacokinetic and pharmacodynamic properties/models might render detailed insights in the area of peptide therapy evolution with tremendous efficacy and safety, and minimal deleterious repercussions [235]. In addition, further experimental and clinical investigations are remarkably required in order to attain an exhaustive knowledge regarding NPs, their analogs, agonists, antagonists, their mode of action in bringing nerve cell protective action and overcoming nerve cell devastation, and to open neoteric, exciting, and magnificent gateways in the therapy of PD. Acknowledgments The authors are thankful to the Chitkara College of Pharmacy, Chitkara University, Punjab, India, for providing the various resources for the completion of the article. Author Contributions Conceptualization, T.B. and P.M.; methodology, T.B. and A.S.; investigation, T.B., P.M. and S.S.; resources, T.B. and S.B.; data curation, T.B., P.M., H.A.M. and M.A.; writing—original draft preparation, T.B. and P.M.; writing—review and editing, H.A.A. and A.M.M.; visualization, A.S. and T.B.; supervision, T.B. and S.B. All authors have read and agreed to the published version of the manuscript. Funding Funding for the publication of this paper were provided by the University of Oradea, Oradea, Romania, by an Internal Project. Conflicts of Interest The authors declare no conflict of interest. Abbreviations AC, adenyl cyclase; ACE, angiotensin-converting enzyme; ACh, acetylcholine; ADCYAP1, adenylate cyclase activating polypeptide 1; ADME, absorption, distribution, metabolism, and excretion; ALP, autophagy lysosomal pathway; AMPK, 5′ adenosine monophosphate-activated protein kinase; ATP13A2, neuronal P-type adenosine triphosphate (ATP)ase gene; BAs, biogenic amines; Bax, Bcl-2-associated X protein; Bcl-2, B-cell lymphoma-2; BDNF, brain-derived neurotrophic factor; BiP, binding immunoglobulin protein; BNST, bed nucleus of the stria terminalis; C, carboxy; Ca2+, calcium ions; cAMP, cyclic adenosine monophosphate; CArgic, catecholaminergic; CAT, catalase; CCl4, carbon tetrachloride; CD11b, cluster of differentiation molecule 11b; CH3OH, methanol; CNS, central nervous system; C=O, carbonyl; COMT, catechol-O-methyltransferase; CR, caloric restriction; CSF, cerebrospinal fluid; CTSB, cathepsin B; CTSD, cathepsin D; Cu, copper; Cu-Zn SOD, copper-zincsuperoxide dismutase; C57BL/6, C57 black 6; DA, dopamine; DArgic, dopaminergic; DJ-1, protein deglycase; DNA, deoxyribonucleic acid; DPP IV, dipeptidyl peptidase IV; D2R, DA 2 receptors; ECE, endothelin-converting enzyme; ELISA, enzyme-linked immunosorbent assay; ENS, enteric nervous system; ER stress, endoplasmic reticulum stress; ERK, extracellular signal-regulated kinase; ETC, electron transport chain; Fe, iron; GBA, glucocerebrosidase; GH, growth hormone; GHIH, growth hormone-inhibiting hormone; GhrelinR, ghrelin receptor; GHS-R1a, growth hormone secretagogue receptor 1a; Gi, G inhibitory; GPCRs, G-protein coupled receptors; GRP78, 78-kDa glucose-regulated protein; GSH, glutathione; HCN, hyperpolarization-activated cyclic nucleotide-gated channels; HSC70, heat shock cognate protein 70;HSPs, heat shock proteins; HSP35, hereditary spastic paraplegia type 35; HTRA2, high-temperature requirement A2; H2O, water; IFNs, interferons; IFN-γ, interferon-gamma; ILs, interleukins; IL-1β, interleukin-1β; IL-6, interleukin-6; INS, intrinsic nervous system; InsP3, inositol 1,4,5-trisphosphate; K+, potassium; KATP, ATP-sensitive potassium channel; LAMP1, lysosome-associated membrane protein 1; LAMP2A, lysosome-associated membrane protein 2A; LBs, lewy bodies; LC, locus coeruleus; LC3-II, microtubule-associated protein light chain 3-phosphatidylethanolamine conjugate; LDH, lactate dehydrogenase; LOOH, lipid hydroperoxides; LPS, lipopolysaccharide; LRRK2, leucine-rich repeat kinase 2; LY303870, lanepitant; MAO-B, monoamine oxidase B; MAPK, mitogen-activated protein kinase; MDA, malondialdehyde; MES23.5, DArgic nerve cell line; MGO, methylglyoxal; MMP-9, matrix metallopeptidase 9; MPP+, 1-methyl-4-phenylpyridinium ion; MPTP, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; mRNA, messenger ribonucleic acid; MtMP, mitochondrial membrane potential; NADPH, reduced nicotinamide adenine dinucleotide phosphate; NAT, N-acetyl-L-tryptophan; NEP, neutral endopeptidase; NF-κB, Nuclear Factor kappa-light-chain-enhancer of activated B cells; NK, neurokinin; NKA, neurokinin A; NKB, neurokinin B; NK1, neurokinin 1; NK2, neurokinin 2; NK3, neurokinin 3; NK1R, NK1 receptor; NMDA, N-methyl-D-aspartate; NO, nitric oxide; NPs, neuropeptides; NPY, neuropeptide Y; NTR1, neurotensin receptor 1; NTR2, neurotensin receptor 2; NTR3, neurotensin receptor 3; NUCB2, nucleobindin-2; O2, oxygen; PA28, proteasome activator 28;PA700, proteasome activator 700; PAC1, PACAP type I receptor; PACAP, pituitary adenylate cyclase-activating polypeptide; Parkin, Parkin RBR E3 ubiquitin–protein ligase; PCNA, proliferating cell nuclear antigen; p-CREB, phosphorylated cAMP-response element binding protein; PD, Parkinson’s disease; PERC, perchloroethylene; PFF, preformed fibrils; PGCs, progenitor cells; PI3K, phosphoinositide 3-kinase; PINK1, PTEN-induced kinase 1; PKA, protein kinase A; PKB/Akt, protein kinase B signaling pathway; PKCδ, protein kinase C delta; PLC-β, phospholipase C-beta; PNS, peripheral nervous system; PP, pancreatic polypeptide; PYY, peptide YY; RGCs, retinal ganglion cells; ROS, reactive oxygen species; S, sulphur; SH-SY5Y, human neuroblastoma cells; SMPD1, sphingomyelin phosphodiesterase 1,SN, substantia nigra; SNCA, α-synuclein; SN-PC, substantia nigra pars compacta; SOD, superoxide dismutase; SP, substance P; SST, somatostatin; SSTR1, SST receptor 1; SSTR2, SST receptor 2; SSTR3, SST receptor 3; SSTR4, SST receptor 4; SSTR5, SST receptor 5; SVZ, subventricular zone; SYNJ1, synaptojanin 1; TAC1, tachykinin precursor 1 gene; TCE, trichloroethylene; TH, tyrosine hydroxylase; TK, tachykinin; TMEM175, transmembrane protein 175; TNF-α, tumor necrosis factor-alpha; TRAF6, tumor necrosis factor receptor-associated factor 6; UCHL1, ubiquitin carboxy (C)-terminal hydrolase L1; UCP2, uncoupling protein 2; UPS, ubiquitin-proteasome system; VIP, vasoactive intestinal polypeptide; VPAC1, VIP receptor 1; VPAC2, VIP receptor 2; VPS10P, vacuolar protein sorting 10 protein; VPS35, vacuolar protein sorting 35; XBP1s, spliced form of X-box binding protein 1; Y1R, Y1 receptor; Y2R, Y2 receptor; Y3R, Y3 receptor; Y4R, Y4 receptor; Y5R, Y5 receptor; Y6R, Y6 receptor; Zn, zinc; 5-HT, 5-hydroxytryptamine; 6-OHDA, 6-hydroxy DA; 7TM, seven transmembrane domain. Figure 1 The diagram illustrates the expansion in the incidence rate of Parkinson’s disease with aging, and also outlines the motor and nonmotor manifestations associated with Parkinson’s disease. Aging is reckoned as the critical parameter actively engaged in the evolution of PD. The incidence rate of PD escalates with aging, i.e., people ranging under 30–40 years rarely experience PD, 2% of people ranging above 60–70 years experience PD, and 5% of people ranging above 80–90 years experience PD. The motor manifestations of the disease comprehends rigor, postural deformities, tremor, and bradykinesia. On the other hand, the nonmotor manifestations associated with the disease comprehends urinary dysfunction, sexual abnormalities, sleep disturbances, psychosis, dementia, anxiety, apathy, depression, constipation, and erectile dysfunction. Howbeit, the nonmotor manifestations are comparably less noticeable than motor manifestations. PD, Parkinson’s disease; ↑, increasing. Figure 2 Portraying the active participation of genetic mutations, environmental toxins, and pathogenic processes in the progression/evolvement of Parkinson’s disease. The trio, namely genetic mutations (SNCA, Parkin, DJ-1, UCHL1, PINK1, GBA, LRRK2, VPS35, ATP13A2, HTRA2, and SYNJ1), subjection to environmental toxins (MPTP, 6-OHDA, CH3OH, PERC, TCE, CCl4, zineb, paraquat, rotenone, dieldrin, ziram, thiram, nabam, maneb, copper, and iron), and pathogenic mechanisms (oxidative stress, dysfunction of ALP, abnormality in UPS, mitochondrial devastation, nerve cell inflammation, clumping of α-synuclein, and programmed cell death), give rise to dopaminergic nerve cell demise, which as a result culminates in the progression/evolvement of PD. SNCA, α-synuclein; Parkin, Parkin RBR E3 ubiquitin–protein ligase; DJ-1, protein deglycase; UCHL1, ubiquitin carboxy (C)-terminal hydrolase L1; PINK1, PTEN-induced kinase 1; GBA, glucocerebrosidase; LRRK2, leucine-rich repeat kinase 2; VPS35, vacuolar protein sorting 35; ATP13A2, neuronal P-type adenosine triphosphate (ATP)ase gene; HTRA2, high temperature requirement A2; SYNJ1, synaptojanin 1; MPTP, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; 6-OHDA, 6-hydroxy DA; CH3OH, methanol; PERC, perchloroethylene; TCE, trichloroethylene; CCl4, carbon tetrachloride; ALP, autophagy lysosomal pathway; UPS, ubiquitin-proteasome system; PD, Parkinson’s disease. Figure 3 Highlighting the neuropeptides that contribute to significant nerve cell protection in Parkinson’s disease, encompassing substance P, ghrelin, neuropeptide Y, neurotensin, PACAP, nesfatin-1, and somatostatin. PACAP, pituitary adenylate cyclase-activating polypeptide. Figure 4 Highlighting the neuroprotective role of Substance P in Parkinson’s disease. SP upon interaction with NK1 receptor, decreases Ca2+ entry, caspase-3 stimulation, ROS formation, and modulates MtMP, which in turn inhibit programmed cell death and cytotoxicity, and thereby protects MES23.5 cells from MPP+-prompted neurotoxicity. Septide, an analog of SP, upon interaction with NK1 receptor culminates in the suppression of programmed cell death pathways and stimulation of PKB/Akt signaling pathway, thereby protects nerve cells against 6-OHDA-prompted neurotoxicity. Further, senktide, upon interaction with NK3 receptor, reinstated the temporal order memory in the 6-OHDA-lesioned hemiparkinsonian rat model. In addition, the three NK1 receptor antagonists, namely NAT, L-733060, and LY303870, upon introduction, decrease cellular demise provoked by 6-OHDA subjection, upgrade motor operations, and decrease levodopa-precipitated dyskinesia. Finally, by virtue of these mechanisms, SP markedly contributes to neuroprotective action in PD. Subtraction symbol indicates inhibitory/suppressing action, while addition symbol indicates stimulatory action. SP, Substance P; NAT, N-acetyl-L-tryptophan; LY303870, lanepitant; NK, neurokinin; NK1, neurokinin 1; NK3, neurokinin 3; Ca2+, calcium ions; ROS, reactive oxygen species; MtMP, mitochondrial membrane potential; MES23.5, DArgic nerve cell line; MPP+, 1-methyl-4-phenylpyridinium ion; PKB/Akt, protein kinase B signaling pathway; 6-OHDA, 6-hydroxy DA; PD, Parkinson’s disease; ↓, decreasing or reducing. Figure 5 Highlighting the neuroprotective role of ghrelin in Parkinson’s disease. Ghrelin, along with its analogs, namely HM01 and Dpr3ghr, markedly render nerve cell protective actions against MPTP, MPP+, and rotenone-provoked neuronal destruction, via suppressing oxidative stress, programmed cell death, and de-escalating inflammation. Ghrelin, following its interaction with ghrelinR/GHS-R1a, suppresses K+ channels, which, as a consequence, contributes to DA nerve cell excitation. In addition, ghrelin, by means of the canonical Wnt signaling pathway, increases the number or amount of NSCs of the midbrain and in vitro and ex vivo DArgic nerve cell differentiation. Moreover, ghrelin treatment regulates MtMP, suppresses mitochondrial complex-I operation, de-escalates ROS formation and caspase-3 stimulation, and eventually restrains nerve cells from MPP+ or rotenone-instigated detrimental repercussions. Additionally, ghrelin escalates the levels of antioxidant biocatalysts, namely CAT and Cu-Zn SOD, de-escalates the MDA levels, and suppresses the NF-κB translocation, and thereby suppresses the oxidative stress via ceasing the lipid peroxidation and generation of ROS, and finally contributes to nerve cell protective action. Moreover, ghrelin possesses its nerve cell protective action via activating AMPK and escalating the mitophagy, which subsequently culminates in the amplification in the chondriosomal bioenergetics. Furthermore, therapy with the assistance of ghrelin culminates in the suppression of microglia stimulation, which, as a result, de-escalates inflammation (via de-escalating the levels of TNF-α and IL-1β), and finally culminates into nerve cell protective action. Apart from this, the two, namely ghrelin and Dpr3ghr, suppress programmed cell death and possess substantial nerve cell protective action by elevating Bcl-2 expression, and declining Bax expression and the Bax/Bcl-2 ratio. Subtraction symbol indicates inhibitory/suppressing action, while addition symbol indicates stimulatory action. GhrelinR, ghrelin receptor; GHS-R1a, growth hormone secretagogue receptor 1a; K+, potassium; DA, dopamine; NSCs, neural stem cells; DArgic, dopaminergic; MtMP, mitochondrial membrane potential; ROS, reactive oxygen species; MPP+, 1-methyl-4-phenylpyridinium ion; CAT, catalase; Cu-Zn SOD, copper-zincsuperoxide dismutase; NF-κB, Nuclear Factor kappa-light-chain-enhancer of activated B cells; MDA, malondialdehyde; AMPK, 5′ adenosine monophosphate-activated protein kinase; TNF-α,tumor necrosis factor-alpha; IL-1β, interleukin-1β;Bcl-2,B-cell lymphoma-2; Bax, Bcl-2-associated X protein; MPTP, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; ↑, increasing; ↓, decreasing. Figure 6 Highlighting the neuroprotective role of neuropeptide Y in Parkinson’s disease. NPY exhibits its significant nerve cell protective action upon interacting with DA-forming cells by means of Y2R. NPY stimulates the two, namely the Akt signaling pathway and the MAPK signaling pathway and contributes to the enhanced viability of DA-forming nerve cells of the nigral area of the encephalon. In addition, NPY exerts its nerve cell protective action by suppressing the microglia, which in turn suppresses the NO formation and IL-1β liberation, and finally results in the suppression of inflammation. Furthermore, NPY, via stimulating the PI3K-XBP1s-Bip/GRP78 signaling pathway, suppresses the ER stress-provoked nerve cell demise, and ultimately contributes to neuroprotection. Moreover, therapy with the assistance of NPY de-escalated the stimulation/operation of caspase-3 and caspase-4, which consequently suppressed the programmed cell death and ER stress-triggered nerve cell demise and contributed to nerve cell protection in PD. Subtraction symbol indicates inhibitory/suppressing action, while addition symbol indicates stimulatory action. NPY, neuropeptide Y; Y2R,Y2 receptor; DArgic, dopaminergic; SN, substantia nigra; MAPK, mitogen-activated protein kinase; NO, nitric oxide; IL-1β, interleukin-1β; PI3K, phosphoinositide 3-kinase; XBP1s, spliced form of X-box binding protein 1; BiP, binding immunoglobulin protein; GRP78, 78-kDa glucose-regulated protein; ER stress, endoplasmic reticulum stress; DA, dopamine; PD, Parkinson’s disease; ↑, increasing; ↓, decreasing. Figure 7 Highlighting the neuroprotective role of pituitary adenylate cyclase-activating polypeptide in Parkinson’s disease. The duo, namely experimental animal and cellular PD models illustrate that PACAP therapy renders significant nerve cell protective action against rotenone, MPTP, and 1-methyl-6,7-dihydroxy-1,2,3,4-tetrahydroisochinolin-provoked deleterious repercussions through suppressing inflammation, autophagy, and cellular demise. The prior therapy of SH-SY5Y cells with PACAP (1–38) safeguarded these cells against inflammatory-mediated detrimental effects via decreasing the liberation of inflammatory mediators. PACAP regulates the sequestosome-1/p62 protein concentrations and elevates the LC3-II formation, and thereby suppresses the autophagic-operation. PACAP safeguarded against rotenone-precipitated cellular demise via carrying out the stimulation of PKA signaling process, as well as the two downstream signals, viz., p38 MAPK and ERK. In addition, intravenously introduced PACAP27 safeguarded against MPTP-instigated nerve cell demise via altering the DArgic and cholinergic synaptic conveyance, by way of elevating the D2R operation and the KATP subunits expression within the striatal region of the basal nuclei. Further, therapy with the assistance of PACAP/agonists of PACAP receptors elevated the BDNF and p-CREB expression, and suppressed the caspase-3 expression, and, as a consequence, decreased the 1-methyl-6,7-dihydroxy-1,2,3,4-tetrahydroisochinolin-prompted SH-SY5Y cells toxicity. Apart from this, PACAP has the tendency to considerably suppress the LPS-prompted microglia stimulation, which in turn suppresses the formation and liberation of NO and TNF-α, and finally suppresses the inflammation. Subtraction symbol indicates inhibitory/suppressing action, while addition symbol indicates stimulatory action. PACAP, pituitary adenylate cyclase-activating polypeptide; VIP, vasoactive intestinal polypeptide; PAC1, PACAP type I receptor; VPAC1, VIP receptor 1; VPAC2, VIP receptor 2; NO, nitric oxide; TNF-α, tumor necrosis factor-alpha; IL-1β, interleukin-1β; IL-6, interleukin-6; ROS, reactive oxygen species; MMP-9, matrix metallopeptidase 9; LC3-II, microtubule-associated protein light chain 3-phosphatidylethanolamine conjugate; PKA, protein kinase A; MAPK, mitogen-activated protein kinase; ERK, extracellular signal-regulated kinase; D2R, DA 2 receptors; KATP, ATP-sensitive potassium channel; DArgic, dopaminergic; MPTP, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; BDNF, brain-derived neurotrophic factor; p-CREB, phosphorylated cAMP-response element binding protein; SH-SY5Y, human neuroblastoma cells; LPS, lipopolysaccharide; PD, Parkinson’s disease; ↑, increasing; ↓, decreasing. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Palacios-Sánchez L. Nupan M.T. Botero-Meneses J.S. James Parkinson and His Essay on “Shaking Palsy”, Two Hundred Years Later Arq. Neuro-Psiquiatr. 2017 75 671 672 10.1590/0004-282x20170108 28977149 2. Hayes M.T. Parkinson’s Disease and Parkinsonism Am. J. Med. 2019 132 802 807 10.1016/j.amjmed.2019.03.001 30890425 3. Dawson T.M. Parkinson’s Disease: Clinical Manifestations and Treatment Int. Rev. Psychiatry 2000 12 263 269 10.1080/09540260020002488 4. Surmeier D.J. Determinants of Dopaminergic Neuron Loss in Parkinson’s Disease FEBS J. 2018 285 3657 3668 10.1111/febs.14607 30028088 5. Muangpaisan W. Hori H. Brayne C. Systematic Review of the Prevalence and Incidence of Parkinson’s Disease in Asia J. Epidemiol. 2009 19 281 293 10.2188/jea.JE20081034 19801887 6. Zou Y.-M. Liu J. Tian Z.-Y. Lu D. Zhou Y.-Y. Systematic Review of the Prevalence and Incidence of Parkinson’s Disease in the People’s Republic of China Neuropsychiatr. Dis. Treat. 2015 11 1467 1472 10.2147/NDT.S85380 26109861 7. Picillo M. Nicoletti A. Fetoni V. Garavaglia B. Barone P. Pellecchia M.T. The Relevance of Gender in Parkinson’s Disease: A Review J. Neurol. 2017 264 1583 1607 10.1007/s00415-016-8384-9 28054129 8. Pang S.Y.Y. Ho P.W.L. Liu H.F. Leung C.T. Li L. Chang E.E.S. Ramsden D.B. Ho S.L. The Interplay of Aging, Genetics and Environmental Factors in the Pathogenesis of Parkinson’s Disease Transl. Neurodegener. 2019 8 23 10.1186/s40035-019-0165-9 31428316 9. Veldman B.A.J. Wijn A.M. Knoers N. Praamstra P. Horstink M.W.I.M. Genetic and Environmental Risk Factors in Parkinson’s Disease Clin. Neurol. Neurosurg. 1998 100 15 26 10.1016/S0303-8467(98)00009-2 9637199 10. Georgiou A. Demetriou C.A. Christou Y.P. Heraclides A. Leonidou E. Loukaides P. Yiasoumi E. Pantziaris M. Kleopa K.A. Papacostas S.S. Genetic and Environmental Factors Contributing to Parkinson’s Disease: A Case-Control Study in the Cypriot Population Front. Neurol. 2019 10 1047 10.3389/fneur.2019.01047 31681140 11. Wang H. Yang F. Zhang S. Xin R. Sun Y. Genetic and Environmental Factors in Alzheimer’s and Parkinson’s Diseases and Promising Therapeutic Intervention via Fecal Microbiota Transplantation NPJ Park. Dis. 2021 7 70 10.1038/s41531-021-00213-7 34381040 12. Thakur M. Rachamalla M. Niyogi S. Datusalia A.K. Flora S.J.S. Molecular Mechanism of Arsenic-Induced Neurotoxicity including Neuronal Dysfunctions Int. J. Mol. Sci. 2021 22 10077 10.3390/ijms221810077 34576240 13. Patel R.S. Rachamalla M. Chary N.R. Shera F.Y. Tikoo K. Jena G. Cytarabine Induced Cerebellar Neuronal Damage in Juvenile Rat: Correlating Neurobehavioral Performance with Cellular and Genetic Alterations Toxicology 2012 293 41 52 10.1016/j.tox.2011.12.005 22212197 14. Bhardwaj S. Kesari K.K. Rachamalla M. Mani S. Ashraf G.M. Jha S.K. Kumar P. Ambasta R.K. Dureja H. Devkota H.P. CRISPR/Cas9 Gene Editing: New Hope for Alzheimer’s Disease Therapeutics J. Adv. Res. 2021 In press 10.1016/j.jare.2021.07.001 15. Kouli A. Torsney K.M. Kuan W.-L. Parkinson’s Disease: Etiology, Neuropathology, and Pathogenesis Parkinson’s Disease: Pathogenesis and Clinical Aspects Codon Publications Brisbane, Australia 2018 16. Iarkov A. Barreto G.E. Grizzell J.A. Echeverria V. Strategies for the Treatment of Parkinson’s Disease: Beyond Dopamine Front. Aging Neurosci. 2020 12 4 10.3389/fnagi.2020.00004 32076403 17. Armstrong M.J. Okun M.S. Diagnosis and Treatment of Parkinson Disease: A Review JAMA—J. Am. Med. Assoc. 2020 323 548 560 10.1001/jama.2019.22360 18. Kakkar A.K. Dahiya N. Management of Parkinsonߣs Disease: Current and Future Pharmacotherapy Eur. J. Pharmacol. 2015 750 74 81 10.1016/j.ejphar.2015.01.030 25637088 19. Carniglia L. Ramírez D. Durand D. Saba J. Turati J. Caruso C. Scimonelli T.N. Lasaga M. Neuropeptides and Microglial Activation in Inflammation, Pain, and Neurodegenerative Diseases Mediat. Inflamm. 2017 2017 5048616 10.1155/2017/5048616 20. Hökfelt T. Barde S. Xu Z.-Q.D. Kuteeva E. Rüegg J. Le Maitre E. Risling M. Kehr J. Ihnatko R. Theodorsson E. Neuropeptide and Small Transmitter Coexistence: Fundamental Studies and Relevance to Mental Illness Front. Neural Circuits 2018 12 106 10.3389/fncir.2018.00106 30627087 21. Merighi A. Costorage and Coexistence of Neuropeptides in the Mammalian CNS Prog. Neurobiol. 2002 66 161 190 10.1016/S0301-0082(01)00031-4 11943450 22. Merighi A. Salio C. Ferrini F. Lossi L. Neuromodulatory Function of Neuropeptides in the Normal CNS J. Chem. Neuroanat. 2011 42 276 287 10.1016/j.jchemneu.2011.02.001 21385606 23. Surget A. Leman S. Griebel G. Belzung C. Yalcin I. Neuropeptides in Psychiatric Diseases: An Overview with a Particular Focus on Depression and Anxiety Disorders CNS Neurol. Disord.—Drug Targets 2006 5 135 145 10.2174/187152706776359682 16611088 24. Barrett A.J. Proteases Curr. Protoc. Protein Sci. 2000 21 21.1.1 21.1.12 10.1002/0471140864.ps2101s21 18429162 25. Mentlein R. Cell-Surface Peptidases Int. Rev. Cytol. 2004 235 165 213 10.1016/s0074-7696(04)35004-7 15219783 26. Zhang Y. Wang Z. Parks G.S. Civelli O. Novel Neuropeptides as Ligands of Orphan G Protein-Coupled Receptors Curr. Pharm. Des. 2012 17 2626 2631 10.2174/138161211797416110 27. Hökfelt T. Bartfai T. Bloom F. Neuropeptides: Opportunities for Drug Discovery Lancet Neurol. 2003 2 463 472 10.1016/S1474-4422(03)00482-4 12878434 28. Palkovits M. Neuropeptide Messenger Plasticity in the CNS Neurons Following Axotomy Mol. Neurobiol. 1995 10 91 103 10.1007/BF02740669 7576312 29. Catalani E. De Palma C. Perrotta C. Cervia D. Current Evidence for a Role of Neuropeptides in the Regulation of Autophagy BioMed Res. Int. 2017 2017 5856071 10.1155/2017/5856071 28593174 30. Wang S.-Y. Chen L. Xue Y. Xia Y.-J. Substance P Prevents 1-methyl-4-Phenylpyridiniuminduced Cytotoxicity through Inhibition of Apoptosis via Neurokinin-1 Receptors in MES23.5 Cells Mol. Med. Rep. 2012 12 8085 8092 10.3892/mmr.2015.4464 31. Shi L. Du X. Jiang H. Xie J. Ghrelin and Neurodegenerative Disorders—A Review Mol. Neurobiol. 2017 54 1144 1155 10.1007/s12035-016-9729-1 26809582 32. Bayliss J.A. Lemus M. Santos V.V. Deo M. Elsworth J.D. Andrews Z.B. Acylated but not Des-acyl Ghrelin is Neuroprotective in an MPTP Mouse Model of Parkinson’s Disease J. Neurochem. 2016 137 460 471 10.1111/jnc.13576 26872221 33. Li C. Wu X. Liu S. Zhao Y. Zhu J. Liu K. Roles of Neuropeptide Y in Neurodegenerative and Neuroimmune Diseases Front. Neurosci. 2019 13 869 10.3389/fnins.2019.00869 31481869 34. Lazarova M. Popatanasov A. Klissurov R. Stoeva S. Pajpanova T. Kalfin R. Tancheva L. Preventive Effect of Two New Neurotensin Analogues on Parkinson’s Disease Rat Model J. Mol. Neurosci. 2018 66 552 560 10.1007/s12031-018-1171-6 30374780 35. Maasz G. Zrinyi Z. Reglodi D. Petrovics D. Rivnyak A. Kiss T. Jungling A. Tamas A. Pirger Z. Pituitary Adenylate Cyclase-Activating Polypeptide (PACAP) Has Neuroprotective Function in Dopamine-Based Neurodegeneration Developed in Two Parkinsonian Models Dis. Model. Mech. 2017 10 127 139 10.1242/dmm.027185 28067625 36. Dong D. Xie J. Wang J. Neuroprotective Effects of Brain-Gut Peptides: A Potential Therapy for Parkinson’s Disease Neurosci. Bull. 2019 35 1085 1096 10.1007/s12264-019-00407-3 31286411 37. Bai L. Zhang X. Li X. Liu N. Lou F. Ma H. Luo X. Ren Y. Somatostatin Prevents Lipopolysaccharide-Induced Neurodegeneration in the Rat Substantia Nigra by Inhibiting the Activation of Microglia Mol. Med. Rep. 2015 12 1002 1008 10.3892/mmr.2015.3494 25777539 38. Srinivasan E. Chandrasekhar G. Chandrasekhar P. Anbarasu K. Vickram A.S. Karunakaran R. Rajasekaran R. Srikumar P.S. Alpha-Synuclein Aggregation in Parkinson’s Disease Front. Med. 2021 8 736978 10.3389/fmed.2021.736978 39. Coelho M. Ferreira J. Rosa M. Sampaio C. Treatment Options for Non-motor Symptoms in Late-Stage Parkinson’s Disease Expert Opin. Pharmacother. 2008 9 523 535 10.1517/14656566.9.4.523 18312155 40. Yang W. Hamilton J.L. Kopil C. Beck J.C. Tanner C.M. Albin R.L. Dorsey E.R. Dahodwala N. Cintina I. Hogan P. Current and Projected Future Economic Burden of Parkinson’s Disease in the U.S NPJ Park. Dis. 2020 6 15 10.1038/s41531-020-0117-1 41. Cacabelos R. Parkinson’s Disease: From Pathogenesis to Pharmacogenomics Int. J. Mol. Sci. 2017 18 551 10.3390/ijms18030551 42. Hirsch L. Jette N. Frolkis A. Steeves T. Pringsheim T. The Incidence of Parkinson’s Disease: A Systematic Review and Meta-Analysis Neuroepidemiology 2016 46 292 300 10.1159/000445751 27105081 43. Driver J.A. Logroscino G. Gaziano J.M. Kurth T. Incidence and Remaining Lifetime Risk of Parkinson Disease in Advanced Age Neurology 2009 72 432 438 10.1212/01.wnl.0000341769.50075.bb 19188574 44. Crispino P. Gino M. Barbagelata E. Ciarambino T. Politi C. Ambrosino I. Ragusa R. Marranzano M. Biondi A. Vacante M. Gender Differences and Quality of Life in Parkinson’s Disease Int. J. Environ. Res. Public Health 2021 18 198 10.3390/ijerph18010198 45. Cerri S. Mus L. Blandini F. Parkinson’s Disease in Women and Men: What’s the Difference? J. Parkinson’s Dis. 2019 9 501 515 10.3233/JPD-191683 31282427 46. Maiti P. Manna J. Dunbar G.L. Current Understanding of the Molecular Mechanisms in Parkinson’s Disease: Targets for Potential Treatments Transl. Neurodegener. 2017 6 28 10.1186/s40035-017-0099-z 29090092 47. Liu H. Koros C. Strohäker T. Schulte C. Bozi M. Varvaresos S. Ibáñez de Opakua A. Simitsi A.M. Bougea A. Voumvourakis K. A Novel SNCA A30G Mutation Causes Familial Parkinson’s Disease Mov. Disord. 2021 36 1624 1633 10.1002/mds.28534 33617693 48. Magistrelli L. Contaldi E. Comi C. The Impact of SNCA Variations and Its Product Alpha-Synuclein on Non-Motor Features of Parkinson’s Disease Life 2021 11 804 10.3390/life11080804 34440548 49. Kamienieva I. Duszyński J. Szczepanowska J. Multitasking Guardian of Mitochondrial Quality: Parkin Function and Parkinson’s Disease Transl. Neurodegener. 2021 10 5 10.1186/s40035-020-00229-8 33468256 50. Nawaz M.S. Asghar R. Pervaiz N. Ali S. Hussain I. Xing P. Bao Y. Abbasi A.A. Molecular Evolutionary and Structural Analysis of Human UCHL1 Gene Demonstrates the Relevant Role of Intragenic Epistasis in Parkinson’s Disease and Other Neurological Disorders BMC Evol. Biol. 2020 20 130 10.1186/s12862-020-01684-7 33028204 51. Vizziello M. Borellini L. Franco G. Ardolino G. Disruption of Mitochondrial Homeostasis: The Role of PINK1 in Parkinson’s Disease Cells 2021 10 3022 10.3390/cells10113022 34831247 52. Repici M. Giorgini F. DJ-1 in Parkinson’s Disease: Clinical Insights and Therapeutic Perspectives J. Clin. Med. 2019 8 1377 10.3390/jcm8091377 31484320 53. Hur E.-M. Lee B.D. LRRK2 at the Crossroad of Aging and Parkinson’s Disease Genes 2021 12 505 10.3390/genes12040505 33805527 54. Cerri S. Ghezzi C. Ongari G. Croce S. Avenali M. Zangaglia R. Di Monte D.A. Valente E.M. Blandini F. GBA Mutations Influence the Release and Pathological Effects of Small Extracellular Vesicles from Fibroblasts of Patients with Parkinson’s Disease Int. J. Mol. Sci. 2021 22 2215 10.3390/ijms22042215 33672321 55. Luo A. Xu Z. Liao S. VPS35, the Core Component of the Retromer Complex, and Parkinson’s Disease Ibrain 2021 7 318 324 10.1002/ibra.12004 56. Park J.-S. Blair N.F. Sue C.M. The Role of ATP13A2 in Parkinson’s Disease: Clinical Phenotypes and Molecular Mechanisms Mov. Disord. 2015 30 770 779 10.1002/mds.26243 25900096 57. He Y.-C. Huang P. Li Q.-Q. Sun Q. Li D.-H. Wang T. Shen J.-Y. Du J.-J. Cui S.-S. Gao C. Mutation Analysis ofHTRA2Gene in Chinese Familial Essential Tremor and Familial Parkinson’s Disease Park. Dis. 2017 2017 3217474 10.1155/2017/3217474 58. Lesage S. Mangone G. Tesson C. Bertrand H. Benmahdjoub M. Kesraoui S. Arezki M. Singleton A. Corvol J.-C. Brice A. Clinical Variability of SYNJ1-Associated Early-Onset Parkinsonism Front. Neurol. 2021 12 648457 10.3389/fneur.2021.648457 33841314 59. Behl T. Madaan P. Sehgal A. Singh S. Sharma N. Bhatia S. Al-Harrasi A. Chigurupati S. Alrashdi I. Bungau S.G. Elucidating the Neuroprotective Role of PPARs in Parkinson’s Disease: A Neoteric and Prospective Target Int. J. Mol. Sci. 2021 22 10161 10.3390/ijms221810161 34576325 60. Thirugnanam T. Santhakumar K. Chemically Induced Models of Parkinson’s Disease Comp. Biochem. Physiol. Part C Toxicol. Pharmacol. 2022 252 109213 10.1016/j.cbpc.2021.109213 61. Kochmanski J. VanOeveren S.E. Patterson J.R. Bernstein A.I. Developmental Dieldrin Exposure Alters DNA Methylation at Genes Related to Dopaminergic Neuron Development and Parkinson’s Disease in Mouse Midbrain Toxicol. Sci. 2019 169 593 607 10.1093/toxsci/kfz069 30859219 62. Jia Z. Misra H.P. Exposure to Mixtures of Endosulfan and Zineb Induces Apoptotic and Necrotic Cell Death in SH-SY5Y Neuroblastoma Cells, In Vitro J. Appl. Toxicol. 2007 27 434 446 10.1002/jat.1218 17309119 63. Martin C.A. Myers K.M. Chen A. Martin N.T. Barajas A. Schweizer F.E. Krantz D.E. Ziram, a Pesticide Associated with Increased Risk for Parkinson’s Disease, Differentially Affects the Presynaptic Function of Aminergic and Glutamatergic Nerve Terminals at the Drosophila Neuromuscular Junction Exp. Neurol. 2016 275 232 241 10.1016/j.expneurol.2015.09.017 26439313 64. Pouchieu C. Piel C. Carles C. Gruber A. Helmer C. Tual S. Marcotullio E. LeBailly P. Baldi I. Pesticide Use in Agriculture and Parkinson’s Disease in the AGRICAN Cohort Study Int. J. Epidemiol. 2018 47 299 310 10.1093/ije/dyx225 29136149 65. Anderson C.C. Marentette J.O. Rauniyar A.K. Prutton K.M. Khatri M. Matheson C. Reisz J.A. Reigan P. D’Alessandro A. Roede J.R. Maneb Alters Central Carbon Metabolism and Thiol Redox Status in a Toxicant Model of Parkinson’s Disease Free Radic. Biol. Med. 2021 162 65 76 10.1016/j.freeradbiomed.2020.11.028 33279619 66. Lock E.A. Zhang J. Checkoway H. Solvents and Parkinson Disease: A Systematic Review of Toxicological and Epidemiological Evidence Toxicol. Appl. Pharmacol. 2013 266 345 355 10.1016/j.taap.2012.11.016 23220449 67. Goldman S.M. Quinlan P.J. Ross G.W. Marras C. Meng C. Bhudhikanok G.S. Comyns K. Korell M. Chade A.R. Kasten M. Solvent Exposures and Parkinson Disease Risk in Twins Ann. Neurol. 2012 71 776 784 10.1002/ana.22629 22083847 68. Behl T. Madaan P. Sehgal A. Singh S. Anwer K. Makeen H.A. Albratty M. Mohan S. Bungau S. Mechanistic Insights Expatiating the Redox-Active-Metal-Mediated Neuronal Degeneration in Parkinson’s Disease Int. J. Mol. Sci. 2022 23 678 10.3390/ijms23020678 35054862 69. Wang X. Michaelis E.K. Selective Neuronal Vulnerability to Oxidative Stress in the Brain Front. Aging Neurosci. 2010 2 12 10.3389/fnagi.2010.00012 20552050 70. Parker W.D. Jr. Parks J.K. Swerdlow R.H. Complex I Deficiency in Parkinson’s Disease Frontal Cortex Brain Res. 2008 1189 215 218 10.1016/j.brainres.2007.10.061 18061150 71. Anderson R.F. Harris T.A. Dopamine and Uric Acid Act as Antioxidants in the Repair of DNA Radicals: Implications in Parkinson’s Disease Free Radic. Res. 2003 37 1131 1136 10.1080/10715760310001604134 14703803 72. Lotharius J. Brundin P. Pathogenesis of Parkinson’s Disease: Dopamine, Vesicles and α-synuclein Nat. Rev. Neurosci. 2002 3 932 942 10.1038/nrn983 12461550 73. Aoyama K. Glutathione in the Brain Int. J. Mol. Sci. 2021 22 5010 10.3390/ijms22095010 34065042 74. Wei Z. Li X. Li X. Liu Q. Cheng Y. Oxidative Stress in Parkinson’s Disease: A Systematic Review and Meta-Analysis Front. Mol. Neurosci. 2018 11 236 10.3389/fnmol.2018.00236 30026688 75. Wang J.-Y. Zhuang Q.-Q. Zhu L.-B. Zhu H. Li T. Li R. Chen S.-F. Huang C.-P. Zhang X. Zhu J.-H. Meta-Analysis of Brain Iron Levels of Parkinson’s Disease Patients Determined by Postmortem and MRI Measurements Sci. Rep. 2016 6 36669 10.1038/srep36669 27827408 76. De Farias C.C. Maes M. Bonifácio K.L. Bortolasci C.C. de Souza Nogueira A. Brinholi F.F. Matsumoto A.K. do Nascimento M.A. De Melo L.B. Nixdorf S.L. Highly Specific Changes in Antioxidant Levels and Lipid Peroxidation in Parkinson’s Disease and Its Progression: Disease and Staging Biomarkers and New Drug Targets Neurosci. Lett. 2016 617 66 71 10.1016/j.neulet.2016.02.011 26861200 77. Jiménez-Jiménez F. Alonso-Navarro H. Herrero M. García-Martín E. Agúndez J. An Update on the Role of Nitric Oxide in the Neurodegenerative Processes of Parkinson’s Disease Curr. Med. Chem. 2016 23 2666 2679 10.2174/0929867323666160812151356 27776473 78. Subedi L. Gaire B. Parveen A. Kim S.-Y. Nitric Oxide as a Target for Phytochemicals in Anti-Neuroinflammatory Prevention Therapy Int. J. Mol. Sci. 2021 22 4771 10.3390/ijms22094771 33946349 79. Senkevich K. Gan-Or Z. Autophagy Lysosomal Pathway Dysfunction in Parkinson’s Disease; Evidence from Human Genetics Park. Relat. Disord. 2020 73 60 71 10.1016/j.parkreldis.2019.11.015 80. Alvarez-Erviti L. Rodriguez-Oroz M.C. Cooper J.M. Caballero C. Ferrer I. Obeso J.A. Schapira A.H.V. Chaperone-Mediated Autophagy Markers in Parkinson Disease Brains Arch. Neurol. 2010 67 1464 1472 10.1001/archneurol.2010.198 20697033 81. Hou X. Watzlawik J.O. Fiesel F.C. Springer W. Autophagy in Parkinson’s Disease J. Mol. Biol. 2020 432 2651 2672 10.1016/j.jmb.2020.01.037 32061929 82. Lynch-Day M.A. Mao K. Wang K. Zhao M. Klionsky D.J. The Role of Autophagy in Parkinson’s Disease Cold Spring Harb. Perspect. Med. 2012 2 a009357 10.1101/cshperspect.a009357 22474616 83. Beilina A. Cookson M.R. Genes Associated with Parkinson’s Disease: Regulation of Autophagy and Beyond J. Neurochem. 2016 139 91 107 10.1111/jnc.13266 26223426 84. McNaught K.S. Jenner P. Proteasomal Function Is Impaired in Substantia Nigra in Parkinson’s Disease Neurosci. Lett. 2001 297 191 194 10.1016/S0304-3940(00)01701-8 11137760 85. Zheng Q. Huang T. Zhang L. Zhou Y. Luo H. Xu H. Wang X. Dysregulation of Ubiquitin-Proteasome System in Neurodegenerative Diseases Front. Aging Neurosci. 2016 8 303 10.3389/fnagi.2016.00303 28018215 86. Dawson T.M. Dawson V.L. Rare Genetic Mutations Shed Light on the Pathogenesis of Parkinson Disease J. Clin. Investig. 2003 111 145 151 10.1172/JCI200317575 12531866 87. Zucchelli S. Codrich M. Marcuzzi F. Pinto M. Vilotti S. Biagioli M. Ferrer I. Gustincich S. TRAF6 Promotes Atypical Ubiquitination of Mutant DJ-1 and Alpha-Synuclein and Is Localized to Lewy Bodies in Sporadic Parkinson’s Disease Brains Hum. Mol. Genet. 2010 19 3759 3770 10.1093/hmg/ddq290 20634198 88. McNaught K.S.P. Belizaire R. Isacson O. Jenner P. Olanow C.W. Altered Proteasomal Function in Sporadic Parkinson’s Disease Exp. Neurol. 2003 179 38 46 10.1006/exnr.2002.8050 12504866 89. Prasuhn J. Davis R.L. Kumar K.R. Targeting Mitochondrial Impairment in Parkinson’s Disease: Challenges and Opportunities Front. Cell Dev. Biol. 2021 8 615461 10.3389/fcell.2020.615461 33469539 90. Park J.-S. Davis R.L. Sue C.M. Mitochondrial Dysfunction in Parkinson’s Disease: New Mechanistic Insights and Therapeutic Perspectives Curr. Neurol. Neurosci. Rep. 2018 18 21 10.1007/s11910-018-0829-3 29616350 91. Bose A. Beal M.F. Mitochondrial Dysfunction in Parkinson’s Disease J. Neurochem. 2016 139 216 231 10.1111/jnc.13731 27546335 92. Tanner C.M. Kamel F. Ross G.W. Hoppin J.A. Goldman S.M. Korell M. Marras C. Bhudhikanok G.S. Kasten M. Chade A.R. Rotenone, Paraquat, and Parkinson’s Disease Environ. Health Perspect. 2011 119 866 872 10.1289/ehp.1002839 21269927 93. Bogaerts V. Theuns J. Van Broeckhoven C. Genetic Findings in Parkinson’s Disease and Translation into Treatment: A Leading Role for Mitochondria? Genes Brain Behav. 2008 7 129 151 10.1111/j.1601-183X.2007.00342.x 17680806 94. Anglade P. Vyas S. Javoy-Agid F. Herrero M.T. Michel P.P. Marquez J. Mouatt-Prigent A. Ruberg M. Hirsch E.C. Agid Y. Apoptosis and Autophagy in Nigral Neurons of Patients with Parkinson’s Disease Histol. Histopathol. 1997 12 25 31 9046040 95. Hunot S. Brugg B. Ricard D. Michel P.P. Muriel M.-P. Ruberg M. Faucheux B.A. Agid Y. Hirsch E.C. Nuclear Translocation of NF-κB is Increased in Dopaminergic Neurons of Patients with Parkinson Disease Proc. Natl. Acad. Sci. USA 1997 94 7531 7536 10.1073/pnas.94.14.7531 9207126 96. Moujalled D. Strasser A. Liddell J.R. Molecular Mechanisms of Cell Death in Neurological Diseases Cell Death Differ. 2021 28 2029 2044 10.1038/s41418-021-00814-y 34099897 97. Mount M.P. Lira A. Grimes D. Smith P.D. Faucher S. Slack R. Anisman H. Hayley S. Park D.S. Involvement of Interferon-γ in Microglial-Mediated Loss of Dopaminergic Neurons J. Neurosci. 2007 27 3328 3337 10.1523/JNEUROSCI.5321-06.2007 17376993 98. Bellucci A. Bubacco L. Longhena F. Parrella E. Faustini G. Porrini V. Bono F. Missale C. Pizzi M. Nuclear Factor-κB Dysregulation and α-Synuclein Pathology: Critical Interplay in the Pathogenesis of Parkinson’s Disease Front. Aging Neurosci. 2020 12 68 10.3389/fnagi.2020.00068 32265684 99. Hartmann A. Mouatt-Prigent A. Vila M. Abbas N. Perier C. Faucheux B.A. Vyasb S. Hirsch E.C. Increased Expression and Redistribution of the Antiapoptotic Molecule Bcl-xL in Parkinson’s Disease Neurobiol. Dis. 2002 10 28 32 10.1006/nbdi.2002.0494 12079401 100. Badanjak K. Fixemer S. Smajić S. Skupin A. Grünewald A. The Contribution of Microglia to Neuroinflammation in Parkinson’s Disease Int. J. Mol. Sci. 2021 22 4676 10.3390/ijms22094676 33925154 101. Zhang W. Wang T. Pei Z. Miller D.S. Wu X. Block M.L. Wilson B. Zhang W. Zhou Y. Hong J.-S. Aggregated α-synuclein Activates Microglia: A Process Leading to Disease Progression in Parkinson’s Disease FASEB J. 2005 19 533 542 10.1096/fj.04-2751com 15791003 102. He S. Zhong S. Liu G. Yang J. Alpha-Synuclein: The Interplay of Pathology, Neuroinflammation, and Environmental Factors in Parkinson’s Disease Neurodegener. Dis. 2021 20 55 64 10.1159/000511083 33465773 103. Abeliovich A. Schmitz Y. Fariñas I. Choi-Lundberg D. Ho W.-H. Castillo P. Shinsky N. García-Verdugo J.M. Armanini M. Ryan A. Mice Lacking α-Synuclein Display Functional Deficits in the Nigrostriatal Dopamine System Neuron 2000 25 239 252 10.1016/S0896-6273(00)80886-7 10707987 104. Conway K.A. Harper J.D. Lansbury P.T. Accelerated In Vitro Fibril Formation by a Mutant α-synuclein Linked to Early-Onset Parkinson Disease Nat. Med. 1998 4 1318 1320 10.1038/3311 9809558 105. Greenbaum E.A. Graves C.L. Mishizen-Eberz A.J. Lupoli M.A. Lynch D.R. Englander S.W. Axelsen P.H. Giasson B.I. The E46K Mutation in α-Synuclein Increases Amyloid Fibril Formation J. Biol. Chem. 2005 280 7800 7807 10.1074/jbc.M411638200 15632170 106. Volpicelli-Daley L.A. Luk K.C. Lee V.M.-Y. Addition of Exogenous α-Synuclein Preformed Fibrils to Primary Neuronal Cultures to Seed Recruitment of Endogenous α-Synuclein to Lewy Body and Lewy Neurite–like Aggregates Nat. Protoc. 2014 9 2135 2146 10.1038/nprot.2014.143 25122523 107. Luk K.C. Song C. O’Brien P. Stieber A. Branch J.R. Brunden K.R. Trojanowski J.Q. Lee V.M.-Y. Exogenous α-Synuclein Fibrils Seed the Formation of Lewy Body-like Intracellular Inclusions in Cultured Cells Proc. Natl. Acad. Sci. USA 2009 106 20051 20056 10.1073/pnas.0908005106 19892735 108. Meier F. Abeywardana T. Dhall A. Marotta N.P. Varkey J. Langen R. Chatterjee C. Pratt M.R. Semisynthetic, Site-Specific Ubiquitin Modification of α-Synuclein Reveals Differential Effects on Aggregation J. Am. Chem. Soc. 2012 134 5468 5471 10.1021/ja300094r 22404520 109. Tanik S.A. Schultheiss C.E. Volpicelli-Daley L.A. Brunden K.R. Lee V.M.Y. Lewy Body-like α-Synuclein Aggregates Resist Degradation and Impair Macroautophagy J. Biol. Chem. 2013 288 15194 15210 10.1074/jbc.M113.457408 23532841 110. Euler U.S.V. Gaddum J.H. An Unidentified Depressor Substance in Certain Tissue Extracts J. Physiol. 1931 72 74 87 10.1113/jphysiol.1931.sp002763 16994201 111. Chang M.M. Leeman S.E. Niall H.D. Amino-acid Sequence of Substance P Nat. New Biol. 1971 232 86 87 10.1038/newbio232086a0 5285346 112. Nicoll R.A. Schenker C. Leeman S.E. Substance P as a Transmitter Candidate Annu. Rev. Neurosci. 1980 3 227 268 10.1146/annurev.ne.03.030180.001303 6158283 113. Warden M.K. Young W.S. Distribution of Cells Containing mRNAs Encoding Substance P and Neurokinin B in the Rat Central Nervous System J. Comp. Neurol. 1988 272 90 113 10.1002/cne.902720107 2454979 114. Mashaghi A. Marmalidou A. Tehrani M. Grace P.M. Pothoulakis C. Dana R. Neuropeptide Substance P and the Immune Response Cell Mol. Life Sci. 2016 73 4249 4264 10.1007/s00018-016-2293-z 27314883 115. Rana T. Behl T. Sehgal A. Singh S. Sharma N. Abdeen A. Ibrahim S.F. Mani V. Iqbal M.S. Bhatia S. Exploring the Role of Neuropeptides in Depression and Anxiety Prog. Neuro-Psychopharmacol. Biol. Psychiatry 2022 114 110478 10.1016/j.pnpbp.2021.110478 116. Mauborgne A. Javoy-Agid F. Legrand J.C. Agid Y. Cesselin F. Decrease of Substance P-like Immunoreactivity in the Substantia Nigra and Pallidum of Parkinsonian Brains Brain Res. 1983 268 167 170 10.1016/0006-8993(83)90403-1 6190539 117. Tenovuo O. Rinne U.K. Viljanen M.K. Substance P Immunoreactivity in the Post-mortem Parkinsonian Brain Brain Res. 1984 303 113 116 10.1016/0006-8993(84)90217-8 6203617 118. Schröder J.B. Marian T. Claus I. Muhle P. Pawlowski M. Wiendl H. Suntrup-Krueger S. Meuth S.G. Dziewas R. Ruck T. Substance P Saliva Reduction Predicts Pharyngeal Dysphagia in Parkinson’s Disease Front. Neurol. 2019 10 386 10.3389/fneur.2019.00386 31040820 119. Lindefors N. Brodin E. Tossman U. Segovia J. Ungerstedt U. Tissue Levels and in Vivo Release of Tachykinins and GABA in Striatum and Substantia Nigra of Rat Brain after Unilateral Striatal Dopamine Denervation Exp. Brain Res. 1989 74 527 534 10.1007/BF00247354 2468514 120. Thornton E. Vink R. Treatment with a Substance P Receptor Antagonist Is Neuroprotective in the Intrastriatal 6-Hydroxydopamine Model of Early Parkinson’s Disease PLoS ONE 2012 7 e34138 10.1371/journal.pone.0034138 22485158 121. Zheng Y. Zhang L. Xie J. Shi L. The Emerging Role of Neuropeptides in Parkinson’s Disease Front. Aging Neurosci. 2021 13 90 10.3389/fnagi.2021.646726 122. Thornton E. Tran T.T.B. Vink R. A Substance P Mediated Pathway Contributes to 6-hydroxydopamine Induced Cell Death Neurosci. Lett. 2010 481 64 67 10.1016/j.neulet.2010.06.057 20599590 123. Chu J.M.T. Chen L.W. Chan Y.S. Yung K.K.L. Neuroprotective Effects of Neurokinin Receptor One in Dopaminergic Neurons Are Mediated through Akt/PKB Cell Signaling Pathway Neuropharmacology 2011 61 1389 1398 10.1016/j.neuropharm.2011.08.027 21907219 124. Amadoro G. Pieri M. Ciotti M.T. Carunchio I. Canu N. Calissano P. Zona C. Severini C. Substance P Provides Neuroprotection in Cerebellar Granule Cells through Akt and MAPK/Erk Activation: Evidence for the Involvement of the Delayed Rectifier Potassium Current Neuropharmacology 2007 52 1366 1377 10.1016/j.neuropharm.2007.01.020 17397881 125. Chao O.Y. Wang A.-L. Nikolaus S. De Souza Silva M.A. NK3 Receptor Agonism Reinstates Temporal Order Memory in the Hemiparkinsonian Rat Behav. Brain Res. 2015 285 208 212 10.1016/j.bbr.2014.06.006 24928770 126. Thornton E. Hassall M.M. Corrigan F. Vink R. The NK1 Receptor Antagonist N-acetyl-l-tryptophan Reduces Dyskinesia in a Hemi-parkinsonian Rodent Model Park. Relat. Disord. 2014 20 508 513 10.1016/j.parkreldis.2014.02.008 127. Yang X. Zhao H. Shi H. Wang X. Zhang S. Zhang Z. Zu J. Zhang W. Shen X. Cui G. Intranigral Administration of Substance P Receptor Antagonist Attenuated Levodopa-Induced Dyskinesia in a Rat Model of Parkinson’s Disease Exp. Neurol. 2015 271 168 174 10.1016/j.expneurol.2015.05.007 26001615 128. Kim W.-G. Mohney R.P. Wilson B. Jeohn G.-H. Liu B. Hong J.-S. Regional Difference in Susceptibility to Lipopolysaccharide-Induced Neurotoxicity in the Rat Brain: Role of Microglia J. Neurosci. 2000 20 6309 6316 10.1523/JNEUROSCI.20-16-06309.2000 10934283 129. Kojima M. Hosoda H. Date Y. Nakazato M. Matsuo H. Kangawa K. Ghrelin Is a Growth-Hormone-Releasing Acylated Peptide from Stomach Nature 1999 402 656 660 10.1038/45230 10604470 130. Serrenho D. Santos S.D. Carvalho A.L. The Role of Ghrelin in Regulating Synaptic Function and Plasticity of Feeding-Associated Circuits Front. Cell. Neurosci. 2019 13 205 10.3389/fncel.2019.00205 31191250 131. St-Pierre D.H. Wang L. Taché Y. Ghrelin: A Novel Player in the Gut-Brain Regulation of Growth Hormone and Energy Balance Physiology 2003 18 242 246 10.1152/nips.01460.2003 132. Frago L.M. Ebaquedano E. Eargente J. Chowen J.A. Neuroprotective Actions of Ghrelin and Growth Hormone Secretagogues Front. Mol. Neurosci. 2011 4 23 10.3389/fnmol.2011.00023 21994488 133. Rees D. Johnson A.L. Lelos M. Smith G. Roberts L.D. Phelps L. Dunnett S.B. Morgan A.H. Brown R.M. Wells T. Acyl-Ghrelin Attenuates Neurochemical and Motor Deficits in the 6-OHDA Model of Parkinson’s Disease bioRxiv 2022 10.1101/2022.01.31.478447 134. He X. Yuan W. Liu F. Feng J. Guo Y. Acylated Ghrelin is Protective Against 6-OHDA-induced Neurotoxicity by Regulating Autophagic Flux Front. Pharmacol. 2021 11 586302 10.3389/fphar.2020.586302 33584263 135. Wagner J. Vulinović F. Grünewald A. Unger M.M. Möller J.C. Klein C. Michel P.P. Ries V. Oertel W.H. Alvarez-Fischer D. Acylated and Unacylated Ghrelin Confer Neuroprotection to Mesencephalic Neurons Neuroscience 2017 365 137 145 10.1016/j.neuroscience.2017.09.045 28987508 136. Song N. Wang W. Jia F. Du X. Xie A. He Q. Shen X. Zhang J. Rogers J. Xie J. Assessments of Plasma Ghrelin Levels in the Early Stages of Parkinson’s Disease Mov. Disord. 2017 32 1487 1491 10.1002/mds.27095 28681931 137. Andrews Z.B. Erion D. Beiler R. Liu Z.-W. Abizaid A. Zigman J. Elsworth J.D. Savitt J.M. DiMarchi R. Tschoep M. Ghrelin Promotes and Protects Nigrostriatal Dopamine Function via a UCP2-Dependent Mitochondrial Mechanism J. Neurosci. 2009 29 14057 14065 10.1523/JNEUROSCI.3890-09.2009 19906954 138. Suda Y. Kuzumaki N. Sone T. Narita M. Tanaka K. Hamada Y. Iwasawa C. Shibasaki M. Maekawa A. Matsuo M. Down-Regulation of Ghrelin Receptors on Dopaminergic Neurons in the Substantia Nigra Contributes to Parkinson’s Disease-like Motor Dysfunction Mol. Brain 2018 11 6 10.1186/s13041-018-0349-8 29458391 139. Jiang H. Li L.-J. Wang J. Xie J.-X. Ghrelin Antagonizes MPTP-induced Neurotoxicity to the Dopaminergic Neurons in Mouse Substantia Nigra Exp. Neurol. 2008 212 532 537 10.1016/j.expneurol.2008.05.006 18577498 140. Wang H. Dou S. Zhu J. Shao Z. Wang C. Cheng B. Ghrelin Protects Dopaminergic Neurons against MPTP Neurotoxicity through Promoting Autophagy and Inhibiting Endoplasmic Reticulum Mediated Apoptosis Brain Res. 2020 1746 147023 10.1016/j.brainres.2020.147023 32710901 141. Moon M. Kim H.G. Hwang L. Seo J.-H. Kim S. Hwang S. Kim S. Lee D. Chung H. Oh M.S. Neuroprotective Effect of Ghrelin in the 1-Methyl-4-Phenyl-1,2,3,6-Tetrahydropyridine Mouse Model of Parkinson’s Disease by Blocking Microglial Activation Neurotox. Res. 2009 15 332 347 10.1007/s12640-009-9037-x 19384567 142. Yu J. Xu H. Shen X. Jiang H. Ghrelin Protects MES23.5 Cells against Rotenone via Inhibiting Mitochondrial Dysfunction and Apoptosis Neuropeptides 2016 56 69 74 10.1016/j.npep.2015.09.011 26459609 143. Liu S. Chen S. Ren J. Li B. Qin B. Ghrelin Protects Retinal Ganglion Cells against Rotenone via Inhibiting Apoptosis, Restoring Mitochondrial Function, and Activating AKT-mTOR Signaling Neuropeptides 2018 67 63 70 10.1016/j.npep.2017.11.007 29174113 144. Bayliss J.A. Lemus M.B. Stark R. Santos V.V. Thompson A. Rees D.J. Galic S. Elsworth J.D. Kemp B.E. Davies J.S. Ghrelin-AMPK Signaling Mediates the Neuroprotective Effects of Calorie Restriction in Parkinson’s Disease J. Neurosci. 2016 36 3049 3063 10.1523/JNEUROSCI.4373-15.2016 26961958 145. Shi L. Bian X. Qu Z. Ma Z. Zhou Y. Wang K. Jiang H. Xie J. Peptide Hormone Ghrelin Enhances Neuronal Excitability by Inhibition of Kv7/KCNQ Channels Nat. Commun. 2013 4 1435 10.1038/ncomms2439 23385580 146. Chang X. Ma Z. Shi L. Xie J. Effects of Ghrelin on the Electrical Activities of Substantia Nigra Dopaminergic Neurons Treated with MPP+ Neurochem. Int. 2020 138 104780 10.1016/j.neuint.2020.104780 32569790 147. Gong B. Jiao L. Du X. Li Y. Bi M. Jiao Q. Jiang H. Ghrelin Promotes Midbrain Neural Stem Cells Differentiation to Dopaminergic Neurons through Wnt/β-catenin Pathway J. Cell. Physiol. 2020 235 8558 8570 10.1002/jcp.29699 32329059 148. Minalyan A. Gabrielyan L. Pietra C. Taché Y. Wang L. Multiple Beneficial Effects of Ghrelin Agonist, HM01 on Homeostasis Alterations in 6-Hydroxydopamine Model of Parkinson’s Disease in Male Rats Front. Integr. Neurosci. 2019 13 13 10.3389/fnint.2019.00013 31031602 149. Popelová A. Kákonová A. Hrubá L. Kunes J. Maletínská L. Železná B. Potential Neuroprotective and Anti-Apoptotic Properties of a Long-Lasting Stable Analog of Ghrelin: An In Vitro Study Using SH-SY5Y Cells Physiol. Res. 2018 67 339 346 10.33549/physiolres.933761 29303606 150. Morgan A.H. Rees D.J. Andrews Z.B. Davies J.S. Ghrelin Mediated Neuroprotection—A Possible Therapy for Parkinson’s Disease? Neuropharmacology 2018 136 317 326 10.1016/j.neuropharm.2017.12.027 29277488 151. Dong J. Song N. Xie J. Jiang H. Ghrelin Antagonized 1-Methyl-4-Phenylpyridinium (MPP+)-Induced Apoptosis in MES23.5 Cells J. Mol. Neurosci. 2009 37 182 189 10.1007/s12031-008-9162-7 19052922 152. Liu L. Xu H. Jiang H. Wang J. Song N. Xie J. Ghrelin Prevents 1-methyl-4-phenylpyridinium Ion-induced Cytotoxicity through Antioxidation and NF-κB Modulation in MES23.5 cells Exp. Neurol. 2010 222 25 29 10.1016/j.expneurol.2009.11.009 19931250 153. Bayliss J.A. Andrews Z.B. Ghrelin Is Neuroprotective in Parkinson’s Disease: Molecular Mechanisms of Metabolic Neuroprotection Ther. Adv. Endocrinol. Metab. 2013 4 25 36 10.1177/2042018813479645 23515333 154. Tatemoto K. Carlquist M. Mutt V. Neuropeptide Y—A Novel Brain Peptide with Structural Similarities to Peptide YY and Pancreatic Polypeptide Nature 1982 296 659 660 10.1038/296659a0 6896083 155. Gehlert D.R. Introduction to the Reviews on Neuropeptide Y Neuropeptides 2004 38 135 140 10.1016/j.npep.2004.07.002 15337366 156. Tatemoto K. Neuropeptide Y: Complete Aamino Acid Sequence of the Brain Peptide Proc. Natl. Acad. Sci. USA 1982 79 5485 5489 10.1073/pnas.79.18.5485 6957876 157. Shiozaki K. Kawabe M. Karasuyama K. Kurachi T. Hayashi A. Ataka K. Iwai H. Takeno H. Hayasaka O. Kotani T. Neuropeptide Y Deficiency Induces Anxiety-like Behaviours in Zebrafish (Danio rerio) Sci. Rep. 2020 10 5913 10.1038/s41598-020-62699-0 32246073 158. Pedragosa-Badia X. Stichel J. Beck-Sickinger A.G. Neuropeptide Y Receptors: How to Get Subtype Selectivity Front. Endocrinol. 2013 4 5 10.3389/fendo.2013.00005 159. Lee C.C. Miller R.J. Is There Really an NPY Y3 Receptor? Regul. Pept. 1998 75–76 71 78 10.1016/S0167-0115(98)00054-8 160. Starbäck P. Wraith A. Eriksson H. Larhammar D. Neuropeptide Y Receptor Gene y6: Multiple Deaths or Resurrections? Biochem. Biophys. Res. Commun. 2000 277 264 269 10.1006/bbrc.2000.3656 11027673 161. dos Santos V.V. Santos D.B. Lach G. Rodrigues A.L.S. Farina M. De Lima T.C.M. Prediger R. Neuropeptide Y (NPY) Prevents Depressive-like Behavior, Spatial Memory Deficits and Oxidative Stress Following Amyloid-β (Aβ1–40) Administration in Mice Behav. Brain Res. 2013 244 107 115 10.1016/j.bbr.2013.01.039 23396168 162. Malva J.O. Xapelli S. Baptista S. Valero J. Agasse F. Ferreira R. Silva A.P. Multifaces of Neuropeptide Y in the Brain—Neuroprotection, Neurogenesis and Neuroinflammation Neuropeptides 2012 46 299 308 10.1016/j.npep.2012.09.001 23116540 163. Stoddard S.L. Tyce G.M. Ahlskog J. Zinsmeister A.R. Nelson D.K. Carmichael S.W. Decreased Levels of [Met]Enkephalin, Neuropeptide Y, Substance P, and Vasoactive Intestinal Peptide in Parkinsonian Adrenal Medulla Exp. Neurol. 1991 114 23 27 10.1016/0014-4886(91)90080-V 1915731 164. Martignoni E. Blandini F. Petraglia F. Pacchetti C. Bono G. Nappi G. Cerebrospinal Fluid Norepinephrine, 3-Methoxy-4-Hydroxyphenylglycol and Neuropeptide Y Levels in Parkinson’s Disease, Multiple System Atrophy and Dementia of the Alzheimer Type J. Neural Transm. 1992 4 191 205 10.1007/BF02260903 1320891 165. Cannizzaro C. Tel B.C. Rose S. Zeng B.-Y. Jenner P. Increased Neuropeptide Y mRNA Expression in Striatum in Parkinson’s Disease Mol. Brain Res. 2003 110 169 176 10.1016/S0169-328X(02)00555-7 12591154 166. Goto S. Kawarai T. Morigaki R. Okita S. Koizumi H. Nagahiro S. Munoz E.L. Lee L.V. Kaji R. Defects in the Striatal Neuropeptide Y System in X-Linked Dystonia-Parkinsonism Brain 2013 136 1555 1567 10.1093/brain/awt084 23599389 167. Svenningsson P. Pålhagen S. Mathé A.A. Neuropeptide Y and Calcitonin Gene-Related Peptide in Cerebrospinal Fluid in Parkinson’s Disease with Comorbid Depression versus Patients with Major Depressive Disorder Front. Psychiatry 2017 8 102 10.3389/fpsyt.2017.00102 28659833 168. Obuchowicz E. Antkiewicz-Michaluk L. Romańska I. Herman Z.S. Increased Striatal Neuropeptide Y Immunoreactivity and Its Modulation by Deprenyl, Clonidine and L-dopa in MPTP-treated Mice J. Neural Transm. 2003 110 1375 1391 10.1007/s00702-003-0047-1 14666410 169. Decressac M. Pain S. Chabeauti P.-Y. Frangeul L. Thiriet N. Herzog H. Vergote J. Chalon S. Jaber M. Gaillard A. Neuroprotection by Neuropeptide Y in Cell and Animal Models of Parkinson’s Disease Neurobiol. Aging 2012 33 2125 2137 10.1016/j.neurobiolaging.2011.06.018 21816512 170. Pain S. Vergote J. Gulhan Z. Bodard S. Chalon S. Gaillard A. Inflammatory Process in Parkinson Disease: Neuroprotection by Neuropeptide Y Fundam. Clin. Pharmacol. 2019 33 544 548 10.1111/fcp.12464 30866091 171. Ferreira R. Xapelli S. Santos T. Silva A.P. Cristóvão A. Cortes L. Malva J.O. Neuropeptide Y Modulation of Interleukin-1β (IL-1β)-induced Nitric Oxide Production in Microglia J. Biol. Chem. 2010 285 41921 41934 10.1074/jbc.M110.164020 20959451 172. Lee D.Y. Hong S.H. Kim B. Lee D.S. Yu K. Lee K.-S. Neuropeptide Y Mitigates ER Stress–induced Neuronal Cell Death by Activating the PI3K–XBP1 Pathway Eur. J. Cell Biol. 2018 97 339 348 10.1016/j.ejcb.2018.04.003 29650257 173. Sendtner M. Holtmann B. Kolbeck R. Thoenen H. Barde Y. Brain-Derived Neurotrophic Factor Prevents the Death of Motoneurons in Newborn Rats after Nerve Section Nature 1992 360 757 759 10.1038/360757a0 1465147 174. Fumagalli F. Racagni G. Riva M.A. Shedding Light into the Role of BDNF in the Pharmacotherapy of Parkinson’s Disease Pharm. J. 2006 6 95 104 10.1038/sj.tpj.6500360 16402079 175. Carraway R. Leeman S.E. The Isolation of a New Hypotensive Peptide, Neurotensin, from Bovine Hypothalami J. Biol. Chem. 1973 248 6854 6861 10.1016/S0021-9258(19)43429-7 4745447 176. Russjan E. Kaczyńska K. Beneficial Effects of Neurotensin in Murine Model of Hapten-Induced Asthma Int. J. Mol. Sci. 2019 20 5025 10.3390/ijms20205025 177. Carraway R. Leeman S.E. The Amino Acid Sequence of a Hypothalamic Peptide, Neurotensin J. Biol. Chem. 1975 250 1907 1911 10.1016/S0021-9258(19)41780-8 1167549 178. St.-Gelais F. Jomphe C. Trudeau L.É.. The Role of Neurotensin in Central Nervous System Pathophysiology: What Is the Evidence? J. Psychiatry Neurosci. 2006 31 229 245 16862241 179. Wu Z. Martinez-Fong D. Trédaniel J. Forgez P. Neurotensin and Its High Affinity Receptor 1 as a Potential Pharmacological Target in Cancer Therapy Front. Endocrinol. 2013 3 184 10.3389/fendo.2012.00184 180. Mazella J. Béraud-Dufour S. Devader C. Massa F. Coppola T. Neurotensin and Its Receptors in the Control of Glucose Homeostasis Front. Endocrinol. 2012 3 143 10.3389/fendo.2012.00143 181. Jomphe C. Lemelin P.-L. Okano H. Kobayashi K. Trudeau L.-E. Bidirectional Regulation of Dopamine D2 and Neurotensin NTS1 Receptors in Dopamine Neurons Eur. J. Neurosci. 2006 24 2789 2800 10.1111/j.1460-9568.2006.05151.x 17116165 182. Fernandez A. de Ceballos M.L. Rose S. Jenner P. Marsden C.D. Alterations in Peptide Levels in Parkinson’s Disease and Incidental Lewy Body Disease Brain 1996 119 823 830 10.1093/brain/119.3.823 8673494 183. Fernandez A. Jenner P. Marsden C.D. de Ceballos M.L. Characterization of Neurotensin-like Immunoreactivity in Human Basal Ganglia: Increased Neurotensin Levels in Substantia Nigra in Parkinson’s Disease Peptides 1995 16 339 346 10.1016/0196-9781(94)00141-3 7784265 184. Schimpff R.-M. Avard C. Fénelon G. Lhiaubet A.-M. Tennezé L. Vidailhet M. Rostène W. Increased Plasma Neurotensin Concentrations in Patients with Parkinson’s Disease J. Neurol. Neurosurg. Psychiatry 2001 70 784 786 10.1136/jnnp.70.6.784 11385014 185. Yamada M. Richelson E. Heterogeneity of Melanized Neurons Expressing Neurotensin Receptor Messenger RNA in the Substantia Nigra and the Nucleus Paranigralis of Control and Parkinson’s Disease Brain Neuroscience 1995 64 405 417 10.1016/0306-4522(94)00395-L 7700529 186. Uhl G.R. Whitehouse P.J. Price D.L. Tourtelotte W.W. Kuhar M.J. Parkinson’s Disease: Depletion of Substantia Nigra Neurotensin Receptors Brain Res. 1984 308 186 190 10.1016/0006-8993(84)90935-1 6089953 187. Fernandez A. de Ceballos M.L. Jenner P. Marsden C.D. Neurotensin, Substance P, Delta and Mu Opioid Receptors Are Decreased in Basal Ganglia of Parkinson’s Disease Patients Neuroscience 1994 61 73 79 10.1016/0306-4522(94)90061-2 7969897 188. Chinaglia G. Probst A. Palacios J. Neurotensin Receptors in Parkinson’s Disease and Progressive Supranuclear Palsy: An Autoradiographic Study in Basal Ganglia Neuroscience 1990 39 351 360 10.1016/0306-4522(90)90273-7 1965015 189. Rivest R. St-Pierre S. Jolicoeur F.B. Structure-Activity Studies of Neurotensin on Muscular Rigidity and Tremors Induced by 6-Hydroxydopamine Lesions in the Posterolateral Hypothalamus of the Rat Neuropharmacology 1991 30 47 52 10.1016/0028-3908(91)90041-9 1904561 190. Antonelli T. Fuxe K. Tomasini M.C. Mazzoni E. Agnati L.F. Tanganelli S. Ferraro L. Neurotensin Receptor Mechanisms and Its Modulation of Glutamate Transmission in the Brain: Relevance for Neurodegenerative Diseases and Their Treatment Prog. Neurobiol. 2007 83 92 109 10.1016/j.pneurobio.2007.06.006 17673354 191. Gilmartin M.R. Ferrara N.C. Pituitary Adenylate Cyclase-Activating Polypeptide in Learning and Memory Front. Cell. Neurosci. 2021 15 663418 10.3389/fncel.2021.663418 34239418 192. Miyata A. Arimura A. Dahl R.R. Minamino N. Uehara A. Jiang L. Culler M.D. Coy D.H. Isolation of a Novel 38 Residue-Hypothalamic Polypeptide Which Stimulates Adenylate Cyclase in Pituitary Cells Biochem. Biophys. Res. Commun. 1989 164 567 574 10.1016/0006-291X(89)91757-9 2803320 193. Liao C. De Molliens M.P. Schneebeli S.T. Brewer M. Song G. Chatenet D. Braas K.M. May V. Li J. Targeting the PAC1 Receptor for Neurological and Metabolic Disorders Curr. Top. Med. Chem. 2019 19 1399 1417 10.2174/1568026619666190709092647 31284862 194. Shioda S. Takenoya F. Wada N. Hirabayashi T. Seki T. Nakamachi T. Pleiotropic and Retinoprotective Functions of PACAP Anat. Sci. Int. 2016 91 313 324 10.1007/s12565-016-0351-0 27324639 195. Lamine-Ajili A. Fahmy A.M. Létourneau M. Chatenet D. Labonté P. Vaudry D. Fournier A. Effect of the Pituitary Adenylate Cyclase-Activating Polypeptide on the Autophagic Activation Observed in in Vitro and in Vivo Models of Parkinson’s Disease Biochim. Biophys. Acta (BBA)—Mol. Basis Dis. 2016 1862 688 695 10.1016/j.bbadis.2016.01.005 26769362 196. Hirabayashi T. Shibato J. Kimura A. Yamashita M. Takenoya F. Shioda S. Potential Therapeutic Role of Pituitary Adenylate Cyclase-Activating Polypeptide for Dry Eye Disease Int. J. Mol. Sci. 2022 23 664 10.3390/ijms23020664 35054857 197. Saklani P. Khan H. Gupta S. Kaur A. Singh T.G. Neuropeptides: Potential Neuroprotective Agents in Ischemic Injury Life Sci. 2022 288 120186 10.1016/j.lfs.2021.120186 34852271 198. Hannibal J. Pituitary Adenylate Cyclase-Activating Peptide in the Rat Central Nervous System: An Immunohistochemical and in Situ Hybridization Study J. Comp. Neurol. 2002 453 389 417 10.1002/cne.10418 12389210 199. Feher M. Gaszner B. Tamas A. Gil-Martinez A.L. Fernandez-Villalba E. Herrero M.T. Reglodi D. Alteration of the PAC1 Receptor Expression in the Basal Ganglia of MPTP-Induced Parkinsonian Macaque Monkeys Neurotox. Res. 2018 33 702 715 10.1007/s12640-017-9841-7 29230633 200. Reglodi D. Renaud J. Tamas A. Tizabi Y. Socías S.B. Del-Bel E. Raisman-Vozari R. Novel Tactics for Neuroprotection in Parkinson’s Disease: Role of Antibiotics, Polyphenols and Neuropeptides Prog. Neurobiol. 2017 155 120 148 10.1016/j.pneurobio.2015.10.004 26542398 201. Reglodi D. Kiss P. Lubics A. Tamas A. Review on the Protective Effects of PACAP in Models of Neurodegenerative Diseases In Vitro and In Vivo Curr. Pharm. Des. 2011 17 962 972 10.2174/138161211795589355 21524257 202. Reglodi D. Lubics A. Tamás A. Szalontay L. Lengvári I. Pituitary Adenylate Cyclase Activating Polypeptide Protects Dopaminergic Neurons and Improves Behavioral Deficits in a Rat Model of Parkinson’s Disease Behav. Brain Res. 2004 151 303 312 10.1016/j.bbr.2003.09.007 15084446 203. Reglődi D. Tamás A. Lubics A. Szalontay L. Lengvári I. Morphological and Functional Effects of PACAP in 6-Hydroxydopamine-Induced Lesion of the Substantia Nigra in Rats Regul. Pept. 2004 123 85 94 10.1016/j.regpep.2004.05.016 15518897 204. Deguil J. Chavant F. Lafay-Chebassier C. Pérault-Pochat M.-C. Fauconneau B. Pain S. Neuroprotective Effect of PACAP on Translational Control Alteration and Cognitive Decline in MPTP Parkinsonian Mice Neurotox. Res. 2010 17 142 155 10.1007/s12640-009-9091-4 19626386 205. Wang G. Qi C. Fan G.-H. Zhou H.-Y. Chen S.-D. PACAP Protects Neuronal Differentiated PC12 Cells against the Neurotoxicity Induced by a Mitochondrial Complex I Inhibitor, Rotenone FEBS Lett. 2005 579 4005 4011 10.1016/j.febslet.2005.06.013 16004991 206. Brown D. Tamas A. Reglodi D. Tizabi Y. PACAP Protects Against Salsolinol-Induced Toxicity in Dopaminergic SH-SY5Y Cells: Implication for Parkinson’s Disease J. Mol. Neurosci. 2013 50 600 607 10.1007/s12031-013-0015-7 23625270 207. Shivers K.-Y. Nikolopoulou A. Machlovi S.I. Vallabhajosula S. Figueiredo-Pereira M.E. PACAP27 Prevents Parkinson-like Neuronal Loss and Motor Deficits but Not Microglia Activation Induced by Prostaglandin J2 Biochim. Biophys. Acta (BBA)—Mol. Basis Dis. 2014 1842 1707 1719 10.1016/j.bbadis.2014.06.020 208. Brown D. Tamás A. Reglodi D. Tizabi Y. PACAP Protects Against Inflammatory-Mediated Toxicity in Dopaminergic SH-SY5Y Cells: Implication for Parkinson’s Disease Neurotox. Res. 2014 26 230 239 10.1007/s12640-014-9468-x 24740430 209. Wang G. Pan J. Tan Y.-Y. Sun X.-K. Zhang Y.-F. Zhou H.-Y. Ren R.-J. Wang X.-J. Chen S.-D. Neuroprotective Effects of PACAP27 in Mice Model of Parkinson’s Disease Involved in the Modulation of KATP Subunits and D2 Receptors in the Striatum Neuropeptides 2008 42 267 276 10.1016/j.npep.2008.03.002 18440632 210. Yang S. Yang J. Yang Z. Chen P. Fraser A. Zhang W. Pang H. Gao X. Wilson B. Hong J.-S. Pituitary Adenylate Cyclase-Activating Polypeptide (PACAP) 38 and PACAP4–6 Are Neuroprotective through Inhibition of NADPH Oxidase: Potent Regulators of Microglia-Mediated Oxidative Stress J. Pharmacol. Exp. Ther. 2006 319 595 603 10.1124/jpet.106.102236 16891616 211. Said S.I. Mutt V. Polypeptide with Broad Biological Activity: Isolation from Small Intestine Science 1970 169 1217 1218 10.1126/science.169.3951.1217 5450698 212. Korkmaz O.T. Tunçel N. Advantages of Vasoactive Intestinal Peptide for the Future Treatment of Parkinson’s Disease Curr. Pharm. Des. 2019 24 4693 4701 10.2174/1381612825666190111150953 30636594 213. Delgado M. Ganea D. Neuroprotective Effect of Vasoactive Intestinal Peptide (VIP) In a Mouse Model of Parkinson’s Disease by Blocking Microglial Activation FASEB J. 2003 17 1 18 10.1096/fj.02-0799fje 12522106 214. Broome S.T. Musumeci G. Castorina A. PACAP and VIP Mitigate Rotenone-Induced Inflammation in BV-2 Microglial Cells J. Mol. Neurosci. 2022 2022 1 13 10.1007/s12031-022-01968-1 35199308 215. Oh I.S. Shimizu H. Satoh T. Okada S. Adachi S. Inoue K. Eguchi H. Yamamoto M. Imaki T. Hashimoto K. Identification of Nesfatin-1 as a Satiety Molecule in the Hypothalamus Nature 2006 443 709 712 10.1038/nature05162 17036007 216. Rupp S.K. Wölk E. Stengel A. Nesfatin-1 Receptor: Distribution, Signaling and Increasing Evidence for a G Protein-Coupled Receptor—A Systematic Review Front. Endocrinol. 2021 12 740174 10.3389/fendo.2021.740174 217. Goebel-Stengel M. Wang L. Central and Peripheral Expression and Distribution of NUCB2/nesfatin-1 Curr. Pharm. Des. 2013 19 6935 6940 10.2174/138161281939131127124814 23537079 218. Shimizu H. Mori M. Nesfatin-1: Its Role in the Diagnosis and Treatment of Obesity and Some Psychiatric Disorders Calcium-Bind. Proteins RAGE 2013 963 327 338 10.1007/978-1-62703-230-8_20 219. Chen H. Li X. Ma H. Zheng W. Shen X. Reduction in Nesfatin-1 Levels in the Cerebrospinal Fluid and Increased Nigrostriatal Degeneration Following Ventricular Administration of Anti-nesfatin-1 Antibody in Mice Front. Neurosci. 2021 15 621173 10.3389/fnins.2021.621173 33613183 220. Price T.O. Samson W.K. Niehoff M.L. Banks W.A. Permeability of the Blood–Brain Barrier to a Novel Satiety Molecule Nesfatin-1 Peptides 2007 28 2372 2381 10.1016/j.peptides.2007.10.008 18006117 221. Emir G.K. Ünal Y. Yılmaz N. Tosun K. Kutlu G. The Association of Low Levels of Nesfatin-1 and Glucagon-like Peptide-1 with Oxidative Stress in Parkinson’s Disease Neurol. Sci. 2019 40 2529 2535 10.1007/s10072-019-03975-4 31280388 222. Tang C.-H. Fu X.-J. Xu X.-L. Wei X.-J. Pan H.-S. The Anti-inflammatory and Anti-apoptotic Effects of Nesfatin-1 in the Traumatic Rat Brain Peptides 2012 36 39 45 10.1016/j.peptides.2012.04.014 22561242 223. Özsavcí D. Erşahin M. Şener A. Özakpinar Ö.B. Toklu H.Z. Akakín D. Şener G. Yeğen B. The Novel Function of Nesfatin-1 as an Anti-inflammatory and Antiapoptotic Peptide in Subarachnoid Hemorrhage–Induced Oxidative Brain Damage in Rats Neurosurgery 2011 68 1699 1708 10.1227/NEU.0b013e318210f258 21336215 224. Tan Z. Xu H. Shen X. Jiang H. Nesfatin-1 Antagonized Rotenone-Induced Neurotoxicity in MES23.5 Dopaminergic Cells Peptides 2015 69 109 114 10.1016/j.peptides.2015.04.019 25937598 225. Erfani S. Moghimi A. Aboutaleb N. Khaksari M. Protective Effects of Nesfatin-1 Peptide on Cerebral Ischemia Reperfusion Injury via Inhibition of Neuronal Cell Death and Enhancement of Antioxidant Defenses Metab. Brain Dis. 2019 34 79 85 10.1007/s11011-018-0323-2 30269302 226. Shen X.-L. Song N. Du X.-X. Li Y. Xie J.-X. Jiang H. Nesfatin-1 Protects Dopaminergic Neurons against MPP+/MPTP-Induced Neurotoxicity through the C-Raf–ERK1/2-Dependent Anti-Apoptotic Pathway Sci. Rep. 2017 7 40961 10.1038/srep40961 28106099 227. Brazeau P. Vale W. Burgus R. Guillemin R. Isolation of Somatostatin (a Somatotropin Release Inhibiting Factor) of Ovine Hypothalamic Origin Can. J. Biochem. 1974 52 1067 1072 10.1139/o74-148 4609582 228. Patel Y.C. Somatostatin and Its Receptor Family Front. Neuroendocr. 1999 20 157 198 10.1006/frne.1999.0183 229. Kumar U. Grant M. Somatostatin and Somatostatin Receptors Results and Problems in Cell Differentiation Rehfeld J. Bundgaard J. Springer Berlin, Germany 2010 Volume 50 137 184 10.1007/400_2009_29 230. O’Toole T.J. Sharma S. Physiology, Somatostatin StatPearls StatPearls Publishing Treasure Island, FL, USA 2019 231. Weckbecker G. Lewis I. Albert R. Schmid H.A. Hoyer D. Bruns C. Opportunities in Somatostatin Research: Biological, Chemical and Therapeutic Aspects Nat. Rev. Drug Discov. 2003 2 999 1017 10.1038/nrd1255 14654798 232. Reubi J.C. Somatostatin Receptors in the Gastrointestinal Tract in Health and Disease Yale J. Biol. Med. 1992 65 493 536 1340064 233. Reichlin S. Somatostatin (Second of Two Parts) N. Engl. J. Med. 1983 309 1556 1563 10.1056/NEJM198312223092506 6140639 234. Reichlin S. Somatostatin (First of Two Parts) N. Engl. J. Med. 1983 309 1495 1501 10.1056/NEJM198312153092406 6139753 235. Behl T. Rachamalla M. Najda A. Sehgal A. Singh S. Sharma N. Bhatia S. Al-Harrasi A. Chigurupati S. Vargas-De-La-Cruz C. Applications of Adductomics in Chemically Induced Adverse Outcomes and Major Emphasis on DNA Adductomics: A Pathbreaking Tool in Biomedical Research Int. J. Mol. Sci. 2021 22 10141 10.3390/ijms221810141 34576304

PMC009xxxxxx/PMC9099670.txt

==== Front Int J Mol Sci Int J Mol Sci ijms International Journal of Molecular Sciences 1422-0067 MDPI 10.3390/ijms23094789 ijms-23-04789 Article Differences in Tau Seeding in Newborn and Adult Wild-Type Mice https://orcid.org/0000-0001-9888-8754 Ferrer Isidro 123* https://orcid.org/0000-0003-3000-0338 Andrés-Benito Pol 123 Garcia-Esparcia Paula 123 López-Gonzalez Irene 123 https://orcid.org/0000-0002-0924-8454 Valiente Diego 123 Jordán-Pirla Mónica 123 https://orcid.org/0000-0002-9275-9000 Carmona Margarita 123 Sala-Jarque Julia 456 https://orcid.org/0000-0002-7782-5575 Gil Vanessa 456 https://orcid.org/0000-0002-5214-4909 del Rio José Antonio 456 García-Ayllón María-Salud Academic Editor 1 Department of Pathology and Experimental Therapeutics, University of Barcelona, Feixa Llarga sn, 08907 Hospitalet de Llobregat, Spain; pol.andres.benito@gmail.com (P.A.-B.); p.garcies@gmail.com (P.G.-E.); lopez.gonzalez.irene@gmail.com (I.L.-G.); diego.valiente.cerro@outlook.com (D.V.); mjordanpirla@gmail.com (M.J.-P.); mcarmona@idibell.cat (M.C.) 2 Bellvitge Biomedical Research Centre—IDIBELL, Feixa Llarga sn, 08907 Hospitalet de Llobregat, Spain 3 Network Centre of Biomedical Research of Neurodegenerative Diseases—CIBERNED, Institute of Health Carlos III, Feixa Llarga sn, 08907 Hospitalet de Llobregat, Spain 4 Molecular and Cellular Neurobiotechnology, Institute of Bioengineering of Catalonia, Barcelona Institute for Science and Technology, Parc Científic de Barcelona, Baldiri Reixac sn, 08020 Barcelona, Spain; jsala@ibecbarcelona.eu (J.S.-J.); vgil@ibecbarcelona.eu (V.G.); jadelrio@ibecbarcelona.eu (J.A.d.R.) 5 Department of Cell Biology, Physiology and Immunology, Faculty of Biology, University of Barcelona, Baldiri Reixac sn, 08020 Barcelona, Spain 6 Centro de Investigación Biomédica en Red Sobre Enfermedades Neurodegenerativas (CIBERNED), 28031 Madrid, Spain * Correspondence: 8082ifa@gmail.com; Tel.: +34-93-4035808 26 4 2022 5 2022 23 9 478905 3 2022 25 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Alzheimer’s disease (AD) and other tauopathies are common neurodegenerative diseases in older adults; in contrast, abnormal tau deposition in neurons and glial cells occurs only exceptionally in children. Sarkosyl-insoluble fractions from sporadic AD (sAD) containing paired helical filaments (PHFs) were inoculated unilaterally into the thalamus in newborn and three-month-old wild-type C57BL/6 mice, which were killed at different intervals from 24 h to six months after inoculation. Tau-positive cells were scanty and practically disappeared at three months in mice inoculated at the age of a newborn. In contrast, large numbers of tau-positive cells, including neurons and oligodendrocytes, were found in the thalamus of mice inoculated at three months and killed at the ages of six months and nine months. Mice inoculated at the age of newborn and re-inoculated at the age of three months showed similar numbers and distribution of positive cells in the thalamus at six months and nine months. This study shows that (a) differences in tau seeding between newborn and young adults may be related to the ratios between 3Rtau and 4Rtau, and the shift to 4Rtau predominance in adults, together with the immaturity of connections in newborn mice, and (b) intracerebral inoculation of sAD PHFs in newborn mice does not protect from tau seeding following intracerebral inoculation of sAD PHFs in young/adult mice. tau seeding and spreading newborn Alzheimer’s disease thalamus ==== Body pmc1. Introduction Alzheimer’s disease (AD) and tauopathies are neurodegenerative diseases with abnormal accumulation of hyper-phosphorylated tau in neurons and glial cells [1,2]. All of them are disorders characterised by selective cellular vulnerability and stereotyped patterns of disease progression of variable complexity, as exemplified in AD [3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18]. Tau pathology is an initiating factor in sAD [19]. The formation of abnormal protein aggregates in particular cell types, and the progression of the disease, are active and very complex processes involving several steps [20,21,22,23,24,25,26]. In addition to specific neuronal, glial, and regional vulnerability, tau transmission from one neuron to another may occur trans-synaptically [21,24,25,27,28,29,30,31]. Free release of tau to the extracellular space and vesicle-associated tau exocytosis are other mechanisms of tau secretion [21,24,32,33,34,35,36]. Tau can also be transmitted through tunnelling nanotubes which are actin-based nanotubular channels that connect one cell to another [37]. This latter mechanism may apply particularly to glial cells. A curious situation is the extremely rare occurrence of tau pathology in children, despite the active tau phosphorylation during brain development [38,39]. The reason for the underlying resistance to tauopathy in children is not known. Moreover, the initiation of tau pathology in transgenic mice bearing tau mutations, including P301S transgenic mice which express the T34 isoform of microtubule-associated protein tau with one N-terminal insert and four microtubule-binding repeats encoding the human P301S mutation, all driven by prnp promoter, always occurs in young/adult animals, never in newborns [40]. Tauopathy can also be experimentally induced following intracerebral inoculation of tau species in mice under appropriate conditions. Thus, intracerebral inoculation of pre-formed synthetic tau fibrils in transgenic mice expressing human mutant tau induces tau pathology in connected brain regions [41,42,43,44]. Seeding and spreading of abnormal tau also arises following inoculation of brain homogenates from P301S transgenic mice, sporadic AD (sAD), and other tauopathies into the brain of transgenic mice overexpressing human 4Rtau or human mutated tau [29,45,46,47,48]. Tau seeding and propagation may also occur in WT mice following intracerebral inoculation of sarkosyl-insoluble fractions obtained from sAD and various tauopathies [26,49,50,51,52,53,54,55]. Therefore, intracerebral inoculation of tau at different developmental stages would afford us the opportunity to learn about age-dependent vulnerability to developing tau pathology. The present study was geared to learn about the (i) differences in the vulnerability of tau seeding between newborn and young/adult mice aged 3 months following thalamic inoculation of sarkosyl-insoluble fractions of sAD, and (ii) possible protective effects of brain inoculation of sarkosyl-insoluble fractions of sAD homogenates at the age of newborn in mice re-inoculated with the same fractions at the age of three months. To these ends, and to avoid differences linked to particularities of different inoculums, all the animals were inoculated with the same sAD brain homogenate and processed in the same way. 2. Results 2.1. Tau during Normal Brain Development (Group 1) 2.1.1. Tau Species as Revealed by Western Blotting Western blots of total brain homogenates showed high levels of a wide band of tau-5 with a molecular weight of from 50 kDa to 68 kDa in mice aged 15 days. The expression levels of tau-5 were reduced in mice aged 3 months and 12 months. A similar pattern was observed in Western blots incubated with anti-3Rtau antibodies. High levels of 3Rtau of about 50 kDa were identified in mice aged 15 days. 3Rtau levels decreased in mice aged 3 months and again in mice aged 12 months. The expression of 4Rtau differed from the others. Two bands of 68 kDa and 64 kDa occurred in mice aged 15 days, but three main bands of 68 kDa, 64 kDa, and 60 kD, together with a weaker band of about 50 kDa, were visible at the age of 3 months and 12 months. Tau phosphorylation also differed with age. Tau in mice aged 15 days was mainly phosphorylated at Thr231, to decrease in mice aged 3 months and 12 months. However, the levels of phospho-tau at Thr181 and Ser202/Thr305 (antibody AT8) were similar in the three age groups (Figure 1). Quantification of Western blots is shown in the diagram in Figure 1. The levels of Tau 5, 3Rtau, and phospho-tau Thr231 were significantly higher in mice aged 15 days when compared with mice aged 3 months and 12 months, p < 0.001. In contrast, the levels of 4Rtau were significantly higher in mice aged 3 months and 12 months when compared with 15-day-old mice, p < 0.001. 2.1.2. Tau Immunohistochemistry Tau immunohistochemistry reproduced the pattern observed in Western blots. Antibodies against Tau5, 3Rtau, and 4Rtau showed a diffuse pattern in the neuropil, which was more marked in mice aged 15 days when compared with mice aged 3 months and 12 months. The antibody anti-tau Thr231 decorated the cytoplasm and dendrites of neurons in young adults, whereas the antibody AT8 stained the nuclei at every age. Deposits of tau were not found in the cytoplasm of neurons or glial cells at any age. Although 3Rtau decreased and 4Rtau was predominant in the young/adult murine brain, selected neuronal subpopulations expressed 3Rtau in the young/adult brain. 3Rtau-immunoreactive neurons were localised in the inner layer of the granule cells in the dentate gyrus, olfactory bulb, and periventricular layer of the lateral ventricles. In addition, a few 3Rtau-immunoreactive neurons were found in the entorhinal cortex, periventricular hypothalamic nuclei, basal forebrain, and thalamus; very rarely, 3Rtau-immunoreactive neurons were also encountered in the amygdala and cerebral cortex (Figure 2). 2.2. Human Sample: Characterization of Sarkosyl-Insoluble Fraction The characteristics of the patient with sAD are detailed in the section Material and methods. Western blotting of the sarkosyl-insoluble fraction processed with anti-phospho-tau Ser422 antibody revealed three bands of 68 kDa, 64 kDa, and 60 kDa, together with a low upper band of 73 kDa, several bands of about 50 kDa, several bands between 30 kDa and 40 kDa, and two lower bands of truncated tau at the C-terminal, one of which was of about 20 kDa, as well as a smear of higher molecular weight represented tau oligomers (Figure 3A). TEM of the same sarkosyl-insoluble fraction revealed the presence of typical paired helical filaments (Figure 3B). Incubation of sarkosyl-insoluble fraction with thioflavin showed increased fluorescence peaking at three hours and a decrease thereafter (Figure 3C), indicating the presence of amyloid fibrils. Western blotting of sarkosyl-soluble fractions stained with anti-3Rtau and anti-4Rtau antibodies showed bands at the expected molecular weights of the six main tau isoforms present in the brain (between 68 kDa and 60 kDa); a lower band of about 37 kDa was also identified with anti-3Rtau antibodies. No bands were stained with anti-phospho-tau Ser422 antibodies. TEM of sarkosyl-soluble fractions revealed no fibrillar structures (data not shown). 2.3. Inoculation of Sarkosyl-Insoluble and Soluble Fractions from sAD into the Thalamus of Newborn WT Mice (Group 2) Diffuse AT8-positive staining at the site of the injection was observed 24 h after inoculation (Figure 4A). Small phospho-tau-immunoreactive dots, as revealed with the AT8 antibody, were seen at the site of the injection of sarkosyl-insoluble fractions 48 h and 72 h after inoculation (Figure 4B,C). Very small numbers of neurons containing phospho-tau were observed at 1 month (Figure 4D,E); several serial sections were needed to see a single positive neuron per section, if any, at the age of 3 months (Figure 4F,G). AT8-immunoreactive neurons were distributed in the ipsilateral ventral lateral, ventral posterolateral, and ventral posteromedial thalamic nuclei. Positive neurons showed fine granular tau-immunoreactive deposits in the cytoplasm and proximal dendrites (Figure 4D,F,G); round and dense aggregates were only exceptionally observed (Figure 4E). No tau-immunoreactive neurons were seen in mice killed at the age of six months. No tau-immunoreactive neuronal inclusions were identified in mice inoculated with sarkosyl-soluble fractions at the newborn age and killed at the age of 1 month (data not shown). Common reactions in inoculated mice were the presence of phagocytes during the first week and the presence of a few reactive astrocytes along the trajectory of the needle at the age of one month (data not shown). 2.4. Inoculation of Sarkosyl-Insoluble and Sarkosyl-Soluble Fractions from sAD into the Thalamus of Adult WT Mice (Group 3) WT mice inoculated with sarkosyl-insoluble fractions at the age of 3 months and killed 24 h later showed diffuse AT8 immunoreactivity at the inoculation site (Figure 5A). Phospho-tau-immunoreactive granules were observed surrounding the membrane of local neurons at 48 h and 72 h after inoculation (Figure 5B,C). Small granules, threads, and positive cytoplasmic inclusions were identified at some distance from the injection site 1 month after inoculation (Figure 5D,E). Inoculated mice at the age of 3 months and killed at the age of 6 months (Figure 6A–D) or 9 months (Figure 6E–I) showed numerous cells with cytoplasmic AT8-immunoreactive deposit and threads in the ipsilateral ventral lateral, ventral posterolateral, ventral posteromedial, lateral dorsal, and reticular nuclei of the thalamus. In addition, AT8-positive neurons were observed in the habenula and caudate/putamen, together with positive fibres and glial cells with tau-positive inclusions in the internal capsule and fimbria (Figure 6). No AT8-immunoreactive deposits were observed in the somatosensory cortex, hippocampal complex, amygdala, or hippocampus. Similar deposits were seen in mice inoculated with sarkosyl-insoluble fractions and killed at the age of 9 months (6 months’ survival). No tau deposits were found in mice inoculated with sAD sarkosyl-soluble fractions. Reactive astrocytes and microglia were present along the trajectory of the needle in mice inoculated with sAD fractions and with vehicle alone; a few macrophages were clustered at the inoculation site, and a few reactive astrocytes and residual macrophages were still present in some cases months after the inoculation. In no case did microglial or astroglial responses extend beyond the restricted territory of the traumatic lesion produced by the needle. 2.5. Brain Inoculation of sAD Sarkosyl-Insoluble Fractions in Newborn Mice Does Not Prevent Tau Seeding following Re-Inoculation of sAD Sarkosyl-Insoluble Fractions at the Age of Three Months (Group 4) Newborn mice inoculated with sarkosyl-insoluble fractions into the thalamus, re-inoculated in the ipsilateral thalamus with sAD sarkosyl-insoluble fractions at the age of 3 months, and killed at the age of 6 months showed large numbers of cells with tau-immunoreactive deposits in the cytoplasm. Positive cells were distributed in the ipsilateral ventral lateral, ventral posterolateral, ventral posteromedial, lateral dorsal, and reticular nuclei of the thalamus; positive cells were also observed in the caudate/putamen and corpus callosum (Figure 7A–F). Since some cells had the appearance of glial cells, double-labelling immunofluorescence and confocal microscopy with AT8 and Olig2 or GFAP were used for assessment. Tau-positive deposits were localised in oligodendrocytes (Figure 7G). Tau deposits were not seen in GFAP-immunoreactive astrocytes (data not shown). The distribution of positive cells in mice inoculated in the thalamus with sAD sarkosyl-insoluble fractions (a) at the age of newborn and killed at one month and three months; (b) at the age of three months and killed at the age of 6 months or nine months, and (c) at the age of newborn, re-inoculated at the age of three months, and killed at the age of six months or nine months, is illustrated in Figure 8. Quantitative studies of AT8-immunoreactive cells in an arbitrary area of the thalamus measuring 0.045 mm2 in mice inoculated at 3 months (group 3) and in mice inoculated with sarkosyl-insoluble fractions at the age of newborn and then re-inoculated at 3 months (group 4) and surviving 3 months was 5.292 ± 0.42 and 5.167 ± 0.49 (t-test, p-value: 0.843). The number of AT8-immunoreactive cells in mice surviving 6 months was 6.208 ± 0.42 and 5.889 ± 0.41, respectively, for mice inoculated at 3 months (group 3) and mice inoculated with sarkosyl-insoluble fractions at the age of newborn and re-inoculated at 3 months (group 4) (t-test, p-value: 0.5994) (Figure 9). 3. Discussion Alzheimer’s disease and tauopathies are characteristically adult neurodegenerative diseases. The first neurofibrillary tangles in the human brain appear in the third decade of life in the locus coeruleus and raphe nuclei, entorhinal cortex and transentorhinal cortex, and olfactory bulb and tracts, to increase in numbers and extension to other brain regions with brain ageing and Alzheimer’s disease [4,5,6,56,57,58,59]. Hyper-phosphorylated tau occurs at the early stages of normal human brain development, but it never forms aggregates similar to those seen in tauopathies [38,39]. Tau phosphorylation has been reported in adolescents exposed to high levels of contamination compared with children from nonpolluted areas [60]. Neurofibrillary tangles in young patients are seen only in rare conditions such as Niemann–Pick’s disease type C, subacute sclerosing panencephalitis, and genetic syndromes linked to PLA2G6 and SLC9A6 mutations [2]. The present observations show that newborn mice are resistant to tau seeding and spreading following inoculation of sarkosyl-insoluble fractions from sAD into the thalamus. In contrast, deposits of phosphorylated tau are abundant in mice inoculated at the age of 3 months with the same inocula in the thalamus and surviving 3 and 6 months. This observation is in line with previous findings showing that old mice are more prone than young mice to tau seeding and spreading [55]. The differences are even more striking between newborn and young mice aged 3 months than those described in young versus old mice [61]. Protein tau in the human brain is expressed in six isoforms arising from alternative splicing of exons 2 and 3, which encode N-terminal sequences, and exon 10, which encodes a microtubule-binding repeat domain; isoforms with 352 (3R/0N), 381 (3R/1N), and 410 (3R/2N) amino acids are 3Rtau, and isoforms with 383 (4R/0N), 412 (4R/1N), and 441 (4R/2N) amino acids are 4Rtau [62]. In human and rodent fetal brains, the smallest 3Rtau isoform that lacks sequences from exons 2, 3, and 10 (3R/0N) is predominant [63,64,65,66,67]. All six isoforms are expressed in the adult human brain at a 1:1 ratio; in contrast, 4Rtau isoforms predominate in the brain of adult mice [63,64,67,68,69,70,71,72,73,74]. A shift from fetal to adult tau isoform expression occurs in most species, including mice and humans. This is manifested by a predominance of 3Rtau isoforms during the early stages of development and then increased 4Rtau isoforms in the adult brain, with high levels of phosphorylated tau in the developing brain when compared with the adult brain. These changes occur before the end of weaning in mice and are delayed for longer periods in humans [65,67,75,76,77]. However, subpopulations expressing 3Rtau persist in the hippocampus in the adult murine brain [78,79,80]. 3Rtau is also expressed in the olfactory bulb and periventricular regions and in a few neurons in the periventricular hypothalamus, thalamus, basal forebrain, amygdala, and cerebral cortex in adult mice. This may explain the presence of 3Rtau bands in Western blots of adult mice. Differential regulation of microtubule dynamics by 3Rtau and 4Rtau has been implicated in the onset of neurodegenerative diseases [81]. It can be suggested that the profile of tau isoforms during development may contribute to the reduced production of tau aggregate in the youngest. Unilateral inoculation of sAD sarkosyl-insoluble fractions into the thalamus in mice aged 3 months results in phosphorylated tau seeding in cells and threads in the ipsilateral ventral lateral, ventral posterolateral, ventral posteromedial, lateral dorsal, and reticular nuclei of the thalamus. Tau-immunoreactive neurons are also observed in the habenula and caudate/putamen, together with positive fibers and glial cells in the internal capsule and fimbria, thus suggesting local spreading to the striatum and habenula [82,83]. Yet differences in tau deposition are not marked when comparing survival times of 3 and 6 months after inoculation. These observations are in line with previous data suggesting reduced tau spreading in the thalamus when compared with the hippocampus [51]. Indeed, neurons of the somatosensory cortex and primary motor cortex did not contain hyper-phosphorylated tau following thalamic inoculation despite the robust connectivity between the thalamus and the somatosensory and motor cortex [84,85,86,87,88,89,90]. Reduced tau seeding and spreading in newborn mice compared with mice aged 3 months may also be related to the immaturity of thalamic connections at this age [91,92], in addition to seeding resistance. In any case, tau seeding and spreading are not accompanied by astrocytic and microglial responses. Brain inoculation of sarkosyl-insoluble fractions in the newborn mouse does not protect from tau seeding after intracerebral re-inoculation of the same fractions at the age of 3 months. The amount and distribution of tau deposits in re-inoculated mice are similar to those seen in mice inoculated at the age of 3 months and surviving 3 or 6 months. sAD sarkosyl-insoluble fractions contain different tau species along with many other molecules that are components of human neurofibrillary tangles [93]. Moreover, murine tau differs from human tau in a number of ways, including in the N-terminal domain (residues 18 to 28) and three amino acid residues in the C-terminal domain [62,66,68,94,95]. Despite these differences between human and murine tau, in addition to the presence of additional components in sarkosyl-insoluble fractions, there is a lack of immune responses after brain inoculation of human brain homogenates containing paired helical filaments. 4. Materials and Methods 4.1. Animals Wild-type C57BL/6 mice from our colony were used. All animal procedures were carried out following the guidelines of the European Communities Council Directive 2010/63/EU and with the approval of the local ethical committee (C.E.E.A: Comitè Ètic d’Experimentació Animal; University of Barcelona, Spain; ref. 426/18). The animals were maintained under standard conditions of 12 h light/dark cycles, constant temperature, and free access to food and water. Newborn mice were maintained with their mothers in individual cages, one mother per cage, until weaning. The animals analysed were of both sexes, following the criteria used in our institute to avoid sex discrimination and the use of the minimal number of animals necessary for the present study. The following four groups of mice were assessed:WT mice aged 15 days, 3 months, and 12 months of both sexes were used for Western blot (n = 4 for every time point; total 12) and immunohistochemical studies (n = 6 per age, total = 18). Twelve animals of the eighteen were used only for Western blotting; Newborn WT mice aged 1–5 days were inoculated in the right thalamus with sarkosyl-insoluble fractions of brain homogenates from sAD and killed at the following times postinoculation: 24 h (n = 2), 48 h (n = 2), 72 h (n = 2), 1 month (n = 3), 3 months (n = 6), and 6 months (n = 4). In addition, four newborn mice aged 5 days were inoculated with sarkosyl-soluble fractions from sAD and killed at the age of 1 month. The total number of mice was 23, indistinctly males or females; WT mice aged 3 months were inoculated into the right thalamus with sarkosyl-insoluble fractions of brain homogenates from sAD and killed at the following times after inoculation: 0 h (n = 2), 24 h (n = 2), 48 h (n = 2), 72 h (n = 2), 7 days (n = 2), 3 months (n = 4), and 6 months (n = 4) after inoculation. Four mice aged 3 months were inoculated with sarkosyl-soluble fractions from sAD, and killed at the age of 6 months (survival 3 months). The total number of mice was 22, including equal numbers of males and females. In parallel, two WT mice aged 3 months were injected with 50 mM Tris-HCl (pH 7.4) as vehicle (negative) controls; Newborn WT mice aged 1–5 days (n = 6) were inoculated in the right thalamus with sarkosyl-insoluble fractions of brain homogenates from sAD, and re-inoculated in the right thalamus with sarkosyl-insoluble fractions of sAD at the age of 3 months. Mice were killed at the age of 6 months (survival 3 months) or 9 months (survival 6 months). The total number of mice in this group was six. 4.2. Western Blotting of Total Mouse Brain Homogenates of Noninoculated Mice Noninoculated WT mice aged 15 days, 3 months, and 12 months (n = 4 for every time-point) (group 1) were killed under anaesthesia, and the brains were rapidly removed from the skull. The left hemisphere was immediately frozen on dry ice and stored at −80 °C until used for Western blot studies. The right hemisphere was rapidly fixed with 4% paraformaldehyde in phosphate buffer and embedded in paraffin for immunohistochemistry. Brain homogenates were lysed in RIPA buffer (50 mM Tris-HCl, pH 7.0; 150 mM NaCl, 1% Nonidet P-40; 0.5% Na-deoxycholate; 0.1% SDS) supplemented with protease and phosphatase inhibitors (Roche, Basel, Switzerland). After centrifugation at 20,000× g for 20 min at 4 °C (Ultracentrifuge Beckman with 70Ti rotor, Barcelona, Spain), supernatants were quantified with BCA reagent (Pierce, Waltham, MA, USA). Protein samples were mixed with loading sample buffer and heated at 95 °C for 5 min. 20 µug of protein was separated by electrophoresis in SDS-PAGE gels and transferred to nitrocellulose membranes (200 mA per membrane, 120 min). Nonspecific binding was blocked by incubation in 5% nonfatty milk in Tris-buffered saline (TBS) containing 0.2% Tween (TBS-T) for 1 h at room temperature. After washing, the membranes were incubated at 4 °C overnight with one of the primary antibodies (Table 1) in TBS containing 3% albumin and 0.2% Tween. Membranes were washed with TBS-T and incubated for 1 h at room temperature with the appropriate horseradish peroxidase-conjugated secondary antibody (1:2000; Dako, Glostrup, DE). Immune complexes were revealed by incubating the membranes with a chemiluminescence reagent (Electrochemiluminescence; Amersham, GE Healthcare, Buckinghamshire, UK). Densitometry was carried out with Totallab software (TL100 v.2006b), and values were normalised using β-actin. The normality of distribution of fold change values was analysed with the Kolmogorov–Smirnov test. The Unpaired t-test was used. Statistical analysis and graphic design were performed with GraphPad Prism version 5.01 (La Jolla, CA, USA). The data were expressed as mean ± SEM, and significance levels were set at p < 0.05, p < 0.01, and p < 0.001. 4.3. Tissue Processing for Immunohistochemistry and Immunofluorescence Group 1 WT mice aged 15 days (n = 6), 3 months (n = 6), and 12 months (n = 6), total n = 18, and group 2 (n = 23), group 3 (n = 25), and group 4 (n = 6) inoculated mice were killed under anaesthesia, and the brains were rapidly fixed with 4% paraformaldehyde in phosphate buffer, and embedded in paraffin. Consecutive serial sections 4 μm thick were obtained with a sliding microtome. Dewaxed sections were stained with haematoxylin and eosin or processed for immunohistochemistry with antibodies listed in Table 1. Following incubation with the primary antibody, the sections were incubated with EnVision + system peroxidase for 30 min at room temperature. The peroxidase reaction was visualised with diaminobenzidine and H2O2. Control of the immunostaining included omission of the primary antibody; no signal was obtained following incubation with only the secondary antibody. Quantification of AT8-positive cells in the thalamus was carried out in inoculated mice aged 3 months (group 3) and in newborn inoculated mice re-inoculated at the age of 3 months (group 4) with sarkosyl-insoluble fractions. Counts were made at survival times of 3 months and 6 months (n = 8 and n = 6, respectively, for group 3 and group 4) in random areas of 0.045 mm2 in every case. Values were expressed as mean values ± SEM. The t-test was used to assess differences between paired groups. Double-labelling immunofluorescence was carried out on dewaxed sections, 4 μm thick. The sections were boiled in citrate buffer to enhance antigenicity and blocked for 30 min at room temperature with 10% fetal bovine serum diluted in 0.1 M phosphate-buffered saline (PBS). The sections were stained with a saturated solution of Sudan black B (Merck, Kenilworth, NJ, USA) for 15 min to block autofluorescence of putative lipofuscin granules present in cell bodies and then rinsed in 70% ethanol and washed in distilled water. Then the sections were incubated at 4 °C overnight with AT8 and rabbit polyclonal Olig2, or rabbit polyclonal anti-GFAP or Iba1. After washing, the sections were incubated with Alexa488 or Alexa546 fluorescence secondary antibodies against the corresponding host species. Nuclei were stained with DRAQ5TM. Then the sections were mounted in Immuno-FluoreTM mounting medium (MP Biomedicals, CA, USA), sealed, and dried overnight. Sections were examined with a Leica TCS-SL confocal microscope. Identification of brain regions was made following Paxinos and Franklin, 2019, and Schröder et al., 2019 [96,97]. 4.4. sAD Sarkosyl-Insoluble and Soluble Fractions Used for Brain Inoculation Samples of the frontal cortex (10 g) from one man, aged 68 years, with sporadic AD (sAD) stage VI/C of Braak and Braak and phase 4 of Thal without comorbidities were obtained 4 h postmortem and immediately frozen at −80 °C until use. The samples were provided by the Institute of Neuropathology Brain Bank, now a branch of the HUB-ICO-IDIBELL Biobank, following the guidelines of the Real Decreto 1716/2011 of the Spanish legislation and the approval of the local ethics committee. The amount of 10 g of the brain tissue was lysed in 10 vol ) with cold suspension buffer (10 mM Tris-HCl, pH 7.4, 0.8 M NaCl, 1 mM EGTA) supplemented with 10% sucrose, protease, and phosphatase inhibitors (Roche, Basel, Switzerland). The homogenates were centrifuged at 20,000× g for 20 min (Ultracentrifuge Beckman with 70Ti rotor,) and the supernatant (S1) was saved. The pellet was rehomogenised in a 5 vol buffer and recentrifuged at 20,000× g for 20 min. The two supernatants (S1 + S2) were mixed and incubated with 0.1% N-lauroylsarkosynate (sarkosyl) for 1 h at room temperature while being shaken. Samples were then centrifuged at 100,000× g for 1 h. Sarkosyl-insoluble pellets were resuspended (0.2 mL/g) in 50 mM Tris-HCl (pH 7.4). Protein concentrations were quantified with the bicinchoninic acid assay (BCA) assay (Pierce, Waltham, MA, USA). Sarkosyl-insoluble fractions were processed for Western blotting. Samples were mixed with loading sample buffer and heated at 95 °C for 5 min. 60 µug of protein was separated by electrophoresis in SDS-PAGE gels and transferred to nitrocellulose membranes (200 mA per membrane, 90 min). The membranes were blocked for 1 h at room temperature with 5% nonfat milk in TBS containing 0.2% Tween and were then incubated with the phospho-specific antibody anti-tau Ser422 (diluted 1:1000; Thermo Fisher, Waltham, MA, USA). After washing with TBS-T, blots were incubated with anti-rabbit IgG conjugated with horseradish peroxidase (diluted at 1:2000; Agilent, Santa Clara, CA, USA) for 45 min at room temperature. Immune complexes were revealed by incubating the membranes with a chemiluminescence reagent (Amersham, Germany Healthcare, Life Sciences, Buckinghamshire, UK) [55]. Western blotting of sarkosyl-soluble fractions included, in addition to anti-tau Ser422, the incubation with anti-4Rtau and anti-3Rtau antibodies (dilutions 1:1000 and 1:2000, respectively) to control the presence of tau. In addition, sarkosyl-insoluble and sarkosyl-soluble fractions were analyzed with transmission electron microscopy (TEM) procedures and thioflavin T (ThT) amyloid quantification assay. Sarkosyl-insoluble fractions were fixed to carbon-forward-coated copper supports, and negative staining was performed using 2% uranyl acetate (pH 7.4). The samples were then placed in silica-based desiccant for a minimum of 2 h and examined with a Jeol JEM-1010 transmission electron microscope. ThT stock solution was prepared at 2.5 mM (dissolved in 10 mM phosphate buffer, 150 mM NaCl, pH 7.0) and preserved in a single-aliquot at −80 °C. The ThT assay was performed by dissolving 0.2 µL of sarkosyl-insoluble fraction in 0.2 mL of freshly prepared ThT (final concentration 30 µM), followed by quantification using an absorbance/excitation (445/485) microplate reader (Tecan Infinite M200Pro, Männedorf, Switzerland) in 96 flat-bottom polystyrol plates (Nunclor, ThermoFisher Scientific, Germany). Plates were prepared and incubated at 37 °C. Readings were taken each hour over the 15–16 h period. 4.5. Brain Inoculation of sAD Sarkosyl-Insoluble, Sarkosyl-Soluble Fractions, and Vehicle Alone Newborn mice aged 1–5 days were anesthetised on a cold plate surface and fixed manually. Intrathalamic injections were administered using a Hamilton syringe. A volume of 1.2 µL (amount of protein, about 3.5 µg/µL) was injected at a rate of 0.05 µL/min. The syringe was withdrawn slowly over a period of 10 min to avoid leakage of the inoculum. Following surgery, the animals were kept in a warm blanket and monitored until they recovered and then returned to the cages with their mother. Mice aged 3 months were deeply anaesthetised by intraperitoneal ketamine/xylazine/buprenorphine cocktail injection and placed in a stereotaxic frame after assuring lack of reflexes. Intracerebral injections were administered using a Hamilton syringe. The coordinates for thalamic injections were −1.3 mm AP, −1.2/−1.4 mm ML relative to Bregma (interaural 2.46 mm), and −3/−3.5 mm DV from the dural surface [96]. A volume of 1.5 µL was injected at a rate of 0.05µL/min. The syringe was withdrawn slowly over a period of 10 min to avoid leakage of the inoculum. Following surgery, the animals were kept in a warm blanket and monitored until they recovered from the anaesthesia. Carprofen analgesia was administered immediately after surgery and once a day during the two following days. Animals were housed individually with full access to food and water. 5. Conclusions Our study concludes that: (i) tau seeding following intracerebral inoculation of sAD PHFs in newborn mice is very limited in contrast to the widespread tau seeding following intracerebral inoculation of the same inocula in mice aged 3 months; (ii) differences in tau seeding between newborn and young adults may be related to the ratios between 3Rtau and 4Rtau and the shift to 4Rtau predominance in adults, together with the immaturity of connections in newborn mice; and (iii) intracerebral inoculation of sAD PHFs in newborn mice does not protect from tau seeding following intracerebral inoculation of sAD PHFs in young/adult mice. Acknowledgments We wish to Tom Yohannan for editorial help. Author Contributions P.A.-B., P.G.-E., D.V. and M.J.-P. prepared the inoculum, and inoculated the mice; I.L.-G. assessed tau species with Western blotting in the mice during development; P.A.-B. and M.C. carried out immunohistochemistry; P.A.-B. quantified immunohistochemical signals; J.S.-J. and V.G. analyzed sarkosyl-insoluble and sarkosyl-soluble fractions P.H.F. fractions with electron microscopy and amyloid testing; I.F. and J.A.d.R. designed the experiments; I.F. verified the results. All the authors wrote the manuscript. All authors have read and agreed to the published version of the manuscript. Funding The project leading to these results received funding from the “la Caixa” Foundation (ID 100010434) under the agreement LCF/PR/HR19/52160007, co-financed by ERDF under the program Interreg Poctefa: RedPrion 148/16 and the Intra-CIBERNED 2019 collaborative project. We thank the CERCA programme of the Generalitat de Catalunya for institutional support. Institutional Review Board Statement All animal procedures were carried out following the guidelines of the European Communities Council Directive 2010/63/EU and with the approval of the local ethical committee (C.E.E.A: Comitè Ètic d’Experimentació Animal; University of Barcelona, Spain; ref. 426/18). Data Availability Statement All the data are in the text. Conflicts of Interest The authors declare that they have no conflict of interest. Figure 1 Tau expression in the developing and young/adult brain at the time-points of 15 days, 3 months, and 12 months (4 animals per group) as shown in Western blots of total brain homogenates. Total tau levels are higher at 15 days than at 3 months and 12 months. 3Rtau levels are high at the age of 15 days but decrease markedly at 3 months and 12 months. 4Rtau is expressed in mice aged 15 days, 3 months, and 12 months, but the band pattern differs in mice aged 15 days when compared with young/adult mice. Tau phosphorylation at Thr231 is more marked in mice aged 15 days than in mice aged 3 months and 12 months, but the levels of tau phosphorylation at Thr181 and Ser202/Thr305 (antibody AT8) are similar in the three groups. β-actin was used as a marker of protein loading. Unpaired t-test compared 3 months and 12 months with 15 days: *** statistically significant p < 0.001. Figure 2 3Rtau-immunoreactive cells persist in the adult murine brain in the internal polymorphic layer of the dentate gyrus (A), olfactory bulb (B), and a periventricular layer of the lateral ventricle (C). Scattered positive neurons are also found in the entorhinal cortex (D), periventricular hypothalamus (E), nuclei of the basal forebrain (F), and thalamus (G); very rarely, 3Rtau-immunoreactive neurons are encountered in the amygdala (H) and cerebral cortex (I). Paraffin sections, lightly counterstained with hematoxylin, bar = 35 µm. Figure 3 Characteristics of sarkosyl-insoluble fraction from sAD. Western blotting processed with anti-phospho-tau Ser422 (P-tau Ser422) reveals three bands of 68 kDa, 64 kDa, and 60 kDa, a weak upper band of 73 kDa, and oligomeric smears, in addition to several bands of about 50 kDa, between 30 kDa and 40 kDa, and lower bands of truncated tau at the C-terminal, one of them of about 20 kDa (A). Transmission electron microscopy of fibres in AD sarkosyl-insoluble fraction showing typical paired helical filaments, bar = 50 nm (B). Thioflavin T incubation of sarkosyl-insoluble fractions shows increased fluorescence with time, indicating amyloid fibril formation (C). Figure 4 Phospho-tau immunoreactive deposits, as revealed with the antibody AT8 in mice unilaterally inoculated with sAD sarkosyl-insoluble fractions in the ventral thalamus at the age of newborn and killed 24 h, 48 h, 72 h, 1 month, and 3 months later. Diffuse AT8 immunoreactivity is seen at the injection site after 24 h (A). Positive dots are seen at the injection site 48 h (B) and 72 h after inoculation (B,C). Very rarely, isolated positive neurons with granular cytoplasm and fine radiating dendrites and exceptional dense cytoplasmic inclusions are seen in mice aged 1 month (D,E,G) and 3 months (F). Paraffin sections slightly counterstained with haematoxylin, bar = 25 µm. Figure 5 Phospho-tau immunoreactive deposits, as revealed with the antibody AT8 in mice unilaterally inoculated with sAD sarkosyl-insoluble fractions in the ventral thalamus at the age of 3 months and killed 24 h, 48 h, 72 h, 1 month, and 3 months after injection. Diffuse AT8 immunoreactivity is seen at the injection site after 24 h (A). Tau-immunoreactive local dot-like and fine thread deposits are localised around the cytoplasm of cells at 48 h (B) and 72 h (C). Positive dots, threads, and isolated cells show AT8 immunoreactivity at a distance from the injection site one month after inoculation (D,E). Paraffin sections slightly counterstained with haematoxylin, bar = 25 µm. Figure 6 Large numbers of phospho-tau deposits, as revealed with AT8 antibody, in cells and threads in different thalamic nuclei of mice unilaterally inoculated with sAD sarkosyl-insoluble fraction in the ventral thalamus at the age 3 months and killed at the age of 6 months (A–D) or 9 months (E–I) (A–I). Tau-immunoreactive inclusions are round or elongated deposits in the cytoplasm and fine neurites in the neuropil (threads); proximal dendrites do not contain phosphorylated tau. Paraffin sections slightly counterstained with haematoxylin, bar = 25 µm. Figure 7 Large numbers of AT8-positive neurons, glial cells, and threads in different thalamic nuclei in mice inoculated in the ventral thalamus with sAD sarkosyl-insoluble fractions at the age of newborn and then re-inoculated with sAD sarkosyl-insoluble fractions in the thalamus at the age of 3 months. Mice were killed at the age of six months (A–C) or 9 months (D–F). Paraffin sections slightly counterstained with haematoxylin, bar = 50 µm. Double-labelling immunofluorescence with AT8 and Olig2 shows localization of tau deposits in oligodendroglia. Paraffin sections, nuclei stained with DRAQ5TM, bar = 20 µm (G). Figure 8 Distribution of phospho-tau deposits in mice following unilateral inoculation in the ventral thalamus with a sarkosyl-insoluble fraction from sAD in newborn mice killed at 1 month and 3 months (group 2), inoculated at the age of 3 months and killed at the age of 6 months or 9 months (group 3), and inoculated at the age of newborn, re-inoculated at the age of 3 months (group 4), and killed at the age of 6 months or 9 months. Diagrams modified from Paxinos and Franklin, 2019 (see Methods section). In inoculated newborn mice, a few AT8-containing cells are dispersed in the ipsilateral ventral lateral, ventral posterolateral, and ventral posteromedial thalamic nuclei at 1 month; tau-containing cells are barely present in mice aged 3 months. In contrast, large numbers of tau-containing cells and threads are seen in the ipsilateral ventral lateral, ventral posterolateral, ventral posteromedial, lateral dorsal, and reticular nuclei of the thalamus and in the habenula and caudate/putamen in mice inoculated at the age of 3 months and killed at the age of 6 months or 9 months. In addition, positive fibers and glial cells are seen in the internal capsule and fimbria. A similar amount and distribution of AT8-immunoreactive cells occur in inoculated mice at the age of newborn, re-inoculated at 3 months, and killed at the age of 6 months and 9 months. Similar patterns are seen in mice surviving 3 months and 6 months following inoculation. Red dots are representative of the distribution of seeding. The vertical lines indicate the injection site. Figure 9 Quantification of AT8-immunoreactive cells in the thalamus in mice inoculated with sarkosyl-insoluble fraction at the age of newborn and re-inoculated at 3 months (group 4), and killed at 6 months and 9 months (3 mpi and 6 mpi: survival 3 months and 6 months, respectively), and in mice inoculated with a sarkosyl-insoluble fraction the age of 3 months (group 3) and killed 3 months and 6 months later (3 mpi and 6 mpi, respectively). No differences in the number of AT8-immunoreactive cells are seen among the four lots of mice. Values correspond to the number of positive cells in an arbitrary area of 0.045 mm2. ijms-23-04789-t001_Table 1 Table 1 Antibodies used for Western blotting (wb) and immunohistochemistry (ih) in the present study. Dil—dilution. Antibody Reference Supplier Host Dil wb Dil ih β-actin A5316 Sigma-Aldrich (St Louis, MO, USA) Ms 1:30,000 - Tau 5 MA5-12808 Thermo Fisher (Waltham, MA, USA) Ms 1:1000 1:100 3RTau 05-803 Millipore (Darmstadt, Germany) Ms 1:2000 1:800 4RTau 05-804 Millipore (Darmstadt, Germany) Ms 1:1000 1:50 P-tau Thr181 11107 Signalway (College Park, MA, USA) Rb 1:1000 1:50 P-tau AT8 MN1020 Invitrogen, Thermo Fisher (Waltham, MA, USA) Ms 1:250 1:50 P-tau Thr231 577813 Calbiochem (Darmstadt, Germany) Rb 1:1000 1:50 GFAP GA-524 Dako (Glostrup, Denmark) Rb - 1:500 Olig2 Ab109186 Abcam (Cambridge, UK) Rb - 1:500 Iba1 019-19741 Wako (Richmond, VA, USA) Rb - 1:1:1000 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Duyckaerts C. Alzheimer’s Disease Neuropathology of Neurodegenerative Diseases: A Practical Guide Kovacs G.G. Cambridge University Cambridge, UK 2015 80 108 2. Kovacs G.G. Tauopathies Neuropathology of Neurodegenerative Diseases: A Practical Guide Cambridge University Cambridge, UK 2015 109 148 3. Braak H. Braak E. Neuropathological Stageing of Alzheimer-Related Changes Acta Neuropathol. 1991 82 239 259 10.1007/BF00308809 1759558 4. Ferrer I. Defining Alzheimer as a common age-related neurodegenerative process not inevitably leading to dementia Progr. Neurobiol. 2012 97 38 51 10.1016/j.pneurobio.2012.03.005 22459297 5. Braak H. Del Tredici K. The Preclinical Phase of the Pathological Process Underlying Sporadic Alzheimer’s Disease Brain 2015 138 2814 2833 10.1093/brain/awv236 26283673 6. Šimić G. Leko M.B. Wray S. Harrington C. Delalle I. Milošević N.J. Bažadona D. Buee L. de Silva R. Di Giovanni G. Monoaminergic Neuropathology in Alzheimer’s Disease Prog. Neurobiol. 2017 151 101 138 10.1016/j.pneurobio.2016.04.001 27084356 7. Masters C. Bateman R. Blennow K. Rowe C.C. Sperling R.A. Cummings J.L. Alzheimer’s Disease Nat. Rev. Dis. Prim. 2015 1 15056 10.1038/nrdp.2015.56 27188934 8. DeTure M.A. Dickson D.W. The Neuropathological Diagnosis of Alzheimer’s Disease Mol. Neurodegener. 2019 14 32 10.1186/s13024-019-0333-5 31375134 9. Jellinger K.A. Neuropathology of the Alzheimer’s Continuum: An Update Free Neuropathol. 2020 1 32 10. Avila J. Perry G. A Multilevel View of the Development of Alzheimer’s Disease Neuroscience 2021 457 283 293 10.1016/j.neuroscience.2020.11.015 33246061 11. Ferrer I. Santpere G. Van Leeuwen F.W. Argyrophilic Grain Disease Brain 2008 131 1416 1432 10.1093/brain/awm305 18234698 12. Ahmed Z. Bigio E.H. Budka H. Dickson D.W. Ferrer I. Ghetti B. Giaccone G. Hatanpaa K.J. Holton J.L. Josephs K.A. Globular Glial Tauopathies (GGT): Consensus Recommendations Acta Neuropathol. 2013 126 537 544 10.1007/s00401-013-1171-0 23995422 13. Crary J.F. Trojanowski J.Q. Schneider J.A. Abisambra J.F. Abner E.L. Alafuzoff I. Arnold S.E. Attems J. Beach T.G. Bigio E.H. Primary Age-Related Tauopathy (PART): A Common Pathology Associated with Human Aging Acta Neuropathol. 2014 128 755 766 10.1007/s00401-014-1349-0 25348064 14. Kovacs G.G. Ferrer I. Grinberg L.T. Alafuzoff I. Attems J. Budka H. Cairns N.J. Crary J.F. Duyckaerts C. Ghetti B. Aging-Related Tau Astrogliopathy (ARTAG): Harmonized Evaluation Strategy Acta Neuropathol. 2016 131 87 102 10.1007/s00401-015-1509-x 26659578 15. Armstrong M.J. Progressive Supranuclear Palsy: An Update Curr. Neurol. Neurosci. Rep. 2018 18 12 10.1007/s11910-018-0819-5 29455271 16. Saranza G.M. Whitwell J.L. Kovacs G.G. Lang A.E. Corticobasal Degeneration Int. Rev. Neurobiol. 2019 149 87 136 10.1016/bs.irn.2019.10.014 31779825 17. Kovacs G.G. Ghetti B. Goedert M. Classification of Diseases with Accumulation of Tau Protein Neuropathol. Appl. Neurobiol. 2022 48 e12792 10.1111/nan.12792 35064600 18. Goedert M. Spillantini M.G. Ordered Assembly of Tau Protein and Neurodegeneration Adv. Exp. Med. Biol. 2019 184 3 21 10.1007/978-981-32-9358-8_1 19. Arnsten A.F.T. Datta D. Del Tredici K. Braak H. Hypothesis: Tau Pathology Is an Initiating Factor in Sporadic Alzheimer’s Disease Alzheimer’s Dement. 2021 17 115 124 10.1002/alz.12192 33075193 20. Jucker M. Walker L.C. Pathogenic Protein Seeding in Alzheimer Disease and Other Neurodegenerative Disorders Ann. Neurol. 2011 70 532 540 10.1002/ana.22615 22028219 21. Lewis J. Dickson D.W. Propagation of Tau Pathology: Hypotheses, Discoveries, and Yet Unresolved Questions from Experimental and Human Brain Studies Acta Neuropathol. 2016 131 27 48 10.1007/s00401-015-1507-z 26576562 22. Arendt T. Stieler J. Holzer M. Tau and Tauopathies Brain Res. Bull. 2016 126 238 292 10.1016/j.brainresbull.2016.08.018 27615390 23. Goedert M. Eisenberg D.S. Crowther R.A. Propagation of Tau Aggregates and Neurodegeneration Annu. Rev. Neurosci. 2017 40 189 210 10.1146/annurev-neuro-072116-031153 28772101 24. Goedert M. Spillantini M.G. Propagation of Tau Aggregates Mol. Brain 2017 10 18 10.1186/s13041-017-0298-7 28558799 25. Gibbons G.S. Lee V.M.Y. Trojanowski J.Q. Mechanisms of Cell-to-Cell Transmission of Pathological Tau JAMA Neurol. 2019 76 101 108 10.1001/jamaneurol.2018.2505 30193298 26. Andrés-Benito P. Carmona M. Jordán M. Fernández-Irigoyen J. Santamaría E. del Rio J.A. Ferrer I. Host Tau Genotype Specifically Designs and Regulates Tau Seeding and Spreading and Host Tau Transformation Following Intrahippocampal Injection of Identical Tau AD Inoculum Int. J. Mol. Sci. 2022 23 718 10.3390/ijms23020718 35054902 27. Liu L. Drouet V. Wu J.W. Witter M.P. Small S.A. Clelland C. Duff K. Trans-Synaptic Spread of Tau Pathology In Vivo PLoS ONE 2012 7 e31302 10.1371/journal.pone.0031302 22312444 28. De Calignon A. Polydoro M. Suárez-Calvet M. William C. Adamowicz D.H. Kopeikina K.J. Pitstick R. Sahara N. Ashe K.H. Carlson G.A. Propagation of Tau Pathology in a Model of Early Alzheimer’s Disease Neuron 2012 73 685 697 10.1016/j.neuron.2011.11.033 22365544 29. Ahmed Z. Cooper J. Murray T.K. Garn K. McNaughton E. Clarke H. Parhizkar S. Ward M.A. Cavallini A. Jackson S. A Novel in Vivo Model of Tau Propagation with Rapid and Progressive Neurofibrillary Tangle Pathology: The Pattern of Spread Is Determined by Connectivity, Not Proximity Acta Neuropathol. 2014 127 667 683 10.1007/s00401-014-1254-6 24531916 30. Dujardin S. Lecolle K. Caillierez R. Bégard S. Zommer N. Lachaud C. Carrier S. Dufour N. Aurégan G. Winderickx J. Neuron-to-Neuron Wild-Type Tau Protein Transfer through a Trans-Synaptic Mechanism: Relevance to Sporadic Tauopathies Acta Neuropathol. Commun. 2014 2 14 10.1186/2051-5960-2-14 24479894 31. Mudher A.A. Colin M. Dujardin S. Medina M. Dewachter I. Naini S.M.A.S. Mandelkow E.-M.E. Mandelkow E.E. Buee L. Goedert M.M. What Is the Evidence That Tau Pathology Spreads through Prion-Like Propagation? Acta Neuropathol. Commun. 2017 5 99 10.1186/s40478-017-0488-7 29258615 32. Dujardin S. Bégard S. Caillierez R. Lachaud C. Delattre L. Carrier S. Loyens A. Galas M.-C. Bousset L. Melki R. Ectosomes: A New Mechanism for Non-Exosomal Secretion of Tau Protein PLoS ONE 2014 9 e100760 10.1371/journal.pone.0100760 24971751 33. Wang Y. Balaji V. Kaniyappan S. Krüger L. Irsen S. Tepper K. Chandupatla R. Maetzler W. Schneider A. Mandelkow E. The Release and Trans-Synaptic Transmission of Tau via Exosomes Mol. Neurodegener. 2017 12 5 10.1186/s13024-016-0143-y 28086931 34. Guix F.X. Corbett G.T. Cha D.J. Mustapic M. Liu W. Mengel D. Chen Z. Aikawa E. Young-Pearse T. Kapogiannis D. Detection of Aggregation-Competent Tau in Neuron-Derived Extracellular Vesicles Int. J. Mol. Sci. 2018 19 663 10.3390/ijms19030663 35. Polanco J.C. Li C. Durisic N. Sullivan R. Götz J. Exosomes Taken up by Neurons Hijack the Endosomal Pathway to Spread to Interconnected Neurons Acta Neuropathol. Commun. 2018 6 10 10.1186/s40478-018-0514-4 29448966 36. Peng C. Trojanowski J.Q. Lee V.M. Protein transmission in neurodegenerative diseases Nat. Rev. Neurol. 2020 16 199 212 10.1038/s41582-020-0333-7 32203399 37. Tardivel M. Bégard S. Bousset L. Dujardin S. Coens A. Melki R. Buée L. Colin M. Tunneling Nanotube (TNT)-Mediated Neuron-to Neuron Transfer of Pathological Tau Protein Assemblies Acta Neuropathol. Commun. 2016 4 117 10.1186/s40478-016-0386-4 27809932 38. Tuerde D. Kimura T. Miyasaka T. Furusawa K. Shimozawa A. Hasegawa M. Ando K. Hisanaga S.-I. Isoform-Independent and -Dependent Phosphorylation of Microtubule-Associated Protein Tau in Mouse Brain During Postnatal Development J. Biol. Chem. 2018 293 1781 1793 10.1074/jbc.M117.798918 29196605 39. Hefti M.M. Kim S. Bell A.J. Betters R.K. Fiock K. Iida M.A. Smalley M.E. Farrell K. Fowkes M.E. Crary J.F. Tau Phosphorylation and Aggregation in the Developing Human Brain J. Neuropathol. Exp. Neurol. 2019 78 930 938 10.1093/jnen/nlz073 31504677 40. López-González I. Aso E. Carmona M. Armand-Ugon M. Blanco R. Naudí A. Cabré R. Portero-Otin M. Pamplona R. Ferrer I. Neuroinflammatory Gene Regulation, Mitochondrial Function, Oxidative Stress, and Brain Lipid Modifications with Disease Progression in Tau P301S Transgenic Mice as a Model of Frontotemporal Lobar Degeneration-Tau J. Neuropathol. Exp. Neurol. 2015 74 975 999 10.1097/NEN.0000000000000241 26360374 41. Iba M. Guo J.L. McBride J.D. Zhang B. Trojanowski J.Q. Lee V.M.-Y. Synthetic Tau Fibrils Mediate Transmission of Neurofibrillary Tangles in a Transgenic Mouse Model of Alzheimer’s-Like Tauopathy J. Neurosci. 2013 33 1024 1037 10.1523/JNEUROSCI.2642-12.2013 23325240 42. Iba M. McBride J.D. Guo J.L. Zhang B. Trojanowski J.Q. Lee V.M.-Y. Tau Pathology Spread in PS19 Tau Transgenic Mice Following Locus Coeruleus (LC) Injections of Synthetic Tau Fibrils Is Determined by the LC’S Afferent and Efferent Connections Acta Neuropathol. 2015 130 349 362 10.1007/s00401-015-1458-4 26150341 43. Peeraer E. Bottelbergs A. Van Kolen K. Stancu I.-C. Vasconcelos B. Mahieu M. Duytschaever H. Donck L.V. Torremans A. Sluydts E. Intracerebral Injection of Preformed Synthetic Tau Fibrils Initiates Widespread Tauopathy and Neuronal Loss in the Brains of Tau Transgenic Mice Neurobiol. Dis. 2015 73 83 95 10.1016/j.nbd.2014.08.032 25220759 44. Stancu I.-C. Vasconcelos B. Ris L. Wang P. Villers A. Peeraer E. Buist A. Terwel D. Baatsen P. Oyelami T. Templated Misfolding of Tau by Prion-Like Seeding Along Neuronal Connections Impairs Neuronal Network Function and Associated Behavioral Outcomes in Tau Transgenic Mice Acta Neuropathol. 2015 129 875 894 10.1007/s00401-015-1413-4 25862635 45. Clavaguera F. Bolmont T. Crowther R.A. Abramowski D. Frank S. Probst A. Fraser G. Stalder A.K. Beibel M. Staufenbiel M. Transmission and Spreading of Tauopathy in Transgenic Mouse Brain Nat. Cell Biol. 2009 11 909 913 10.1038/ncb1901 19503072 46. Clavaguera F. Akatsu H. Fraser G. Crowther R.A. Frank S. Hench J. Probst A. Winkler D.T. Reichwald J. Staufenbiel M. Brain Homogenates from Human Tauopathies Induce Tau Inclusions in Mouse Brain Proc. Natl. Acad. Sci. USA 2013 110 9535 9540 10.1073/pnas.1301175110 23690619 47. Clavaguera F. Lavenir I. Falcon B. Frank S. Goedert M. Tolnay M. “Prion-Like” Templated Misfolding in Tauopathies Brain Pathol. 2013 23 342 349 10.1111/bpa.12044 23587140 48. Boluda S. Iba M. Zhang B. Raible K.M. Lee V.M.-Y. Trojanowski J.Q. Differential Induction and Spread of Tau Pathology in Young PS19 Tau Transgenic Mice Following Intracerebral Injections of Pathological Tau from Alzheimer’s Disease or Corticobasal Degeneration Brains Acta Neuropathol. 2015 129 221 237 10.1007/s00401-014-1373-0 25534024 49. Guo J.L. Narasimhan S. Changolkar L. He Z. Stieber A. Zhang B. Gathagan R. Iba M. McBride J.D. Trojanowski J.Q. Unique Pathological Tau Conformers from Alzheimer’s Brains Transmit Tau Pathology in Nontransgenic Mice J. Exp. Med. 2016 213 2635 2654 10.1084/jem.20160833 27810929 50. Audouard E. Houben S. Masaracchia C. Yilmaz Z. Suain V. Authelet M. De Decker R. Buée L. Boom A. Leroy K. High–Molecular-Weight Paired Helical Filaments from Alzheimer Brain Induces Seeding of Wild-Type Mouse Tau into an Argyrophilic 4R Tau Pathology in Vivo Am. J. Pathol. 2016 186 2709 2722 10.1016/j.ajpath.2016.06.008 27497324 51. Narasimhan S. Guo J.L. Changolkar L. Stieber A. McBride J.D. Silva L.V. He Z. Zhang B. Gathagan R.J. Trojanowski J.Q. Pathological tau strains from human brains recapitulate the diversity of tauopathies in non-transgenic mouse brain J. Neurosci. 2017 37 11406 11423 10.1523/JNEUROSCI.1230-17.2017 29054878 52. Ferrer I. García M.A. González I.L. Lucena D.D. Villalonga A.R. Tech M.C. Llorens F. Garcia-Esparcia P. Martinez-Maldonado A. Mendez M.F. Aging-Related Tau Astrogliopathy (ARTAG): Not Only Tau Phosphorylation in Astrocytes Brain Pathol. 2018 28 965 985 10.1111/bpa.12593 29396893 53. Ferrer I. García M.A. Carmona M. Benito P.A. Torrejón-Escribano B. Garcia-Esparcia P. Del Rio J.A. Involvement of Oligodendrocytes in Tau Seeding and Spreading in Tauopathies Front. Aging Neurosci. 2019 11 112 10.3389/fnagi.2019.00112 31191295 54. Ferrer I. Benito P.A. Zelaya M.V. Aguirre M.E.E. Carmona M. Ausín K. Lachén-Montes M. Fernández-Irigoyen J. Santamaría E. Del Rio J.A. Familial Globular Glial Tauopathy Linked to MAPT Mutations: Molecular Neuropathology and Seeding Capacity of a Prototypical Mixed Neuronal and Glial Tauopathy Acta Neuropathol. 2020 139 735 771 10.1007/s00401-019-02122-9 31907603 55. Ferrer I. Zelaya M.V. García M.A. Carmona M. López-González I. Andrés-Benito P. Lidón L. Gavín R. Garcia-Esparcia P. Del Rio J.A. Relevance of Host Tau in Tau Seeding and Spreading in Tauopathies Brain Pathol. 2020 30 298 318 10.1111/bpa.12778 31397930 56. Braak H. Del Tredici K. The Pathological Process Underlying Alzheimer’s Disease in Individuals under Thirty Acta Neuropathol. 2010 121 171 181 10.1007/s00401-010-0789-4 21170538 57. Braak H. Thal D.R. Ghebremedhin E. del Tredici K. Stages of the Pathologic Process in Alzheimer Disease: Age Categories from 1 to 100 Years J. Neuropathol. Exp. Neurol. 2011 70 960 969 10.1097/NEN.0b013e318232a379 22002422 58. Duyckaerts C. Tau Pathology in Children and Young Adults: Can You Still Be Unconditionally Baptist? Acta Neuropathol. 2011 121 145 147 10.1007/s00401-010-0794-7 21225273 59. Elobeid A. Soininen H. Alafuzoff I. Hyperphosphorylated Tau in Young and Middle-Aged Subjects Acta Neuropathol. 2011 123 97 104 10.1007/s00401-011-0906-z 22160320 60. Calderón-Garcidueñas L. Franco-Lira M. Mora-Tiscareño A. Medina-Cortina H. Torres-Jardón R. Kavanaugh M. Early Alzheimer’s and Parkinson’s Disease Pathology in Urban Children: Friend Versus Foe responses—It Is Time to Face the Evidence BioMed Res. Int. 2013 2013 161687 10.1155/2013/161687 23509683 61. Wegmann S. Bennett R.E. Delorme L. Robbins A.B. Hu M. MacKenzie D. Kirk M.J. Schiantarelli J. Tunio N. Amaral A.C. Experimental Evidence for the Age Dependence of Tau Protein Spread in the Brain Sci. Adv. 2019 5 eaaw6404 10.1126/sciadv.aaw6404 31249873 62. Goedert M. Wischik C.M. Crowther R.A. Walker J.E. Klug A. Cloning and Sequencing of the CDNA Encoding a Core Protein of the Paired Helical Filament of Alzheimer Disease: Identification as the Microtubule-Associated Protein Tau Proc. Natl. Acad. Sci. USA 1988 85 4051 4055 10.1073/pnas.85.11.4051 3131773 63. Goedert M. Spillantini M. Jakes R. Rutherford D. Crowther R. Multiple Isoforms of Human Microtubule-Associated Protein Tau: Sequences and Localization in Neurofibrillary Tangles of Alzheimer’s Disease Neuron 1989 3 519 526 10.1016/0896-6273(89)90210-9 2484340 64. Kosik K.S. Orecchio L.D. Bakalis S. Neve R.L. Developmentally Regulated Expression of Specific Tau Sequences Neuron 1989 2 1389 1397 10.1016/0896-6273(89)90077-9 2560640 65. Brion J.-P. Smith C. Couck A.-M. Gallo J.-M. Anderton B.H. Developmental Changes in τ Phosphorylation: Fetal τ Is Transiently Phosphorylated in a Manner Similar to Paired Helical Filament-τ Characteristic of Alzheimer’s Disease J. Neurochem. 1993 61 2071 2080 10.1111/j.1471-4159.1993.tb07444.x 8245963 66. Janke C. Beck M. Stahl T. Holzer M. Brauer K. Bigl V. Arendt T. Phylogenetic Diversity of the Expression of the Microtubule-Associated Protein Tau: Implications for Neurodegenerative Disorders Mol. Brain Res. 1999 68 119 128 10.1016/S0169-328X(99)00079-0 10320789 67. Takuma H. Arawaka S. Mori H. Isoforms Changes of Tau Protein During Development in Various Species Dev. Brain Res. 2003 142 121 127 10.1016/S0165-3806(03)00056-7 12711363 68. Lee G. Cowan N. Kirschner M. The Primary Structure and Heterogeneity of Tau Protein from Mouse Brain Science 1988 239 285 288 10.1126/science.3122323 3122323 69. Goedert M. Spillantini M. Cairns N. Crowther R. Tau Proteins of Alzheimer Paired Helical Filaments: Abnormal Phosphorylation of All Six Brain Isoforms Neuron 1992 8 159 168 10.1016/0896-6273(92)90117-V 1530909 70. Goode B.L. Feinstein S.C. Identification of a Novel Microtubule Binding and Assembly Domain in the Developmentally Regulated Inter-Repeat Region of Tau J. Cell Biol. 1994 124 769 782 10.1083/jcb.124.5.769 8120098 71. Gustke N. Trinczek B. Biernat J. Mandelkow E.-M. Domains of Tau Protein and Interactions with Microtubules Biochemistry 1994 33 9511 9522 10.1021/bi00198a017 8068626 72. Brion J.-P. Tremp G. Octave J.-N. Transgenic Expression of the Shortest Human Tau Affects Its Compartmentalization and Its Phosphorylation as in the Pretangle Stage of Alzheimer’s Disease Am. J. Pathol. 1999 154 255 270 10.1016/S0002-9440(10)65272-8 9916940 73. Hernández F. Pérez M. De Barreda E.G. Goñi-Oliver P. Avila J. Tau as a Molecular Marker of Development, Aging and Neurodegenerative Disorders Curr. Aging Sci. 2008 1 56 61 10.2174/1874609810801010056 20021373 74. Spillantini M.G. Goedert M. Tau Pathology and Neurodegeneration Lancet Neurol. 2013 12 609 622 10.1016/S1474-4422(13)70090-5 23684085 75. Götz J. Probst A. Spillantini M. Schäfer T. Jakes R. Bürki K. Goedert M. Somatodendritic Localization and Hyperphosphorylation of Tau Protein in Transgenic Mice Expressing the Longest Human Brain Tau Isoform EMBO J. 1995 14 1304 1313 10.1002/j.1460-2075.1995.tb07116.x 7729409 76. Mattsson N. Sävman K. Österlundh G. Blennow K. Zetterberg H. Converging Molecular Pathways in Human Neural Development and Degeneration Neurosci. Res. 2010 66 330 332 10.1016/j.neures.2009.11.012 19961883 77. Hefti M.M. Farrell K. Kim S. Bowles K.R. Fowkes M.E. Raj T. Crary J.F. High-Resolution Temporal and Regional Mapping of MAPT Expression and Splicing in Human Brain Development PLoS ONE 2018 13 e0195771 10.1371/journal.pone.0195771 29634760 78. Bullmann T. de Silva R. Holzer M. Mori H. Arendt T. Expression of embryonic tau protein isoforms persist during adult neurogenesis in the hippocampus Hippocampus 2007 17 98 102 10.1002/hipo.20255 17183532 79. McMillan P. Korvatska E. Poorkaj P. Evstafjeva Z. Robinson L. Greenup L. Leverenz J. Schellenberg G.D. D’Souza I. Tau Isoform Regulation Is Region- and Cell-Specific in Mouse Brain J. Comp. Neurol. 2008 511 788 803 10.1002/cne.21867 18925637 80. Liu C. Götz J. Profiling Murine Tau With 0N, 1N and 2N Isoform-Specific Antibodies in Brain and Peripheral Organs Reveals Distinct Subcellular Localization, With the 1N Isoform Being Enriched in the Nucleus PLoS ONE 2013 8 e84849 10.1371/journal.pone.0084849 24386422 81. Panda D. Samuel J.C. Massie M. Feinstein S.C. Wilson L. Differential Regulation of Microtubule Dynamics by Three- and Four-Repeat Tau: Implications for the Onset of Neurodegenerative Disease Proc. Natl. Acad. Sci. USA 2003 100 9548 9553 10.1073/pnas.1633508100 12886013 82. Pan W.X. Mao T. Dudman J.T. Inputs to the Dorsal Striatum of the Mouse Reflect the Parallel Circuit Architecture of the Forebrain Front. Neuroanat. 2010 4 147 10.3389/fnana.2010.00147 21212837 83. Jesuthasan S. The Thalamo-Habenula Projection Revisited Semin. Cell Dev. Biol. 2018 78 116 119 10.1016/j.semcdb.2017.08.023 28803896 84. Donovan S.L. McCasland J.S. GAP-43 Is Critical for Normal Targeting of Thalamocortical and Corticothalamic, But Not Trigeminothalamic Axons in the Whisker Barrel System Somatosens. Mot. Res. 2008 25 33 47 10.1080/08990220701830696 18344146 85. Graziano A. Liu X.-B. Murray K.D. Jones E.G. Vesicular Glutamate Transporters Define Two Sets of Glutamatergic Afferents to the Somatosensory Thalamus and Two Thalamocortical Projections in the Mouse J. Comp. Neurol. 2008 507 1258 1276 10.1002/cne.21592 18181146 86. Bezdudnaya T. Keller A. Laterodorsal Nucleus of the Thalamus: A Processor of Somatosensory Inputs J. Comp. Neurol. 2008 507 1979 1989 10.1002/cne.21664 18273888 87. Tlamsa A.P. Brumberg J.C. Organization and Morphology of Thalamocortical Neurons of Mouse Ventral Lateral Thalamus Somatosens. Mot. Res. 2010 27 34 43 10.3109/08990221003646736 20141408 88. Hooks B.M. Mao T. Gutnisky D.A. Yamawaki N. Svoboda K. Shepherd G.M.G. Organization of Cortical and Thalamic Input to Pyramidal Neurons in Mouse Motor Cortex J. Neurosci. 2013 33 748 760 10.1523/JNEUROSCI.4338-12.2013 23303952 89. Collins D. Anastasiades P.G. Marlin J.J. Carter A.G. Reciprocal Circuits Linking the Prefrontal Cortex with Dorsal and Ventral Thalamic Nuclei Neuron 2018 98 366 379.e4 10.1016/j.neuron.2018.03.024 29628187 90. Schroeder S.K. Joly-Amado A. Gordon M.N. Morgan D. Tau-Directed Immunotherapy: A Promising Strategy for Treating Alzheimer’s Disease and Other Tauopathies J. Neuroimmune Pharmacol. 2015 11 9 25 10.1007/s11481-015-9637-6 26538351 91. Crandall J.E. Caviness V.S. Jr. Thalamocortical Connections in Newborn Mice J. Comp. Neurol. 1984 228 542 556 10.1002/cne.902280407 6490968 92. Evrard A. Ropert N. Early Development of the Thalamic Inhibitory Feedback Loop in the Primary Somatosensory System of the Newborn Mice J. Neurosci. 2009 29 9930 9940 10.1523/JNEUROSCI.1671-09.2009 19657043 93. Drummond E. Pires G. MacMurray C. Askenazi M. Nayak S. Bourdon M. Safar J. Ueberheide B. Wisniewski T. Phosphorylated Tau Interactome in the Human Alzheimer’s Disease Brain Brain 2020 143 2803 2817 10.1093/brain/awaa223 32812023 94. Kampers T. Pangalos M. Geerts H. Wiech H. Mandelkow E. Assembly of Paired Helical Filaments from Mouse Tau: Implications for the Neurofibrillary Pathology in Transgenic Mouse Models for Alzheimer’s Disease FEBS Lett. 1999 451 39 44 10.1016/S0014-5793(99)00522-0 10356980 95. Sayas C.L. Medina M. Cuadros R. Ollá I. García E. Pérez M. Ferrer I. Hernández F. Avila J. Role of Tau N-Terminal Motif in the Secretion of Human Tau by End Binding Proteins PLoS ONE 2019 14 e0210864 10.1371/journal.pone.0210864 30668577 96. Paxinos G. Franklin K.B.J. The Mouse Brain in Stereotaxic Coordinates 5th ed. Elsevier Amsterdam, The Netherlands 2019 97. Schröder H. Moser N. Huggenberger S. Neuroanatomy of the Mouse Springer International Publishing Berlin, Germany 2020 10.1007/978-3-030-19898-5

PMC009xxxxxx/PMC9099671.txt

==== Front Nutrients Nutrients nutrients Nutrients 2072-6643 MDPI 10.3390/nu14091762 nutrients-14-01762 Article Gut Microbiota and Phenotypic Changes Induced by Ablation of Liver- and Intestinal-Type Fatty Acid-Binding Proteins https://orcid.org/0000-0002-7910-2813 Wu Guojun 12† Tawfeeq Hiba R. 34† Lackey Atreju I. 34 Zhou Yinxiu 3 Sifnakis Zoe 3 https://orcid.org/0000-0003-4380-5119 Zacharisen Sophia M. 3 Xu Heli 34 Doran Justine M. 3 https://orcid.org/0000-0001-7339-7115 Sampath Harini 134 Zhao Liping 12 https://orcid.org/0000-0002-5724-1142 Lam Yan Y. 125* https://orcid.org/0000-0001-5482-1777 Storch Judith 34* Bellizzi Dina Academic Editor 1 New Jersey Institute for Food, Nutrition and Health, Rutgers University, New Brunswick, NJ 08901, USA; gary.guojun.wu@rutgers.edu (G.W.); harini.sampath@rutgers.edu (H.S.); liping.zhao@rutgers.edu (L.Z.) 2 Department of Biochemistry and Microbiology, Rutgers University, New Brunswick, NJ 08901, USA 3 Department of Nutritional Sciences, Rutgers University, New Brunswick, NJ 08901, USA; hrt11@scarletmail.rutgers.edu (H.R.T.); atrejulackey@gmail.com (A.I.L.); yinzhou@sebs.rutgers.edu (Y.Z.); zoesifnakis@gmail.com (Z.S.); sophia.zacharisen@gmail.com (S.M.Z.); heli0115@hotmail.com (H.X.); justine5522@gmail.com (J.M.D.) 4 Rutgers Center for Lipid Research, Rutgers University, New Brunswick, NJ 08901, USA 5 Gut Microbiota and Metabolism Group, Centre for Chinese Herbal Medicine Drug Development, School of Chinese Medicine, Hong Kong Baptist University, Hong Kong, China * Correspondence: yanlam2021@hkbu.edu.hk (Y.Y.L.); storch@sebs.rutgers.edu (J.S.); Tel.: +852-3411-2922 (Y.Y.L.); +1-848-932-1689 (J.S.) † These authors contributed equally to this work. 22 4 2022 5 2022 14 9 176225 3 2022 14 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Intestinal fatty acid-binding protein (IFABP; FABP2) and liver fatty acid-binding protein (LFABP; FABP1) are small intracellular lipid-binding proteins. Deficiency of either of these proteins in mice leads to differential changes in intestinal lipid transport and metabolism, and to markedly divergent changes in whole-body energy homeostasis. The gut microbiota has been reported to play a pivotal role in metabolic process in the host and can be affected by host genetic factors. Here, we examined the phenotypes of wild-type (WT), LFABP−/−, and IFABP−/− mice before and after high-fat diet (HFD) feeding and applied 16S rRNA gene V4 sequencing to explore guild-level changes in the gut microbiota and their associations with the phenotypes. The results show that, compared with WT and IFABP−/− mice, LFABP−/− mice gained more weight, had longer intestinal transit time, less fecal output, and more guilds containing bacteria associated with obesity, such as members in family Desulfovibrionaceae. By contrast, IFABP−/− mice gained the least weight, had the shortest intestinal transit time, the most fecal output, and the highest abundance of potentially beneficial guilds such as those including members from Akkermansia, Lactobacillus, and Bifidobacterium. Twelve out of the eighteen genotype-related bacterial guilds were associated with body weight. Interestingly, compared with WT mice, the levels of short-chain fatty acids in feces were significantly higher in LFABP−/− and IFABP−/− mice under both diets. Collectively, these studies show that the ablation of LFABP or IFABP induced marked changes in the gut microbiota, and these were associated with HFD-induced phenotypic changes in these mice. intestinal fatty acid-binding protein liver fatty acid-binding protein gut microbiota ==== Body pmc1. Introduction Fatty acid-binding proteins (FABPs) are a family of 14–15 kDa intracellular proteins that are thought to transport fatty acids (FAs) and other lipophilic molecules within the cell interior [1,2]. Liver fatty acid-binding protein (LFABP, FABP1), the first member identified [3,4], is highly expressed in the liver and also found abundantly in the proximal small intestine [5]. In contrast to other FABPs, which bind a single molecule of ligand, LFABP binds two molecules of long-chain fatty acids (LCFAs) [6,7] or two molecules of monoacylglycerol [8], in addition to a variety of other hydrophobic ligands including cholesterol, bile acids [9], lysophospholipids [10], and endocannabinoids [11,12]. In addition to LFABP, intestinal fatty acid-binding protein (IFABP, FABP2) [5,13] is also found in the small intestine (SI), its sole tissue of expression. IFABP has a high affinity for both saturated and unsaturated LCFAs with a single ligand-binding site [6,14,15], and has recently been shown to bind endocannabinoids as well [12]. While their precise functions are not entirely known, both of the enterocyte FABPs are considered to be important as reservoirs for cytoplasmic FAs, preventing lipotoxicity caused by elevated intracellular free fatty acid (FFA) levels, and to traffic FAs to enzymes involved in their synthetic incorporation into triglycerides (TGs) and phospholipids (PLs), or in their oxidation [2,16]. It is also suggested that FABPs may traffic their ligands to proteins involved in cellular signaling [17]. Although both IFABP and LFABP are expressed in the same cell type, the proximal intestinal enterocyte, and while both bind LCFAs, we have demonstrated that the two proteins are functionally distinct. In vitro studies revealed markedly different mechanisms of ligand transfer between IFABP or LFABP and membranes [2,18]. Further, it was found using null mice for either of these genes that LFABP is involved in directing intestinal monoglycerides (MGs) toward TG synthesis and FAs to oxidative pathways, while IFABP directs FAs toward synthesis of TG [8,19]. In addition to these local cellular effects, many phenotypic and metabolic differences at the whole-body level have also been observed between the LFABP and IFABP null mice. Specifically, null mice for LFABP become heavier and fatter on a high-fat diet (HFD) than WT mice, with a lower respiratory exchange ratio (RER), indicative of increased fat oxidation [19,20,21], supporting a role of LFABP in regulating whole-body energy homeostasis. The increase in body weight of LFABP−/− mice was, in part, due to higher food intake, which may be secondary to the increase in mucosal levels of the endocannabinoids (ECs) 2-arachidonoylglycerol (2-AG) and arachidonoylethanolamine (AEA) [19]. Despite their obese phenotype, LFABP−/− mice are normoinsulinemic, display higher levels of spontaneous activity than the WT control mice [19], and are protected against the HFD-induced decline in endurance-exercise capacity [22]. Due to these aforementioned metabolic changes, null mice for LFABP are considered an example of a metabolically healthy obese “MHO” phenotype. Conversely to LFABP−/−, we found that ablation of IFABP results in less weight gain upon HFD feeding relative to WT, with IFABP−/− mice having a higher RER, indicative of greater carbohydrate oxidation, and a lower food intake than WT mice [19]. IFABP ablation did not result in higher fecal lipid concentration [19]. However, we recently found that HFD-fed IFABP−/− mice have blunt villi, a thinner muscularis layer, reduced goblet cell and Paneth cell densities, reduced transit time, increased fecal excretion, and increased intestinal permeability [23]. These findings may indicate nutrient malabsorption, including lipid malabsorption, which likely contributes to the leaner phenotype observed in IFABP null mice [23]. The markedly different whole-body phenotypes in LFABP−/− vs. IFABP−/− mice support a growing understanding of gut lipid metabolism and transport as an important regulatory factor in whole-body energy homeostasis. Notably, the phenotypic changes were not due to compensatory changes in the expression of the other FABPs located in the small intestine (SI) of IFABP−/− and LFABP−/− mice [19], further supporting the independent and distinct roles of the proximal SI FABPs, IFABP, and LFABP in intestinal and whole-body homeostasis. It is now well established that gut microbiota plays an essential role in host health and can modulate many host metabolic processes including lipid metabolism [24] and energy homeostasis [25] through multiple direct and indirect biological mechanisms. These include production of a variety of bioactive compounds such as short-chain fatty acids (SCFAs), lipopolysaccharide (LPS), secondary bile acids, and others [25,26,27,28]. The structure of the gut microbiota is dynamic and can be affected by the amount and composition of dietary carbohydrates and fats [29,30,31,32]. While most of the products of dietary lipid digestion are absorbed in the proximal small intestine, a minority will pass through the gastrointestinal tract and directly modulate the gut microbiota composition in the distal intestine, via modulation of bacterial growth and by influencing bacterial metabolism as substrates [26]. Additionally, host genetic factors can affect the gut microbiota composition. For example, using 113 different strains in the Hybrid Mouse Diversity Panel, Org et al. found that 7 host loci were significantly associated with the gut microbiota composition [33]. The genes in the identified loci were involved in processes related to lipid metabolism, innate immune responses, and acute-phase immunological responses to lipopolysaccharides [33]. To gain insight into whether the observed dramatic whole-body phenotypic divergence between IFABP−/− and LFABP−/− mice was associated with the gut microbiota, here we compared the microbiome of WT, IFABP−/−, and LFABP−/− mice before and after an 11-week high-saturated-fat feeding period. Our findings indicate that alterations in bacterial communities as a function of genotype and secondary to HFD feeding are associated with the lean phenotype of the IFABP−/− mice, and with the MHO phenotype of the LFABP−/− mice. 2. Materials and Methods 2.1. Animals and Diets LFABP−/− mice on a C57BL/6N background were generously provided by Binas and coworkers [34]. The mice were additionally backcrossed with C57BL/6J mice from The Jackson Laboratory (Bar Harbor, ME) as described previously [8,22]. IFABP−/− mice used in the present studies were a substrain bred by intercrossing of the original IFABP−/− mice [13], and were also on a C57BL/6J background as described [8,19]. WT C57BL/6J mice from The Jackson Laboratory bred in our facility were used as controls. Six mice were in each genotypic group. Mice were maintained on a 12 h light/dark cycle and allowed ad libitum access to standard rodent chow (Purina Laboratory Rodent Diet 5015). At 2 months of age, male LFABP−/−, IFABP−/−, and WT mice were housed 2–3 per cage and fed a 45% kcal fat HFD (D10080402; Research Diets, Inc., New Brunswick, NJ, USA) for 11 weeks; the lipid sources were cocoa butter (43% kcal) and soybean oil (2% kcal) (Table 1). 2.2. Body Weight and Food Intake At 2 months of age, mice were fed the HFD. The mice were maintained on this diet for 11 weeks, and body weights were measured each week. Food intake was assessed using the Oxymax system (Columbus Instruments, Columbus, OH, USA) during weeks 10–11 of the feeding protocol. Mice were placed in chambers (1 mouse per chamber) with food for 48 h. The first 24 h were used as an acclimation period, while the second 24 h period was used to measure food intake. 2.3. Intestinal Transit Time Transit time measurements were performed on non-fasted mice between weeks 10 and 11 of the HFD period. Prior to the start of the experiment, mice were individually housed. After two hours of acclimation, mice were given 250 μL of 6% carmine red and 0.5% methylcellulose (Sigma-Aldrich, St. Louis, MO, USA) in PBS by oral gavage. After gavaging the mice, the cages were checked every 10 min and the time of appearance of the first red fecal pellet recorded [35,36]. 2.4. Total Fecal Excretion Mice were housed 2–3 per cage. Feces from each cage were collected every 3–4 days between weeks 10 and 11 of the HFD feeding period, dried overnight at 60 °C, and then weighed. The weight of the feces was divided by the number of mice in the cage, and by the number of days of collection. In order to control for differences in food intake, fecal excretion in grams was normalized by dividing it by the respective 24 h food intake. 2.5. Gut Microbiota Analyses Fresh fecal pellets were collected from 6 individual mice per genotype at baseline and again after 11 weeks of HFD feeding. Samples were snap-frozen in liquid nitrogen and stored at −80 °C until analysis. Genomic DNA was extracted using the QIAmp Power Fecal DNA kit (QIAGEN, Germantown, MD, USA), as per manufacturer instructions. The hypervariable region V4 of the 16S rRNA gene was amplified using the 515F and 806R primers modified by Parada et al. [37] and Apprill et al. [38] and sequenced using the Ion GeneStudio S5 (ThermoFisher Scientific, Waltham, MA, USA). Primers were trimmed from the raw reads using Cutadapt [39] in QIIME 2 [40]. Amplicon sequence variants (ASVs) [41] were obtained by denoising using the dada2 denoise-single command in QIIME 2 with parameters --p-trim-left 0 --p-trunc-len 215. Spurious ASVs were further removed by abundance filtering [42]. A phylogenetic tree of ASVs was built using the QIIME 2 commands alignment mafft, alignment mask, phylogeny fastree, and phylogeny midpoint-root to generate weighted UniFrac metrics. Taxonomy assignment was performed using the q2-feature-classifier plugin [43] in QIIME 2 based on the silva database (release 132) [44]. The data were rarified to 17,000 reads/sample for subsequent analyses. Overall gut microbiota structure was evaluated using alpha diversity indices (Shannon index and observed ASVs) and beta diversity distance metric (weighted UniFrac). Principal coordinates analysis (PCoA) was performed using the R “ape” package [45] to visualize differences in gut microbiota structure between treatment groups along principal coordinates that accounted for most of the variations. Random Forest analysis was performed and cross-validated using the R “randomForest” package [46] and the “rfcv” function, respectively, to test for correlations between gut microbiota composition and body weight. Figures were visualized by the R “ggplot2” package [47] and “pheatmap” package [48]. ASV shared by >25% of the samples were considered prevalent and selected for the guild-level analysis. Pairwise correlations among the ASVs were calculated using the method described by Bland and Altman [49]. The correlation values were converted to a correlation distance (1 − correlation value) and the ASVs were clustered using the Ward clustering algorithm. From the top of the clustering tree, permutational multivariate analysis of variance (PERMANOVA; 9999 permutations with a p < 0.001 cut-off) was used to sequentially determine whether the two clades were significantly different and accordingly clustered the prevalent ASVs into guilds [50]. 2.6. Statistical Analysis Body weight, body weight change, transit time, and fecal output were analyzed using one-way ANOVA with Tukey’s post hoc between groups and repeated-measures ANOVA with Tukey’s post hoc between time points. Shannon index, ASV number, and distance to WT mice were analyzed using a Mann–Whitney test between groups and Wilcoxon matched-pairs signed-ranks test between time points. At each time point, differential guilds between the groups were identified by using a Kruskal–Wallis test with Dunn’s post hoc. A Random Forest (RF) regression model with leave-one-out cross-validation was used to regress body weight on the guild abundance by using R randomForest packages. All statistical analyses were performed using GraphPad Prism (version 9.0.1 for Mac, GraphPad Software, San Diego, CA, USA) and R version 4.1.1. 2.7. GC/MS Analysis of SCFAs in Fecal Samples SCFA species, including acetate, propionate, isobutyrate, butyrate, isovalerate, and valerate from fecal samples of WT, IFABP−/−, and LFABP−/− mice were analyzed by GC/MS as described previously [51], at the core facility of the New Jersey Institute for Food, Nutrition, and Health of Rutgers University. 3. Results 3.1. Body Weight Gain Differs in WT, IFABP−/−, and LFABP−/− Mice after Chronic HF Feeding After 11 weeks of 45% kcal HF feeding, IFABP−/− mice gained less weight and remained lean when compared to both WT and LFABP−/− mice, in agreement with previous results (Figure 1A,B) [19,23]. Compared with WT mice, LFABP−/− mice had a significantly higher body weight at both week 0 and week 11 and had a similar body weight change (%) after the 11 weeks of HFD (Figure 1A,B). 3.2. Intestinal Transit Time and Total Fecal Excretion Are Altered in Mice Lacking IFABP and LFABP As we have reported [23], IFABP−/− mice displayed faster intestinal transit time on the HFD, and higher fecal output normalized for food intake, suggesting some malabsorption of lipid and other nutrients (Figure 1C,D). Interestingly, the 45% kcal HF-fed LFABP−/− mice displayed significantly slower intestinal transit times than their WT controls (Figure 1C), and a significant decrease in total fecal excretion normalized for total food intake (Figure 1D). These changes may contribute, in part, to the increased body weight gain relative to the WT [19]. The observed changes in the intestinal transit time and fecal excretion in both IFABP−/− and LFABP−/− mice, relative to WT and to each other, prompted us to examine potential differential impacts of these genetic modifications on the gut microbiota [52]. 3.3. The Microbiota Composition Is Altered by IFABP and LFABP Ablation and Shows Different Responses to HFD To explore whether IFABP−/− and LFABP−/− mice displayed alterations in the gut microbiota, we collected fecal samples from WT, IFABP−/−, and LFABP−/− mice (n = 6 per group) at both week 0 (8 weeks of age, prior to the HFD feeding period) and at week 11 of HF feeding, to profile gut microbiota composition via 16S rRNA gene V4 sequencing. In total, 785 bacterial amplicon sequence variants (ASVs) [41] were identified from the 36 samples. At week 0, LFABP−/− mice had a significantly higher Shannon index than WT (Figure 2A). At week 11, the differences between LFABP−/− and WT mice remained and LFABP−/− mice also had a significantly higher Shannon index than IFABP−/− mice. Within each genotype, there was no change in Shannon index from week 0 to week 11. Both knockout groups had significantly more ASVs than WT mice at week 0 (Figure 2B). At week 11, the ASV number showed the same differential pattern as the Shannon index among the three groups. Only in WT mice was there a significant increase in ASV number from week 0 to week 11. These results show that IFABP−/− mice have increased gut microbiota diversity relative to WT only under normal chow, while LFABP−/− mice have increased diversity relative to WT under both normal chow and following prolonged HF feeding, and to IFABP−/− mice under HFD only. In contrast to the WT mice, the HFD treatment had no effect on the alpha diversity of the gut microbiota in either of the FABP knockout mice. To compare the overall structure of the gut microbiota across groups, scatter plots of principal coordinate analysis based on weighted UniFrac distance were constructed (Figure 2C). The HFD significantly changed the gut microbiota structure in all groups, with clear segregations observed between week 0 and week 11 within each genotype (PERMANOVA test p = 0.004 in each group, R2 = 0.75 for WT, R2 = 0.78 for IFABP−/−, R2 = 0.65 for LFABP−/−). At week 0, both knockout groups were significantly different from WT (IFABP−/− vs. WT p = 0.009, R2 = 0.33; LFABP−/− vs. WT p = 0.004, R2 = 0.43), while there was no significant difference between IFABP−/− and LFABP−/− mice (p = 0.072, R2 = 0.17). The dissimilarity between LFABP−/− and WT was significantly greater than that between IFABP−/− and WT (Figure 2D). After the 11-week HF feeding, the three genotypes were significantly different from each other (IFABP−/− vs. WT p = 0.041, R2 = 0.29; LFABP−/− vs. WT p = 0.028, R2 = 0.22; IFABP−/− vs. LFABP−/− p = 0.004, R2 = 0.58), and the dissimilarity between IFABP−/− and WT was similar to that between LFABP−/− and WT. These results show that compared with WT, both IFABP−/− and LFABP−/− significantly altered the overall gut microbiota structure under both normal chow and HFD treatment. In addition, the effect of HFD on the gut microbiota structure was more profound in IFABP−/− and WT than in LFABP−/− mice. Bacteria in the gut ecosystem form complex interactions as functional groups rather than existing in isolation [53]. Members that exploit the same class of resources in a similar way can be considered as a guild [54], in which the guild members typically show co-abundance patterns. Thus, to identify potential guilds, we explored the co-abundance relationships among the 202 prevalent and dominant ASVs which were shared in at least 25% of the samples and accounted for ~90% of the total abundance. The 202 ASVs were grouped into 24 different guilds (Table S1). As shown in Figure 3, at week 0, the abundance of three guilds (Guilds #19, 23, and 24) was significantly higher and that of two guilds (Guilds #14 and 15) was significantly lower in IFABP−/− mice compared with WT. A comparison of LFABP−/− and WT mice revealed even more significantly different guilds, i.e., 11 (Guilds #1, 3, 4, 7, 8, 11, 18, 19, 20, 22, and 24) were higher in abundance and 3 (Guilds #14, 15, and 17) were lower in the LFABP−/− mice. Among the 24 differentially regulated guilds, Guilds #19 and 24 increased, and Guilds #14 and 15 decreased in both knockout groups. These results show that, under a low-fat chow diet, both FABP gene knockouts affected several functional guilds. IFABP−/− changed fewer guilds than LFABP−/−, consistent with the aforementioned results that the dissimilarity between IFABP−/− and WT was smaller than that between LFABP−/− and WT. Over the HFD feeding period from week 0 to week 11, 12 guilds (Guilds #1, 2, 3, 4, 5, 6, 7, 8, 10, 11, 12, and 13) increased and 5 (Guilds #14, 17, 19, 20, and 23) decreased significantly in LFABP−/− mice; 9 guilds (Guilds #1, 3, 4, 5, 7, 8, 9, 10, and 12) increased and 5 (Guild #15, 17, 19, 20 and 23) decreased significantly in IFABP−/− mice; 11 guilds (Guilds #1, 3, 4, 6, 7, 8, 9, 10, 11, 12, and 21) increased and 5 (Guild #14, 15, 17, 19, and 23) decreased significantly in WT mice (Figure 3). Among these 20 HF-responding guilds, 10 (Guilds #1, 3, 4, 7, 8, 10, and 12 which increased; Guilds #17, 19, and 23 which decreased) displayed changes in the same direction in all the groups, while the other 10 changed in two or only one of the groups. These results indicate that while some of the HFD-induced changes were independent of genotype, the gut microbiota of WT, IFABP−/−, and LFABP−/− mice also displayed differential responses to the HFD. At week 11 of the HFD, only Guild #18 showed a significant difference between IFABP−/− and WT, being higher in IFABP−/−. Compared with WT, the LFABP−/− had five guilds (Guilds #1, 2, 4, 18, and 20) that were significantly higher and two guilds (Guilds #7 and 14) that were lower. Under both normal chow and HFD, no guilds showed consistent differences between WT and IFABP−/−. However, Guilds #1, 4, 18, and 20 were consistently higher in LFABP−/− compared with WT mice. These results indicate that, at the guild level, the differences between the two knockout groups and WT remain but become smaller after HFD feeding as the number of different guilds decreased. Notably, however, the differences in four guilds between LFABP−/− and WT mice were present regardless of the diet. 3.4. Associations between Gut Microbiota and Body Weight To explore the associations between the gut microbiota and body weight, we applied a Random Forest regression model to correlate the 24 guilds and body weight. Using the data at week 0, based on the leave-one-out cross-validation and feature-selection process, the best regression model with minimum mean squared error for body weight contained eight guilds (Figure 4A,B), all of which showed differences between the three genotype groups. Each of the eight guilds also showed significant correlation with body weight based on Spearman correlations (Table S2). Particularly, among them, three guilds showed very large differences between the three genotype groups (Figure S1A). Guild #17 accounted for 47.97% of the total abundance in WT, 22.09% in IFABP−/−, and 5.29% in LFABP−/− mice. Similarly, Guild #15 was the most abundant in WT (27.19%), followed by IFABP−/− (6.82%) and LFABP−/− (1.60%). In contrast, the abundance of Guild #20 was the lowest in WT (0.42%) but higher in IFABP−/− (2.23%) and markedly higher in LFABP−/− (18.55%). The predicted body weights from cross-validation were significantly correlated with the measured values (r = 0.721, p = 0.001) (Figure 4C). This result indicates that the genotype-related guilds associate with the host body weight under normal chow at 8 weeks of age. To determine associations between guilds and body weight following the HFD period, we applied the Random Forest regression model to correlate the 24 guilds and body weight at week 11 (Figure 5A,B). Ten guilds were included in the best model with a minimum mean squared error. Individually, based on Spearman correlations, 5 out of the 10 showed significant correlation with body weight as well (Table S2). Among these 10, 5 guilds had >5% differences between the three genotypes (Figure S1B). Guild #12 was most dominant in the IFABP−/− mice (20.21% in WT, 43.81% in IFABP−/−, and 6.23% in LFABP−/−). In contrast, the abundance of Guild #10 was the lowest in IFABP−/− mice (13.70% in WT, 5.69% in IFABP−/−, and 12.41% in LFABP−/−). Guilds #1 and 3 were most abundant in LFABP−/− mice (Guild #1: 2.15% in WT, 0.98% in IFABP−/−, and 11.63% in LFABP−/−; Guild #3: 7.53% in WT, 5.96% in IFABP−/−, and 13.43% in LFABP−/−). Guild #9 had the lowest abundance in LFABP−/− mice (9.27% in WT, 8.32% in IFABP−/−, and 3.80% in LFABP−/−). As shown in Figure 5C, the predicted body weights from cross-validation were significantly correlated with the measured values (r = 0.734, p = 0.001). In addition, we found four common guilds (Guilds #1, 3, 4, and 22) in the two Random Forest regression models built on the data at week 0 and week 11. The predicted body weight values, which were based on the week 0 model and the week 11 guilds, were significantly correlated with the measured body weights at week 11 as well (r = 0.519, p = 0.0207) (Figure 4D). These results indicate that the associations between guilds and body weight identified under normal chow are retained, in part, after HFD feeding. The contribution of some guilds to the body weight, by contrast, were manifested only after HFD feeding. 3.5. IFABP and LFABP Ablation and HFD Feeding Alter Fecal SCFA Levels Among the guilds associated with body weight variation, we noted that some of the members were SCFA-producing bacteria (Table S1). It has been reported that SCFAs are important gut microbial metabolites related to host energy homeostasis [25]. To explore whether the different genotypic mice had differences in SCFA levels, we measured SCFAs in the fecal samples. At week 0, prior to starting the HFD, the levels of all measured SCFAs, including acetate, propionate, isobutyrate, butyrate, isovalerate, and valerate showed significant differences between the three genotypes (Figure 6). Acetate, propionate, butyrate, and valerate were significantly higher in both IFABP−/− and LFABP−/− mice compared to their control counterparts (Figure 6A,B,D,F). Isobutyrate and isovalerate were significantly higher than the WT group only in the LFABP−/− mice (Figure 6C,E). After 11 weeks of HF feeding, the concentrations of SCFAs remained different between the three genotypes. In all three genotypes, acetate, propionate, and butyrate levels were significantly decreased when compared to week 0 (Figure 6A,B,D), while valerate was significantly increased after HF feeding (Figure 6F). In keeping with what was observed at week 0, all of the SCFA levels were significantly greater in both IFABP−/− and LFABP−/− mice when compared to the WT mice at week 11 of the HFD period. Both butyrate and valerate were higher in IFABP−/− mice when compared to LFABP−/− (Figure 6D,F), while isovalerate was higher in LFABP−/− compared to IFABP−/− mice (Figure 6E). These results indicate that differences in the levels of SCFAs are primarily due the genetic ablation of IFABP and LFABP, and persisted after chronic HF feeding. 4. Discussion In the present study, we found divergent effects of IFABP vs. LFABP gene knockout on intestinal transit times and fecal output which had significant differential impact on the gut microbiota. The LFABP−/− mice had significantly slower transit, i.e., longer transit times, and lower fecal output per gram consumed. In agreement with our prior findings [23], the opposite was found in the IFABP−/− mice. Thus, the opposing body weight phenotypes of the IFABP and LFABP null mice are likely due, in part, to increased energy harvest in LFABP−/− and decreased energy harvest in IFABP−/− mice. In recent years, it has been shown that FABPs, including both LFABP and IFABP, bind not only LCFAs, but also have high-affinity binding for the ECs 2-arachidonoylglycerol (2-AG) and anandamide [11,12,55,56]. ECs are involved in the regulation of food intake and intestinal motility through activation of the cannabinoid receptor1 (CB1R) on vagal afferent neurons [57,58,59]. It has been shown that activation of the CB1R by receptor agonists such as 2-AG inhibits peristalsis and can increase appetite [60,61,62]. Indeed, we previously showed that mucosal levels of 2-AG were lower in IFABP−/− mice, whereas they were significantly higher in LFABP−/− mice, when compared to their WT counterparts [19]. Thus, the highly divergent phenotypes that have been observed in body weight, in the amount of fecal output, and in the intestinal motility of both IFABP−/− and LFABP−/− mice could, in part, be secondary to altered CB1R activation caused by different mucosal EC levels. Compared with WT mice, both IFABP−/− and LFABP−/− had altered overall gut microbiota structure under a normal chow diet. Although shifting from normal chow to HFD changed the gut microbiota structure dramatically in all the three genotypes, the responses of the gut microbiota in each genotype were different. Such differences together with their different gut microbiotas after HFD feeding may be associated with the aforementioned variations in intestinal motility. Transit time is related to bacterial composition and metabolism in the gut [52,63]. Interestingly, the LFAPB−/− mice had the longest transit time and highest number of bacterial ASVs among the three genotypes, which is consistent with the finding that a long transit time associates with high microbial richness [52,63]. To assess the functional significance of the alterations in the gut microbiota, we applied guild analysis, which overcomes the pitfalls of commonly used taxonomy analysis and is a more ecologically sound approach for finding host phenotype-associated gut microbial members [50]. Under normal chow, among the eight guilds that were associated with body weight, Guilds #15, 17, and 20 showed remarkable and significant differences between the three genotypes. Guild #15 was negatively correlated with body weight and had one ASV from Akkermansia; the species Akkermansia muciniphila in this genus has been characterized as beneficial in whole-body energy metabolism [64]. Guild #17, which was negatively correlated with body weight, had ASVs from genera including Lactobacillus and Lachnoclostridium. Many members of Lactobacillus are considered as probiotics and are associated with host health [65], and members in Lachnoclostridium have been reported to be associated with an anti-obesity function [66]. Guild #20, which was positively correlated with body weight, contained one ASV from Tyzzerella, which have been reported as pro-inflammatory bacteria and to be related to obesity [67,68]. Compared with WT and IFABP−/−, LFABP−/− mice not only had the lowest abundance of the potentially beneficial Guilds #15 and 17 but also highest abundance of the potentially obesogenic Guild #20. After HF feeding, among the 10 guilds that were associated with body weight, Guild #12 was most dominant in the IFABP−/− mice and negatively correlated with body weight. This guild had two ASVs from Lactobacillus, one from Bifidobacterium, one from Ileibacterium, one from Lactococcus, one from Dubosiella newyorkensis, two from Lachnospiraceae, two from Ruminococcaceae, one from Enterorhabdus, and one from Streptococcus. Several members in Lactobacillus, Bifidobacterium, and Lactococcus have been reported to attenuate HFD-induced obesity [69,70,71]. Guild #10, which had three ASVs from Desulfovibrionaceae, had the lowest abundance in the IFABP−/− mice and positively correlated with body weight. Members of Desulfovibrionaceae, which produce endotoxin and hydrogen sulfide, are considered pro-inflammatory and have been reported to be positively associated with obesity and inflammation [72,73]. Guilds #1 and 3, which had ASVs from Odoribacter, had the highest abundance in LFABP−/− mice and were positively correlated with body weight. Odoribacter has been reported to be positively correlated with body weight [74]. Indeed, under both diets, the LFABP−/− mice had the highest body weight among the three genotypes. Overall, the LFABP−/− mice had more potentially obesity-promoting guilds including bacteria such as those from Tyzzerella, Desulfovibrionaceae, and Odoribacter, and fewer anti-obesity guilds including bacteria from Akkermansia, Lactobacillus, Lachnoclostridium, and Bifidobacterium [64,65,66,69,70,71]. The IFABP−/− mice, by contrast, had more anti-obesity and fewer obesity-promoting guilds after HFD feeding, which appears associated with their lean phenotype relative to WT and LFABP−/− mice. In addition to body weight, which was focused on here, our previous studies showed that LFABP−/− mice can be considered an example of being metabolically healthy obese “MHO”, with higher levels of spontaneous activity [19] and protection against the HFD-induced decline in endurance-exercise capacity [22]. Recent human studies have highlighted that exercise can stimulate changes in the gut microbiota associated with higher SCFA production [75,76]. Thus, in addition to the different transit time noted above, higher levels of endurance activity may be considered as another factor which potentially contributes to the significant differences in the gut microbiota between WT and LFABP−/− mice. Previously, we found that LFABP−/− mice had higher muscle glycogen levels and an increased FA oxidation rate when compared with WT mice [22]. Here, we found that SCFAs were significantly higher in LFABP−/− mice compared with WT mice. These findings are consistent with the recently proposed “gut–muscle axis” [77,78], in which SCFAs are considered as potential regulators, via increasing skeletal muscle glycogen and promoting FA uptake and oxidation [79]. Other studies have also shown that high levels of plasma and fecal acetate and propionate are associated with endurance-exercise improvement [80,81]. Thus, the gut microbiota may play an essential role in the “MHO” features of LFABP−/− mice. In previous studies, we showed that IFABP−/− mice remained lean (Figure 1A), a result that was also found in the present studies, and we also showed that the IFABP−/− mice had lower plasma glucose levels than their WT counterparts, and a normoinsulinimic phenotype after chronic HF feeding [19]. Here, we showed that IFABP−/− mice maintain a high level of fecal SCFAs. Many studies have indicated beneficial effects of SCFAs, specifically, acetate, propionate, and butyrate, on energy homeostasis and metabolism, and their crucial role in preventing HFD-induced obesity and improving insulin sensitivity [82,83,84]. Fecal SCFA levels can be modulated by several mechanisms including colonic absorption, colonic transit time, dietary intake, and the microbiota [85]. Though a lower abundance of SCFA-producing bacteria was found in the LFABP−/− mice than in the IFABP−/− mice, their higher level of fecal SCFAs may be related to their longer transit time, which increases the fermentation time [86]. The higher level of fecal SCFAs in the IFABP−/− mice may be related to increased SCFA production or reduced absorption. A higher abundance of SCFA-producing bacteria was identified in IFABP−/− mice, which suggests the possibility of increased SCFA production. However, as IFABP−/− mice have reduced transit time, this may result in reduced absorption of SCFAs. In order to dissect the contributions of the observed SCFA changes to the IFABP−/− and LFABP−/− mice phenotypes, the measurement of SCFA absorption will be of interest. In summary, our results show that the gut microbiota is associated with the high-fat-diet-induced whole-body phenotypes of IFABP−/− and LFABP−/− mice. To determine whether the structure of the microbiota is an essential mediator of the effects of these gene knockouts on host phenotypes, future studies will assess the impact of transplanting the gut microbiota from the IFABP−/− and LFABP−/− mice to germ-free or antibiotic-treated WT mice. The FABP gene family includes a group of diverse proteins that have important roles in regulating host metabolism and have been shown to be related to several metabolic diseases [87]. Better understanding of their functions and mechanisms of action, which may be mediated by the gut microbiota, will facilitate FABP-related drug development and therapeutic approaches for metabolic diseases. Acknowledgments The authors thank Luis Agellon (McGill University) for helpful discussions. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/nu14091762/s1, Figure S1: The average abundance of the guilds which were included in the best model of Random Forest analysis and had >5% difference among the three genotypes; Table S1: 202 ASVs were grouped into 24 different guilds; Table S2: Spearman correlation between Guilds and body weight. Click here for additional data file. Author Contributions G.W., H.R.T., L.Z., Y.Y.L. and J.S. designed the experiments; H.R.T., A.I.L., Y.Z., Z.S., S.M.Z., H.X. and J.M.D. performed animal procedures and sample collections; G.W. and Y.Y.L. performed fecal DNA and gut microbiota analyses; H.S. performed metabolite analysis. H.R.T., G.W., J.S., L.Z. and Y.Y.L. wrote and edited the manuscript. All authors have read and agreed to the published version of the manuscript. Funding This research was funded by National Institutes of Health grant DK-38389 (J.S.) and funds from the New Jersey Agricultural Experiment Station (J.S., L.Z.), and the Health@InnoHK Initiative Fund of the Hong Kong Special Administrative Region Government (ITC RC/IHK/4/7) (Y.Y.L.). Institutional Review Board Statement All animal experiments were approved by the Institutional Animal Care and Use Committee of Rutgers University (protocol ID 999900318). Informed Consent Statement Not applicable. Data Availability Statement The raw gut microbiome sequencing data have been deposited to the sequence read archive at NCBI under the BioProject ID PRJNA772007. Conflicts of Interest L.Z. is a co-founder of Notitia Biotechnologies Company. Y.Y.L. is an employee of Hong Kong Baptist University and an academic staff seconded to the Centre for Chinese Herbal Medicine Drug Development (CDD) for conducting academic research. CDD is a company 100% owned by Hong Kong Baptist University and is hosted by its School of Chinese Medicine. Figure 1 Effect of IFABP and LFABP knockout on body weight (A), body weight change (B), intestinal transit time (C), and total fecal output (D). Repeated-measures ANOVA with Tukey’s post hoc was applied in (A). One-way ANOVA with Tukey’s post hoc was applied in (B–D). * p < 0.05, *** p < 0.001. (B,C) were from a separate group of mice with the same genotypes and fed the same HFD. N = 6 for each group. LFABP; FABP1: Intestinal fatty acid-binding protein (IFABP; FABP2) and liver fatty acid-binding protein. Figure 2 Effect of IFABP and LFABP knockout and a HF diet on the gut microbiota. (A) Shannon Index; (B) ASV number; (C) Principal coordinate plot based on weighted UniFrac distance; (D) Weighted UniFrac distance from IFABP−/− and LFABP−/− to WT at each time point. Data at different timepoints within the same genotype group were compared using the Wilcoxon matched-pairs signed-ranks test (two-tailed) and data at the same timepoint between the groups were compared using the Mann–Whitney test (two-tailed). * p < 0.05, ** p < 0.01. Boxes show the medians and the interquartile ranges (IQRs), and the whiskers denote the lowest and highest values within 1.5 times the IQR from the 1st and 3rd quartiles. N = 6 for each group. Figure 3 Differences and changes in the guilds of the three different genotypes. The heatmap shows the log10-transformed relative abundance of each guild. At each time point, guilds were compared among the groups using the Kruskal–Wallis test and post hoc Dunn’s test. Values not sharing common letters are significantly different from one another (p < 0.05). Wilcoxon matched-pairs signed-ranks test (two-tailed) was used to test the same guild between week 0 and week 11 within each genotype. p < 0.05 was considered as significant. N = 6 for each group. Figure 4 The association between the gut microbiota and body weight at week 0, prior to HF feeding (chow fed from weaning until 8 weeks of age). Random Forest (RF) model regressing body weight on the guild abundance at week 0. (A) shows the number of variables and mean squared error of the corresponding model. (B) The RF assigns a mean error rate, or feature-importance score, to each feature; this value indicates the extent to which each predictor contributes to the accuracy of the model. (C) Significantly positive correlation between the measured body weight and the predicted values from leave-one-out cross-validation based on RF model. (D) Significantly positive correlation between the measured body weight and the predicted values from guild abundance at week 11 based on the model trained in (A). Pearson correlation was applied. Figure 5 The association between the gut microbiota and body weight following 11 weeks of the HF diet. Random Forest (RF) model regressing body weight on the guild abundance at week 11. (A) shows the number of variables and mean squared error of the corresponding model. (B) The RF assigns a mean error rate, or feature-importance score, to each feature; this value indicates the extent to which each predictor contributes to the accuracy of the model. (C) Scatter plot of the measured body weight and the predicted values from leave-one-out cross-validation. Pearson correlation was applied. Figure 6 Analysis of SCFAs in WT, IFABP−/−, and LFABP−/− mice at week 0 (chow diet from weaning until 8 weeks of age) and after 11 weeks of the HF diet. (A) Acetate; (B) Propionate; (C) Isobutyrate; (D) Butyrate; (E) Isovalerate; (F) Valerate. Feces were pooled from six mice in each genotype. Two-way ANOVA with Tukey’s post hoc was applied. * p < 0.05, **p < 0.01, *** p < 0.001. The error bars are from technical replicates. nutrients-14-01762-t001_Table 1 Table 1 FA composition of high-saturated-fat diet [19]. HFS Grams/4057 kcal C16 49.9 C16:1 0.4 C18 64.3 C18:1 65.2 C18:2 10.7 C18:3 1.0 % Saturated fatty acids 60.0 Monounsaturated fatty acids 33.9 Polyunsaturated fatty acids 6.1 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Zimmerman A.W. Veerkamp J.H. New insights into the structure and function of fatty acid-binding proteins Cell. Mol. Life Sci. CMLS 2002 59 1096 1116 10.1007/s00018-002-8490-y 12222958 2. Storch J. Thumser A.E. Tissue-specific functions in the fatty acid-binding protein family J. Biol. Chem. 2010 285 32679 32683 10.1074/jbc.R110.135210 20716527 3. Mishkin S. Stein L. Gatmaitan Z. Arias I.M. The binding of fatty acids to cytoplasmic proteins: Binding to Z protein in liver and other tissues of the rat Biochem. Biophys. Res. Commun. 1972 47 997 1003 10.1016/0006-291X(72)90931-X 5029869 4. Ockner R.K. Manning J.A. Poppenhausen R.B. Ho W.K. A binding protein for fatty acids in cytosol of intestinal mucosa, liver, myocardium, and other tissues Science 1972 177 56 58 10.1126/science.177.4043.56 5041774 5. Gordon J.I. Elshourbagy N. Lowe J.B. Liao W.S. Alpers D.H. Taylor J.M. Tissue specific expression and developmental regulation of two genes coding for rat fatty acid binding proteins J. Biol. Chem. 1985 260 1995 1998 10.1016/S0021-9258(18)89501-1 2579065 6. Richieri G.V. Ogata R.T. Zimmerman A.W. Veerkamp J.H. Kleinfeld A.M. Fatty acid binding proteins from different tissues show distinct patterns of fatty acid interactions Biochemistry 2000 39 7197 7204 10.1021/bi000314z 10852718 7. He Y. Yang X. Wang H. Estephan R. Francis F. Kodukula S. Storch J. Stark R.E. Solution-state molecular structure of apo and oleate-liganded liver fatty acid-binding protein Biochemistry 2007 46 12543 12556 10.1021/bi701092r 17927211 8. Lagakos W.S. Gajda A.M. Agellon L. Binas B. Choi V. Mandap B. Russnak T. Zhou Y.X. Storch J. Different functions of intestinal and liver-type fatty acid-binding proteins in intestine and in whole body energy homeostasis Am. J. Physiol.-Gastrointest. Liver Physiol. 2011 300 G803 G814 10.1152/ajpgi.00229.2010 21350192 9. Thumser A.E. Wilton D.C. The binding of cholesterol and bile salts to recombinant rat liver fatty acid-binding protein Biochem. J. 1996 320 Pt 3 729 733 10.1042/bj3200729 9003356 10. Thumser A.E. Voysey J.E. Wilton D.C. The binding of lysophospholipids to rat liver fatty acid-binding protein and albumin Biochem. J. 1994 301 Pt 3 801 806 10.1042/bj3010801 8053904 11. Huang H. McIntosh A.L. Martin G.G. Landrock D. Chung S. Landrock K.K. Dangott L.J. Li S. Kier A.B. Schroeder F. FABP1: A Novel Hepatic Endocannabinoid and Cannabinoid Binding Protein Biochemistry 2016 55 5243 5255 10.1021/acs.biochem.6b00446 27552286 12. Lai M.P. Katz F.S. Bernard C. Storch J. Stark R.E. Two fatty acid-binding proteins expressed in the intestine interact differently with endocannabinoids Protein Sci. A Publ. Protein Soc. 2020 29 1606 1617 10.1002/pro.3875 13. Vassileva G. Huwyler L. Poirier K. Agellon L.B. Toth M.J. The intestinal fatty acid binding protein is not essential for dietary fat absorption in mice FASEB J. 2000 14 2040 2046 10.1096/fj.99-0959com 11023988 14. Lowe J.B. Sacchettini J.C. Laposata M. McQuillan J.J. Gordon J.I. Expression of rat intestinal fatty acid-binding protein in Escherichia coli. Purification and comparison of ligand binding characteristics with that of Escherichia coli-derived rat liver fatty acid-binding protein J. Biol. Chem. 1987 262 5931 5937 10.1016/S0021-9258(18)45664-5 3553183 15. Richieri G.V. Ogata R.T. Kleinfeld A.M. Equilibrium constants for the binding of fatty acids with fatty acid-binding proteins from adipocyte, intestine, heart, and liver measured with the fluorescent probe ADIFAB J. Biol. Chem. 1994 269 23918 23930 10.1016/S0021-9258(19)51026-2 7929039 16. Gajda A.M. Storch J. Enterocyte fatty acid-binding proteins (FABPs): Different functions of liver and intestinal FABPs in the intestine Prostaglandins Leukot. Essent. Fat. Acids 2015 93 9 16 10.1016/j.plefa.2014.10.001 17. Makowski L. Hotamisligil G.S. Fatty acid binding proteins--the evolutionary crossroads of inflammatory and metabolic responses J. Nutr. 2004 134 2464S 2468S 10.1093/jn/134.9.2464S 15333743 18. Hsu K.T. Storch J. Fatty acid transfer from liver and intestinal fatty acid-binding proteins to membranes occurs by different mechanisms J. Biol. Chem. 1996 271 13317 13323 10.1074/jbc.271.23.13317 8662836 19. Gajda A.M. Zhou Y.X. Agellon L.B. Fried S.K. Kodukula S. Fortson W. Patel K. Storch J. Direct comparison of mice null for liver or intestinal fatty acid-binding proteins reveals highly divergent phenotypic responses to high fat feeding J. Biol. Chem. 2013 288 30330 30344 10.1074/jbc.M113.501676 23990461 20. Atshaves B.P. McIntosh A.L. Storey S.M. Landrock K.K. Kier A.B. Schroeder F. High dietary fat exacerbates weight gain and obesity in female liver fatty acid binding protein gene-ablated mice Lipids 2010 45 97 110 10.1007/s11745-009-3379-2 20035485 21. Newberry E.P. Xie Y. Kennedy S.M. Luo J. Davidson N.O. Protection against Western diet-induced obesity and hepatic steatosis in liver fatty acid-binding protein knockout mice Hepatology 2006 44 1191 1205 10.1002/hep.21369 17058218 22. Xu H. Gajda A.M. Zhou Y.X. Panetta C. Sifnakis Z. Fatima A. Henderson G.C. Storch J. Muscle metabolic reprogramming underlies the resistance of liver fatty acid-binding protein (LFABP)-null mice to high-fat feeding-induced decline in exercise capacity J. Biol. Chem. 2019 294 15358 15372 10.1074/jbc.RA118.006684 31451493 23. Lackey A.I. Chen T. Zhou Y.X. Bottasso Arias N.M. Doran J.M. Zacharisen S.M. Gajda A.M. Jonsson W.O. Corsico B. Anthony T.G. Mechanisms underlying reduced weight gain in intestinal fatty acid-binding protein (IFABP) null mice Am. J. Physiol.-Gastrointest. Liver Physiol. 2020 318 G518 G530 10.1152/ajpgi.00120.2019 31905021 24. Ghazalpour A. Cespedes I. Bennett B.J. Allayee H. Expanding role of gut microbiota in lipid metabolism Curr. Opin. Lipidol. 2016 27 141 147 10.1097/MOL.0000000000000278 26855231 25. Cani P.D. Van Hul M. Lefort C. Depommier C. Rastelli M. Everard A. Microbial regulation of organismal energy homeostasis Nat. Metab. 2019 1 34 46 10.1038/s42255-018-0017-4 32694818 26. Schoeler M. Caesar R. Dietary lipids, gut microbiota and lipid metabolism Rev. Endocr. Metab. Disord. 2019 20 461 472 10.1007/s11154-019-09512-0 31707624 27. Fei N. Zhao L. An opportunistic pathogen isolated from the gut of an obese human causes obesity in germfree mice ISME J. 2013 7 880 884 10.1038/ismej.2012.153 23235292 28. Fei N. Bruneau A. Zhang X.J. Wang R.R. Wang J.X. Rabot S. Gerard P. Zhao L.P. Endotoxin Producers Overgrowing in Human Gut Microbiota as the Causative Agents for Nonalcoholic Fatty Liver Disease Mbio 2020 11 e03263-19 10.1128/mBio.03263-19 32019793 29. Caesar R. Tremaroli V. Kovatcheva-Datchary P. Cani P.D. Backhed F. Crosstalk between Gut Microbiota and Dietary Lipids Aggravates WAT Inflammation through TLR Signaling Cell Metab. 2015 22 658 668 10.1016/j.cmet.2015.07.026 26321659 30. Lam Y.Y. Ha C.W.Y. Hoffmann J.M.A. Oscarsson J. Dinudom A. Mather T.J. Cook D.I. Hunt N.H. Caterson I.D. Holmes A.J. Effects of dietary fat profile on gut permeability and microbiota and their relationships with metabolic changes in mice Obesity 2015 23 1429 1439 10.1002/oby.21122 26053244 31. Devkota S. Wang Y.W. Musch M.W. Leone V. Fehlner-Peach H. Nadimpalli A. Antonopoulos D.A. Jabri B. Chang E.B. Dietary-fat-induced taurocholic acid promotes pathobiont expansion and colitis in Il10−/− mice Nature 2012 487 104 108 10.1038/nature11225 22722865 32. Singh R.K. Chang H.W. Yan D. Lee K.M. Ucmak D. Wong K. Abrouk M. Farahnik B. Nakamura M. Zhu T.H. Influence of diet on the gut microbiome and implications for human health J. Transl. Med. 2017 15 73 10.1186/s12967-017-1175-y 28388917 33. Org E. Parks B.W. Joo J.W. Emert B. Schwartzman W. Kang E.Y. Mehrabian M. Pan C. Knight R. Gunsalus R. Genetic and environmental control of host-gut microbiota interactions Genome Res. 2015 25 1558 1569 10.1101/gr.194118.115 26260972 34. Martin G.G. Danneberg H. Kumar L.S. Atshaves B.P. Erol E. Bader M. Schroeder F. Binas B. Decreased liver fatty acid binding capacity and altered liver lipid distribution in mice lacking the liver fatty acid-binding protein gene J. Biol. Chem. 2003 278 21429 21438 10.1074/jbc.M300287200 12670956 35. Nagakura Y. Ito H. Kiso T. Naitoh Y. Miyata K. The selective 5-hydroxytryptamine (5-HT)4-receptor agonist RS67506 enhances lower intestinal propulsion in mice Jpn. J. Pharmacol. 1997 74 209 212 10.1016/S0021-5198(19)31411-8 9243330 36. Kelly J.R. Borre Y. O’Brien C. Patterson E. El Aidy S. Deane J. Kennedy P.J. Beers S. Scott K. Moloney G. Transferring the blues: Depression-associated gut microbiota induces neurobehavioural changes in the rat J. Psychiatr. Res. 2016 82 109 118 10.1016/j.jpsychires.2016.07.019 27491067 37. Parada A.E. Needham D.M. Fuhrman J.A. Every base matters: Assessing small subunit rRNA primers for marine microbiomes with mock communities, time series and global field samples Environ. Microbiol. 2016 18 1403 1414 10.1111/1462-2920.13023 26271760 38. Apprill A. McNally S. Parsons R. Weber L. Minor revision to V4 region SSU rRNA 806R gene primer greatly increases detection of SAR11 bacterioplankton Aqaut. Microb. Ecol. 2015 75 129 137 10.3354/ame01753 39. Martin M. Cutadapt removes adapter sequences from high-throughput sequencing reads EMBnet J. 2011 17 10 12 10.14806/ej.17.1.200 40. Bolyen E. Rideout J.R. Dillon M.R. Bokulich N.A. Abnet C.C. Al-Ghalith G.A. Alexander H. Alm E.J. Arumugam M. Asnicar F. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2 Nat. Biotechnol. 2019 37 852 857 10.1038/s41587-019-0209-9 31341288 41. Callahan B.J. McMurdie P.J. Rosen M.J. Han A.W. Johnson A.J. Holmes S.P. DADA2: High-resolution sample inference from Illumina amplicon data Nat. Methods 2016 13 581 583 10.1038/nmeth.3869 27214047 42. Wang J. Zhang Q. Wu G. Zhang C. Zhang M. Zhao L. Minimizing spurious features in 16S rRNA gene amplicon sequencing PeerJ Prepr. 2018 6 2167 9843 10.7287/peerj.preprints.26872v1 43. Bokulich N.A. Kaehler B.D. Rideout J.R. Dillon M. Bolyen E. Knight R. Huttley G.A. Gregory Caporaso J. Optimizing taxonomic classification of marker-gene amplicon sequences with QIIME 2’s q2-feature-classifier plugin Microbiome 2018 6 90 10.1186/s40168-018-0470-z 29773078 44. Quast C. Pruesse E. Yilmaz P. Gerken J. Schweer T. Yarza P. Peplies J. Glockner F.O. The SILVA ribosomal RNA gene database project: Improved data processing and web-based tools Nucleic Acids Res. 2013 41 D590 D596 10.1093/nar/gks1219 23193283 45. Paradis E. Claude J. Strimmer K. APE: Analyses of Phylogenetics and Evolution in R language Bioinformatics 2004 20 289 290 10.1093/bioinformatics/btg412 14734327 46. Liaw A. Wiener M. Classification and regression by randomForest R News 2002 2 18 22 47. Wickham H. Ggplot2: Elegant Graphics for Data Analysis Springer Berlin/Heidelberg, Germany 2016 10.18637/jss.v077.b02 48. Kolde R. Pheatmap: Pretty Heatmaps; R Package Version 1.0. 8 2015 Available online: https://CRAN.R-project.org/package=pheatmap (accessed on 24 March 2022) 49. Bland J.M. Altman D.G. Calculating correlation coefficients with repeated observations: Part 1—Correlation within subjects BMJ 1995 310 446 10.1136/bmj.310.6977.446 7873953 50. Wu G. Zhao N. Zhang C. Lam Y.Y. Zhao L. Guild-based analysis for understanding gut microbiome in human health and diseases Genome Med. 2021 13 22 10.1186/s13073-021-00840-y 33563315 51. García-Villalba R. Giménez-Bastida J.A. García-Conesa M.T. Tomás-Barberán F.A. Carlos Espín J. Larrosa M. Alternative method for gas chromatography-mass spectrometry analysis of short-chain fatty acids in faecal samples J. Sep. Sci. 2012 35 1906 1913 10.1002/jssc.201101121 22865755 52. Muller M. Hermes G.D.A. Canfora E.E. Smidt H. Masclee A.A.M. Zoetendal E.G. Blaak E.E. Distal colonic transit is linked to gut microbiota diversity and microbial fermentation in humans with slow colonic transit Am. J. Physiol.-Gastrointest. Liver Physiol. 2020 318 G361 G369 10.1152/ajpgi.00283.2019 31869241 53. Faust K. Raes J. Microbial interactions: From networks to models Nat. Rev. Microbiol. 2012 10 538 550 10.1038/nrmicro2832 22796884 54. Simberloff D. Dayan T. The guild concept and the structure of ecological communities Annu. Rev. Ecol. Syst. 1991 22 115 143 10.1146/annurev.es.22.110191.000555 55. Kaczocha M. Glaser S.T. Deutsch D.G. Identification of intracellular carriers for the endocannabinoid anandamide Proc. Natl. Acad. Sci. USA 2009 106 6375 6380 10.1073/pnas.0901515106 19307565 56. Lagakos W.S. Guan X. Ho S.Y. Sawicki L.R. Corsico B. Kodukula S. Murota K. Stark R.E. Storch J. Liver fatty acid-binding protein binds monoacylglycerol in vitro and in mouse liver cytosol J. Biol. Chem. 2013 288 19805 19815 10.1074/jbc.M113.473579 23658011 57. Partosoedarso E.R. Abrahams T.P. Scullion R.T. Moerschbaecher J.M. Hornby P.J. Cannabinoid1 receptor in the dorsal vagal complex modulates lower oesophageal sphincter relaxation in ferrets J. Physiol. 2003 550 149 158 10.1113/jphysiol.2003.042242 12879865 58. Vianna C.R. Donato J. Jr. Rossi J. Scott M. Economides K. Gautron L. Pierpont S. Elias C.F. Elmquist J.K. Cannabinoid receptor 1 in the vagus nerve is dispensable for body weight homeostasis but required for normal gastrointestinal motility J. Neurosci. 2012 32 10331 10337 10.1523/JNEUROSCI.4507-11.2012 22836266 59. DiPatrizio N.V. Endocannabinoids in the Gut Cannabis Cannabinoid Res. 2016 1 67 77 10.1089/can.2016.0001 27413788 60. Cristino L. Becker T. Di Marzo V. Endocannabinoids and energy homeostasis: An update BioFactors 2014 40 389 397 10.1002/biof.1168 24752980 61. Wang J. Ueda N. Biology of endocannabinoid synthesis system Prostaglandins Other Lipid Mediat. 2009 89 112 119 10.1016/j.prostaglandins.2008.12.002 19126434 62. Koch M. Cannabinoid Receptor Signaling in Central Regulation of Feeding Behavior: A Mini-Review Front. Neurosci. 2017 11 293 10.3389/fnins.2017.00293 28596721 63. Roager H.M. Hansen L.B.S. Bahl M.I. Frandsen H.L. Carvalho V. Gobel R.J. Dalgaard M.D. Plichta D.R. Sparholt M.H. Vestergaard H. Colonic transit time is related to bacterial metabolism and mucosal turnover in the gut Nat. Microbiol. 2016 1 16093 10.1038/nmicrobiol.2016.93 27562254 64. Xu Y. Wang N. Tan H.Y. Li S. Zhang C. Feng Y.B. Function of Akkermansia muciniphila in Obesity: Interactions With Lipid Metabolism, Immune Response and Gut Systems Front. Microbiol. 2020 11 219 10.3389/fmicb.2020.00219 32153527 65. Heeney D.D. Gareau M.G. Marco M.L. Intestinal Lactobacillus in health and disease, a driver or just along for the ride? Curr. Opin. Biotechnol. 2018 49 140 147 10.1016/j.copbio.2017.08.004 28866243 66. Wang P. Li D. Ke W. Liang D. Hu X. Chen F. Resveratrol-induced gut microbiota reduces obesity in high-fat diet-fed mice Int. J. Obes. 2020 44 213 225 10.1038/s41366-019-0332-1 30718820 67. Zhou L. Ni Z.X. Yu J. Cheng W. Cai Z.L. Yu C.Q. Correlation Between Fecal Metabolomics and Gut Microbiota in Obesity and Polycystic Ovary Syndrome Front. Endocrinol. 2020 11 628 10.3389/fendo.2020.00628 68. Chen L.W. Xu J. Soh S.E. Aris I.M. Tint M.T. Gluckman P.D. Tan K.H. Shek L.P.C. Chong Y.S. Yap F. Implication of gut microbiota in the association between infant antibiotic exposure and childhood obesity and adiposity accumulation Int. J. Obes. 2020 44 1508 1520 10.1038/s41366-020-0572-0 69. Won S.M. Chen S. Lee S.Y. Lee K.E. Park K.W. Yoon J.H. Lactobacillus sakei ADM14 Induces Anti-Obesity Effects and Changes in Gut Microbiome in High-Fat Diet-Induced Obese Mice Nutrients 2020 12 3703 10.3390/nu12123703 70. Wang J. Tang H. Zhang C. Zhao Y. Derrien M. Rocher E. van-Hylckama Vlieg J.E. Strissel K. Zhao L. Obin M. Modulation of gut microbiota during probiotic-mediated attenuation of metabolic syndrome in high fat diet-fed mice ISME J. 2015 9 1 15 10.1038/ismej.2014.99 24936764 71. Naudin C.R. Maner-Smith K. Owens J.A. Wynn G.M. Robinson B.S. Matthews J.D. Reedy A.R. Luo L. Wolfarth A.A. Darby T.M. Lactococcus lactis Subspecies cremoris Elicits Protection Against Metabolic Changes Induced by a Western-Style Diet Gastroenterology 2020 159 639 651.e5 10.1053/j.gastro.2020.03.010 32169430 72. Loubinoux J. Bronowicki J.P. Pereira I.A.C. Mougenel J.L. Le Faou A.E. Sulfate-reducing bacteria in human feces and their association with inflammatory bowel diseases Fems Microbiol. Ecol. 2002 40 107 112 10.1111/j.1574-6941.2002.tb00942.x 19709217 73. Zhang C.H. Zhang M.H. Wang S.Y. Han R.J. Cao Y.F. Hua W.Y. Mao Y.J. Zhang X.J. Pang X.Y. Wei C.C. Interactions between gut microbiota, host genetics and diet relevant to development of metabolic syndromes in mice Isme J. 2010 4 312 313 10.1038/ismej.2009.144 74. Li J. Pang B. Shao D. Jiang C. Hu X. Shi J. Artemisia sphaerocephala Krasch polysaccharide mediates lipid metabolism and metabolic endotoxaemia in associated with the modulation of gut microbiota in diet-induced obese mice Int. J. Biol. Macromol. 2020 147 1008 1017 10.1016/j.ijbiomac.2019.10.069 31726144 75. Allen J.M. Mailing L.J. Niemiro G.M. Moore R. Cook M.D. White B.A. Holscher H.D. Woods J.A. Exercise Alters Gut Microbiota Composition and Function in Lean and Obese Humans Med. Sci. Sports Exerc. 2018 50 747 757 10.1249/MSS.0000000000001495 29166320 76. Barton W. Penney N.C. Cronin O. Garcia-Perez I. Molloy M.G. Holmes E. Shanahan F. Cotter P.D. O’Sullivan O. The microbiome of professional athletes differs from that of more sedentary subjects in composition and particularly at the functional metabolic level Gut 2018 67 625 633 10.1136/gutjnl-2016-313627 28360096 77. Ticinesi A. Lauretani F. Milani C. Nouvenne A. Tana C. Del Rio D. Maggio M. Ventura M. Meschi T. Aging Gut Microbiota at the Cross-Road between Nutrition, Physical Frailty, and Sarcopenia: Is There a Gut-Muscle Axis? Nutrients 2017 9 1303 10.3390/nu9121303 29189738 78. Grosicki G.J. Fielding R.A. Lustgarten M.S. Gut Microbiota Contribute to Age-Related Changes in Skeletal Muscle Size, Composition, and Function: Biological Basis for a Gut-Muscle Axis Calcif. Tissue Int. 2018 102 433 442 10.1007/s00223-017-0345-5 29058056 79. Frampton J. Murphy K.G. Frost G. Chambers E.S. Short-chain fatty acids as potential regulators of skeletal muscle metabolism and function Nat. Metab. 2020 2 840 848 10.1038/s42255-020-0188-7 32694821 80. Okamoto T. Morino K. Ugi S. Nakagawa F. Lemecha M. Ida S. Ohashi N. Sato D. Fujita Y. Maegawa H. Microbiome potentiates endurance exercise through intestinal acetate production Am. J. Physiology. Endocrinol. Metab. 2019 316 E956 E966 10.1152/ajpendo.00510.2018 81. Morita H. Kano C. Ishii C. Kagata N. Ishikawa T. Uchiyama Y. Hara S. Nakamura T. Fukuda S.J.b. Bacteroides uniformis enhances endurance exercise performance through gluconeogenesis BioRxiv 2020 10.1101/2020.03.04.975730 82. Den Besten G. Bleeker A. Gerding A. van Eunen K. Havinga R. van Dijk T.H. Oosterveer M.H. Jonker J.W. Groen A.K. Reijngoud D.J. Short-Chain Fatty Acids Protect Against High-Fat Diet-Induced Obesity via a PPARγ-Dependent Switch From Lipogenesis to Fat Oxidation Diabetes 2015 64 2398 2408 10.2337/db14-1213 25695945 83. Lu Y. Fan C. Li P. Lu Y. Chang X. Qi K. Short Chain Fatty Acids Prevent High-fat-diet-induced Obesity in Mice by Regulating G Protein-coupled Receptors and Gut Microbiota Sci. Rep. 2016 6 37589 10.1038/srep37589 27892486 84. Gao Z. Yin J. Zhang J. Ward R.E. Martin R.J. Lefevre M. Cefalu W.T. Ye J. Butyrate improves insulin sensitivity and increases energy expenditure in mice Diabetes 2009 58 1509 1517 10.2337/db08-1637 19366864 85. Wong J.M. de Souza R. Kendall C.W. Emam A. Jenkins D.J. Colonic health: Fermentation and short chain fatty acids J. Clin. Gastroenterol. 2006 40 235 243 10.1097/00004836-200603000-00015 16633129 86. Tottey W. Feria-Gervasio D. Gaci N. Laillet B. Pujos E. Martin J.F. Sebedio J.L. Sion B. Jarrige J.F. Alric M. Colonic Transit Time Is a Driven Force of the Gut Microbiota Composition and Metabolism: In Vitro Evidence J. Neurogastroenterol. 2017 23 124 134 10.5056/jnm16042 87. Smathers R.L. Petersen D.R. The human fatty acid-binding protein family: Evolutionary divergences and functions Hum. Genom. 2011 5 170 191 10.1186/1479-7364-5-3-170

PMC009xxxxxx/PMC9099672.txt

==== Front Animals (Basel) Animals (Basel) animals Animals : an Open Access Journal from MDPI 2076-2615 MDPI 10.3390/ani12091206 animals-12-01206 Article Evaluation of the Nutritive Value and Digestibility of Sprouted Barley as Feed for Growing Lambs: In Vivo and In Vitro Studies https://orcid.org/0000-0003-3994-9216 Al-Baadani Hani H. 1 Alowaimer Abdullah N. 1 Al-Badwi Mohammed A. 1 Abdelrahman Mutassim M. 1 https://orcid.org/0000-0002-6273-2050 Soufan Walid H. 2 https://orcid.org/0000-0003-2547-4395 Alhidary Ibrahim A. 1* Bovera Fulvia Academic Editor Piccolo Giovanni Academic Editor 1 Department of Animal Production, College of Food and Agriculture Science, King Saud University, P.O. Box 2460, Riyadh 11451, Saudi Arabia; hsaeed@ksu.edu.sa (H.H.A.-B.); aowaimer@ksu.edu.sa (A.N.A.); malbadwi@ksu.deu.sa (M.A.A.-B.); amutassim@ksu.edu.sa (M.M.A.) 2 Department of Plant Production, College of Food and Agriculture Science, King Saud University, P.O. Box 2460, Riyadh 11451, Saudi Arabia; wsoufan@ksu.edu.sa * Correspondence: ialhidary@ksu.edu.sa; Tel.: +966-555339331 07 5 2022 5 2022 12 9 120619 4 2022 03 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Simple Summary Sprouted barley has recently been used in animal feed, but studies on its nutritional and digestibility values for growing lambs are very limited, especially when compared with traditional forages. However, the aim of the current research is to determine the practical level and nutritional properties of sprouted barley as a replacement for the traditional feed for growing lambs by ascertaining its efficacy of performance, nutritional value, digestibility, and fermentation characteristics in vivo and in vitro. The current research helps both workers and researchers to determine the level and the most effective way to utilize sprouted barley as an alternative strategy to traditional feeding systems. In conclusion, sprouted barley with traditional feed improves digestibility and fermentation characteristics. Further studies are needed to increase nutrient requirements for optimal lamb growth performance. Abstract The main objective of this study was to investigate the effects of freshly sprouted barley on the growth of lambs, in addition to its nutritional value and digestibility. In addition, sprouted barley digestibility and rumen fermentation were studied in vitro on a dry matter (DM) basis. A total of 45 three-month-old Awassi lambs were randomly assigned to five treatments of sprouted barley (0, 25, 50, 75, 100%) diets. Bodyweight, weight gain, feed intake and feed efficiency were recorded every two weeks. Nutrient analyses were performed on feed, faecal, and urine samples. DM and non-fibrous carbohydrates were measured. Digestibility of DM, organic matter (OM), and neutral detergent fiber (NDF), as well as gas production, pH value, ammonia-N, and volatile fatty acids (VFAs), were determined in vitro using continuous culture. The results showed that final bodyweight was lower (p < 0.05), while feed intake and the feed-to-gain ratio were increased (p < 0.05) in sprouted barley treatments. Nutrient analysis indicators of sprouted barley treatments (25 to100%) were lower (p < 0.05) for DM, crude protein, acid detergent fiber, lignin and ash, and higher for total digestible nutrients, NDF, fat, phosphorus, zinc, copper, and net energy than the traditional diet. In the in vivo study, the digestibility of nutrients in sprouted barley treatments was improved (p < 0.05), while the diet (sprouted barley 100%) had the lowest digestibility of DM, OM, and NDF compared with the other treatments in the in vitro study. In conclusion, the addition of sprouted barley improved digestibility, and fermentation characteristics, while having a negative effect on growth. Further studies are recommended for optimal growth performance. growing lambs nutrient digestibility performance rumen fermentation sprouted barley National Plan for Science, Technology and Innovation (MAARIFAH), King Abdulaziz City for Science and Technology, Kingdom of Saudi Arabia2-17-04-001-0030 This Project was funded by the National Plan for Science, Technology and Innovation (MAARIFAH), King Abdulaziz City for Science and Technology, Kingdom of Saudi Arabia, Award Number (2-17-04-001-0030). ==== Body pmc1. Introduction Lambs are an important resource that support food security in many countries, which are mostly raised on extensive grazing systems. These systems differ in terms of quality and quantity, which depend mainly on climatic variations, including temperature and precipitation, and are also seasonal [1]. However, the nutrient requirements of grazing animals are not often met under such systems to allow them to reach their productive efficiency [2]. Due to the lack of rangeland fodder, people have been compelled to switch to alternative feed sources and, as a result, their production patterns have shifted to semi-intensive systems using traditional feeds such as grain and rough fodder. At the same time, there are obstacles that herd owners face in using these systems, including high prices globally and waste from water source consumption [3]. Sprouted barley is a new way of producing feed forages without using soil, with a high germination rate and a fast-growing period. This method could be especially important in regions where water shortages and the seasonality of forages are common challenges for livestock producers [4]. Sprouted grains are efficiently digested compared to grain seeds because of the high activation of hydrolytic enzymes as a result of germination [5]. Consequently, Fazaeli et al. [6] reported that hydrolytic enzymes convert proteins, starch and fat into simple forms of amino acids, sugars, and fatty acids. Furthermore, the sprouted process increases the content of crude fiber [7], chelates of minerals [8], and decreases the content of phytic acid and protease inhibitors, as well as many other anti-nutrients [9]. In addition, the important benefit of producing sprouted barley is the minimal water consumption compared to the conventional production system. Germination has been demonstrated to be an inexpensive (low-cost process) and sustainable process that improves nutrient quality and the content of functional compounds of grains, as well as their palatability, digestibility, and bioavailability [10,11]. However, the magnitude of changes caused by germination depends on the grain variety and germination conditions [12]. Dung et al. [13] indicated that the benefit of sprouted barley in lamb feed may be negated by the total dry matter (DM) loss, with no improvement in nutrient concentrations or digestibility. Several studies suggest that feeding sprouted barley increases performance only in animals that do not receive adequate protein, energy, or minerals [14], or that the readily available nutrients in sprouted barley may stimulate enhanced utilization of poor-quality feed [15]. To our knowledge, the nutritional value and digestibility of sprouted barley, especially compared to traditional forages such as barley grain and alfalfa hay, have not been extensively demonstrated in previous studies. The hypothesis of this study is that sprouted barley could be an alternative strategy to traditional feed for growing lambs, with the identification of factors resulting from the substitution of animals. Therefore, the main objective of the current study is to investigate the effects of freshly sprouted barley levels with traditional feed on growth performance, nutritional value, and digestibility in growing lambs. Additionally, to evaluate sprouted barley levels with traditional feed (DM basis) on digestibility, CO2 production, and fermentation characteristics in vitro using a continuous culture fermentation. 2. Materials and Methods Animal husbandry and sampling were carried out in accordance with the procedures established by the Scientific Research Ethics Committee, King Saud University, Saudi Arabia (Ethics Reference No: KSU-SE-22-01), also with the care and use of farm animals in research in the United States of America, Animal Science Association [16]. 2.1. Sprouted Barley Production The use of hydroponics for the cultivation of sprouted barley was carried out according to the method described by Al-Saadi and Al-Zubiadi [17]. Briefly, the sprouted barley production plan was carried out in a hydroponic steel chamber (3.0 m length × 2.5 m height × 2.5 m width) in the Department of Food and Agriculture Science, King Saud University, Saudi Arabia. The hydroponic chamber was designed to accommodate 140 trays (70 cm length × 30 cm width) with a capacity for seven growth stages (7 days). Each tray had an automated sprinkler watering system, and the conditions inside the chamber were managed in terms of ventilation, heat, and relative humidity by conditioning and circulating the air to maintain a constant temperature range of 18–20 °C and relative humidity of around 75%. Fluorescent lighting of about 1 watt/cm3 was used throughout the day in a vertical position for leaf development. Barley seeds were purchased locally. About 2 kg of seeds were placed in each tray after cleaning and washing and soaking in water for 24 h. On the seventh day of the growth stage, when they had reached a height of 18–20 cm, the carpets were removed and aired for 24 h to dry them further before being cut and presented to the animals, with the production cycle continuing daily during the study period. 2.2. Diets Sampling and Analysis During the study period, every 15 days of the study (five replicates), feed samples were collected in the same levels of sprouted barley given to the treated lambs, dried (60 °C) to determine the initial moisture content and then ground to a fine powder. According to the previously described method [2], the forage powder samples were analyzed in triplicate to estimate the content of nutrients such as dry matter (DM) by drying overnight at 105 °C in a drying oven (Sanyo convection oven, Osaka, Japan), crude protein (CP; Kjeldahl method, using an N conversion factor of 6.6), crude fat, ash, total digestible nutrients (TDN), and net energy (NE) according to the methods of the Association of Official Analytical Chemists [18]. According to Van Soest et al. [19], fiber fractions such as neutral detergent fiber (NDF), acid detergent fiber (ADF), and lignin were determined. Organic matter (OM) was calculated as OM % = 100-ash [20]. Non-fibrous carbohydrates (NFC) were calculated using the equation NFC (%) = 100% − (CP + total fiber + crude fat + ash) according to Bachmann et al. [21]. The content of macro and micro minerals such as calcium, phosphorus, magnesium, potassium, sulfur, sodium, zinc, and copper were digested in a mixture of sulfuric acid and hydrogen peroxide (105 °C) in a closed microwave using the method previously described [6,22]. All minerals were determined by an atomic absorption spectrometer (PerkinElmer, instruments, Analyst, Waltham, MA, USA) using the Association of Official Analytical Chemists [18]. 2.3. Housing Lambs and Experimental Design A total of 45 growing Awassi male lambs (27.85 ± 2.5 kg) were used for the present study for 75 days. They were purchased at the age of 3 months from local trustworthy farms, and then brought to the Experimental farm of the Department of Animal Production Department, the University of Food and Agriculture Sciences, King Saud University, Riyadh. Lambs were weighed individually and then randomly divided into 45 individual pens (150 × 120 cm) in five feeding treatments, each pen representing one experimental unit (nine replicates per treatment with one lamb per replicate), based on a completely randomized design under the natural winter environmental conditions of the region. Feed treatments were distributed as follows: T1: 100% added traditional feed (control; 70% barley grain + 30% alfalfa hay), T2: 25% added barley sprouts with 75% traditional feed, T3: 50% added barley sprouts with 50% traditional feed, T4: 75% added barley sprouts with 25% traditional feed and T5: 100% added barley sprouts. The feed ingredients and nutrient composition of all dietary treatments are listed in Table 1. Before the start of the actual study, all lambs were acclimatized to the used diet for 14 days, vaccinated against enterotoxaemia and septicemia, and PPR subcutaneously inoculated with an agent against ecto-/endoparasites according to the recommendations of the Directorate of Animal Resources of the Ministry of Environment, Water and Agriculture [23] in Saudi Arabia, Veterinary Vaccine Centre (manufactured by Ibrize Co. Riyadh, Saudi Arabia). During the study period, all animals were provided ad libitum access to feed and water, as well as up to 5% additional feed daily to reach refusal. 2.4. Parameters of Growth Performance All lambs were weighed to determine initial body weight and final body weight (1 and 75 days) to determine average weight gain and daily gain (final body weight − initial body weight). Average daily feed intake (amount of feed rejected − feed offered) was also measured to determine the feed-to-gain ratio (feed intake/weight gain) according to Pereira et al. [24]. Relative growth (RG) was calculated using the following equation: RG = 2 (final body weight − initial body weight)/(final body weight + initial body weight) × 100 according to Goiri et al. [25]. 2.5. Digestibility Procedure and Analysis In Vivo At the end of the 75-day feeding experiment, five lambs were randomly selected from each treatment and housed in metabolism cages (120 × 80 cm) to perform a digestibility experiment according to the methodology described by Omar [26]; Al-Saadi and Al-Zubiadi [17]. The digestion experiment lasted eight days, beginning with a four-day acclimation phase for all animals in the cages, followed by a four-day collection period, during which daily feed intake and fecal and urine excretion were recorded. Feed samples (both offered and rejected), feces, and urine samples were collected and stored at −20 °C until nutrient analysis. Urine was analyzed for nitrogen content, while fecal samples were subjected to the same tests as the feed samples. Apparent digestibility (AD) was calculated on a DM basis using the following formula: AD % = [(intake − fecal excretion)/intake] × 100 for each animal according to Bachmann et al. [21]. 2.6. Digestibility Procedures and Gas Production of Diets In Vitro Diet samples (five dietary treatments; T1:T5) were collected and dried to process DM and stored in sterile bags at 5 °C until apparent digestibility could be estimated using the method of Embaby et al. [27]. In four ANKOM gas jars per treatment as replicates (20 gas jars/treatment), 70 mL of strained rumen fluid was combined with 130 mL of pre-warmed buffer medium (as batch rumen cultures). Then, 200 g of finely powdered feed (based on DM) was weighed into a Dacron bag (Ankom Inc., Fairport, NY, USA) and then added to each of the ANKOM jars, and then incubated for 24 h to determine the apparent digestibility of DM, OM, and NDF. The samplers and feed treatments were analyzed for DM calculation, ash analysis for calculation of OM (100-Ash), and NDF according to the methods of the Association of Official Analytical Chemists [18]. All jars of dietary treatments, which were previously sealed, were filled with CO2 and then connected to a Tedlar gas collection bag (Santa Ana, CA, USA). The jars were placed in a water bath (Thermo Fisher Scientific, model 2873, Waltham, MA, USA) at 39 °C for 24 h until the release process of the gasses in the collection bag was completed. During this process, the jars were shaken for 30 s every 2 h. All CO2 gas production values were expressed as milliliters per time (2 h) and total gas production (24 h) [28]. 2.7. Measurement of Fermentation Characteristics of Diets In Vitro To determine the volatile fatty acids (VFA) and Ammonia-N (NH3-N) in the culture jar at the end of each experiment. From each jar, 5 mL of sample was taken and placed directly in an ice bath and then stored at −20 °C until analysis. Samples were mixed with 1 mL of 25% meta-phosphoric acid and centrifuged at 20,000× g for 10 min at 4 °C to produce a clear supernatant. One ml of the supernatant was filtered with a PTFE syringe filter (0.2 μm) and transferred to a 1.5 mL glass chromatography vial (Agilent). Acetic acid (C2), propionic acid (C3), butyric acid (C4), iso-butyric acid (iso-C4), valeric acid (Val), and iso-valeric acid (iso-Val) were analyzed on a gas chromatograph (Shimadzu Scientific Instruments Inc., Columbia, MD, USA) using 2-ethylbutyric acid as an internal standard [29]. Separation of VFAs was performed as previously described [30]. VFAs were expressed as mM/1 mL samples. A sample was taken from each of the jars and centrifuged at 12,000× g for 15 min at 4 °C to measure the ammonia N concentration (NH3-N) using a spectrophotometer (Perkin Elmer, Waltham, MA, USA) after being acidified with 0.5 mL of 0.1 N HCl, according to the method of Abdelrahman et al. [31]. The pH values of the culture jar sample were determined directly during sampling using a digital pH meter (Model pH 211; Hanna Instruments, Woonsocket, RI, USA). 2.8. Statistical Analysis For both studies, all data were analyzed as a completely randomized design using general linear model procedures of Statistical Analysis System software [32] based on animal and jar as the experimental unit. Only the use of the collected samples for chemical composition analysis of the diet precluded statistical comparison. Statistical differences (p < 0.05) between the means of the dietary treatments were determined using Duncan’s multiple range test and were also analyzed for linear or quadratic responses with orthogonal contrasts (T1 vs. T2 + T3 + T4 + T5). All means were reported with the standard error of the means (means ± SEM). 3. Results The chemical composition of the dietary treatments on a DM basis is shown in Table 1. DM was lower with an increasing proportion of sprouted barley at 25 to 100% with 75 to 0% traditional feed (T2 to T5) compared to 100% added traditional feeding system (T1; 70% barley grains + 30% alfalfa hay). However, a numerical increase in NDF, fat, phosphorus, zinc, copper, TDN, and NE concentrations of 2.69%, 1.01%, 0.16%, 37.0 ppm, 1.0 ppm, 3.40%, and 0.03 Mcal/lb, respectively, was observed in sprouted barley (T5) compared to the traditional feeding system (T1). CP, ADF, lignin, ash, and some minerals (calcium, magnesium, potassium, sulphur, and sodium) were numerically lower in the sprouted barley than in the traditional feeding system. The general growth performance of growing lambs fed on dietary treatments with sprouted barley is shown in Table 2. The results of the in vivo study show that the addition of sprouted barley at 25 to 100% with 75 to 0% traditional diet (T2 to T5; respectively) had the lowest (p < 0.05) final body weight (FBW) at 75 days of the study period compared to 100% added traditional feeding system (T1; 70% barley grains + 30% alfalfa hay). Bodyweight gain (BWG) and average daily gain (ADG) were lower at T5 (p < 0.05), while there were no significant effects at T2, T3 and T4 compared to T1. The comparison (T1 vs. T2 + T3 + T4 + T5) showed a linear effect (p < 0.05) in FBW, BWG, and ADG with increasing sprouted barley content. There was a linear and quadratic (p < 0.05) increase in average daily intake and feed to gain ratio with increasing sprouted barley content (T2 to T5) compared to T1 (1 to 75 days), while the feed-to-gain ratio was not significantly affected between T1, T2, and T3. The daily dry feed intake of growing lambs fed on dietary treatments with sprouted barley is shown in Table 3. The current results show that the daily dry feed intake of DM, OM, CP, crude fat, ash, nitrogen-free extract, fiber fractions (NDF and ADF), lignin, and mineral content (calcium, phosphorus, magnesium, sulfur, sodium) decreased linearly with increasing sprouted barley content (T2 to T5) from 1 to 75 days of the study period, compared to lambs fed with the traditional feeding system (T1) (p < 0.05). For apparent digestibility for growing lambs fed on dietary treatments with sprouted barley, different levels are shown in Table 4. The digestibility of DM, OM, CP, fat, ash, NFC, NDF, ADF, and minerals such as magnesium and sulfur were linearly increased (p < 0.05) with sprouted barley levels (T2 to T5) compared to the lambs receiving the traditional feeding system (T1). The addition of sprouted barley at 50 to 100% with 50 to 0% traditional feed (T3 to T5; respectively) had the highest (p < 0.05) digestibility of CP and fat compared to the lambs receiving T1 and T2 but were not affected either linearly or quadratically. Other items’ digestibility was not affected by sprouted barley levels (p > 0.05), either linearly or quadratically, at 75 days of the study period. Estimates of apparent digestibility for dietary treatments with sprouted barley at different levels in vitro are shown in Table 5. The results of the in vitro study showed that a diet of 100% sprouted barley (T5) had the quadratically lowest (p < 0.05) digestibility of DM, OM, and NDF compared to the traditional diet (T1) and other sprouted barley levels (T2, T3, and T4). The pH value, Ammonia-N (NH3-N), and gas production during the in vitro digestion and fermentation for dietary treatments with different levels of sprouted barley are shown in Table 6. The culture pH value was linearly decreased (p < 0.05) with sprouted barley levels (T2 to T5) compared to the traditional diet (T1). NH3-N was not affected by fermentation of dietary treatments with different levels of sprouted barley (p > 0.05), either linearly or quadratically. The results of the in vitro study show that gas production (2 to 24 h) by digestion and fermentation of a diet of T5 (100% sprouted barley) had the highest (p < 0.05) compared to other treatments but was not affected either linearly or quadratically (p > 0.05) (Table 6; Figure 1). VFAs during in vitro fermentation for dietary treatments with different levels of sprouted barley are shown in Table 7. The current results indicate that acetic acid (C2) concentration was higher (p < 0.05) in the diet of T2 (25% sprouted barley with 75% traditional diet) compared to other treatments, except for T3 (50% sprouted barley with 50% traditional diet). There was no significant difference and it was not affected either linearly or quadratically (p > 0.05) between treatments. Additionally, T2 was linearly higher (p < 0.05) in propionic acid (C3) and total VFA compared to T1 and T4 but there was no significant difference with other treatments (T3 and T5). On the other hand, diets of 75 and 100% sprouted barley with 25 and 0% traditional diet (T4 and T5; respectively) were lower in iso-butyric acid (Iso-C4) and iso-valeric acid (Iso-Val) compared to other treatments. Butyric acid (C4) and valeric acid (Val) were not affected by sprouted barley levels (p > 0.05), either linearly or quadratically. Acetic acid-to-propionic acid ratio (C2:C3) had, linearly, the lowest (p < 0.05) in dietary treatments with sprouted barley (T2 to T5) compared to traditional diet (T1). 4. Discussion The results of the chemical composition of dietary treatments showed that the content of sprouted barley DM, CP, ADF, lignin, ash, and some minerals decreased compared to traditional feeding, while NDF, fat, phosphorus, zinc, copper, TDN, and NE increased numerically. Previous studies have reported similar changes in the chemical composition of sprouted barley [2]. Fazaeli. et al. [33] and Girma and Gebremariam [7] reported that any reduction in starch content (53–67% of the dry weight of barley seeds) causes an equal reduction in OM, DM, and NFC in sprouted barley. This is in agreement with Al-Saadi and Al-Zubiadi [17], who showed that crude protein content was higher in sprouted barley compared to untreated grain, which could be due to the use of carbohydrates to provide energy to the seeds by respiration. NDF and ADF were also increased, but NFC decreased in sprouted barley compared to barley grain on a DM basis [7]. In the current study, used sprouted barley at 25–100% compared to 75–0% traditional diet had the lowest FBW, while BWG and ADG was lower only in lambs fed 100% sprouted barley. In addition, our results showed that the average daily intake and feed-to-gain ratio were increased with sprouted barley compared to lambs fed with the traditional diet (T1). These results agree with the reports of Muhammad et al. [34] that feeding sprouted barley with a traditional diet may have a negative effect on growth performance compared to a concentrate diet. This could be due to the lower daily dry feed intake which allows animals to become full without meeting the nutrient requirements for optimal performance. Morales et al. [35] reported that sprouted barley contains a lower DM, so animals may not be able to meet their dry feed intake requirements, and this may have a negative effect on growth. In addition, the lower DM in sprouted barley resulted in a lower OM value as it constitutes a major part of the DM. The OM, especially starch, may be consumed to support metabolism and energy requirements during sprout [36]. A study by Saidi and Jamal [4] showed that sprouted barley did not affect the bodyweight of ewes. However, Ata [1] found that lambs fed 62% sprouted barley as part of the total mixed ratio (TMR) had a higher total weight gain, average daily gain, average daily intake, and better feed-to-gain ratio than TMR (control diet). In another study, it was shown that goats fed sprouted barley had a higher total intake of DM and weight gain compared to concentrate diets [37]. Moreover, Fayed [3] found that feeding sprouted barley on rice straw had a positive effect on the growth performance of lambs. The current results show that the apparent digestibility of DM, OM, CP, fat, ash, NFC, NDF, ADF, and minerals such as Mg and sulfur increased linearly with the level of sprouted barley compared to lambs fed traditional diets may be due to the easy degradation of sprouted barley in the rumen. These results agree with those of Al-Saadi and Al-Zubiadi [17], who reported that digestibility of DM, OM, CP, and fat was higher in sprouted barley than cereal barley. Similarly, Morgan et.al. [36] indicated that the digestibility of OM and DM was higher with the addition of sprouted barely. Nutrient digestibility was increased in lambs fed 33, 66, and 100% sprouted barley compared to traditional feeding [38]. In contrast, Dung et al. [13] reported that sprouted barely did not affect nutrient digestibility in ruminants due to total loss DM, while the lowest rumen pH value and the highest VFAs. Fayed [3] showed that apparent DM, OM, CP, and fat digestibility were higher when sprouted barley was combined with rice straw in the diet of lambs. On the other hand, the estimation of apparent digestibility in vitro for the different feed treatments with sprouted barley showed that the diet with 100% sprouted barley had the lowest digestibility of DM, OM and NDF compared to the other diets. These results agree with Fazaeli et al. [6] who reported that nutrient digestibility of sprouted barley was lower than that of pure cereals in vitro through rumen fluid in glass syringes. The high nutrient digestibility in vivo could be due to the high content of leaves and roots, which are easily digested and hydrolyzed by the enzymes of the microflora in the rumen of the lambs, and enzymatic digestion in the lytic vacuoles of the plant cells, which may lead to differences in the digestibility of the feed in vivo and in vitro, which was digested as DM. Ikram et al. [39] showed that many biochemical alterations occur during germination, affecting digestibility due to enzymes that split up carbohydrates and proteins into basic compounds in barley seed. Our findings that sprouted barley lowers the pH of the culture in vitro. These results agree with the in vivo study of Al-Saadi and Al-Zubiadi [17] who found that lambs fed the sprouted barley had lower rumen fluid pH than grain barley. The low pH at different levels of sprouted barley could also indicate a change in the rumen ecosystem, such as microflora activity or production of VFAs due to the utilization of sprouted barley. However, the concentration of VFAs was higher in lambs fed fresh sprouted barley [40]. Our results are consistent with the fact that 25% sprouted barley with 75% traditional diet (T2) had higher propionic acid and total VFA concentration and high acetic acid concentration in the diet of T2 compared to T1 and T4, which may indicate that sprouted barley enhances carbohydrate fermentation in the rumen and absorption of VFAs [3,9]. In addition, our results suggest that the levels of VFAs during in vitro fermentation for the sprouted barley feed treatments are within the ranges found in continuous culture fermentation studies [41,42]. The sprouted barley forage treatments had numerically higher NFC values compared to the traditional forages, which could be due to the higher concentration of total VFA and propionic acid. The acetic acid concentration was higher in the diet of T2, which may be due to the high NDF content that contributed to higher acetate concentration compared to the traditional diet (T1). 5. Conclusions From the results of the current study, it can be concluded that the addition of sprouted barley with traditional diet improved the nutritional value, the digestibility of diet composition, and fermentation characteristics. On the other hand, the addition of sprouted barley has a negative effect on growth performance and lower daily dry feed intake. So, it may be a good strategy if further studies are conducted to increase nutrient requirements for optimal growth performance. Author Contributions H.H.A.-B. and I.A.A.: Conceptualization, Methodology, Formal analysis, Writing—original draft. W.H.S., A.N.A. and M.M.A.: Conceptualization, Methodology M.A.A.-B., I.A.A.: Data curation, Investigation, review and editing also project administration. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study complied with the King Saudi Arabia standards on animal use (KSU-SE-22-01) and were approved by the local animal care and welfare committee of King Saud University. Informed Consent Statement Not applicable. Data Availability Statement All data sets collected and analyzed during the current study are available from the corresponding author on fair request. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Cumulative CO2 gas production during the in vitro digestion for dietary treatments from 2 to 24 h. animals-12-01206-t001_Table 1 Table 1 Chemical composition of the diets fed growing lambs on a DM basis 1. Items 2 Treatments 3 T1 T2 T3 T4 T5 DM, % 95.5 76.7 57.8 38.9 20.1 CP, % 15.01 14.64 14.43 14.23 13.86 NDF, % 34.21 35.00 30.26 31.05 36.90 ADF, % 19.78 18.25 16.03 17.57 16.82 Lignin, % 5.78 5.41 4.96 4.12 2.10 NFC, % 43.86 43.60 48.77 48.34 43.50 Fat, % 1.68 1.81 1.90 2.01 2.69 Ash, % 5.24 4.95 4.64 4.37 3.05 Calcium, % 0.65 0.65 0.53 0.45 0.19 Phosphorus, % 0.23 0.25 0.27 0.30 0.39 Magnesium, % 0.18 0.19 0.19 0.18 0.14 Potassium, % 1.47 0.40 1.20 1.04 0.51 Sulfur, % 0.22 0.23 0.22 0.21 0.18 Sodium, % 0.20 0.19 0.20 0.18 0.15 Zinc, ppm 33.00 37.00 39.00 41.00 70.00 Copper, ppm 5.00 7.00 6.00 5.00 6.00 TDN 80.00 81.80 84.30 82.50 83.40 NE, Mcal/lb 0.84 0.86 0.88 0.86 0.87 1 The chemical composition analysis was performed in triplicate; 2 DM = Dry matter; CP = Crude protein; NDF = Neutral detergent fiber; ADF = Acid detergent fiber; NFC = Non-fibrous carbohydrates [NFC = 100% − (CP + total fiber + fat + ash)]; TDN = Total digestible nutrients [TDN = dig CP + dig fiber + dig NFC + (2.25 × dig fat)]; NE = Net energy. 3 Treatments, T1: 100% traditional diet (Barley 70: Alfalfa hay 30); T2: 75% traditional diet with 25% sprouted barley; T3: 50% traditional diet with 50% sprouted barley; T4: 25% traditional diet with 75% sprouted barley and T5: 100% sprouted barley. animals-12-01206-t002_Table 2 Table 2 Growth performance for growing lambs fed on dietary treatments with different levels of sprouted barley from 1 to 75 days. Items 2 Treatments 1 SEM 3 p-Value T1 T2 T3 T4 T5 Treat. Linear Quadratic IBW (kg) 27.8 27.6 27.6 27.8 28.10 0.91 0.986 0.903 0.597 FBW (kg) 41.0 a 38.2 b 38.1 b 38.2 b 31.0 c 1.88 0.041 0.048 0.407 BWG (kg) 13.18 a 10.60 a 10.50 a 10.35 a 2.90 b 1.46 0.009 0.014 0.185 ADG (g/d) 175.8 a 140.6 a 143.6 a 137.8 a 39.0 b 19.5 0.009 0.014 0.185 ADI (g/d) 1117 d 1338 c 1587 b 1874 a 1672 b 63.1 <0.0001 <0.0001 0.013 FI: WG (g:g) 7.2 c 10.2 c 11.8 bc 17.6 b 43.8 a 2.15 <0.0001 <0.0001 <0.0001 RG % 41.5 a 31.6 a 32.3 a 35.8 a 10.0 b 3.29 <0.0001 <0.0001 0.156 a–d Means values within rows for each item with clarification of the significant difference in the form of superscripts (p < 0.05). 1 Treatments, T1: 100 % traditional diet (Barley 70: Alfalfa hay 30); T2: 75% traditional diet with 25% sprouted barley; T3: 50% traditional diet with 50% sprouted barley; T4: 25% traditional diet with 75% sprouted barley and T5: 100% sprouted barley. 2 IBW = Initial body weight; FBW = Final body weight; BWG = Weight gain; ADG = Average daily gain; ADI = Average daily intake (g/d); FI: WG = Feed-to-gain ratio; RG = Relative growth. 3 SEM = Standard error of means for treatments effect. animals-12-01206-t003_Table 3 Table 3 Daily dry feed intake (g/day) for growing lambs fed on dietary treatments with different levels of sprouted barley from 1 to 75 days. Items 2 Treatments 1 SEM 3 p-Value T1 T2 T3 T4 T5 Treat. Linear Quadratic DM 1465 a 1314 a 953 b 777 b 232 c 97.7 <0.0001 <0.0001 0.112 OM 1388 a 1249 a 909 b 743 b 224 c 93.1 <0.0001 <0.0001 0.107 CP 219.8 a 192.4 a 137.5 b 110.6 b 32.2 c 14.1 <0.0001 <0.0001 0.175 NDF 501.1 a 460.1 a 288.4 b 241.4 b 85.6 c 31.4 <0.0001 <0.0001 0.404 ADF 289.7 a 239.9 b 152.8 c 136.6 c 39.0 d 16.5 <0.0001 <0.0001 0.713 Lignin 84.6 a 71.1 a 47.2 b 32.0 c 4.8 d 4.91 <0.0001 <0.0001 0.314 NFC 642.5 a 573.1 ab 464.8 bc 375.9 c 100.9 d 45.8 <0.0001 <0.0001 0.033 Fat 24.6 a 23.7 a 18.1 b 15.6 b 6.2 c 1.84 <0.0001 0.0008 0.056 Ash 76.7 a 65.0 a 44.2 b 33.9 b 7.0 c 4.61 <0.0001 <0.0001 0.263 Calcium 9.52 a 8.54 a 5.05 b 3.50 b 0.44 c 0.55 <0.0001 <0.0001 0.317 Phosphorus 3.36 a 3.28 a 2.57 ab 2.33 b 0.90 c 0.26 <0.0001 0.001 0.035 Magnesium 2.63 a 2.49 a 1.81 b 1.40 b 0.32 c 0.18 <0.0001 <0.0001 0.029 Sulfur 3.22 a 3.02 a 2.09 b 1.63 b 0.41 c 0.21 <0.0001 <0.0001 0.067 Sodium 2.92 a 2.49 a 1.90 b 1.40 b 0.34 c 0.19 <0.0001 <0.0001 0.107 a–d Means values within rows for each item with clarification of the significant difference in the form of superscripts (p < 0.05). 1 Treatments, T1: 100 % traditional diet (Barley 70: Alfalfa hay 30); T2: 75% traditional diet with 25% sprouted barley; T3: 50% traditional diet with 50% sprouted barley; T4: 25% traditional diet with 75% sprouted barley and T5: 100% sprouted barley. 2 DM = Dry matter; OM = Organic matter; CP = Crude protein; NDF = Neutral detergent fiber; ADF = Acid detergent fiber; NFC = Non-fibrous carbohydrates. 3 SEM = Standard error of means for treatments effect. animals-12-01206-t004_Table 4 Table 4 Apparent digestibility (%) for growing lambs fed on dietary treatments with different levels of sprouted barley at 75 days. Items 2 Treatments 1 SEM 3 p-Value T1 T2 T3 T4 T5 Treat. Linear Quadratic DM, % 81.80 b 81.59 b 87.33 a 88.15 a 90.89 a 1.69 0.004 0.015 0.882 OM, % 82.73 b 82.47 b 88.34 a 88.88 a 91.46 a 1.63 0.004 0.014 0.955 CP, % 72.41 b 69.61 b 79.19 a 83.46 a 86.77 a 3.46 0.015 0.077 0.601 NDF, % 65.31 c 67.51 bc 74.49 bc 78.40 ab 86.71 a 3.84 0.009 0.017 0.511 ADF, % 67.96 bc 66.38 c 73.03 bc 74.68 b 82.25 a 2.46 0.003 0.042 0.162 Lignin, % 72.40 74.42 80.17 77.29 73.59 3.33 0.494 0.302 0.112 NFC, % 99.54 98.79 99.26 98.51 99.06 0.63 0.812 0.388 0.560 Fat, % 65.21 c 61.04 c 73.92 b 81.05 ab 89.53 a 5.95 0.025 0.114 0.379 Ash, % 64.93 64.56 66.71 72.41 72.67 3.80 0.392 0.344 0.728 Calcium, % 40.58 42.03 43.86 59.74 36.23 7.57 0.279 0.572 0.246 Phosphorus, % 47.85 57.03 55.34 50.80 66.78 5.35 0.175 0.128 0.614 Magnesium, % 39.91 b 41.84 b 58.50 a 68.08 a 67.76 a 4.65 0.0008 0.002 0.527 Sulfur, % 65.83 b 65.90 b 76.15 ab 78.54 a 79.10 a 3.27 0.017 0.025 0.582 Sodium, % 72.97 75.32 82.08 83.65 75.68 5.22 0.547 0.303 0.186 a–c Means values within rows for each item with clarification of the significant difference in the form of superscripts (p < 0.05). 1 Treatments, T1: 100 % traditional diet (Barley 70: Alfalfa hay 30); T2: 75% traditional diet with 25% sprouted barley; T3: 50% traditional diet with 50% sprouted barley; T4: 25% traditional diet with 75% sprouted barley and T5: 100% sprouted barley. 2 DM = Dry matter; OM = Organic matter; CP = Crude protein; NDF = Neutral detergent fiber; ADF = Acid detergent fiber; NFC = Non-fibrous carbohydrates. 3 SEM = Standard error of means for treatments effect. animals-12-01206-t005_Table 5 Table 5 Apparent digestibility for dietary treatments with different levels of sprouted barley in vitro. Items 2 Treatments 1 SEM 3 p-Value T1 T2 T3 T4 T5 Treat. Linear Quadratic DM, % 79.4 a 79.9 a 82.7 a 82.5 a 72.6 b 1.16 0.0007 0.979 0.001 OM, % 80.0 a 80.4 a 82.8 a 82.8 a 74.3 b 0.98 0.0007 0.924 0.001 NDF, % 53.3 a 55.4 a 53.6 a 54.1 a 46.4 b 1.74 0.030 0.634 0.020 a,b Means values within rows for each item with clarification of the significant difference in the form of superscripts (p < 0.05). 1 Treatments, T1: 100 % traditional diet (Barley 70: Alfalfa hay 30); T2: 75% traditional diet with 25% sprouted barley; T3: 50% traditional diet with 50% sprouted barley; T4: 25% traditional diet with 75% sprouted barley and T5: 100% sprouted barley. 2 DM = dry matter; OM = Organic matter; NDF = Neutral detergent fiber. 3 SEM = Standard error of means for treatments effect. animals-12-01206-t006_Table 6 Table 6 The pH value, Ammonia-N and CO2 gas production during the in vitro digestion and fermentation for dietary treatments with different levels of sprouted barley. Items 2 Treatments 1 SEM 3 p-Value T1 T2 T3 T4 T5 Treat. Linear Quadratic Culture pH 6.13 a 6.02 b 6.02 b 6.01 b 5.92 b 0.03 0.010 0.002 0.669 NH3-N, mM 4.12 4.54 3.99 4.28 4.85 0.68 0.902 0.707 0.643 TGP, mL 179.3 b 176.1 b 185.4 b 176.7 b 196.0 a 3.13 0.005 0.255 0.108 a,b Means values within rows for each item with clarification of the significant difference in the form of superscripts (p < 0.05). 1 Treatments, T1: 100 % traditional diet (Barley 70: Alfalfa hay 30); T2: 75% traditional diet with 25% sprouted barley; T3: 50% traditional diet with 50% sprouted barley; T4: 25% traditional diet with 75% sprouted barley and T5: 100% sprouted barley. 2 NH3 = Ammonia-N; TGP = total CO2 gas production. 3 SEM = Standard error of means for treatments effect. animals-12-01206-t007_Table 7 Table 7 Volatile fatty acids (VFA) during the in vitro fermentation for dietary treatments with sprouted barley different levels. Items 2 Treatments 1 SEM 3 p-Value T1 T2 T3 T4 T5 Treat. Linear Quadratic C2, mM 30.5 b 33.8 a 32.2 ab 29.3 b 30.8 b 0.90 0.042 0.331 0.301 C3, mM 12.1 c 15.0 a 13.9 ab 12.7 bc 14.1 ab 0.53 0.019 0.011 0.289 Iso-C4, mM 0.46 a 0.49 a 0.45 ab 0.37 c 0.39 bc 0.02 0.010 0.170 0.472 C4, mM 9.3 10.5 10.1 9.1 10.5 0.64 0.423 0.294 0.918 Iso-Val, mM 0.91 a 0.90 a 0.86 a 0.74 b 0.82 ab 0.03 0.028 0.072 0.598 Val, mM 0.95 1.07 1.03 0.91 1.08 0.06 0.351 0.358 0.936 C2:C3 2.51 a 2.24 b 2.31 b 2.30 b 2.18 b 0.05 0.016 0.002 0.367 Total VFA, mM 54.2 b 61.9 a 58.6 ab 53.2 b 57.8 ab 1.93 0.050 0.123 0.409 a–c Means values within rows for each item with clarification of the significant difference in the form of superscripts (p < 0.05). 1 Treatments, T1: 100 % traditional diet (Barley 70: Alfalfa hay 30); T2: 75% traditional diet with 25% sprouted barley; T3: 50% traditional diet with 50% sprouted barley; T4: 25% traditional diet with 75% sprouted barley and T5: 100% sprouted barley. 2 C2 = acetic acid; C3 = propionic acid; C4 = butyric acid; Val = valeric acid; Total VFA = total volatile fatty acids. 3 SEM = Standard error of means for treatments effect. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Ata M. Effect of hydroponic barley fodder on Awassi lamb’s performance J. Biol. Agric. Healthc. 2016 6 60 64 2. Hafla A.N. Soder K.J. Brito A.F. Rubano M.D. Dell C.J. Effect of sprouted barley grain supplementation of an herbage-based or haylage-based diet on ruminal fermentation and methane output in continuous culture J. Dairy Sci. 2014 97 7856 7869 10.3168/jds.2014-8518 25262180 3. Fayed M. Comparative study and feed evaluation of sprouted barley grains on rice straw versus tamarix mannifera on performance of growing Barki lambs in Sinai J. Am. Sci. 2011 7 954 961 4. Saidi A.M.A. Jamal A.J. The biological and economical feasibility of feeding barley green fodder to lactating awassi ewes J. Anim. Sci. 2015 5 99 105 10.4236/ojas.2015.52012 5. Lemmens E. Moroni A.V. Pagand J. Heirbaut P. Ritala A. Karlen Y. Delcour J.A. Impact of cereal seed sprouting on its nutritional and technological properties: A critical review Compr. Rev. Food Sci. Food Saf. 2019 18 305 328 10.1111/1541-4337.12414 33337026 6. Fazaeli H. Golmohammadi H.A. Tabatabayee S.N. Asghari-Tabrizi M. Productivity and nutritive value of barley green fodder yield in hydroponic system World Appl. Sci. J. 2012 16 531 539 7. Girma F. Gebremariam B. Review on hydroponic feed value to livestock production J. Sci. Innov. Res. 2018 7 106 109 8. Shipard I. How Can I Grow and Use Sprouts as Living Food? Stewart Publishing Toronto, ON, Canada 2005 9. Farghaly M.M. Abdullah M.A. Youssef I.M. Abdel-Rahim I.R. Abouelezz K. Effect of feeding hydroponic barley sprouts to sheep on feed intake, nutrient digestibility, nitrogen retention, rumen fermentation and ruminal enzymes activity Livest. Sci. 2019 228 31 37 10.1016/j.livsci.2019.07.022 10. Cáceres P.J. Martínez-Villaluenga C. Amigo L. Frias J. Maximising the phytochemical content and antioxidant activity of Ecuadorian brown rice sprouts through optimal germination conditions Food Chem. 2014 152 407 414 10.1016/j.foodchem.2013.11.156 24444955 11. Cáceres P.J. Peñas E. Martinez-Villaluenga C. Amigo L. Frias J. Enhancement of biologically active compounds in germinated brown rice and the effect of sun-drying J. Cereal Sci. 2017 73 1 9 10.1016/j.jcs.2016.11.001 12. Rico D. Peñas E. García M.D.C. Martínez-Villaluenga C. Rai D.K. Birsan R.I. & Martín-Diana A.B. Sprouted barley flour as a nutritious and functional ingredient Foods 2020 9 296 10.3390/foods9030296 32150936 13. Dung D.D. Godwin I.R. Nolan J.V. Nutrient content and in sacco digestibility of barley grain and sprouted barley J. Anim. Vet. Adv. 2010 9 2485 2492 10.3923/javaa.2010.2485.2492 14. Sneath R. McIntosh F. Review of Hydroponic Fodder Production for Beef Cattle Department of Primary Industries Dalby, QLD, Australia 2003 Volume 84 1 54 15. Tudor G. Darcy T. Smith P. Shallcross F. The Intake and Live Weight Change of Drought Thomas, T.A., 1977. An automated procedure for the determination of soluble carbohydrate in herbage J. Sci. Food Agric. 2003 28 639 642 16. American Dairy Science Association American Society of Animal Science and Poultry Science Association Guide for the Care and Use of Agricultural Animals in Research and Teaching ADSA ASAS PSA American Dairy Science Association Champaign, IL, USA 2010 17. Al-Saadi M.J. Al-Zubiadi I.A. Effects of substitution barley by 10%, 30% of sprouted barley on rumen characters, digestibility and feed efficiency in diet of awassi male lambs Int. J. Sci. Res. 2016 5 2228 2233 18. AOAC Official Methods of Analysis 19th ed. Association of Official Analytical Chemists Rockville, MD, USA 2012 19. Van Soest P.J. Robertson J.B. Lewis B.A. Methods fordietary fiber, neutral detergent fiber and non-starch polysaccharides in relation to animal nutrition J. Dairy Sci. 1991 74 3583 3597 10.3168/jds.S0022-0302(91)78551-2 1660498 20. Haddad S.G. Younis H.M. The effect of adding ruminally protected fat in fattening diets on nutrient intake, digestibility and growth performance of Awassi lambs Anim. Feed Sci. Technol. 2004 113 61 69 10.1016/j.anifeedsci.2003.10.015 21. Bachmann M. Kuhnitzsch C. Okon P. Martens S.D. Greef J.M. Steinhöfel O. Zeyner A. Ruminal in vitro protein degradation and apparent digestibility of energy and nutrients in sheep fed native or ensiled+ toasted pea (Pisum sativum) grains Animals 2019 9 401 10.3390/ani9070401 31266144 22. Kaur K. Kaushal S. Rani R. Chemical composition, antioxidant and antifungal potential of clove (Syzygium aromaticum) essential oil, its major compound and its derivatives J. Essent. Oil-Bear. Plants 2019 22 1195 1217 10.1080/0972060X.2019.1688689 23. MEWA Small Ruminant Vaccine Protocol and Programs Veterinary Vaccine Centre Riyadh, Saudi Arabia 2015 24. Pereira T.L. Fernandes A.R.M. Oliveira E.R. Cônsolo N.R.B. Marques O.F.C. Maciel T.P. Gandra J.R. Serum metabolomic fingerprints of lambs fed chitosan and its association with performance and meat quality traits Animal 2020 14 1987 1998 10.1017/S1751731120000749 32290896 25. Goiri I. Zubiria I. Lavín J.L. Benhissi H. Atxaerandio R. Ruiz R. García-Rodríguez A. Evaluating the inclusion of cold-pressed rapeseed cake in the concentrate for dairy cows upon ruminal biohydrogenation process, ruminal microbial community and milk production and acceptability Animals 2021 11 2553 10.3390/ani11092553 34573519 26. Omar J.A. Effects of feeding different levels of sesame oil cake on performance and digestibility of Awassi lambs Small Rumin. Res. 2002 46 187 190 10.1016/S0921-4488(02)00173-6 27. Embaby M.G. Günal M. AbuGhazaleh A. Effect of unconventional oils on in vitro rumen methane production and fermentation Cienc. Investig. Agrar. 2019 46 276 285 10.7764/rcia.v46i3.2062 28. Johnson M.C. Devine A.A. Ellis J.C. Grunden A.M. Fellner V. Effects of antibiotics and oil on microbial profiles and fermentation in mixed cultures of ruminal microorganisms J. Dairy Sci. 2009 92 4467 4480 10.3168/jds.2008-1841 19700708 29. Kenichi I. Yumeto F. Preparation of fatty acid methyl esters for gas-liquid chromatography J. Lipid Res. 2010 51 635 640 19759389 30. Cui K. Qi M. Wang S. Diao Q. Zhang N. Dietary energy and protein levels influenced the growth performance, ruminal morphology and fermentation and microbial diversity of lambs Sci. Rep. 2019 9 16612 10.1038/s41598-019-53279-y 31719633 31. Abdelrahman M.M. Alhidary I. Alyemni A.H. Khan R.U. Bello A.R.S. Al-Saiady M.Y. Amran R.A. Effect of alfalfa hay on rumen fermentation patterns and serum biochemical profile of growing Naemi lambs with ad libitum access to total mixed rations Pak. J. Zool. 2017 49 1519 1522 10.17582/journal.pjz/2017.49.4.sc6 32. SAS Institute SAS Users Guide: Statistics SAS Institute Inc. Cary, NC, USA 2008 33. Fazaeli H. Golmohammadi H.A. Shoayee A.A. Montajebi N. Mosharraf S.H. Performance of feedlot calves fed hydroponics fodder barley J. Agric. Sci. Technol. 2011 13 365 375 34. Muhammad S. Afzal H. Mudassar S. Use of sprouted grains in the diets of poultry and ruminants, Pakistan Indian Res. J. 2013 2 20 27 35. Morales M.A. Juárez M. Ávila E. Effect of substituting hydroponic green barley forage for a commercial feed on performance of growing rabbits World Rabbit Sci. 2009 17 35 38 10.4995/wrs.2009.668 36. Raeisi Z. Tahmasbi R. Dayani O. Ayatollahi Mehrgardi A. Tavassolian I. Digestibility, microbial protein synthesis, rumen and blood parameters in sheep fed diets containing hydroponic barley fodder J. Lives Sci. Technol. 2018 6 9 17 37. Gebremedhin W. Nutritional benefit and economic value of feeding hydroponically grown maize and barley fodder for Konkan Kanyal goats J. Agric. Vet. Sci. 2015 8 24 30 38. Moghaddam A.S. Mehdipour M. Dastar B. The determining of digestible energy and digestibility coefficients of protein, calcium and phosphorus of malt (Germinated Barley) in broilers Int. J. Poult. Sci. 2009 8 788 791 10.3923/ijps.2009.788.791 39. Ikram A. Saeed F. Afzaal M. Imran A. Niaz B. Tufail T. Anjum F.M. Nutritional and end-use perspectives of sprouted grains: A comprehensive review Food Sci. Nutr. 2021 9 4617 4628 10.1002/fsn3.2408 34401108 40. Dung D.D. Godwin I.R. Nolan J.V. Digestive characteristics, rumen ammonia nitrogen and volatile fatty acids levels in sheep fed commercial pellets supplemented with grimmett barley grain or freeze-dried or fresh barley sprouts J. S. Pac. Agric. 2012 16 1 10 41. Bargo F. Varga G.A. Muller L.D. Kolver E.S. Pasture intake and substitution rate effects on nutrient digestion and nitrogen metabolism during continuous culture fermentation J. Dairy Sci. 2003 86 1330 1340 10.3168/jds.S0022-0302(03)73718-7 12741559 42. Soder K.J. Brito A.F. Rubanom M.D. Effect of supplementing orchard grass herbage with a total mixed ration or flaxseed on fermentation profile and bacterial protein synthesis in continuous culture J. Dairy Sci. 2013 96 3228 3237 10.3168/jds.2012-6307 23522677

PMC009xxxxxx/PMC9099673.txt

==== Front Molecules Molecules molecules Molecules 1420-3049 MDPI 10.3390/molecules27092911 molecules-27-02911 Article Comparison of Extraction Techniques for the Determination of Volatile Organic Compounds in Liverwort Samples https://orcid.org/0000-0003-2772-6197 Guzowska Małgorzata * Wasiak Wiesław https://orcid.org/0000-0001-9765-8287 Wawrzyniak Rafał Carrillo Daniel Academic Editor Tabanca Nurhayat Academic Editor Faculty of Chemistry, Adam Mickiewicz University in Poznań, Uniwersytetu Poznańskiego 8, 61-614 Poznań, Poland; wasiakw@amu.edu.pl (W.W.); rafwawrz@amu.edu.pl (R.W.) * Correspondence: malguz@amu.edu.pl; Tel.: +48-61-829-1713 03 5 2022 5 2022 27 9 291116 4 2022 01 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). This article focuses on the comparison of four popular techniques for the extraction of volatile organic compounds (VOCs) from liverworts of the Calypogeia azurea species. Since extraction is the most important step in the sample analysis of ingredients present in botanical preparations, their strengths, and weaknesses are discussed. In order to determine the VOCs present in plants, selecting the appropriate one is a key step of the extraction technique. Extraction should ensure the isolation of all components present in the oily bodies of Calypogeia azurea without the formation of any artifacts during treatment. The best extraction method should yield the determined compounds in detectable amounts. Hydrodistillation (HD), applying Deryng apparatus and solid-liquid extraction (SLE), microwave-assisted extraction (MAE), and headspace solid-phase microextraction (HS-SPME) were used for volatile extraction. The extracts obtained were analysed by gas chromatography coupled to mass spectrometry (GC-MS) to determine the compounds. Hepaticae liverworts specialized metabolites volatile organic compounds extraction Calypogeia azurea hydrodistillation SLE MAE SPME National Science Centre2017/27/B/NZ8/01091 European Union through the European Social Fund under the Operational Program Knowledge Education DevelopmentPOWR.03.02.00-00-I020/17 This research was funded by the National Science Centre, grant number 2017/27/B/NZ8/01091. The work was partly supported by grant no. POWR.03.02.00-00-I020/17, co-financed by the European Union through the European Social Fund under the Operational Program Knowledge Education Development. ==== Body pmc1. Introduction Plants synthesize and secrete many different volatile organic compounds (VOCs). Biologically active substances that are produced by plants are phytochemicals; however, they are not ubiquitous and are the product of specialized metabolism, narrowed down to specific families and species of plants. A feature that distinguishes plant material is the wealth of compounds that are very diverse in terms of physicochemical properties that occur in a very wide range of concentrations. The sample preparation step is important for the reliability of the results of the determined compounds. The extraction of biologically active compounds depends on many factors, including extraction methods, the type of raw material, and extraction solvent. The selectivity of extraction, the number of analytes in the sample, and the ecological aspects are also of key importance [1]. Specialized metabolites may have a wide variety of polarities, solubility, volatility, thermal stability, and the presence of different functional groups. They are often found in low concentrations in the raw material; therefore, there is no universal and simple method to isolate them. Obtaining VOCs is a difficult process, as they make up a small fraction of the plant’s raw material. Various extraction methods are used. The most common method of extracting essential oil from a plant is hydrodistillation (HD). Most often, it is carried out with a glass apparatus of the Clevenger or Deryng type. Both devices are recommended pharmacopoeia devices for determining the content of essential oils. The first is described by the European Pharmacopoeia (VI) and the latter according to the Polish Pharmacopoeia (VII) [2]. The efficiency of HD is a fairly complex issue. The general disadvantage of distillation methods is that it is difficult to quantitatively determine the essential oil of small amounts of plants, as the yields are typically low [3]. In the publications that have appeared thus far, various methods of extracting VOCs present in liverworts have been used [4,5,6]. The most common extraction methods for liverworts were headspace solid-phase microextraction (HS-SPME) and solid-liquid extraction (SLE) with solvents of different polarities [6]. In this study, SLE was performed with the use of three different solvents: n-hexane, diethyl ether, and methylene chloride. A Likens–Nickerson distillation was also performed for the extraction of VOCs, in which n-pentane was used to collect the extract. Another method of VOC extraction used was HD in a Clevenger apparatus. In this method, n-hexane was used to collect an extract. High temperature, exposure to air, and hot water or hot steam during long distillation times break down many of the unstable aromatic terpenoids. Due to the small plant samples that can usually be collected, these methods are not used very often. Therefore, comparing the ingredients obtained by the HD process as well as those from solvent extracts is essential to recognize whether they are natural products or compounds formed during the extraction [6]. The percentage of essential oil in the plant substance is often not high, which requires additional work to increase it or select other extraction methods, for example, microwave-assisted extraction (MAE) [7]. HS-SPME [5,7,8,9,10,11,12] and conventional methods, such as SLE, can also be used to extract VOCs [9,10,11]. Modern extraction methods are certainly more efficient and environmentally friendly as very little or no solvents are used [13,14,15]. Polar solvents, such as methanol, ethanol, or ethyl acetate, are used to extract hydrophilic compounds. For the extraction of more lipophilic compounds, dichloromethane or a 1:1 mixture of dichloromethane/methanol is used. In some cases, hexane extraction is used to remove chlorophyll [16,17]. Appropriate extraction measures must be taken to ensure that potentially specialized metabolites are not lost, distorted, or destroyed during the preparation of a plant sample extract [16,18]. The widely used SLE technique is versatile and relatively simple. MAE uses more drastic conditions than the widely used SLE method, which in turn is relatively time-consuming. However, extracts from HD, SLE, and MAE can be stored at low temperatures for a long time and analysed as many times as necessary without quality changes. SPME requires sample preparation indirectly before gas chromatography (GC) analysis. Thus, it does not allow for the extract to be stored. However, it should not be forgotten, that the SPME technique also has many important advantages, such as simplicity, sensitivity, and a remarkably short extraction time. This technique is commonly used to isolate VOCs for qualitative purposes, such as monitoring the changes in volatile substances during storage, volatile substances profile, or metabolomic purposes [19,20,21,22,23]. Liverworts are widespread throughout the world, but most often in the tropics. Liverworts are the ancestors of all land plants and abundantly produce specialized metabolites, including monoterpenes, sesquiterpenes, monoterpenoids, sesquiterpenoids, and aromatic compounds, many of which exhibit notable biological activities, such as the inhibitory effects on allergic contact dermatitis, cytotoxicity, antibacterial, and antifungal activity, anti-insect activity, and antioxidant properties [24,25,26]. On the basis of these statements, it can be concluded that it is very important to develop the best method for the extraction of these compounds from plants. In their natural environment, you can find deciduous and thalli liverworts. A leafy liverwort typically has leaves of two sizes, arranged in three rows along the stem. The larger leaves (called lateral leaves) grow in two rows, along opposite sides of the stem. Most leafy liverworts are prostrate plants and grow along some substrate (e.g., soil, bark, leaves, or rock). The coplanar arrangement of the lateral leaves gives leafy liverworts a “flat” look that is rare in mosses. Thallose liverworts, which are branching and ribbon-shaped, grow commonly in moist soil or damp rocks, whereas leafy liverworts are found in similar habitats as well as tree trunks in damp woods. The thallus (body) of thallose liverworts resembles a lobed liver, hence, the common name liverwort (‘liver plant’). Plants are not economically important to humans, but provide food for animals, facilitate the decay of logs, and aid in the disintegration of rocks by their ability to retain moisture. Most liverworts contain oil bodies, which are intracellular organelles surrounded by a single membrane in which a wide variety of specialized metabolites are synthesized and accumulated, such as terpenes, terpenoids, sesquiterpenes, sesquiterpenoids, and aromatics [15,16,17,18,19,20]. The presence of nitrogen, sulfur, or both nitrogen and sulfur compounds in liverworts is very rare. The most characteristic chemical phenomenon of liverworts is that most sesqui- and diterpenoids are enantiomers of those found in higher plants [6]. Furthermore, the determination of the composition of secondary metabolites that fall under the scope of chemotaxonomy can be one of the methods to help identify taxonomically difficult species [27]. In this study, various techniques were used to extract volatile components from Calypogeia azurea. The purpose of the research is to compare the most frequently used methods of extracting volatile organic compounds from liverwort cells. The results allow the provision of information on which extraction technique is most appropriate for this species of liverwort. The collected information may be useful in further research into other liverworts from the Calypogeia genus. 2. Results and Discussion Twenty-two samples of Calypogeia azurea from Poland (Table 1) were analysed for volatile specialized metabolites in the study. Supplementary Tables S1–S7 show the percentage of compounds detected present in liverwort cells. A total of 73 compounds were detected, 42 of which were identified. Depending on the extraction method used, the content of the identified VOCs differed. The study compares four extraction methods: Three (preparative) methods, which used different types of solvents with different polarities (HD, SLE, and MAE), and the non-preparative SPME method. The most common isolation method is SLE, and the extraction efficiency and activity are highly dependent on the type of solvent used. The polarity of the extraction solvent strongly influences the compounds present in the test sample. Therefore, extraction solvents are selected as they are critical to the complex sample matrix. The extraction solvent system is generally selected according to the purpose of extraction, the polarity of the components concerned, the polarity of the undesirable components, the total cost, safety, and environmental concerns [28]. Essential oils, which were prepared with different solvent methods, took different colours. The colour of the oil produced by the HD was slightly blue to purple. The compounds 1,4-dimethylazulene (53) (the bold numbers in the brackets refer to the compounds in Tables S1–S7) were responsible for these colours. Colourless solutions were formed during the maceration from n-hexane and turned dark green on extraction with methanol. Methanol is a very good extractant for chlorophylls, especially for resistant vascular plants and algae [29]. In the case of MAE, the extract obtained took the form of a dark brown liquid, regardless of the solvent used. The main component of the Calypogeia azurea liverwort is 1,4-dimethylazulene (53). This VOC belongs to sesquiterpene hydrocarbons. The relative content of this compound during HD ranged from 16.64% (for n-hexane) to 19.49% (for m-xylene), depending on the solvent used to collect the extract. For this sesquiterpene, the best extraction method is SLE for 24 h n-hexane as the solvent. The relative content of 1,4-dimethylazulene (53) for this extraction method is 59.62%. MAE obtained only 1.37–0.21% of this sesquiterpene (Figure 1). The sum of sesquiterpene hydrocarbons produced is the highest for this research material and ranges from 13.73% to 78.07% of the total essential oil content, depending on the extraction method used. However, for the SPME method, the relative content of 1,4-dimethylazulene (53) is 42.67% of all detected VOCs present in liverwort cells. Another sesquiterpene compound, which is also a characteristic chemical compound for this species of liverwort, is anastreptene (18). The relative content of this compound ranges from 5.98% for MAE using methanol as a solvent to 15.74% for HD using m-xylene to collect the extract. When this sesquiterpene is detected using the SPME method, the content in liverwort cells is 6.92%. 2.1. Comparison of Solvent Extractions The highest percentage of specialized metabolites could be extracted using the method SLE using n-hexane as the solvent for 24 h (99.51%). MAE, using ethyl acetate as the solvent, was the lowest (55.62%). The solvent method would seem to be the best method used, however, with this method, the amount of solvents used and the extraction time are not favorable. Furthermore, for the HD method and for SLE, the amounts of extracted VOCs were calculated (Table 2). The table shows that the best solvent for the SLE method is methanol, which can yield 0.33 mg/kg of VOCs from 1 g of the sample. In the SLE method, 1 g of liverwort was used and extracted with 10 cm3 of solvent. Unfortunately, concentrating the extracted extract to the same volume is a difficult task due to the possible loss of VOCs during the evaporation process. Table 2 shows that the most optimal extraction time for SLE is 48 h, during which the greatest number of specialized metabolites can be isolated, with the exception of n-hexane and methanol. Although, the VOC content does not differ much between the two solvents when extracted within 24 h and 48 h. In the course of the research, the percentage concentrations of isolated VOCs were calculated for the individual extraction methods used in the research. M-xylene and n-hexane were used during HD to collect the extract. It was found that there were no major differences in the HD between the two solvents. The percentage of sesquiterpenes ranged from 50.95% for n-hexane to 55.25% for m-xylene in the HD, and for aromatic compounds from 22.47% to 24.99%. Because of the lack of significant differences in the percentages of individual groups of VOCs, it seems that it is better to use n-hexane during HD because the oil obtained from plant material can be used for further research, e.g., for the isolation of individual compounds with preparative gas chromatography. 2.2. Comparison of HS-SPME with Solvent Methods HS-SPME is a fairly fast technique to determine VOCs in complex matrices, which are undoubtedly specialized metabolites produced by the oil bodies of Calypogeia azurea. Using this technique, it was possible to detect 98.17% of the compounds, including 85.84% identified. As shown in Table 3, a richer composition of VOCs can be obtained using this method (SPME). Compared with the solvent methods, the SPME method can detect a percentage of compounds that are lower, but the number of compounds detected is greater than that of SLE or HD. For comparison, 66 of the 73 compounds could be detected using HS-SPME. Although only 11 of the 73 compounds were extracted with the SLE method, this may suggest that this method is not a good method for isolating VOCs from Calypogeia azurea. In the case of HD, 53 compounds were detected for n-hexane (HD1) as a solvent, and in the case of m-xylene (HD2), it was 52 out of the 73 compounds that could be isolated with HS-SPME. During the analyzing sample of extracts resulting from the HD and SLE extraction methods, VOCs appeared, which could not be determined by SPME. Examples of such relationships are ledene (40), 4,5,9,10-dehydro-isolongifolene (51), (+)–spathulenol (54). It is sometimes possible to detect compounds that cannot be extracted by solvent methods using HS-SPME. Such compounds include α-pinene (5), β-pinene (6), and limonene (7). The MAE1-MAE4 appears to be the least effective method of all the others. Using this method, it is possible to extract 56.71% of the total VOCs for the methanol solvent, and up to 70.09% for the diethyl ether solvent. Too drastic extraction conditions resulted in the decomposition of some organic compounds, while others, e.g., RI = 1710 (64), were extracted in the greatest amount compared to HD, SLE, and MAE. In the case of this method, the conditions were too drastic (exposure to microwave radiation, elevated temperature, and solvent) for a plant with such delicate cell walls. 2.3. Statistical Analysis 2.3.1. Comparison of the Extraction Methods with the Solventless Method Table 3 presents the results of the comparisons using Student’s t-tests for paired samples and Wilcoxon tests, the objective of which was to capture the significance of the differences between the solventless method and individual extraction methods using various solvents. The results obtained from the difference tests, presented in Table 3, show that there are slight differences between the solvent extraction methods compared to the solvent-free method (SPME). However, with Cohen’s d indices, which define the magnitude of the observed difference between the mean and the tested sample, it was found that in the SLE method there were effects representing a small (d > 0.20) or average difference (d > 0.50). In each of the samples in the SLE method, a slightly higher level of volatile compounds was observed compared to that of the method without solvent. On the other hand, no possible differences were observed in the HD and MAE methods (d < 0.20). The differential effects obtained, despite the lack of statistical significance, indicated a potentially higher level of volatile compounds in the case of using solvents with the SLE method compared to the control method without the use of a solvent. In Table 4, Table 5, Table 6 and Table 7, analogues of the Student’s t-tests and Wilcoxon tests were performed with the division into individual volatile compounds. However, because of the occurrence of single measurements in the fields of aliphatic, monoterpene, and monoterpenoid, these were excluded from the analyses. In turn, for sesquiterpenoid (Table 4) and the aromatic compounds (Table 5), only some comparisons were made due to the lack of data in the remaining conditions. When analyzing the effects obtained in the case of sesquiterpene compounds (Table 4), biased effects were found for the SLE under conditions of SLE5-2 and SLE5-3 compared to SPME (p = 0.080), indicating a higher level of sesquiterpene compounds in the extraction of the SLE method under the above conditions. No more biased or statistically significant differences were found; however, possible differences were observed in the sample using Cohen’s d index. The mean level of sesquiterpene compounds did not differ at all in the SPME method compared to HD and SLE under the conditions of SLE4-1, SLE4-2, and SLE4-3 (d < 0.20). On the other hand, the other conditions for the SLE possibly showed higher levels of sesquiterpene compounds compared to the solvent-free condition (d > 0.20). Furthermore, the fourth method possibly showed a higher severity of sesquiterpene compounds under the MAE2 condition compared to SPME, while under other conditions (MAE1, MAE3, MAE4) the severity of sesquiterpene compounds was slightly lower compared to SPME (d > 0.20). In the case of aromatic compounds (Table 5), no statistically significant differences were found under these comparable conditions. However, it was observed that only in the MAE method, Cohen d values greater than 0.20 were obtained, indicating that there was no possible difference between the means. On the other hand, in the case of HD and SLE, medium-sized difference effects (d > 0.50) were found under the tested conditions, which showed a possibly higher level of aromatic compounds compared to the solvent-free method. As in aromatics, the samples tested for sesquiterpenoid compounds did not show statistically significant differences in the SPME condition (Table 6). However, the differences between the measurements were found to be at least weak in each case (d > 0.20). Observing the averages, it was found that a higher level of volatiles is possible in the SLE and MAE for the conditions HD1, HD2, and MAE2 compared to the solventless method. For the MAE conditions, SLE1-2 and SLE1-3 showed lower levels of sesquiterpenoid compounds compared to the SPME condition, while the conditions SLE3-1, SLE3-2, and SLE3-3 showed potentially higher levels of sesquiterpenoid compared to the solvent-free condition. As in the case of general results, no statistically significant differences were found between the individual extraction methods compared to the solvent-free method in the sample of unidentified compounds (Table 7). However, a biased effect of the difference between SPME and SLE4-1 (p = 0.068) was found, indicating a higher level of unidentified compounds in the SPME trial. Furthermore, it was noticed that only in the case of the HD method, Cohen’s d index allowed us to find no difference between the solvent-free method and the HD method (d < 0.20). For the SLE and MAE methods, at least a weak difference effect was observed between the SLE1-1 and MAE4 measurements compared to the SPME method (d > 0.20). This may mean that unidentified compounds may show a lower intensity level with the third and fourth extraction methods. 2.3.2. Differences between Solvents in Hydrodistillation In Figure 2, the mean levels of volatile compounds are presented in the case of using HD. Due to the lack of data, no calculations were performed to compare the compounds of aliphatic, monoterpene, and monoterpenoid using the Student’s t-test for the dependent samples and the Wilcoxon test, which was aimed at confirming the effect obtained. There were no statistically significant differences between the solvents in HD, t (51) = 0.07; p = 0.944; d = 0.01. Furthermore, the lack of differences was confirmed by the rank test (p = 0.722). This means that, regardless of the solvent used, the HD produced the same effect. Based on the analysis with the Wilcoxon test, no significant differences were found in terms of compounds not identified (p = 0.398) and sesquiterpene (p = 0.163). However, the bias effects of the difference were confirmed for aromatic (p = 0.066) and sesquiterpenoid (p = 0.068). It turned out that the level of aromatic compounds was slightly higher with the HD2 method, while the level of sesquiterpenoid was higher with the HD1 method in both cases. The effect size index of the difference between the means indicated a difference in the mean value difference (d > 0.50). 2.3.3. Differences between Solvents in Extraction Using Solvents Table 8 presents the results of the analysis of differences in the scope of the SLE method. The analysis had a two-stage character using the Friedman ANOVA test due to the small size of the groups in the measurements. As it turned out, the analysis of the differences between the groups of solvents in the method with solvents allowed us to find no differences in all the time intervals. This means that regardless of the solvent used, the intensity of the volatile compounds obtained was similar. In the case of differences within the solvent groups, it turned out that there was a border difference effect in the case of the use of methanol (p = 0.050), indicating a higher intensity of volatile compounds in the case of the measurement of 24 h than in the case of 48 h and 72 h. 2.3.4. Differences between Solvents in Microwave-Assisted Extraction To verify the differences in the fourth method, an analysis of variance was performed in conjunction with an auxiliary Friedman ANOVA test, the results of which are shown in Figure 3. The analyses performed with the parametric test did not show statistically significant differences in the mean intensity of VOCs, F (3.33) = 1.53; p = 0.224; η2 = 0.12. However, the nonparametric analysis, which ignored errors in the measurement of the means, showed a statistically significant effect for the differences between individual solvents in the MAE method, F = 20.01; df = 3; p < 0.001. The intensity of VOCs was significantly higher under the MAE1 condition than under the MAE3 (p = 0.007) and MAE4 (p = 0.001) conditions. Furthermore, a higher level of VOCs was observed under condition MAE2 compared to condition MAE4 (p = 0.034). However, no differences were found between the conditions MAE2 and MAE3 (p = 0.197), MAE1 and MAE2 (p = 1.000) and MAE3 and MAE4 (p = 1.000). 3. Materials and Methods 3.1. Plant Material The plant material of the Calypogeia azurea was collected in 2021 at Szklarska Poręba, latitude, 50°47′52.9″ N; longitude, 15°31′41.8″ E, and the altitude ranged from 700–1200 m ASL. The storage and transfer of plant material were carried out in airtight plastic containers. The collection temperature was 10–12 °C (ambient temperature) and the transfer temperature was 15–16 °C; the pressure was approximately 1013 Mpa (ambient pressure). Only green plants that did not show signs of drying were eligible for collection and further research. In natural habitats, liverwort samples are initially identified on the basis of their morphological structure. The research was carried out on fresh material. 3.2. Material and Reagents The following solvents were used during the research: n-hexane, puriss. p.a., ≥99% (GC), ethyl acetate, puriss. p.a., 99.9% (GC), methanol for GC, Sigma-Aldrich (Steinheim, Germany) and m-xylene, diethyl ether, puriss. ≥99.9% (GC), methylene chloride >99.9%, POCH (Gliwice, Poland). Saturated n-alkanes of C7-C40 standard Supelco (Bellefonte, PA, USA) were used to determine the Kovats retention indices. Fused silica fibers coated with divinylbenzene/carboxy/polydimethylsiloxane (DVB/CAR/PDMS) (Supelco, Bellefonte, PA, USA) stationary phases were used for the SPME analysis. Trace 1310 (Thermo Scientific, Waltham, MA, USA) coupled with a mass spectrometer ISQ QD (Thermo Scientific, Waltham, MA, USA) with a 007-5MS column (30 m, 0.25 mm, 0.25 μm) (Quadrex, Woodbridge, CT, USA) were used to analyze the VOC compounds present in the cells of the Calypogeia azurea species. The TriPlus RSH (Thermo Scientific, Waltham, MA, USA) automatic sample injector was used to ensure that the samples were dispensed with sufficient reproducibility. HD was carried out using a Deryng apparatus consisting of a 500 mL round-bottom flask, a condenser, and a heating bowl (Lab-szkło, Kraków, Poland), recommended by the VI edition of the Polish Pharmacopoeia of 2002. The Ethos one (Milestone, Sorisole, Italy) was used for MAE. 3.3. Methods The experimental conditions for the various extraction methods are shown in Table 9. The specific extraction conditions and methods used in this study are outlined below. 3.3.1. Extraction by Using Headspace Solid-Phase Microextraction The conditions of sorption and desorption were optimized by selecting the type of stationary phase coated fibers, the amount of biological material, the time, and the temperature. A fresh amount of 5 mg of Calypogeia azurea was placed in a screw-capped vial with a 1.7 cm3 silicone/Teflon membrane. The vial was then heated at 50 °C and solid-phase microextraction of the headspace was carried out for 60 min. Desorption was performed at 250 °C for 10 min. 3.3.2. Extraction by Using Solvents An amount of 1 g of plant material was weighed and crushed with an agate mortar and pestle. They were placed in glass bottles and 10 cm3 of solvents were added according to the increasing polarity: n-hexane, diethyl ether, methylene chloride, ethyl acetate, and methanol, and were allowed to macerate for 24 h, 48 h, and 72 h. After this time, the solvent was filtered and injected into a GC-MS. 3.3.3. Extraction by Using Microwave-Assisted Extraction An amount of 5 g of fresh plant material was weighed, placed in Teflon bombs and 50 cm3 of diethyl ether was added. The entire process was carried out with the Ethos One microwave-assisted extraction system in 3 steps: ramp time of 10 min to reach 20 °C, a hold time of 20 min at 70 °C, and cooling for 10 min. The final step in the preparation of the sample for analysis was the quantitative transfer of the samples to a 1.7 cm3 screw cap vial. 3.3.4. Hydrodistillation Extraction in the Deryng Apparatus An amount of 5 g of fresh plant material was weighed and placed into a 500 cm3 round-bottom flask; we then added 250 cm3 of distilled water and 1 cm3 of solvent. For HD, two solvents, n-hexane and m-xylene, were used to collect the extract. The sample flask was heated for 3 h after reaching the boiling point. The vapors were condensed by means of a cold refrigerant. After 180 min of extraction to n-hexane, the essential oil was transferred to vials and kept at 5 °C until gas chromatography-mass spectrometry analyses were performed. HD in the Deryng apparatus was carried out according to the Polish Pharmacopoeia VI [30]. 3.4. GC-MS Analysis The analysis of the composition of the compounds present in the extracts was performed by GC-MS. For liquid samples, the injection volume was 1 µL. The sample was injected in split mode (1:25). Samples analyzed with the SPME technique were injected in splitless mode. The injector temperature in both cases was 250 °C. Helium was used as the carrier gas at a flow rate of 1.0 mL/min. The oven temperature was programmed from 60 to 230 °C at 4 °C/min and then kept isothermal at 230 °C for 40 min. The ISQ QD mass detector was operated at 70 eV in the EI mode in the m/z range 30–550; transfer line, 250 °C. The constituents were identified by comparing their MS spectra with those of the literature, reference compounds, computer matching with the NIST 11, and data obtained from the NIST Chemistry WebBook databases, the Mass Finder 4 library, the Adams library databases, and the Pherobase databases [31,32]. The identification of the compounds was verified by Kovats’ retention indices. The Kovats retention indices were determined relative to a homologous series of n-alkanes (C7–C40) under the same operating conditions. The quantitative data of the components were obtained by integrating the TIC chromatogram and calculating the relative percentage of the peak areas. Each sample was analysed three times. 3.5. Statistical Analysis The results obtained from three separate tests were averaged and expressed as a mean ± standard deviation. In order to verify the differences between the extraction methods with respect to the specificity of volatile compounds, statistical analyses were performed using IBM SPSS Statistics 27 software. The statistical methods used included the Student’s t-test for dependent samples and its nonparametric equivalent Wilcoxon test, as well as an analysis of variance for multiple measurements with its nonparametric equivalent Friedman’s ANOVA. Parametric and nonparametric tests were used in parallel as a result of the varying sample sizes of the extraction methods. A threshold of α = 0.05 was used as the significance level. 4. Conclusions The presented studies are the first to concern (VOCs) formed in the oily bodies of Calypogeia azurea. These studies demonstrated the advantages and optimal use of commonly used extraction techniques. Based on the research carried out, it is possible to find that, despite many studies on the effectiveness of individual techniques for the extraction of metabolites from plant material, you cannot find undoubtedly certain and universal information on which of the techniques available today are the most effective in practice. There is no clear information on the scope of the literature on the applications of the described sample/batch preparation procedures for plant material that allows one to maximize the attainable test concentrations of metabolites. Therefore, it is considered advisable to conduct research on the effectiveness of extraction techniques for the most commonly used in order to isolate metabolites from the plant material. The article compares four methods to extract volatile organic compounds present in the oily bodies of Calypogeia azurea liverwort. On the basis of the conducted experiments, it has been established that one of the best methods of analysis for the determination of VOCs found in the species Calypogeia azurea is SPME. Unfortunately, this method can only be used to determine the qualitative composition. SPME extraction allows the identification of low-boiling compounds that co-elute with solvents used in other methods. The main advantage, in addition to simplicity, speed, and low cost, is its “green” character. This technique does not require any toxic solvents, which is especially important nowadays. The SPME method is a method in which the amount of research material is small, which is especially important when analyzing samples, the acquisition of which is quite difficult. On the contrary, if more research is needed using the essential oils obtained, HD is the best extraction method. With a relatively short extraction time, small amounts of solvents are used. Small amounts of samples used to perform the extraction are beneficial for samples that can be difficult to obtain. During HD, the Deryng apparatus is used, the costs of which are not high, and the extraction costs are low. The disadvantage of this process is undoubtedly the possibility of artifacts forming during heating. The SLE method gives the greatest relative amount, and one group of compounds, the sesquiterpenes, is a self-absorbing method that uses quite large amounts of solvents. The extracts obtained in this way are diluted too much to give a reliable result when analyzed using GC-MS. GC-MS analysis allowed for the identification of 43 components, which, depending on the extraction method used, constituted 31.64% to 97.02% of the obtained product. The MAE method was too drastic and resulted in the creation of large amounts of artifacts. Although quick and simple, this extraction technique is too drastic for delicate plants, such as liverworts. It seems justified to further develop the HD process to obtain essential oils from liverworts. Furthermore, HD is indeed the primary technique and SPME is a complementary method for this type of sample. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/molecules27092911/s1, Table S1: Volatile compounds detected in the samples analysed by SPME and HD; Table S2: Volatile compounds detected in the samples extracted by n-hexane; Table S3: Volatile compounds detected in the samples extracted by diethyl ether; Table S4: Volatile compounds detected in the samples extracted by methylene chloride; Table S5: Volatile compounds detected in the samples extracted by ethyl acetate; Table S6: Volatile compounds detected in the samples extracted by methanol; Table S7: Volatile compounds detected in the samples analysed by MAE. Click here for additional data file. Author Contributions Conceptualization, M.G. and R.W.; methodology, M.G. and R.W.; validation, M.G.; formal analysis, M.G. and R.W.; investigation, M.G., R.W. and W.W.; resources, M.G. and R.W.; data curation, M.G. and R.W.; writing—original draft preparation, M.G. and R.W.; writing—review and editing, M.G. and R.W.; visualization, M.G. and R.W.; supervision, W.W.; project administration, R.W.; funding acquisition, M.G., R.W. and W.W. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Percentage of —1,4-dimethyl azulene (53) and —anastreptene (18), depending on the extraction method. Figure 2 The average levels of volatile compounds using the hydrodistillation —HD1 —HD2. Figure 3 Analysis of variance with the auxiliary Friedman ANOVA versus solvents in Method 4 —mean —average rank. molecules-27-02911-t001_Table 1 Table 1 The liverworts sampling data used for studies. No. Sample Code Extraction Method 1 SPME HS-SPME 2 HD1 Hydrodistillation with n-hexane 3 HD2 Hydrodistillation with m-xylene 4 SLE1-1 Maceration with n-hexane (24 h) 5 SLE1-2 Maceration with n-hexane (48 h) 6 SLE1-3 Maceration with n-hexane (72 h) 7 SLE2-1 Maceration with diethyl ether (24 h) 8 SLE2-2 Maceration with diethyl ether (48 h) 9 SLE2-3 Maceration with diethyl ether (72 h) 10 SLE3-1 Maceration with methylene chloride (24 h) 11 SLE3-2 Maceration with methylene chloride (48 h) 12 SLE3-3 Maceration with methylene chloride (72 h) 13 SLE4-1 Maceration with ethyl acetate (24 h) 14 SLE4-2 Maceration with ethyl acetate (48 h) 15 SLE4-3 Maceration with ethyl acetate (72 h) 16 SLE5-1 Maceration with methanol (24 h) 17 SLE5-2 Maceration with methanol (48 h) 18 SLE5-3 Maceration with methanol (72 h) 19 MAE1 Extraction assisted by microwave radiation with diethyl ether 20 MAE2 Extraction assisted by microwave radiation with methylene chloride 21 MAE3 Extraction assisted by microwave radiation with ethyl acetate 22 MAE4 Extraction assisted by microwave radiation with methanol molecules-27-02911-t002_Table 2 Table 2 The content of VOCs (mg/kg) depending on the extraction method. HD SLE n-Hexane SLE Diethyl Ether SLE Methylene Chloride SLE Ethyl Acetate SLE Methanol 24 h 48 h 72 h 24 h 48 h 72 h 24 h 48 h 72 h 24 h 48 h 72 h 24 h 48 h 72 h 0.15 0.19 0.23 0.24 0.13 0.12 0.09 0.11 0.10 0.09 0.22 0.20 0.20 0.23 0.32 0.33 molecules-27-02911-t003_Table 3 Table 3 Analysis of the differences between the method without solvent and the individual extraction methods among all volatile compounds. Sample Code Method with Solvents Method without Solvent (SPME) t p d Cohen p Wilcoxon Mean SD Mean SD HD1 1.98 4.88 2.06 6.51 0.11 0.909 0.02 0.852 HD2 1.93 5.09 1.97 6.38 0.06 0.953 0.01 0.987 SLE1-1 9.02 17.90 6.87 12.42 −1.25 0.238 0.38 0.534 SLE1-2 6.10 14.10 4.88 10.60 −1.18 0.258 0.29 0.660 SLE1-3 4.97 12.16 4.14 9.83 −1.08 0.295 0.25 0.446 SLE2-1 13.02 17.72 10.52 14.64 −1.29 0.243 0.49 0.310 SLE2-2 13.53 16.84 10.52 14.64 −1.69 0.143 0.64 0.128 SLE2-3 14.02 15.41 10.52 14.64 −1.68 0.145 0.63 0.128 SLE3-1 12.17 18.48 8.78 14.27 −1.76 0.122 0.62 0.123 SLE3-2 12.22 17.40 8.78 14.27 −1.89 0.100 0.67 0.093 SLE3-3 11.98 15.95 8.78 14.27 −1.93 0.094 0.68 0.123 SLE4-1 8.10 14.57 6.50 11.91 −0.81 0.433 0.23 0.875 SLE4-2 7.05 12.62 5.62 11.18 −0.86 0.406 0.23 0.950 SLE4-3 6.78 11.86 5.62 11.18 −0.83 0.419 0.22 0.875 SLE5-1 10.24 14.63 8.32 13.41 −1.34 0.217 0.45 0.260 SLE5-2 9.65 14.47 7.49 12.91 −1.54 0.158 0.49 0.139 SLE5-3 9.67 14.51 7.49 12.91 −1.52 0.164 0.48 0.139 MAE1 4.02 7.46 3.91 9.99 −0.04 0.972 0.01 0.723 MAE2 3.82 6.34 3.67 10.02 −0.05 0.962 0.01 1.000 MAE3 3.24 6.74 3.78 10.01 0.17 0.864 0.04 0.287 MAE4 3.54 6.64 3.89 10.33 0.11 0.915 0.03 0.605 t—Student’s t-test result. p—significance level. d Cohen—Cohen’s average deviation. p Wilcoxon—Wilcoxon’s average deviation. molecules-27-02911-t004_Table 4 Table 4 Analysis of the differences between the solventless method and individual extraction methods among sesquiterpene compounds. Sample Code Method with Solvents Method without Solvent (SPME) t p d Cohen p Wilcoxon Mean SD Mean SD HD1 2.62 5.46 2.06 6.51 0.56 0.585 0.13 0.433 HD2 2.86 5.99 1.97 6.38 0.42 0.677 0.10 0.520 SLE1-1 12.43 23.25 6.87 12.42 −1.13 0.310 0.46 0.753 SLE1-2 11.82 20.94 4.88 10.60 −1.17 0.296 0.48 0.917 SLE1-3 9.69 17.96 4.14 9.83 −1.07 0.324 0.41 0.866 SLE2-1 20.46 23.15 10.52 14.64 −1.75 0.223 1.01 0.285 SLE2-2 21.81 20.88 10.52 14.64 −2.91 0.101 1.68 0.109 SLE2-3 25.25 15.89 10.52 14.64 −2.25 0.153 1.30 0.109 SLE3-1 24.74 26.37 8.78 14.27 −2.46 0.133 1.42 0.109 SLE3-2 24.81 23.54 8.78 14.27 −3.12 0.089 1.80 0.109 SLE3-3 23.49 20.42 8.78 14.27 −2.75 0.111 1.59 0.109 SLE4-1 13.71 17.77 6.50 11.91 −0.30 0.782 0.15 0.465 SLE4-2 13.87 15.36 5.62 11.18 −0.18 0.867 0.09 0.465 SLE4-3 14.23 15.24 5.62 11.18 −0.30 0.787 0.15 0.465 SLE5-1 18.12 18.94 8.32 13.41 −1.83 0.164 0.92 0.144 SLE5-2 15.34 18.49 7.49 12.91 −2.00 0.117 0.89 0.080 SLE5-3 15.40 18.54 7.49 12.91 −1.95 0.123 0.87 0.080 MAE1 2.78 3.92 3.91 9.99 0.98 0.371 0.40 0.600 MAE2 3.88 5.04 3.67 10.02 0.77 0.482 0.35 0.500 MAE3 2.70 3.29 3.78 10.01 0.97 0.386 0.43 0.345 MAE4 2.72 3.62 3.89 10.33 0.97 0.386 0.43 0.345 t—Student’s t-test result. p—significance level. d Cohen—Cohen’s average deviation. p Wilcoxon—Wilcoxon’s average deviation. molecules-27-02911-t005_Table 5 Table 5 Analysis of the differences between the solventless method and the individual extraction methods among aromatic compounds. Sample Code Method with Solvents Method without Solvent (SPME) t p d Cohen p Wilcoxon Mean SD Mean SD HD1 5.62 11.14 3.75 7.43 −1.01 0.387 0.50 0.465 HD2 6.00 11.80 3.75 7.43 −1.03 0.379 0.52 0.285 SLE1-3 11.62 16.38 7.48 10.48 −0.99 0.502 0.70 0.655 SLE4-1 18.89 26.58 7.45 10.52 −1.01 0.498 0.71 0.180 SLE4-2 17.89 25.14 7.45 10.52 −1.01 0.497 0.71 0.180 SLE4-3 15.50 21.67 7.45 10.52 −1.02 0.493 0.72 0.180 MAE3 0.09 0.01 0.10 0.13 0.13 0.921 0.09 0.655 t—Student’s t-test result. p—significance level. d Cohen—Cohen’s average deviation. p Wilcoxon—Wilcoxon’s average deviation. molecules-27-02911-t006_Table 6 Table 6 Analysis of the differences between the solventless method and the individual extraction methods among sesquiterpenoid compounds. Sample Code Method with Solvents Method without Solvent (SPME) t p d Cohen p Wilcoxon Mean SD Mean SD HD1 2.19 3.61 0.26 0.11 −1.05 0.373 0.52 0.273 HD2 1.39 2.03 0.26 0.11 −1.07 0.364 0.53 0.273 SLE1-2 0.20 0.17 0.26 0.13 0.52 0.655 0.30 1.000 SLE1-3 0.18 0.14 0.26 0.13 0.68 0.565 0.39 0.593 SLE3-1 1.48 1.29 0.23 0.16 −1.21 0.439 0.86 0.180 SLE3-2 1.40 1.48 0.23 0.16 −1.01 0.497 0.71 0.180 SLE3-3 1.41 1.55 0.23 0.16 −0.98 0.508 0.69 0.655 MAE2 9.54 13.04 0.23 0.16 −1.00 0.501 0.71 0.655 t—Student’s t-test result. p—significance level. d Cohen—Cohen’s average deviation. p Wilcoxon—Wilcoxon’s average deviation. molecules-27-02911-t007_Table 7 Table 7 Analysis of the differences between the solvent-free method and individual extraction methods among unidentified compounds. Sample Code Method with Solvents Method without Solvent (SPME) t p d Cohen p Wilcoxon Mean SD Mean SD HD1 0.55 1.15 3.49 9.49 0.19 0.855 0.04 0.983 HD2 0.46 0.77 3.49 9.49 0.53 0.603 0.12 0.872 SLE1-1 1.25 0.53 9.16 16.25 0.35 0.787 0.25 0.655 SLE1-2 0.55 0.57 9.16 16.25 0.92 0.409 0.41 0.225 SLE1-3 0.48 0.44 7.92 15.19 1.02 0.353 0.42 0.345 SLE2-1 0.76 0.08 17.73 20.96 1.77 0.327 1.25 0.180 SLE2-2 0.87 0.39 17.73 20.96 2.09 0.284 1.48 0.180 SLE2-3 0.73 0.59 17.73 20.96 2.67 0.228 1.89 0.180 SLE3-1 0.85 0.04 16.49 22.22 1.65 0.347 1.17 0.180 SLE3-2 0.90 0.07 16.49 22.22 1.64 0.349 1.16 0.180 SLE3-3 0.91 0.08 16.49 22.22 1.64 0.349 1.16 0.180 SLE4-1 0.74 0.52 13.40 19.18 2.26 0.109 1.13 0.068 SLE4-2 0.64 0.43 13.40 19.18 1.68 0.153 0.69 0.116 SLE4-3 0.57 0.41 13.40 19.18 2.01 0.101 0.82 0.116 SLE5-1 0.59 0.36 13.46 19.13 1.45 0.283 0.84 0.109 SLE5-2 0.74 0.39 10.77 17.62 1.40 0.297 0.81 0.109 SLE5-3 0.74 0.39 10.77 17.62 1.40 0.297 0.81 0.109 MAE1 4.47 9.14 9.67 15.96 −0.96 0.373 0.36 0.398 MAE2 3.74 6.63 10.73 17.61 −1.04 0.339 0.39 0.612 MAE3 3.29 8.34 10.85 17.56 −0.63 0.555 0.24 0.398 MAE4 3.13 7.35 10.91 17.52 −0.80 0.448 0.28 0.889 t—Student’s t-test result. p—significance level. d Cohen—Cohen’s average deviation. p Wilcoxon—Wilcoxon’s average deviation. molecules-27-02911-t008_Table 8 Table 8 Analysis of differences in the scope of the SLE method. Time 24 h 48 h 72 h Intragroup Tests Solvents M Me SD M Me SD M Me SD n-hexane 8.29 0.66 17.25 5.77 0.35 13.72 4.74 0.29 11.88 F = 0.19; df = 2; p = 0.909 diethyl ether 11.44 3.92 17.01 11.87 5.05 16.29 12.29 8.07 15.08 F = 2.67; df = 2; p = 0.875 methylene chloride 12.17 5.61 18.48 12.22 6.23 17.40 11.98 6.62 15.95 F = 3.68; df = 2; p = 0.159 ethyl acetate 7.52 1.12 14.11 6.62 0.73 12.27 6.38 0.85 11.53 F = 3.22; df = 2; p = 0.199 methanol 9.24 1.38 14.15 8.78 1.11 14.02 8.80 1.11 14.06 F = 6.00; df = 2; p = 0.050 tests between groups F = 2.20; df = 4; p = 0.699 F = 3.35; df = 4; p = 0.500 F = 2.72; df = 4; p = 0.606 M—mean. Me—median. F—the result of the analysis of variance. df—degrees of freedom. p—significance level. molecules-27-02911-t009_Table 9 Table 9 Experimental conditions used during the extraction of the Calypogia azurea. SPME HD SLE MAE the amount of plant material [g] 0.005 5 5 5 solvents not applicable water/n-hexane water/m-xylene n-hexane diethyl ether ethyl acetate, dichloromethane ethanol methanol diethyl ether temperature 50/250 °C 100 °C room temperatures 20–70 °C time 60 min of sorption 10 min desorption 3 h 24 h 48 h 72 h 40 min volume of solvent required not applicable 250 cm3 H2O/0.5 cm3 organic solvent 50 cm3 50 cm3 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Tiwari B.K. Ultrasound: A Clean, Green Extraction Technology TrAC-Trends Anal. Chem. 2015 71 100 109 10.1016/j.trac.2015.04.013 2. Baj T. Sieniawska E. Kowalski R. Wesołowski M. Ulewicz-Magulska B. Effectiveness of the Deryng and Clevenger-Type Apparatus Acta Pol. Pharm.-Drug Res. 2015 72 507 515 0001-6837 3. Ormeño E. Goldstein A. Niinemets Ü. Extracting and Trapping Biogenic Volatile Organic Compounds Stored in Plant Species TrAC-Trends Anal. Chem. 2011 30 978 989 10.1016/j.trac.2011.04.006 4. Asakawa Y. The Isolation, Structure Elucidation, and Bio-and Total Synthesis of Bis-Bibenzyls, from Liverworts and Their Biological Activity Nat. Prod. Commun. 2017 12 1335 1349 10.1177/1934578X1701200845 5. Komala I. Ito T. Yagi Y. Nagashima F. Asakawa Y. Volatile Components of Selected Liverworts, and Cytotoxic, Radical Scavenging and Antimicrobial Activities of Their Crude Extracts Nat. Prod. Commun. 2010 5 1375 1380 10.1177/1934578X1000500908 20922994 6. Asakawa Y. Ludwiczuk A. Bryophytes: Liverworts, Mosses, and Hornworts: Extraction and Isolation Procedures Methods Mol. Biol. 2013 1055 1 20 10.1007/978-1-62703-577-4_1 23963899 7. Chemat S. Lagha A. AitAmar H. Bartels P. Chemat F. Comparison of Conventional and Ultrasound-Assissted Extraction of Carvone and Limonene from Caraway Seeds Flavour Fragr. J. 2004 19 188 195 10.1002/ffj.1339 8. Hallingbäck T. Hodgetts N.G. IUCN/SSC Bryophyte Specialist Group IUCN-The World Conservation Union Sveriges lantbruksuniversitet ArtDatabanken Mosses, Liverworts, and Hornworts: Status Survey and Conservation Action Plan for Bryophytes IUCN in collaboration with the Swedish Threatened Species Unit Gland, Switzerland Cambridge, UK 2000 2831704669 9. Garrido A. Ledezma J.G. Durant-Archibold A.A. Allen N.S. Villarreal A.J.C. Gupta M.P. Chemical Profiling of Volatile Components of the Gametophyte and Sporophyte Stages of the Hornwort Leiosporoceros Dussii (Leiosporocerotaceae) From Panama by HS-SPME-GC-MS Nat. Prod. Commun. 2019 11 1 4 10.1177/1934578X19868875 10. Sakurai K. Tomiyama K. Kawakami Y. Yaguchi Y. Asakawa Y. Characteristic Scent from the Tahitian Liverwort, Cyathodium foetidissimum J. Oleo Sci. 2018 67 1265 1269 10.5650/jos.ess18160 30305559 11. Ghani N.A. Ludwiczuk A. Ismail N.H. Asakawa Y. Volatile Components of the Stressed Liverwort Conocephalum Conicum Nat. Prod. Commun. 2016 11 103 104 10.1177/1934578X1601100130 26996032 12. Rowan D.D. Volatile Metabolites Metabolites 2011 1 41 63 10.3390/metabo1010041 24957243 13. Jha A.K. Sit N. Extraction of Bioactive Compounds from Plant Materials Using Combination of Various Novel Methods: A Review Trends Food Sci. Technol. 2022 119 579 591 10.1016/j.tifs.2021.11.019 14. Madani A. Mazouni N. Nedjhioui M. Plants’ Bioactive Metabolites and Extraction Methods Intechopen Limited London, UK 2021 Volume 3 10.5772/intechopen.96698 15. Huie C.W. A Review of Modern Sample-Preparation Techniques for the Extraction and Analysis of Medicinal Plants Anal. Bioanal. Chem. 2002 373 23 30 10.1007/s00216-002-1265-3 12012169 16. Sasidharan S. Chen Y. Saravanan D. Sundram K.M. Yoga Latha L. Extraction, Isolation and Characterization of Bioactive Compounds from Plants’ Extracts Afr. J. Tradit. Complement. Altern. Med. 2011 8 1 10 10.4314/ajtcam.v8i1.60483 17. Cos P. Vlietinck A.J. Berghe D.V. Maes L. Anti-Infective Potential of Natural Products: How to Develop a Stronger in Vitro “Proof-of-Concept” J. Ethnopharmacol. 2006 106 290 302 10.1016/j.jep.2006.04.003 16698208 18. Fabricant D.S. Farnsworth N.F. The Value of Plants Used in Traditional Medicine for Drug Discovery Environ. Health Perspect. 2001 109 69 75 10.1289/ehp.01109s169 11250806 19. Majcher M. Jeleń H.H. Comparison of Suitability of SPME, SAFE and SDE Methods for Isolation of Flavor Compounds from Extruded Potato Snacks J. Food Compos. Anal. 2009 22 606 612 10.1016/j.jfca.2008.11.006 20. Routray W. Orsat V. Microwave-Assisted Extraction of Flavonoids: A Review Food Bioprocess Technol. 2012 5 409 424 10.1007/s11947-011-0573-z 21. Sajid M. Khaled Nazal M. Rutkowska M. Szczepańska N. Namieśnik J. Płotka-Wasylka J. Solid Phase Microextraction: Apparatus, Sorbent Materials, and Application Crit. Rev. Anal. Chem. 2019 49 271 288 10.1080/10408347.2018.1517035 30582727 22. Vas G. Vékey K. Solid-Phase Microextraction: A Powerful Sample Preparation Tool Prior to Mass Spectrometric Analysis J. Mass Spectrom. 2004 39 233 254 10.1002/jms.606 15039931 23. Cong-Cong X.U. Wang B. Yi-Qiong P.U. Jian-Sheng T. Tong Z. Advances in Extraction and Analysis of Phenolic Compounds from Plant Materials Chin. J. Nat. Med. 2017 15 721 731 29103457 24. Nagashima F. Asakawa Y. Terpenoids and Bibenzyls from Three Argentine Liverworts Molecules 2011 16 10471 10478 10.3390/molecules161210471 22179426 25. Linde J. Combrinck S. van Vuuren S. van Rooy J. Ludwiczuk A. Mokgalaka N. Volatile Constituents and Antimicrobial Activities of Nine South African Liverwort Species Phytochem. Lett. 2016 16 61 69 10.1016/j.phytol.2016.03.003 26. Veljić M. Ćirić A. Soković M. Janaćković P. Marin P.D. Antibacterial and Antifungal Activity of the Liverwort (Ptilidium pulcherrimum) Methanol Extract Arch. Biol. Sci. 2010 62 381 386 10.2298/ABS1002381V 27. Asakawa Y. Ludwiczuk A. Nagashima F. Phytochemical and Biological Studies of Bryophytes Phytochemistry 2013 91 52 80 10.1016/j.phytochem.2012.04.012 22652242 28. Tan M.C. Tan C.P. Ho C.W. Effects of Extraction Solvent System, Time and Temperature on Total Phenolic Content of Henna (Lawsonia Inermis) Stems Int. Food Res. J. 2013 20 3117 3123 29. Ritchie R.J. Consistent Sets of Spectrophotometric Chlorophyll Equations for Acetone, Methanol and Ethanol Solvents Photosynth. Res. 2006 89 27 41 10.1007/s11120-006-9065-9 16763878 30. Wesołowska A. Grzeszczuk M. Jadczak D. Comparison of Chemical Compositions of Essential Oils Isolated by Hydrodistillation from Wild Thyme (Thymus Serpyllum L.) with Use of Deryng and Clevenger Apparatus Herba Pol. 2014 60 7 17 10.2478/hepo-2014-0006 31. Wawrzyniak R. Wasiak W. Jasiewicz B. Bączkiewicz A. Buczkowska K. Chemical Fingerprinting of Cryptic Species and Genetic Lineages of Aneura Pinguis (L.) Dumort. (Marchantiophyta, Metzgeriidae) Molecules 2021 26 1180 10.3390/molecules26041180 33671838 32. Wawrzyniak R. Wasiak W. Bączkiewicz A. Buczkowska K. Volatile Compounds in Cryptic Species of the Aneura pinguis Complex and Aneura Maxima (Marchantiophyta, Metzgeriidae) Phytochemistry 2014 105 115 122 10.1016/j.phytochem.2014.06.010 25034615

PMC009xxxxxx/PMC9099674.txt

==== Front Nutrients Nutrients nutrients Nutrients 2072-6643 MDPI 10.3390/nu14091937 nutrients-14-01937 Article Effects of Moringa oleifera Lam. Supplementation on Inflammatory and Cardiometabolic Markers in Subjects with Prediabetes https://orcid.org/0000-0002-0923-495X Díaz-Prieto Ligia E. 1 https://orcid.org/0000-0002-3281-0118 Gómez-Martínez Sonia 1 Vicente-Castro Iván 1 Heredia Carlos 2 González-Romero Elena A. 2 Martín-Ridaura María del Carmen 3 Ceinos Mercedes 3 https://orcid.org/0000-0001-7057-7962 Picón María J. 4 Marcos Ascensión 1 https://orcid.org/0000-0002-1995-2955 Nova Esther 1* Ramdath Dan Academic Editor 1 Immunonutrition Research Group, Department of Metabolism and Nutrition, Institute of Food Science and Technology and Nutrition (ICTAN)—CSIC, 28040 Madrid, Spain; ldiaz@ictan.csic.es (L.E.D.-P.); sgomez@ictan.csic.es (S.G.-M.); i.vicente@ictan.csic.es (I.V.-C.); amarcos@ictan.csic.es (A.M.) 2 Cea Bermúdez Primary Health Care Centre, Madrid Health Service, 28003 Madrid, Spain; cheredia@salud.madrid.org (C.H.); elenaaurora.gonzalez@salud.madrid.org (E.A.G.-R.) 3 Madrid-Health, Madrid City Hall, 28007 Madrid, Spain; martinrmc@madrid.es (M.d.C.M.-R.); ceinosamm@madrid.es (M.C.) 4 Virgen de la Victoria University Hospital, 29010 Málaga, Spain; mjpiconcesar@gmail.com * Correspondence: enova@ictan.csic.es; Tel.: +34-913-938-021 (ext. 436322) 05 5 2022 5 2022 14 9 193708 4 2022 02 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Different parts of the Moringa oleifera Lam. (MO) tree are consumed as food or food supplements for their nutritional and medicinal value; however, very few human studies have been published on the topic. The current work was aimed to provide ancillary analysis to the antidiabetic effects previously reported in a double-blind, randomized, placebo-controlled, parallel group intervention conducted in patients with prediabetes. Thus, the effect of MO leaves on blood and fecal inflammatory markers, serum lipid profile, plasma antioxidant capacity and blood pressure was studied in participants who consumed 6 × 400 mg capsule/day of MO dry leaf powder (MO, n = 31) or placebo (PLC, n = 34) over 12 weeks. Differences between groups were assessed using each biomarker’s change score with, adjustment for fat status and the baseline value. In addition, a decision tree analysis was performed to find individual characteristics influencing the glycemic response to MO supplementation. No differences in the biomarker’s change scores were found between the groups; however, the decision tree analysis revealed that plasma TNF-α was a significant predictor of the subject’s HbA1c response (improvement YES/NO; 77% correct classification) in the MO group. In conclusion, TNF-α seems to be a key factor to identify potential respondents to MO leaf powder. Moringa oleifera Lam. food supplement prediabetes glycemic control inflammatory markers serum lipid profile blood pressure total antioxidant capacity fecal calprotectin MCIN/AEI/10.13039/501100011033/AGL2017-86044-C2-1-R “ERDF A way of making Europe”AGL2017-86044-C2-1-R This article is part of the R+D+i Project AGL2017-86044-C2-1-R, funded by MCIN/AEI/10.13039/501100011033/ and “ERDF A way of making Europe”. ==== Body pmc1. Introduction Moringa oleifera Lam. (MO) is a tree originally from Asia, grown in most tropical and subtropical areas and with cultivars recently introduced on the Mediterranean coast. MO leaves are nutritionally rich, as they contain high protein levels and abundant fiber, potassium, calcium, magnesium, β-carotene, α-tocopherol and polyphenols [1,2]. MO leaves as well as seeds and pods have been traditionally used as food or food supplements, with medicinal properties including antihypertensive, diuretic, antihyperlipidemic, antispasmodic, antiulcer, hepatoprotective, antidiabetic, antineoplastic, anti-inflammatory, antibacterial and antifungal activities [3,4,5]. Some of these bioactivities have been proven for a variety of leaf compounds, such as peptide fractions [6], isolated polysaccharides [7] and the isothiocyanate (ITC) derivatives of characteristic glucosinolates [8]. In this sense, moringin resulting from myrosinase-hydrolysis of glucomoringin under neutral conditions has been shown to exhibit effective antioxidant, anti-inflammatory and antitumor activities [9]. Other important phytochemicals in MO leaves are flavonoids and phenolic acids [10]. Regarding the antidiabetic effect of MO, a fairly large number of studies have been performed in animal models of hyperglycemia. Most of them show significant improvements in blood glucose, both fasting and in response to a glucose tolerance test [11]. Thus, the activity of MO as a natural antihypoglycemic agent and its potential application for diabetes prevention and treatment has gained considerable interest, since it is regarded as affordable and less prone to induce side effects than other pharmacological treatments [12]. However, only nine clinical trials in humans have been published [13,14,15,16,17,18,19,20,21]. The postprandial studies have shown significant or marginally [13] significant results both in patients with type 2 diabetes mellitus (DM) and healthy subjects [13,15,16], and all longitudinal studies except one [20] reported benefits on fasting blood glucose [17,19] or glycated hemoglobin [18] or both [21]. Prediabetes is characterized by above-normal values of glycemia, although lower than those used for DM diagnosis. Progression of this metabolic alteration is independently associated with abdominal obesity indicators as observed in a 4-year follow-up study [22]. In obesity, excessive adiposity, increased adipose tissue lipolysis, ectopic accumulation of circulating fatty acids in insulin sensitive tissues, insulin resistance and inflammation characterized by augmented production of inflammatory cytokines by macrophages infiltrating the adipose tissue and excessive reactive oxygen species are all related as etiology factors involved in DM development [23,24]. On this basis, plants and herbs such as MO, with the capacity to modify the transcriptional regulation of enzymes involved in lipid and glucose metabolism, can potentially influence the biomarkers associated with metabolic and cardiovascular health and inflammation in at-risk individuals. In this respect, MO has shown prevention of histopathological changes in the liver of diabetes-induced animals, with reduced lipogenic and increased lipolytic gene expressions in this organ and also a significant hypolipidemic effect [11,25,26]. Potent antioxidant and immunomodulatory actions, including an inhibitory effect on proinflammatory mediators such as inducible nitric oxide synthase (iNOS), cyclooxygenase (COX)-2, prostaglandin E (PGE)-2, tumor necrosis factor (TNF)-α and interleukin (IL)-1β and IL-6 have been evidenced in in vitro [27,28] and in vivo experiments [25,29,30]. Moringin is the most abundant MO isothiocyanate, with proven antioxidant and anti-inflammatory properties [31,32,33]. However, other compounds found in specific fractions that have shown inhibition of pro-inflammatory mediator production have been pointed out as potential bioactive molecules. In these sense, seven compounds were identified by liquid chromatography/mass spectrometry analysis of an ethyl acetate extract fraction of MO leaves [28]. Antihypertensive activity has also been attributed to MO extracts or peptide fractions [34,35]. Human studies on hypolipidemic effects [17,36] and the anti-inflammatory and antihypertensive activity [20,37] provide limited evidence due to the few number of published studies, their small sample size and the heterogeneity of the study designs. The effects observed in human studies are highly dependent on the dose and duration of the study and the basal condition of the study participants. Our group recently published a clinical trial showing the beneficial effects of MO on the glycemic control of prediabetic patients [21]; as an ancillary purpose, the potential modifications in inflammatory markers, lipid profile, plasma antioxidant capacity and blood pressure of the studied participants were also assessed. The corresponding results are presented here. In addition, the influence of the baseline value of percentage body fat as well as inflammatory and cardiometabolic markers on the control of glycemia by MO supplementation was evaluated. 2. Materials and Methods 2.1. Study Design The design and protocol of this nutritional intervention study has been published elsewhere [21] and can be consulted for information not reported here. A double-blind, randomized, placebo-controlled, parallel group study was conducted in patients with prediabetes. The intervention with 6 capsules containing 400 mg of MO dry leaf powder or placebo was carried out over 12 weeks. Eligible participants were randomized using a simple block randomization of 1:1. The study was registered in www.ClinicalTrials.gov (accessed on 4 May 2022) (Identifier: NCT04734132). 2.2. Study Participants Subjects with prediabetes who had never used drugs for glycemic control and were within the age of 40 to 70 y were included. Prediabetes was diagnosed following the American Diabetic Association (ADA) criteria [38]: HbA1c: 5.7–6.4%, or fasting glucose between 100 and 125 mg/dL, or 2 h glucose tolerance test between 140 and 199 mg/dL. The recruitment procedures and exclusion criteria have been described previously [21]. The participants were recruited by practitioners performing the screening of potential candidates in two primary health care centers. Some participants were also recruited through dissemination of the study by board advertisements. Seventy-three enrolled participants were randomized to the Placebo (PLC, n = 35) or Moringa (MO, n = 38) groups. Sample size calculation and randomization details are provided in [21]. This study followed the principles established in the Declaration of Helsinki and the guidelines of the Spanish law 14/2007 on Biomedical Research. Moreover, the study procedures were approved by the Puerta de Hierro-Majadahonda University Hospital Ethics Committee as well as the Bioethics committee of the Spanish National Research Council (CSIC). Prior to study entry, written informed consent was obtained from all participants. 2.3. Intervention The MO leaves used in the manufacture of the capsule supplement were obtained from an organic cultivar grown in the Mediterranean region of Spain. Nutrient composition and polyphenol content of the dry leaves are presented in Table S1. Further details can be found in [21]. Patients were instructed to take 2 capsules before each main meal (breakfast, lunch and dinner; 2.4 g/day) for 12 weeks. They were asked to make no changes in their diet or lifestyle. Three visits were programmed, at baseline (0 weeks), 6 weeks and 12 weeks. Compliance with capsule intake was good, and additional data have been already published [21]. On each visit, the participants arrived early in the morning in fasting condition. A blood sample was withdrawn from the antecubital vein, which was collected in vacutainer tubes for different biomarker analyses. A first-void urine sample collected at home was also brought on each visit. Participants were given a list of forbidden polyphenol-rich foods and recommended dishes to apply to the previous day’s dinner. 2.4. Outcomes All biomarkers studied in the current work were included as secondary outcomes in the study conception and protocol. 2.5. Blood Lipid Profile and Inflammatory Biomarker Analyses Within two hours from collection, blood was centrifuged at 1300× g for 15 min, and several aliquots were kept at −80 °C until analysis. Lipid profile and C-reactive protein (hsCRP) were analyzed in serum of freshly collected samples (Unilabs Laboratory, Madrid, Spain). Inflammatory cytokines (IL-1β, TNF-α, IL-6 and macrophage chemoattractant protein (MCP)-1) and adipokines (leptin and adiponectin) were analyzed at ICTAN laboratory by xMAP Luminex® technology (Luminex Corporation, Austin, TX, USA) with the Human High Sensitivity T cell panel and the Metabolic Hormone magnetic bead panels (Merck Millipore Burlington, MA, USA). The protease inhibitors AEBSF (4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride; 1 mg/mL final concentration) and dipeptidyl peptidase (DPP)-IV inhibitor (10 µL per 1 mL blood) (Merck KGaA, Darmstadt, Germany) were added to 1 mL of blood before centrifugation and aliquoting for preserving proteins from degradation. The sensitivity (minimum detectable concentration) of these measurements was as follows: IL-1β, 0.14 pg/mL; TNF-α, 0.16 pg/mL; IL-6, 0.11 pg/mL; MCP-1, 14 pg/mL; leptin, 41 pg/mL; adiponectin, 11 pg/mL. 2.6. TAC Assessment The measurement of TAC was performed following the application note for Photochem (Analytikjena, Jena, Germany) for the determination of lipid-soluble antioxidant capacity in blood plasma. In brief, a sample of 200 μL plasma was mixed with 200 μL dH2O and 400 μL ethanol. Then, 800 μL hexane was added, the mixture was shaken for 1 min and centrifuged 1000× g for 5 min. Then, 200 μL of the lipidic phase was collected and dried under nitrogen flow and stored in the freezer (−80 °C) until analysis. The dried extract was dissolved in 200 μL methanol and centrifuged 5000× g 1 min. Measurement was performed in the supernatant following Photochem application (Analytikjena) using an ACL reagent kit containing Trolox as the standard for the calibration curve and a photosensitizer and 120 µL of sample. The pipetted sample volume was used for final TAC calculation with the formula: TAC = [Quantity (nmol) × Dilution × Trolox Molarity (ng/nmol)]/pipetted volume (µL) 2.7. Blood Pressure Blood pressure was measured with an OMRON M6 device (Omron Healthcare, Kyoto, Japan) twice each visit, and the lowest measure was used in the analysis. Data from several patients at 6 weeks and 12 weeks were missing due to the COVID pandemic and the compulsory distance measures to avoid virus propagation. 2.8. Fecal Sample Biomarkers Frozen fecal samples (−80 °C) were slightly defrosted, and several weighed aliquots were prepared using a scalpel, for different analysis. A 150–200 mg aliquot was prepared for the analysis of calprotectin and secretory immunoglobulin A (sIgA) and kept at −80 °C until analysis. sIgA and calprotectin were analyzed with commercial ELISA kits by Immundiagnostik AG (IDK®, Bensheim, Germany) following the manufacturers’ protocols. Fecal sampling prior to analytical procedures was performed with the Stool Sample Preparation System (ref# K 6998SAS, Bensheim, Germany), which allows the extraction from a standardized amount of 15 mg. The analyses were carried out in duplicate. ELISA plates were read on the PowerWawe XS Spectrophotometer (BioTek, Santa Clara, CA, USA) by reading at 450 nm and fitting the standard curve with non-linear 4-parameter regression. Both kits included high and low quality controls. The detection limit of the kits was 6.947 ng/mL and 2.267 ng/mL for sIgA and calprotectin, respectively. 2.9. MO Metabolites in Urine Samples Collected urine samples were centrifuged at 3000 rpm for 5 min, and 1.5 mL supernatant was filtered through a 0.45 μm pore (13 mm filter), vacuum dried and stored at −80 °C until analysis. Dried samples were reconstituted in 0.2 mL ammonium acetate 13 Mm Ph 4/0.1% formic acid in acetonitrile (50:50, v/v), vortexed for 2 min and sonicated for an additional 10 min. After sonication, samples were vortexed again (2 min) and centrifuged at 10,000× g rpm for 5 min at 4 °C. Supernatants were collected and filtered through a Millex-HV13 0.22 μm pore membrane (Millipore Corp., Bedford, MA, USA). Identification and quantification of glucosinolates and their metabolic derivatives in the urine of participants was performed by ultra-high pressure liquid chromatography coupled to electrospray ionization and a 6460 tandem mass spectrometer with triple quadrupole technology (UHPLC/MS/MS, Agilent Technologies, Waldron, Germany). Chromatographic separation was carried out using a ZORBAX Eclipse Plus C18 column (2.1 × 50 mm2, 1.8 μm) with a chromatographic gradient created with the solvents (A) 13 mM ammonium acetate pH 7 and (B) acetonitrile/formic acid (99.99:0.01, v/v) according to the method specified in [39], which separates intact glucosinolates and indoles, updated for the analysis of the compounds specifically present in the administered matrix. 2.10. Diet and Anthropometry Assessments A three-day dietary registry form was used for dietary assessment, as published elsewhere [21]. Anthropometrical measurements collected from the participants included weight, height, waist and hip circumferences and bioimpedance analysis (InnerScan BC-545; TANITA, Tokyo, Japan). All measurements were taken barefoot and with standard methods. Height was measured only at the basal visit with a stadiometer (0.5 cm precision) and weight at each visit with the TANITA scale to the nearest 0.1 kg. BMI was calculated as Weight (kg)/Height (m)2. Circumferences were measured with an inelastic tape (SECA, precision 0.5 cm) using standard procedures. A dichotomic variable was created and named as Fat_Status based in the fat mass percentage obtained from the bioimpedance analysis and using the upper threshold for normal body fat percentage established for the TANITA bioelectrical impedance device for men (40–59 year, <22% and 60–70 year, <25%) and women (40–59 year, <34% and 60–70 year, <36%) [40]. This variable classified participants as “Normal body fat” or “Above normal body fat”. Other questionnaires were also filled in by trained nutritionists on each occasion, including the MEDAS questionnaire of adherence to the Mediterranean diet [41] and the Minnesota Leisure Time Physical Activity Questionnaire (MLTFAQ, Spanish version). The latter was administered by the interviewer at the basal visit, and at subsequent visits the volunteers went through their responses and highlighted any differences in their previously recorded activities. For each individual, METs (metabolic standard units) were estimated using the coefficients published in the Compendium of Physical Activities [42] and transformed to kcal/week by considering minutes spent in each physical activity per week, weight and the equivalence 0.0175 kcal/minute/kg. 2.11. Statistical Analysis Prior to data analyses, the Kolmogorov–Smirnov test and box-plot representation was used to check the distribution of the variables. Normalization of several variables was obtained after log transformation, including leptin, adiponectin, calprotectin, sIgA, TAG and VLDL-C. The effect of the MO supplementation was assessed by comparison of the change score between the groups. The rate of change for blood and fecal biomarkers was calculated with the following formula: [(value 12 weeks − value 0 weeks)/value 0 weeks] × 100 The difference between the groups in the variable’s rate of change was assessed by ANCOVA, with the fixed factors “treatment” and “body fat mass percentage status” and using the basal value as a covariate. In addition, a MIXED linear model with the repeated factor “visit”, the fixed factors “treatment” and “Fat_Status” and the interactions “visit × treatment” and “visit × Fat_Status” was used to compute the effect of the MO food supplement versus PLC along the intervention. For lipid profile variables’ analysis, having a prescription of lipid lowering agents was also used as a dichotomic variable (yes or no) to further correct the model. Similarly, for blood pressure analyses, the confounding factor “antihypertensive agent” (YES/NO) was also included to adjust the model. In addition, two individuals starting anti-hypercholesterolemic drug treatment during the intervention as well as another two starting antihypertensive treatment were excluded from the respective analyses. In order to assess if the individual characteristics related to the cardiometabolic and inflammatory status influenced the glycemic response to MO supplementation of the patients with prediabetes, a decision tree approach was undertaken. This statistical analysis aimed to classify cases into Respondent (improves HbA1c) or Non-respondent (does not improve HbA1c) based on values of independent (predictor) variables. The tree-based classification model was created using the following as independent variables (potential predictors): sex and Fat_Status (categorical), age, BMI, total cholesterol, HDL-C, VLDL-C, LDL-C, TAG, TAC, systolic pressure, diastolic pressure, leptin, adiponectin, MCP-1, TNF-α, IL1-β, IL-6, calprotectin and sIgA (continuous variables); the nominal variable HbA1c improvement (YES/NO) was used as a dependent variable and CHAID (CHi-square Automatic Interaction Detection) as the growing method. In addition, the correlations between the basal levels of biomarkers and the change in HbA1C during the intervention were also analyzed. 3. Results Figure 1 presents a flow chart of study participants, showing that 65 participants finished the study, distributed as 34 in PLC (18 female) and 31 in MO (18 female) groups. Table 1 presents their basal characteristics. No differences were found between PLC and MO groups in the proportion of subjects with an excess of body fat, which had a high prevalence in both groups. No differences were found either in the prevalence of medication prescription for cardiovascular risk factors, the MEDAS score or the amount of energy expenditure in leisure time physical activity. The demographic, anthropometrical and routine biochemistry test results in the study participants, classified according to their fat mass percentage, are presented in Table 2. In addition to the anthopometrical differences, plasma routine biochemical parameters such as those related to glycemia (glucose and HbA1c), TAG, VLDL-C and hsCRP were different between participants with a different Fat_Status classification. The results showed no significant differences between groups in the change of cytokines, adipokines and hsCRP during the intervention and no significant interaction of time × treatment on these markers’ values (Table 3). No significant visit × treatment interactions were observed in blood lipid profile variables or TAC (Table 4), and no differences in the rate of change were observed between the groups. Two participants that started anti-hypertensive medication during the intervention were excluded from the analysis. A total of 44 valid participants had basal and final measurements of blood pressure taken, and the statistical analysis showed no significant differences in the change score of either SBP or DBP between treatment groups and no significant interaction of visit × treatment was observed during the intervention (Table 5). No effect of treatment was observed on variables of intestinal health, i.e., calprotectin and sIgA (Table 6). The number of participants with positive values for calprotectin (>100 µg/mL) was not different either between groups at baseline or end of treatment (27 vs. 17% at 0 weeks and 18 vs. 17% at 12 weeks, in PLC and MO, respectively). The decision tree analyses showed that TNF-α plasma concentration significantly contributed to 77% correct classification of participants as respondents or non-respondents to MO supplementation (HBA1c improvement, YES/NO) (Table 7), with a discriminant threshold at 7.330 pg/mL (Figure 2). In addition, a significant correlation was found in the MO supplemented group between the change in HbA1c and the basal TNF-α value (r = 0.361; p = 0.050; r = 0.375; p = 0.045 in partial correlation with BMI adjustment). On the contrary, in the PLC group, no variable was successful in the decision tree at correctly classifying the subjects that improved HbA1c during the intervention, and no significant correlation was found. Regarding the identification of MO metabolites in urine, only in 36% of supplemented participants was glucomoringin detected in urine, while moringin was detected in 71%. No relationship was observed between the presence of these MO characteristic molecules and improvement of HbA1c during the study. Glucomoringin was identified in 30% of the participants that improved HbA1c levels during the intervention and in 46% of the non-respondent participants (Chi2 test; p = 0.346); for the ITC moringin, these percentages were 71% and 77%, respectively (Chi2 test; p = 697). 4. Discussion Despite a vast body of evidence on the in vitro antioxidant and anti-inflammatory activities of MO leaf extracts and polyphenolic compound enriched fractions [9,31], as well as evidence from animal studies showing antidiabetic and antihyperlipidemia effects [43], provision of antioxidant stability [44], modification of the expression of enzymes involved in carbohydrate and lipid metabolism [44,45] and of inflammatory markers [29,30,45], the evidence of similar effects and activities in human studies is scarce. This nutritional intervention study was designed to test the hypoglycemic effect of MO dry leaf powder in patients with prediabetes, and a moderate but significant effect was found on glycemia markers [21]. Here, the effects of this intervention on inflammatory and cardiometabolic status markers were assessed as secondary outcomes. The reported results of the inflammation markers measured in serum or plasma (CRP, MCP-1, TNF-α, IL1-β, IL-6) or in fecal samples (calprotectin), the leptin and adiponectin levels, the lipid profile, the plasma antioxidant capacity and the blood pressure showed no significant effects. The dose of MO used, added to individual heterogeneity in the biomarker basal values, might have contributed to the lack of significant effects observed. There are no clinical trials in the literature studying changes in plasma inflammatory markers following MO leaf intake. Many animal model studies have shown that MO leaf extracts in daily amounts of 200–300 mg/kg BW(body weight) decrease the expression of proinflammatory cytokine genes and the protein levels in different organs [11,30]. However, at higher doses, different results have also been reported. Gao et al. [46], in an experiment with mice, when administering 750 mg/kg BW of an aqueous MO leaf extract for 4 weeks, showed impaired colon intestinal barrier and increased inflammatory response as suggested by increased serum LPS and expression of pro-inflammatory cytokines in the colon and liver, although with inconsistent cytokine expression levels in the ileum. This concurred with alterations in bacterial groups in the cecum samples. In humans, the effect of MO on inflammatory conditions has been tested in individuals with gingival inflammation. A cross-over study including 20 subjects with mild to moderate gingivitis showed a greater reduction in mean gingival index scores and plaque scores with the MO-based dentifrice compared to a miswak-based dentifrice [47]. In addition, another study with MO leaf extract-based lozenges reduced inflammation and gingivitis in smokers [48]. In support of these findings, a model of periodontitis as a chronic inflammatory disease proved that MO leaf aqueous extract at 0.5 and 1 g/kg BW for 30 days provides anti-periodontitis activity in rats. The treatment significantly decreased serum TNF-α, IL-1β, and IL-6 compared with the control group, whereas IL-1Ra and IL-10 were increased and alveolar bone resorption was significantly reduced [49]. Despite the fact that no significant effects of MO consumption were found on inflammatory markers in the current study, no definitive conclusions can be drawn regarding the anti-inflammatory potential of MO supplementation because of the variable inflammatory status of the participating subjects. A population with homogeneously high inflammatory markers might have facilitated finding significant anti-inflammatory effects. This, together with the fact that the administered dose was in the low range of those used among the published studies might explain why the anti-inflammatory activity of MO was not observed in the current study. Two clinical trials have assessed the results of MO leaf tablets or powder consumption on the lipid profile. The first one included 35 hyperlipidemic individuals and showed that the consumption of 4.6 g tablets of MO leaves for 50 days resulted in a small (1.6%) but significant (p < 0.05) decrease in total cholesterol and in the total cholesterol/HDL-C ratio (p < 0.001), with a non-significant increase in HDL-C (6.3%) and unchanged TAG, LDL-C and VLDL-C [36]. The second study by Kumari et al. [17] included non-insulin dependent diabetic patients (23 in the experimental group and 9 in the control group) and showed a significant reduction in total cholesterol, LDL-C and VLDL-C after a 40-day intervention with 8 g MO leaf powder per day, without a significant change in HDL-C, which increased by 9%. The difference with the lack of effect observed in the current study might be due to the lower dose used here and the difference in the mean initial values of total cholesterol between studies, which were 198 ± 32 mg/dL in the experimental group of the current study and 261 ± 20 mg/dL in the MO supplemented subjects with diabetes studied by Kumari et al. [17]. In this sense, it is worth mentioning that the subjects with prediabetes studied in the current clinical trial had variable nutritional status as assessed by BMI and body composition measures. As a result, differences were found in plasma routine biochemical parameters such as those related to glycemia (glucose and HbA1c), lipemia (TAG, VLDL-C) and inflammation (CRP) between participants with normal and above-normal fat percentage, which supports the need to include the categorical variable of fat status (based in excessive body fat percentage relative to normal values) as a potential confounding factor in this type of studies. The radical scavenging and antioxidant activity of MO leaves has been amply demonstrated in vitro using different solvent extractions [50]. However, evidence from human studies is scarce. In a study by Ngamukote et al. [51], 20 healthy participants were assigned to receive a single dose of either 200 mL of warm water or 200 mL of MO aqueous leaf extract (500 mg). The increase in ferric reducing ability of plasma (FRAP) and decrease in malondialdehyde (MDA), the main product of lipid peroxidation, were significant 30 min after ingestion of the MO extract. In addition, a negative correlation was observed between both parameters. This study suggests the potential to beneficially modify plasma antioxidant status; however, it provides limited evidence due to the small sample size and does not provide results for long-term consumption. Another study supplemented MO powder (7 g) within daily menus for 3 months in postmenopausal women, showing a significant improvement in the blood antioxidant levels, including serum retinol, ascorbic acid, glutathione peroxidase and superoxide dismutase, whereas MDA was decreased [19]. The participating women were from the Indian state of Punjab, where MO is a traditionally consumed food. Depending on the dryness degree of the leaf powder, the concentration and actual ingested amount of bioactive compounds can vary; however, the dose used in the study by Kushwaha et al. was larger than the one in our study, and this might explain why we were not able to observe an improvement in the plasma antioxidant capacity. The method we used to measure plasma antioxidant capacity analyzed exclusively the circulating antioxidants in the plasma lipophilic fraction, which was chosen to exclude the effect of uric acid, known as the main contributor to plasma total antioxidant capacity as measured using various methods i.e., FRAP [52]. We did not measure MDA, and we cannot exclude that MO exerted an effect on this marker or the enzyme-dependent antioxidant systems. Another study by Taweerutchana et al. [20], testing the hypoglycemic activity in therapy-naïve diabetic patients failed to find a significant effect of MO leaf powder compared to the placebo as evaluated through fasting and postprandial glucose levels monitored with glucometers. According to the authors, this might be related to an insufficient amount of bioactive compounds, like moringin or chlorogenic acid, in the 4 g of MO dry leaf powder provided per day. In addition, in that study, the 28-day duration of the intervention might have been too short. Interestingly, despite no change in antihypertensive agents, the MO leaf group had a reduction of SBP and DBP by 5 mmHg compared to the baseline, whereas the placebo group showed an increase in blood pressure by 2 mmHg; however, these differences had no statistical significance. Similarly, in the current study, blood pressure did not show significant changes following MO supplementation compared to the PLC group; however, due to missing data in the final visit, the sample size for this parameter was smaller than for the rest of the parameters. This lack of effect contrasts with previous observations in animal models [34,35] and human studies [37,53] that substantiate the use of the plant in traditional medicine. Evidence has been published of the capacity of MO to lower 2 h postprandial blood pressure in healthy participants that consumed 120 g of cooked MO for a week [37]. However, long-term randomized placebo-controlled studies have not been published. Thus, a bigger study in therapy-naïve hypertensive subjects with long-term supplementation is warranted. Despite the lack of significant changes in all secondary outcomes reported in this study, an interesting finding was the difference in baseline TNF-α between participants improving their glucose control and those who did not improve with MO supplementation. In the MO supplemented group, 58% of patients improved their HbA1c values during the intervention, compared to 38% in the PLC group [21]. One hundred percent of participants were correctly classified as respondents based on basal TNF-α ≤ 7.330 pg/mL. However, although the threshold of 7.330 pg/mL for basal value TNF-α is the best predictor of HbA1c response, 28% of those subjects with values under this threshold will still be non-respondents, probably because of other factors not measured in this study, which might include genetic polymorphisms and epigenetic regulation of genes involved in carbohydrate and lipid metabolism [54] or microbiome differences between subjects [55,56]. On the other hand, considering the predictive character of this test, this screening method would facilitate the identification of 19% of subjects with prediabetes that would not respond due to high basal TNF-α values, under the conditions used for supplementation in this study. So far, the findings indicate that the presence of chronic low-grade inflammation hinders the benefits of low-dose MO supplementation on glycemic control. Thus, it is possible to speculate that higher doses might work more efficiently to improve both glycemic control and the inflammatory cytokines. Further studies with higher doses adapted to the subjects’ inflammatory condition would be necessary. This study is limited by the low dose used compared to other human studies and the mild baseline metabolic damage observed in the participating subjects as a group, although individual characteristics showed heterogeneity as observed in variable fat excess and differences in baseline laboratory values. The study had several advantages, such as a higher sample size than previously published studies, the blinded design of the intervention and the amount of different parameters measured, which allowed the screening of baseline characteristics to point to TNF-α as a good predictor of the participants’ glycemic response to MO. In conclusion, although this study with MO supplementation improved the glycemic control of participants with prediabetes, as previously published [21], no further effects were evidenced in inflammatory and cardiometabolic markers. The relationship observed between higher basal TNF-α values and failure to improve glycemic control suggests that doses higher than 2.4 g per day might be necessary to increase the number of subjects with a favorable response in all of the biomarkers studied. Acknowledgments The authors acknowledge Miguel Godino from the Polytechnic University of Madrid, leader of a cooperation project on Moringa in Colombia, who provided the initiative for this study, for his advice and for facilitating contact with MO producers on the Iberian Peninsula. The authors are grateful to the Metabolomics Service at CEBAS-CSIC for supporting in the instrumental analyses of this work. The authors acknowledge support of the publication fee by the CSIC Open Access Publication Support Initiative through its Unit of Information Resources for Research (URICI). The authors are grateful to all the participants for their interest in and support of scientific research. Supplementary Materials The following are available online at https://www.mdpi.com/article/10.3390/nu14091937/s1, Table S1: Nutrient and polyphenol content in the MO (Moringa oleifera Lam.) dry leaf powder. Click here for additional data file. Author Contributions Conceptualization, E.N., S.G.-M., A.M. and M.J.P.; methodology, L.E.D.-P., S.G.-M. and E.N.; participant interviews and laboratory procedures, S.G.-M., E.N., I.V.-C., L.E.D.-P., C.H., E.A.G.-R., M.d.C.M.-R., M.C. and M.J.P.; formal analysis, E.N., I.V.-C., L.E.D.-P. and S.G.-M.; data curation, I.V.-C., S.G.-M., L.E.D.-P. and E.N.; writing—original draft preparation, E.N.; writing—review and editing, all authors; supervision, E.N. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of CSIC (protocol code 115/2017) and the Ethics Committee of Hospital Universitario Puerta de Hierro-Majadahonda (code 156-17, Acta 08.18). Informed Consent Statement Informed consent was obtained from all participants involved in the study. Data Availability Statement The data presented in this study are available on request from the corresponding author, due to privacy restriction. The study is registered in www.ClinicalTrials.gov (Identifier: NCT04734132; accessed on 4 May 2022). Conflicts of Interest The authors declare no conflict of interest. Figure 1 Flow chart of study participants. Figure 2 Decision tree on HbA1c improvement as dependent variable. Classification is 77% correct using basal TNF-α as predictor variable. nutrients-14-01937-t001_Table 1 Table 1 Basal characteristics and medication prescribed in the study participants. PLC (n = 34) MO (n = 31) p # BMI < 25, n (%) 6 (18) 4 (13) 0.561 25–29.9, n (%) 17 (50) 13 (42) ≥30, n (%) 11 (32) 14 (45) Fat_Status 0.683 Normal, n (%) 8 (23.5) 6 (19.4) Above normal, n (%) 26 (76.5) 25 (80.6) Body fat excess (%) a 5.1 ± 6.2 5.7 ± 4.4 0.327 Lipid lowering agents, n (%) 8 (23.5) 10 (32.3) 0.432 Antihypertensive agents, n (%) 9 (26.5) 7 (22.6) 0.716 MEDAS score 9.5 ± 2.5 10.2 ± 1.8 0.208 Physical activity (kcal/week) 4591 ± 2831 3890 ± 2176 0.279 # Chi2 or Student’s t test (independent samples). a Body fat excess is calculated as body fat percentage minus the high threshold of the normal body fat percentage range for men and women according to Gallagher et al. (2000) [40] for the corresponding age group (40–59 year or 60–70 year). PLC: placebo; MO: Moringa oleifera Lam. nutrients-14-01937-t002_Table 2 Table 2 Basal characteristics of the participants with prediabetes according to their body fat percentage. Normal Body Fat % (n = 14) High Body Fat % (n = 51) p Male:Female #, n (%) 7:7 (50:50) 22:29 (43:57) 0.647 Age (year) 53.3 ± 11.4 56.9 ± 9.6 0.116 BMI 24.3 ± 1.9 30.2 ± 3.3 <0.001 <25 # (n) 10 0 <0.001 WC (men) 87.6 ± 5.3 101.8 ± 9.7 <0.001 (women) 84.7 ± 5.8 94.9 ± 9.7 <0.001 Glucose (mg/mL) 98.3 ± 9.6 105.3 ± 14.1 0.042 HbA1c (%) 5.7 ± 0.3 5.9 ± 0.3 0.014 Uric acid (mg/dL) 5.4 ± 0.7 5.6 ± 1.4 0.246 GOT (UI/L) Φ 21.50 (17.75–25.50) 22.00 (19.00–27.00) 0.835 GPT (UI/L) Φ 20.50 (15.00–31.50) 26.00 (21.00–33.00) 0.089 GGT (UI/L) Φ 20.00 (14.75–29.75) 24.00 (19.00–36.00) 0.177 Total Cholesterol (mg/dL) 207.0 ± 38.1 202.1 ± 33.7 0.327 TAG (mg/dL) 83.6 ± 18.9 111.8 ± 47.3 0.003 HDL-C (mg/dL) 58.2 ± 12.3 58.1 ± 13.3 0.488 LDL-C (mg/dL) 131.9 ± 34.2 121.7 ± 31.5 0.147 VLDL-C (mg/dL) 16.9 ± 3.9 22.4 ± 9.4 0.004 hsCRP (mg/dL) Φ 0.03 (0.006–0.09) 0.16 (0.07–0.51) 0.001 Data are Mean ± SD or Median (IQR, interquartile range). Mean values between groups were compared by independent sample t test, except for the variables specified with the symbols. # Chi2 test was used for categorical variables and Φ Mann-Whitney U test was used for variables not fitting normal distribution. The log-transformed variables were used for group comparison in the case of HbA1c, total cholesterol, TAG and VLDL-C. WC, waist circumference; TAG, Triacylglycerides. GGT: Gamma-glutamyl Transferase. nutrients-14-01937-t003_Table 3 Table 3 Inflammatory markers and adipokines in patients with prediabetes of the PLC and MO groups during the intervention. 0 Weeks 6 Weeks 12 Weeks MIXED Model p # Rate of Change a 0 Weeks–12 Weeks MCP-1 (pg/mL) PLC 67, 82 ± 57 69, 77 ± 42 69, 77 ± 36 0.328 0.063 ± 0.418 MO 63, 71 ± 35 70, 64 ± 24 57, 69 ± 45 −0.002 ± 0.312 NS TNF-α (pg/mL) PLC 6.0, 7.4 ± 3.7 6.4, 7.1 ± 3.5 5.5, 6.7 ± 3.1 0.291 −0.057 ± 0.232 MO 5.7, 6.4 ± 2.4 6.5, 6.8 ± 2.8 5.5, 6.0 ± 2.3 −0.034 ± 0.279 NS IL-6 (pg/mL) PLC 1.8, 2.6 ± 3.0 1.7, 3.7 ± 7.3 1.8, 2.3 ± 2.1 0.607 −0.012 ± 0.599 MO 1.3, 3.6 ± 7.9 1.7, 4.9 ± 11.5 1.2, 3.2 ± 9.1 −0.158 ± 0.368 NS IL-1β (pg/mL) PLC 1.4, 1.7 ± 1.1 1.3, 1.7 ± 1.0 1.3, 1.6 ± 0.9 0.908 −0.008 ± 0.372 MO 1.4, 1.5 ± 0.8 1.3, 1.5 ± 0.8 1.1, 1.3 ± 0.6 −0.061 ± 0.328 NS hsCRP (pg/mL) PLC 0.15, 0.40 ± 0.64 0.11, 0.31 ± 0.67 0.12, 0.30 ± 0.50 0.359 0.457 ± 2.457 MO 0.10, 0.23 ± 0.41 0.11, 0.19 ± 0.21 0.11, 0.19 ± 0.22 1.189 ± 3.423 NS Leptin (ng/mL) PLC 5.63, 7.59 ± 7.97 6.61, 7.75 ± 8.29 5.66, 6.72 ± 5.45 0.343 0.072 ± 0.454 MO 6.48, 7.15 ± 4.68 8.35, 7.96 ± 4.73 5.87, 6.99 ± 4.95 −0.017 ± 0.387 NS Adiponectin (µg/mL) PLC 23.1, 27.7 ± 21.7 21.09, 24.5 ± 16.6 15.42, 18.5 ± 12.0 0.871 −0.242 ± 0.338 MO 23.4, 26.5 ± 16.1 15.91, 23.2 ± 16.4 15.60, 18.8 ± 10.9 −0.262 ± 0.252 NS Median, Mean ± SD. MIXED linear model with the repeated factor “visit” and the fixed factors “treatment” and “Fat_Status” and the interaction “visit × treatment” and “Fat_Status × treatment”; p # corresponds to “visit × treatment”. a ANCOVA with the fixed factors “treatment” and “Fat_Status” and using the basal value as covariate; NS, not significant. PLC: placebo; MO: Moringa oleifera Lam.; MCP: macrophage chemoattractant protein. nutrients-14-01937-t004_Table 4 Table 4 Serum lipid profile and plasma TAC in patients with prediabetes of the PLC and MO groups during the intervention. 0 Weeks 6 Weeks 12 Weeks MIXED Model p # Rate of Change 0 Weeks–12 Weeks a Total Cholesterol (mg/dL) PLC 206.4 ± 36.4 207.7 ± 36.0 211.2 ± 34.6 0.494 0.033 ± 0.122 MO 197.9 ± 31.5 208.6 ± 29.1 203.1 ± 35.0 0.011 ± 0.123 NS TAG (mg/dL) PLC 102.2 ± 46.5 106.7 ± 43.2 113.0 ± 51.2 0.824 0.143 ± 0.352 MO 110.2 ± 43.3 115.82 ± 49.5 121.8 ± 72.8 0.110 ± 0.297 NS HDL-C (mg/dL) PLC 57.8 ± 12.0 59.6 ± 13.4 61.8 ± 15.0 0.608 0.073 ± 0.157 MO 57.8 ± 14.3 59.1 ± 14.7 58.9 ± 14.9 0.026 ± 0.176 NS LDL-C (mg/dL) PLC 128.1 ± 34.4 126.8 ± 33.2 126.7 ± 34.5 0.307 0.001 ± 0.155 MO 118.1 ± 27.9 126.4 ± 27.4 119.8 ± 30.0 0.000 ± 0.184 NS VLDL-C (mg/dL) PLC 20.4 ± 9.3 21.3 ± 8.6 22.7 ± 10.1 0.776 0.145 ± 0.349 MO 22.1 ± 8.5 23.1 ± 9.9 24.4 ± 14.6 0.096 ± 0.296 NS TAC PLC 1.34 ± 0.32 1.29 ± 0.37 1.30 ± 0.45 0.134 −0.01 ± 0.32 MO 1.22 ± 0.42 1.27 ± 0.39 1.09 ± 0.44 0.03 ± 0.57 NS Mean ± SD. MIXED linear model with the repeated factor “visit” and the fixed factors “treatment”, “Fat_Status” and “anti-hypercholesterolemic treatment” and the interaction “visit × treatment” and “visit × Fat_status”; p # corresponds to “visit × treatment”. a ANCOVA with the fixed factor “treatment”, the confounding factors “Fat_Status” and “anti-hypercholesterolemic treatment” and the basal value as covariate. TAG, triacylglycerides; NS, not significant. PLC: placebo; MO: Moringa oleifera Lam. nutrients-14-01937-t005_Table 5 Table 5 Systolic and diastolic blood pressure in patients with prediabetes of the PLC and MO groups during the intervention. 0 Weeks PLC, n = 34; MO, n = 29 6 Weeks PLC, n = 29; MO, n = 22 12 Weeks PLC, n = 26; MO, n = 18 MIXED Model p # Rate of Change 0 Weeks–12 Weeks a PLC, n = 26; MO, n = 18. SBP (mmHg) PLC 129 ± 15 127 ± 187 128 ± 16 0.807 −0.005 ± 0.099 MO 129 ± 15 125 ± 11 126 ± 11 −0.011 ± 0.077 NS DBP (mmHg) PLC 79 ± 9 78 ± 11 81 ± 11 0.441 0.007 ± 0.083 MO 80 ± 9 76 ± 8 77 ± 8 −0.031 ± 0.063 NS Mean ± SD. MIXED linear model with the repeated factor “visit” and the fixed factors “treatment”, “Fat_Status” and “anti-hypertensive treatment” and the interaction “visit × treatment” and “visit × Fat_status”; p # corresponds to “visit × treatment”. a ANCOVA with the fixed factor “treatment”, the confounding factors “Fat_Status” and “anti-hypertensive treatment” and the basal value as covariate. SBP, systolic blood pressure; DBP, diastolic blood pressure; NS, not significant. PLC: placebo; MO: Moringa oleifera Lam. nutrients-14-01937-t006_Table 6 Table 6 Intestinal health markers in fecal samples of patients with prediabetes of the PLC and MO groups during the intervention. 0 Weeks 6 Weeks 12 Weeks MIXED Model p # Rate of Change 0 Weeks–12 Weeks a Calprotectin (µg/mL) PLC 25; 73 ± 95 - 30; 58 ± 70 0.851 0.0; 0.5 ± 1.6 MO 32; 58 ± 69 - 32; 59 ± 63 0.1; 0.6 ± 1.6 NS sIgA (µg/mL) PLC 1448; 2017 ± 1856 - 1402; 2208 ± 2437 0.941 −0.1; 1.5 ± 5.8 MO 1343; 1663 ± 1478 - 1245; 2187 ± 2154 −0.2; 1.0 ± 3.5 NS Median, Mean ± SD. MIXED linear model with the repeated factor “visit” and the fixed factors “treatment” and “Fat_Status” and the interaction “visit × treatment” and “Fat_Status × treatment”; p # corresponds to “visit × treatment”. a ANCOVA with the fixed factor “treatment”, “Fat_Status” as confounder and the basal value as covariate; NS, not significant. PLC: placebo; MO: Moringa oleifera Lam. nutrients-14-01937-t007_Table 7 Table 7 Classification of participants in the decision tree analysis based in the threshold value of TNF-α. Observed Predicted NO YES Percent Correct NO 6 7 46.2% YES 0 18 100.0% Overall Percentage 19.4% 80.6% 77.4% Growing Method: CHAID. Dependent Variable: HbA1c improvement. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Leone A. Spada A. Battezzati A. Schiraldi A. Aristil J. Bertoli S. Cultivation, Genetic, Ethnopharmacology, Phytochemistry and Pharmacology of Moringa oleifera Leaves: An Overview Int. J. Mol. Sci. 2015 16 12791 12835 10.3390/ijms160612791 26057747 2. Taher M. Bin Nyeem M. Ahammed M. Hossain M. Islam M.N. Moringa oleifera (Shajna): The wonderful indigenous medicinal plant Asian J. Med. Biol. Res. 2017 3 20 30 10.3329/ajmbr.v3i1.32032 3. Anwar F. Latif S. Ashraf M. Gilani A.H. Moringa oleifera: A food plant with multiple medicinal uses Phytother. Res. 2007 21 17 25 10.1002/ptr.2023 17089328 4. Mbikay M. Therapeutic Potential of Moringa oleifera Leaves in Chronic Hyperglycemia and Dyslipidemia: A Review Front. Pharmacol. 2012 3 24 10.3389/fphar.2012.00024 22403543 5. Meireles D. Gomes J. Lopes L. Hinzmann M. Machado J. A review of properties, nutritional and pharmaceutical applications of Moringa oleifera: Integrative approach on conventional and traditional Asian medicine Adv. Tradit. Med. 2020 20 495 515 10.1007/s13596-020-00468-0 6. Garza N.G.G. Koyoc J.A.C. Castillo J.A.T. Zambrano E.A.G. Ancona D.B. Guerrero L.C. García S.R.S. Biofunctional properties of bioactive peptide fractions from protein isolates of moringa seed (Moringa oleifera) J. Food Sci. Technol. 2017 54 4268 4276 10.1007/s13197-017-2898-8 29184233 7. Tian H. Liang Y. Liu G. Li Y. Deng M. Liu D. Guo Y. Sun B. Moringa oleifera polysaccharides regulates caecal microbiota and small intestinal metabolic profile in C57BL/6 mice Int. J. Biol. Macromol. 2021 182 595 611 10.1016/j.ijbiomac.2021.03.144 33836198 8. Li S. Li W. Wu R. Yin R. Sargsyan D. Raskin I. Kong A.N. Epigenome and transcriptome study of moringa isothiocyanate in mouse kidney mesangial cells induced by high glucose, a potential model for diabetic-induced nephropathy AAPS J. 2019 22 8 10.1208/s12248-019-0393-z 31807911 9. Rajan T.S. Giacoppo S. Iori R. De Nicola G.R. Grassi G. Pollastro F. Bramanti P. Mazzon E. Anti-inflammatory and antioxidant effects of a combination of cannabidiol and moringin in LPS-stimulated macrophages Fitoterapia 2016 112 104 115 10.1016/j.fitote.2016.05.008 27215129 10. Amaglo N.K. Bennett R.N. Lo Curto R.B. Rosa E.A. Lo Turco V. Giuffrida A. Lo Curto A. Crea F. Timpo F.M. Profiling selected phytochemicals and nutrients in different tissues of the multipurpose tree Moringa oleifera L., grown in Ghana Food Chem. 2010 122 1047 1054 10.1016/j.foodchem.2010.03.073 11. Nova E. Redondo-Useros N. Martínez-García R.M. Gómez-Martínez S. Díaz-Prieto L.E. Marcos A. Potential of Moringa oleifera to Improve Glucose Control for the Prevention of Diabetes and Related Metabolic Alterations: A Systematic Review of Animal and Human Studies Nutrients 2020 12 2050 10.3390/nu12072050 32664295 12. Islam Z. Islam S.M.R. Hossen F. Mahtab-Ul-Islam K. Hasan M.R. Karim R. Moringa oleifera is a Prominent Source of Nutrients with Potential Health Benefits Int. J. Food Sci. 2021 2021 6627265 10.1155/2021/6627265 34423026 13. Ahmad J. Khan I. Johnson S.K. Alam I. Din Z.U. Effect of Incorporating Stevia and Moringa in Cookies on Postprandial Glycemia, Appetite, Palatability, and Gastrointestinal Well-Being J. Am. Coll. Nutr. 2018 37 133 139 10.1080/07315724.2017.1372821 29272206 14. Anthanont P. Lumlerdkij N. Akarasereenont P. Vannasaeng S. Sriwijitkamol A. Moringa oleifera Leaf Increases Insulin Secretion after Single Dose Administration: A Preliminary Study in Healthy Subjects J. Med. Assoc. Thail. 2016 99 308 313 15. Fombang E.N. Saa W.R. Antihyperglycemic Activity of Moringa oleifera Lam leaf functional tea in rat models and human subjects Food Nutr. Sci. 2016 7 1021 1032 10.4236/fns.2016.711099 16. Leone A. Bertoli S. Di Lello S. Bassoli A. Ravasenghi S. Borgonovo G. Forlani F. Battezzati A. Effect of Moringa oleifera Leaf Powder on Postprandial Blood Glucose Response: In Vivo Study on Saharawi People Living in Refugee Camps Nutrients 2018 10 1494 10.3390/nu10101494 30322091 17. Kumari D.J. Hypoglycemic effect of Moringa oleifera and Azadirachta indica in type-2 diabetes Bioscan 2010 5 211 214 18. Giridhari V.A. Malathi D. Geetha K. Anti diabetic property of drumstick (Moringa oleifera) leaf tablets Int. J. Health Nutr. 2011 2 1 5 19. Kushwaha S. Chawla P. Kochhar A. Effect of supplementation of drumstick (Moringa oleifera) and amaranth (Amaranthus tricolor) leaves powder on antioxidant profile and oxidative status among postmenopausal women J. Food Sci. Technol. 2014 51 3464 3469 10.1007/s13197-012-0859-9 26396347 20. Taweerutchana R. Lumlerdkij N. Vannasaeng S. Akarasereenont P. Sriwijitkamol A. Effect of Moringa oleifera Leaf Capsules on Glycemic Control in Therapy-Naïve Type 2 Diabetes Patients: A Randomized Placebo Controlled Study Evid.-Based Complement. Altern. Med. 2017 2017 6581390 10.1155/2017/6581390 21. Gómez-Martínez S. Díaz-Prieto L.E. Vicente Castro I. Jurado C. Iturmendi N. Martín-Ridaura M.C. Calle N. Dueñas M. Picón M.J. Marcos A. Moringa oleifera Leaf Supplementation as a Glycemic Control Strategy in Subjects with Prediabetes Nutrients 2021 14 57 10.3390/nu14010057 35010932 22. Sangrós F.J. Torrecilla J. Giráldez-García C. Carrillo L. Mancera J. Mur T. Franch J. Díez J. Goday A. Serrano R. Association of General and Abdominal Obesity with Hypertension, Dyslipidemia and Prediabetes in the PREDAPS Study Rev. Esp. Cardiol. 2018 71 170 177 10.1016/j.recesp.2017.04.010 28789915 23. Morigny P. Houssier M. Mouisel E. Langin D. Adipocyte lipolysis and insulin resistance Biochimie 2016 125 259 266 10.1016/j.biochi.2015.10.024 26542285 24. Henderson G.C. Plasma Free Fatty Acid Concentration as a Modifiable Risk Factor for Metabolic Disease Nutrients 2021 13 2590 10.3390/nu13082590 34444750 25. Joung H. Kim B. Park H. Lee K. Kim H.H. Sim H.C. Do H.J. Hyun C.K. Do M.S. Fermented Moringa oleifera decreases hepatic adiposity and ameliorates glucose intolerance in high-fat diet-induced obese mice J. Med. Food 2017 20 439 447 10.1089/jmf.2016.3860 28504910 26. Bao Y. Xiao J. Weng Z. Lu X. Shen X. Wang F. A phenolic glycoside from Moringa oleifera Lam. improves the carbohydrate and lipid metabolisms through AMPK in db/db mice Food Chem. 2020 311 125948 10.1016/j.foodchem.2019.125948 31877545 27. Arulselvan P. Tan W.S. Gothai S. Muniandy K. Fakurazi S. Esa N.M. Alarfaj A.A. Kumar S.S. Anti-Inflammatory Potential of Ethyl Acetate Fraction of Moringa oleifera in Downregulating the NF-κB Signaling Pathway in Lipopolysaccharide-Stimulated Macrophages Molecules 2016 21 1452 10.3390/molecules21111452 28. Luetragoon T. Pankla Sranujit R. Noysang C. Thongsri Y. Potup P. Suphrom N. Nuengchamnong N. Usuwanthim K. Bioactive Compounds in Moringa oleifera Lam. Leaves Inhibit the Pro-Inflammatory Mediators in Lipopolysaccharide-Induced Human Monocyte-Derived Macrophages Molecules 2020 25 191 10.3390/molecules25010191 29. Tang Y. Choi E.J. Han W.C. Oh M. Kim J. Hwang J.Y. Park P.J. Moon S.H. Kim Y.S. Kim E.K. Moringa oleifera from Cambodia ameliorates oxidative stress, hyperglycemia, and kidney dysfunction in type 2 diabetic mice J. Med. Food 2017 20 502 510 10.1089/jmf.2016.3792 28467233 30. Omodanisi E.I. Aboua Y.G. Chegou N.N. Oguntibeju O.O. Hepatoprotective, Antihyperlipidemic, and Anti-inflammatory Activity of Moringa oleifera in Diabetic-induced Damage in Male Wistar Rats Pharmacogn. Res. 2017 9 182 187 10.4103/0974-8490.204651 31. Waterman C. Cheng D.M. Rojas-Silva P. Poulev A. Dreifus J. Lila M.A. Raskin I. Stable, water extractable isothiocyanates from Moringa oleifera leaves attenuate inflammation in vitro Phytochemistry 2014 103 114 122 10.1016/j.phytochem.2014.03.028 24731259 32. Cheng D. Gao L. Su S. Sargsyan D. Wu R. Raskin I. Kong A.-N. Moringa isothiocyanate activates Nrf2: Potential role in diabetic nephropathy AAPS J. 2019 21 31 10.1208/s12248-019-0301-6 30783799 33. Sailaja B.S. Aita R. Maledatu S. Ribnicky D. Verzi M.P. Raskin I. Moringa isothiocyanate-1 regulates Nrf2 and NF-κB pathway in response to LPS-driven sepsis and inflammation PLoS ONE 2021 16 e0248691 10.1371/journal.pone.0248691 33793581 34. Aekthammarat D. Tangsucharit P. Pannangpetch P. Sriwantana T. Sibmooh N. Moringa oleifera leaf extract enhances endothelial nitric oxide production leading to relaxation of resistance artery and lowering of arterial blood pressure Biomed. Pharmacother. 2020 130 110605 10.1016/j.biopha.2020.110605 32781358 35. Ma K. Wang Y. Wang M. Wang Z. Wang X. Ju X. He R. Antihypertensive activity of the ACE-renin inhibitory peptide derived from Moringa oleifera protein Food Funct. 2021 12 8994 9006 10.1039/D1FO01103K 34382048 36. Nambiar V.S. Guin P. Parnami S. Daniel M. Impact of antioxidants from drumstick leaves on the lipid profile of hyperlipidemics J. Herb. Med. Toxicol. 2010 4 165 172 0973-4643 37. Chan Sun M. Ruhomally Z.B. Boojhawon R. Neergheen-Bhujun V.S. Consumption of Moringa oleifera Lam. Leaves Lowers Postprandial Blood Pressure J. Am. Coll. Nutr. 2020 39 54 62 10.1080/07315724.2019.1608602 31063434 38. American Diabetes Association Standards of medical care in diabetes Diabetes Care 2013 36 S11 S66 10.2337/dc13-S011 23264422 39. Domínguez-Perles R. Medina S. Moreno D.A. García-Viguera C. Ferreres F. Gil-Izquierdo A. A new ultra-rapid UHPLC/MS/MS method for assessing glucoraphanin and sulforaphane bioavailability in human urine Food Chem. 2014 143 132 138 10.1016/j.foodchem.2013.07.116 24054222 40. Gallagher D. Heymsfield S.B. Heo M. Jebb S.A. Murgatroyd P.R. Sakamoto Y. Healthy percentage body fat ranges: An approach for developing guidelines based on body mass index Am. J. Clin Nutr. 2000 72 694 701 10.1093/ajcn/72.3.694 10966886 41. Martínez-González M.A. García-Arellano A. Toledo E. Salas-Salvadó J. Buil-Cosiales P. Corella D. Covas M.I. Schröder H. Arós F. Gómez-Gracia E. A 14-Item Mediterranean Diet Assessment Tool and Obesity Indexes among High-Risk Subjects: The PREDIMED Trial PLoS ONE 2012 7 e43134 10.1371/journal.pone.0043134 22905215 42. Ainsworth B.E. Haskell W.L. Herrmann S.D. Meckes N. Bassett D.R. Tudor-Locke C. Greer J.L. Vezina J. Whitt-Glover M.C. Leon A.S. Compendium of physical activities: A second update of codes and MET values Med. Sci. Sports Exerc. 2011 43 1575 1581 10.1249/MSS.0b013e31821ece12 21681120 43. Khan W. Parveen R. Chester K. Parveen S. Ahmad S. Hypoglycemic Potential of Aqueous Extract of Moringa oleifera Leaf and In Vivo GC-MS Metabolomics Front. Pharmacol. 2017 8 577 10.3389/fphar.2017.00577 28955221 44. Abd Eldaim M.A. Shaban Abd Elrasoul A. Abd Elaziz S.A. An aqueous extract from Moringa oleifera leaves ameliorates hepatotoxicity in alloxan-induced diabetic rats Biochem. Cell Biol. 2017 95 524 530 10.1139/bcb-2016-0256 28423281 45. Waterman C. Rojas-Silva P. Tumer T.B. Kuhn P. Richard A.J. Wicks S. Stephens J.M. Wang Z. Mynatt R. Cefalu W. Isothiocyanate-rich Moringa oleifera extract reduces weight gain, insulin resistance, and hepatic gluconeogenesis in mice Mol. Nutr. Food Res. 2015 59 1013 1024 10.1002/mnfr.201400679 25620073 46. Gao X. Xie Q. Liu L. Kong P. Sheng J. Xiang H. Metabolic adaptation to the aqueous leaf extract of Moringa oleifera Lam.-supplemented diet is related to the modulation of gut microbiota in mice Appl. Microbiol. Biotechnol. 2017 101 5115 5130 10.1007/s00253-017-8233-5 28382453 47. Duarte K. Thomas B. Varma S.R. Kamath V. Shetty B. Kuduruthullah S. Nambiar M. Antiplaque Efficacy of a Novel Moringa oleifera Dentifrice: A Randomized Clinical Crossover Study Eur. J. Dent. 2022 10.1055/s-0041-1736418 48. Luetragoon T. Sranujit R.P. Noysang C. Thongsri Y. Potup P. Somboonjun J. Maichandi N. Suphrom N. Sangouam S. Usuwanthim K. Evaluation of Anti-Inflammatory Effect of Moringa oleifera Lam. and Cyanthillium cinereum (Less) H. Rob. Lozenges in Volunteer Smokers Plants 2021 10 1336 10.3390/plants10071336 34208842 49. Wang F. Long S. Zhang J. Moringa oleifera Lam. leaf extract safely inhibits periodontitis by regulating the expression of p38α/MAPK14-OPG/RANKL Arch. Oral Biol. 2021 132 105280 10.1016/j.archoralbio.2021.105280 34678605 50. Siddhuraju P. Becker K. Antioxidant properties of various solvent extracts of total phenolic constituents from three different agroclimatic origins of drumstick tree (Moringa oleifera Lam.) leaves J. Agric. Food Chem. 2003 51 2144 2155 10.1021/jf020444+ 12670148 51. Ngamukote S. Khannongpho T. Siriwatanapaiboon M. Sirikwanpong S. Dahlan W. Adisakwattana S. Moringa Oleifera leaf extract increases plasma antioxidant status associated with reduced plasma malondialdehyde concentration without hypoglycemia in fasting healthy volunteers Chin. J. Integr. Med. 2016 10.1007/s11655-016-2515-0 28035536 52. Costantini D. On the measurement of circulating antioxidant capacity and the nightmare of uric acid Methods Ecol. Evol. 2011 2 321 325 10.1111/j.2041-210X.2010.00080.x 53. Fombang N.E. BlaiseBouba B. Ngaruoa Management of Hypertension in Normal and Obese Hypertensive Patients through Supplementation with Moringa Oleifera Lam Leaf Powder Indian J. Nutr. 2016 3 3 7 54. Pigsborg K. Magkos F. Metabotyping for Precision Nutrition and Weight Management: Hype or Hope? Curr. Nutr. Rep. 2022 10.1007/s13668-021-00392-y 35025088 55. Zheng J.S. Ordovás J.M. Precision nutrition for gut microbiome and diabetes research: Application of nutritional n-of-1 clinical trials J. Diabetes 2021 13 1059 1061 10.1111/1753-0407.13220 34453774 56. Jardon K.M. Canfora E.E. Goossens G.H. Blaak E.E. Dietary macronutrients and the gut microbiome: A precision nutrition approach to improve cardiometabolic health Gut 2022 10.1136/gutjnl-2020-323715

PMC009xxxxxx/PMC9099675.txt

==== Front Int J Environ Res Public Health Int J Environ Res Public Health ijerph International Journal of Environmental Research and Public Health 1661-7827 1660-4601 MDPI 10.3390/ijerph19095000 ijerph-19-05000 Article Impact of the COVID-19 Pandemic on Tobacco Sales and National Smoking Cessation Services in Korea Kim Jinyoung 1 https://orcid.org/0000-0002-6419-2086 Lee Sungkyu 12* Berrigan David Academic Editor 1 Korea Center for Tobacco Control Research and Education, Seoul 06136, Korea; jy9651@daum.net 2 Department of Epidemiology and Health Promotion, Institute for Health Promotion, Graduate School of Public Health, Yonsei University, Seoul 03722, Korea * Correspondence: wwwvince77@gmail.com 20 4 2022 5 2022 19 9 500020 2 2022 14 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). This study aimed to describe the effect of the COVID-19 pandemic, combined with the Korean government’s response to the pandemic on tobacco consumption and national smoking cessation services among the Korean population. We obtained tobacco sale data from the Ministry of Finance and analysed the data on smokers’ visits to national smoking cessation clinics during the pandemic from a member of the National Assembly. We also conducted an online search to understand smokers’ thoughts about their tobacco use during the pandemic. We found that after the emergence of COVID-19 in 2020, the sale of conventional cigarettes increased from 3063.70 to 3209.70 million packs (4.77%). The number of smokers who visited clinics sharply decreased in the first half of 2020. The six-month quit rate decreased from 38.5% in 2017 to 22.3% in early 2020. We also found that smokers increased their consumption and began to switch from conventional cigarettes to heated tobacco products. The COVID-19 pandemic has threatened tobacco control policies and programs in Korea in the last two years; however, based on our experience during this period and considering the WHO recommendation, we should sustain and reinforce tobacco control policies and national smoking cessation services today and in the future. tobacco smoking tobacco sale smoking cessation COVID-19 South Korea ==== Body pmc1. Introduction During the first half of 2020, more than one million smokers in the United Kingdom (UK) successfully quit smoking because of the COVID-19 pandemic [1]. The pandemic provided an unprecedented opportunity for UK smokers to quit smoking. Before the COVID-19 pandemic, in 2019, 14.1% of people aged 19 years and above in the UK said they smoked cigarettes; however, after the pandemic, it dropped to 12.3% in 2020 [2]. In India and Italy, the sales of tobacco products were banned when the countries went into lockdown to prevent the spread of COVID-19, because the virus could reportedly spread while smoking [3]. In Hong Kong, the number of quitline participants has increased since the COVID-19 outbreak. Many participants (68%) in Hong Kong’s quitline did not realize that tobacco use potentially increased their risk of developing and spreading COVID-19; however, 43% agreed that the pandemic motivated their intention to quit, while 83% changed their smoking behavior during the pandemic [4]. Given the possible association between smoking and COVID-19 [5], the pandemic has provided an opportunity to reinforce tobacco control policies worldwide. The World Health Organization (WHO) emphasized, on the basis of the results of a review of studies convened by the WHO, that smokers are more likely to develop severe COVID-19 than non-smokers [6]. The WHO recommends that smokers take immediate steps to quit by using proven methods, such as quitline and nicotine replacement therapies, and requests all governments to sustain and reinforce tobacco control [7]. In the Republic of Korea (hereafter Korea), President Jae-in Moon raised the first issue related to COVID-19 and smoking. On 2 February 2020, a Korean evacuation plane returned to Seoul from Wuhan, China, and 360 evacuated Korean citizens were isolated at two different public facilities in Asan and Jincheon, the country’s central region. A week later, President Moon visited Jincheon to comfort the evacuees. During his visit, the Minister of Public Administration and Security, who served as the Minister of Health and Welfare in 2013, reported that smokers vehemently protested against not being allowed to smoke in the rooms of the smoke-free buildings, as well as not being allowed to go out of their rooms to smoke. President Moon addressed this issue by emphasizing the need to do the right thing and offered support to the smokers to quit smoking during such a difficult time [8]. The message from President Moon seemed to have provided a good opportunity to sustain tobacco control policies and comprehensive smoking cessation services during the pandemic. However, unlike President Moon’s thoughts about smoking during the pandemic, the Ministry of Health and Welfare (MOHW) frequently sent official documents to 254 public health centers that provided smoking cessation services all over the country and other national smoking cessation facilities, suggesting that all smoking cessation services be stopped, in order to focus on fighting the spread of COVID-19 [9]. Nevertheless, the recommendation from the WHO was to sustain and reinforce tobacco control policies and comprehensive smoking cessation services during the COVID-19 pandemic, in order to protect smokers, who are more vulnerable to COVID-19 [7]. On the basis of the evidence representing the association between COVID-19 and smoking, the MOHW announced that it included smokers in the populations most vulnerable to COVID-19 in the early stage of the pandemic [10]. However, the MOHW did not sustain smoking cessation services or help smokers quit during the pandemic. It is worth studying the impact of these actions by the MOHW on tobacco use in the Korean population during the pandemic. Hence, this study aimed to describe the effect of the COVID-19 pandemic, combined with the Korean government’s response to the pandemic on tobacco use among the Korean population, especially on tobacco consumption and national smoking cessation services. 2. Materials and Methods The ideal approach to examine the impact of the COVID-19 pandemic and MOHW’s response to the pandemic on tobacco sales or consumption and smoking behavior in Korea is to observe the changes in smoking prevalence before and after the pandemic. However, given that the data on smoking prevalence in 2020 had not yet been released, as of November 2021, we tried to obtain data from three different data sources and triangulate the information to achieve the research aim. Firstly, we considered tobacco sales data as an alternative source to determine the impact of COVID-19. Specifically, we visited the official website of the Ministry of Finance and obtained quarterly tobacco sales data for the 10-year period of 2011–2020, which was the longitudinal quantitative data. The data are publicly available at https://www.moef.go.kr/ (accessed on 1 April 2020). The tobacco sales data included the sales of conventional cigarettes and heated tobacco products (HTPs, since 2017). HTPs are treated as tobacco products, based on the tobacco definition in Tobacco Business Act [11]; thus, they are tightly regulated as conventional cigarettes in Korea. Secondly, in order to obtain the data on the number of smokers who visited national smoking cessation clinics in public health centers during the pandemic, we worked with Mr. Young-in Koh, a member of the National Assembly. The MOHW has a data collecting system for the national smoking cessation services. All the smoking cessation counselors who work for the smoking cessation clinic in public health centers should upload their records during the smoking cessation counselling. This data has been used to evaluate the national smoking cessation service; thus, they are official and accurate. This data has not been regularly opened to the public, and the MOHW normally releases this data to the public for their purpose or upon the request from news reporters or any members of the National Assembly. In these circumstances, we had contacted a member of the National Assembly. There is a particular system in which the National Assembly requests any kind of information and data from the government. Working with a member of the National Assembly, Mr. Koh’s office, we requested to the MOHW the number of visitors in smoking cessation clinics in public health centers between 2017 and 2020, and the six-month success rate of quitting, all of which can be important criteria for smoking cessation evaluation. Before the COVID-19 pandemic, the six-month success rate of quitting in smoking cessation clinics in public health centers was defined as carbon monoxide (CO) verified abstinence or self-reported abstinence from all tobacco products. This was based on point-prevalent abstinence. However, after the pandemic, the public health centers collected self-reported abstinence. In the context of the pandemic, the smoking cessation clinics were not able to conduct CO monitoring. Thirdly, we conducted an broad online search on the biggest internet portal in Korea, Naver (www.naver.com), by using keywords such as “COVID-19 and smoking”, as well as on Internet blogs that had any communication about the COVID-19 and smoking, to understand smokers’ thoughts about smoking and their tobacco use during the COVID-19 pandemic. Throughout the first week of October 2021, we searched news articles that were publicly released from April 2020 to September 2021, smokers’ comments from the news articles that we selected, and Internet blogs related to our research topics. We first reviewed the titles of the news articles and excluded duplicated articles, as well as the articles with titles that were not related to our research topics. Afterwards, we read the news articles and looked through the public comments located under the new articles. Anyone can add their comment to most new articles released on the Internet, and it is very common culture to attach their thoughts and opinion to new articles in Korea. This is a popular method to understand public opinion in Korea. This method was not a systematic and comprehensive search, and we tried to briefly collect public opinion related to the COVID-19 pandemic and smoking in Korea. Without any funding source, we were not able to conduct any former survey, which might be a better way to collect public views; thus, we decided to identify smokers’ views, related to the pandemic and their smoking behavior, with this method. 3. Results 3.1. Sales of Conventional Cigarettes and Heated Tobacco Products Figure 1 shows the sales of conventional cigarettes and HTPs in Korea from 2014 to 2020. The sales of conventional cigarettes decreased from 3663.60 million packs in 2016 to 3063.70 million packs in 2019; however, after the emergence of COVID-19 in 2020, it increased to 3209.70 million packs (4.77%). After an 80% tax increase on conventional cigarettes in 2015, the sale of conventional cigarettes decreased dramatically, and this continued until 2019, although there was rebound in the sale of conventional cigarettes between 2015 and 2016. HTPs were introduced in the Korean market in 2017, and the sale of HTPs increased rapidly from 78.7 million packs in 2017 to 332.0 million packs in 2018; the sales consistently increased in 2020. 3.2. Participants and Smoking Cessation Success Rates in National Smoking Cessation Services Figure 2 shows the number of participants and smoking cessation success rates for national smoking cessation services from 2017 to the first half of 2020. The number of smokers who visited national smoking cessation clinics in public health centers sharply decreased in the first half of 2020. In 2017, 424,636 smokers participated in smoking cessation clinics; however, in the first half of 2020, only 89,283 smokers visited clinics to quit smoking. In addition, the six-month success rate decreased from 38.5% in 2017 to 22.3% in the first half of 2020. Smokers who wanted to quit during the COVID-19 pandemic did not seek smoking cessation services, because such services were halted by the government. This resulted in a sharp decrease in the number of participants in smoking cessation services. In addition, the drop in the smoking cessation success rate from 38.5% in 2017 to 22.3% in early 2020 can be related to the COVID-19 pandemic. Since the pandemic, national smoking cessation services have only provided support by phone or through text messages, instead of conducting face-to-face consultations. 3.3. Smoking Behavior Change during the COVID-19 Pandemic We searched for relevant references on the biggest Internet portal in Korea, Naver, during the first week of September 2021, to examine smokers’ thoughts about smoking and COVID-19 and their smoking behavior change during the pandemic. We found 480 news articles with the keyword, “COVID-19 and smoking” and firstly excluded duplicated articles that had similar article titles. Then, we read the titles and also excluded the articles that were not related to our topics. With this process, we finally selected 25 news articles for the analysis. The articles mainly reported international research outcomes that informed smokers were more valuable to COVID-19 [12,13] and possible COVID-19 infection in smoking areas where smokers gathered for smoking [14,15]. Although there were a few public comments related to smoking behavior change during the COVID-19 pandemic, we found that smokers who had heard about the relationship between smoking and the severity of COVID-19 wanted to quit smoking and tried to visit smoking cessation clinics in public health centers [16]. We also found that smokers hesitated to enter smoking rooms or areas where they smoked every day during the COVID-19 pandemic. They worried that there were too many people in these areas and felt anxious about taking off their masks to smoke [17]. A comment we have found in the Internet blog said due to the COVID-19 pandemic, he or she smoked indoors without any concern of infection of COVID-19 [18]. With the possible spread of COVID-19, smokers tended to smoke indoors, so as to avoid smoking in public areas. We also found that since the emergence of COVID-19, Korean smokers have become more likely to increase their consumption. They have more time to smoke while working at home. Before the pandemic, they had to be mindful of other colleagues when they stepped out to smoke, but during the pandemic, they could smoke whenever they wanted to because they worked from home [19]. Consequently, many smokers may switch from conventional cigarettes to HTPs to reduce the smell of tobacco, which is a serious concern in indoor smoking. With this transition, new tobacco products, especially HTPs, have become popular among Korean smokers, and there was consistent increase of the HTP sales in 2020 (during the pandemic). Smokers use HTPs for indoor smoking and conventional cigarettes for outdoor smoking. Taking advantage of this change, the tobacco industry has been promoting HTPs, claiming that their brands have minimal odor and are suitable for indoor use. 4. Discussion The COVID-19 pandemic has affected the increase sales of both conventional cigarettes and HTP in Korea. In the absence of the COVID-19 pandemic, the sale of conventional cigarettes would have continually decreased, because smokers found it difficult to get an opportunity to smoke because of social norms that denormalized smoking. This increase in tobacco sales can be interpreted in two ways: Korean smokers might have smoked more during the pandemic, more people smoked, or fewer smokers tried to quit during the pandemic. Either way, the pandemic may have harmed tobacco control and public health in Korea. The MOHW released their mass media campaign with the message, “Smokers should quit during the COVID-19 pandemic” from April to May 2021 [20], because smoking was associated with increased severity of COVID-19. The campaign was widespread, and many smokers were exposed to it. However, smokers experienced difficulty in availing smoking cessation services. With the fight against COVID-19 as the top priority, the MOHW stopped public health centers around the country from providing national smoking cessation services. This may have led to fewer smokers attempting to quit or successfully quitting. Smokers who had heard about the relationship between smoking and severity of COVID-19 wanted to quit smoking and tried to visit smoking cessation clinics in public health centers. Unfortunately, the relevant services had been stopped, because the staff in public health centers changed their roles to COVID-19-related jobs. Unlike in Korea, in Hong Kong, the number of quitline participants has increased since the COVID-19 outbreak. Almost half of the participants said that the pandemic motivated their intention to quit, while 83% changed their smoking behavior during the pandemic [4]. The COVID-19 pandemic may serve as a potential tool to motivate smokers to quit smoking. They were concerned about their smoking behavior. For example, Korean smokers worried that there were too many smokers in smoking rooms or areas where they smoked every day, and they felt anxious about taking off their masks to smoke. The pandemic has, thus, presented a good opportunity to encourage smokers to participate in quitting. Therefore, we should prioritize sustaining and reinforcing smoking cessation services and tobacco control policies during the COVID-19 pandemic or when facing other types of infectious respiratory diseases in the future. The tobacco industry has been more active in using the COVID-19 pandemic for marketing HTPs. The industry claims that HTPs do not create any tobacco smoke or odor and they generate aerosols that are much less harmful than conventional cigarette smoke [21]. Given the situation wherein smokers are spending more time at home during the COVID-19 pandemic, they can be more interested in using new tobacco products, following the industry’s claims. However, it is important to mention that this type of smoking behavior causes dual use of conventional cigarettes and HTPs [22]. Smokers smoke conventional cigarettes when they are outside but smoke HTPs when they are at home or indoors. The government should closely monitor the tobacco industry’s marketing activities and educate smokers about the dangers of using HTPs use. Related to the tobacco industry’s activities during the pandemic, the ninth session of the Conference of the Parties (COP) to the WHO Framework Convention on Tobacco Control (FCTC), which was held on 8–13 November 2021 in Geneva, Switzerland, adopted a declaration in the WHO FCTC and recovery from the COVID-19 pandemic. The declaration calls on 182 FCTC parties to take appropriate measures to prevent tobacco industry interference and involvement in COVID-19-related public health policies and actions in accordance with Article 5.3 of the WHO FCTC and its guidelines for implementation [23]. There were strengths and limitations in this study. The use of longitudinal data from large, routinely collected datasets for the analysis of tobacco sales was a strength, however, there were some limitations, including the possibility of confounding variables impacting the tobacco sales, and the lack of currently available data on smoking prevalence in Korea. It was unclear how reliably sales of tobacco reflect real tobacco consumption and whether there was an illicit trade of tobacco. The data we used on the numbers of smokers visiting national smoking cessation clinics did not seem to be comparable, as data from each of the preceding three years were compared with data from the first half of 2020, and it might be difficult for the reader to meaningfully interpret this. Unfortunately, the MOHW did not provide the raw data on the numbers of smokers visiting national smoking cessation clinics to the member’s office of the National Assembly. However, there is a seasonal effect, whereby Korean smokers are more likely to visit smoking cessation clinics in the first half of the year than the other; thus, we believe that there was a huge reduction in the numbers of smokers visiting national smoking cessation clinics during the first half of 2020. The findings on smoking behavior change during the pandemic (from the Internet) did not represent the majority views of Korean smokers; thus, the finding should be carefully interpreted. In addition, as the findings were collected in the cross-sectional analysis, there was a limitation to infer real changes to smoking patterns at the population level. 5. Conclusions This study aimed to describe the effect of the COVID-19 pandemic, combined with the Korean government’s response to the pandemic on tobacco consumption and national smoking cessation services among the Korean population. We found that, after the COVID-19 pandemic, the tobacco sales increased and the number of smokers who visited smoking cessation clinics sharply decreased in the first half of 2020. We also found that smokers increased their tobacco consumption and began to switch from conventional cigarettes to HTPs. Consistent with Article 14 of the WHO FCTC, the ninth session of the COP to the WHO FCTC aimed to explore health system adaptations to support alternative service delivery options, such as e-health and telemedicine consultations, for tobacco dependence and cessation services [23]. Similar to the WHO’s recommendation, there were Internet- and telephone-based smoking cessation services for smokers who wanted to quit smoking during the pandemic in Korea. If they were effective during the pandemic, in the Korean context, they can be useful and valuable and, thus, need to be expanded further in the future. The COVID-19 pandemic has threatened tobacco control policies and programs in Korea in the last two years; however, based on our experience during this period and considering the COP declaration, we should sustain and reinforce tobacco control policies and national smoking cessation services today and in the future. Author Contributions J.K. and S.L. collected and analyzed the data. J.K. prepared the first draft of the manuscript. S.L. reviewed all of the drafts and helped prepare the final manuscript. All authors have read and agreed to the published version of the manuscript. Funding This research received no external funding. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement The data about tobacco sales in the Republic of Korea are publicly available at https://www.moef.go.kr/ (accessed on 1 April 2020). Conflicts of Interest The authors declare no conflict of interest. Figure 1 Sales of conventional cigarettes and heated tobacco products in Korea, 2014–2020 (source: Ministry of Finance). Figure 2 Numbers of smokers who visited national smoking cessation clinics in public health centers from 2017 to the first half of 2020 (source: Ministry of Health and Welfare). Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Action on Smoking and Health A Million People Have Stopped Smoking Since the COVID Pandemic Hit Britain Available online: https://ash.org.uk/media-and-news/press-releases-media-and-news/pandemicmillion/#:~:text=On%20the%20eve%20of%20a,during%20this%20period%20%5B1%5D (accessed on 3 September 2021) 2. Rabiya Nasir and Alison Brookman Smoking Prevalence in the UK and the Impact of Data Collection Changes: 2020 Office for National Statistics Available online: https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/drugusealcoholandsmoking/bulletins/smokingprevalenceintheukandtheimpactofdatacollectionchanges/2020 (accessed on 5 March 2022) 3. Ahluwalia I.B. Myers M. Cohen J.E. COVID-19 pandemic: An opportunity for tobacco use cessation Lancet 2020 5 E577 10.1016/S2468-2667(20)30236-X 4. Ho L. Li W. Cheung A.T. Xia W. Wang M.P. Cheung D. Lam T.H. Impact of COVID-19 on the Hong Kong Youth Quitline Service and Quitting Behaviors of Its Users Int. J. Environ. Res. Public Health 2020 17 8397 10.3390/ijerph17228397 33202764 5. World Health Organization WHO Statement: Tobacco Use and COVID-19 Available online: https://www.who.int/news/item/11-05-2020-who-statement-tobacco-use-and-covid-19 (accessed on 3 September 2021) 6. World Health Organization Smoking and COVID-19: Scientific Brief Available online: https://www.who.int/publications/i/item/WHO-2019-nCoV-Sci_Brief-Smoking-2020.2 (accessed on 3 September 2021) 7. World Health Organization Tobacco Control during the COVID-19 Pandemic: How We Can Help Available online: https://www.who.int/fctc/secretariat/head/statements/2020/tobacco-control-during-covid-19-pandemic/en/ (accessed on 3 September 2021) 8. Segye-Ilbo Evacuees Complain “No Smoking Area”, President Moon Recommends “Time to Quit Smoking during Such a Difficult Time” Available online: http://www.segye.com/newsView/20200209505480?OutUrl=naver (accessed on 15 August 2021) 9. Kukinews Public Health Centers Stop Smoking Cessation Services Available online: http://www.kukinews.com/newsView/kuk202103090318 (accessed on 15 August 2021) 10. Yonhapnews Government Includes Smokers as Vulnerable Group to COVID-19 Available online: https://www.yna.co.kr/view/AKR20200404045800017?input=1195m (accessed on 16 August 2021) 11. Ministry of Finance Chapter 1 General Provisions. Article 2 Definition. Tobacco Business Act (Amemded June 2020) Available online: https://www.law.go.kr/engLsSc.do?menuId=1&subMenuId=21# (accessed on 5 March 2022) 12. News One Smokers Have Twice the Death Rate from COVID-19 as Non-Smokers Available online: https://www.news1.kr/articles/?4315301 (accessed on 15 August 2021) 13. Yonhapnews Smoking Increases the Probability of COVID-19 among Teenagers by Five Times Available online: https://www.yna.co.kr/view/AKR20200812124800009?input=1195m (accessed on 16 August 2021) 14. Kukinews Is Smoking Is Priority over COVID-19? Available online: http://www.kukinews.com/newsView/kuk202101220081 (accessed on 16 August 2021) 15. Insight Korea Disease Control and Prevention Agency Considers Smoking Ban in the Indoor Smoking Room Available online: https://www.insight.co.kr/news/301065 (accessed on 16 August 2021) 16. No Smoke Guide Do you Know That There Is Close Relationship between Smoking and COVID-19? Available online: https://blog.naver.com/nosmokeguide9030/222507866381 (accessed on 3 September 2021) 17. Hungry Boarder There Was a Big Problem in the Smoking Area on the Top of the Skiing Course Available online: http://www.hungryboarder.com/index.php?mid=Free&document_srl=47621706 (accessed on 16 August 2021) 18. Republic of Korea Electronic Cigarette Pure User Café Staying Home Because of the COVID-19 Pandemic Available online: https://cafe.naver.com/sus3778?iframe_url=/MyCafeIntro.nhn%3Fclubid=25572652 (accessed on 16 August 2021) 19. Aju Business Daily COVID Blue. The Number of People Who Comfort Them with Alcohol and Cigarette Has Increase Available online: https://www.ajunews.com/view/20201123154152290 (accessed on 16 August 2021) 20. Yonhapnews MOHW Released the First Anti-Smoking Advertisement Available online: https://www.yna.co.kr/view/AKR20210330035300530?input=1195m (accessed on 16 September 2021) 21. The Korea Times Philip Morris Unveils Smoke-Free Cigarette in Korea Available online: https://www.koreatimes.co.kr/www/tech/2017/05/694_229533.html (accessed on 20 October 2021) 22. Kim J. Lee S. Kimm H. Lee J.A. Lee C.M. Cho H.J. Heated tobacco product use and its relationship to quitting combustible cigarettes in Korean adults PLoS ONE 2021 16 e0251243 10.1371/journal.pone.0251243 33961641 23. World Health Organization Declaration on WHO FCTC and Recovery from the COVID-19 Pandemic. The Ninth Session of Conference of the Parties to the WHO FCTC. FCTC/COP9(10) Available online: https://untobaccocontrol.org/downloads/cop9/decisions/FCTC_COP9_10_EN.pdf (accessed on 20 October 2021)

PMC009xxxxxx/PMC9099676.txt

==== Front Int J Environ Res Public Health Int J Environ Res Public Health ijerph International Journal of Environmental Research and Public Health 1661-7827 1660-4601 MDPI 10.3390/ijerph19095500 ijerph-19-05500 Article Resilience at Work, Burnout, Secondary Trauma, and Compassion Satisfaction of Social Workers Amidst the COVID-19 Pandemic Ratzon Anva 1* Farhi Moshe 2 https://orcid.org/0000-0003-0509-6059 Ratzon Navah 3 https://orcid.org/0000-0002-4358-0872 Adini Bruria 4 Tchounwou Paul B. Academic Editor 1 Independent Researcher, Tel Aviv 6997801, Israel 2 Social Work Department, Tel-Hai College, Qiryat Shemona 1220800, Israel; moshefar@telhai.ac.il 3 Occupational Therapy Department, School of Health Professions, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv 6997801, Israel; navah@tauex.tau.ac.il 4 Department of Emergency & Disaster Management, School of Public Health, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv 6997801, Israel; adini@tauex.tau.ac.il * Correspondence: anvaratzon@gmail.com 01 5 2022 5 2022 19 9 550024 2 2022 25 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Social workers during the COVID-19 pandemic are at risk due to exposure to varied populations in need, which may impact their resilience, burnout, secondary trauma, and compassion satisfaction. The study assessed resilience at work, burnout, secondary trauma, and compassion satisfaction among social workers in Israel during the first wave of the COVID-19 pandemic (May to June 2020). A convenience sample of 332 social workers (291 women (87.6%)) filled out an online, structured questionnaire that included demographics, a professional quality of life scale (ProQOL) (including three subscales), and resilience at work (RAW) (including seven subscales). The overall mean of the RAW was medium (M = 71, SD ± 8.9) compared to standardized scores. The mean scores of two of the subscales of the RAW, maintaining perspective and staying healthy, were low. The mean scores of the sub-scales of ProQOL were: compassion satisfaction was close to the 50th percentile (M = 48.25); burnout (M = 30.18) and secondary trauma (M = 26.27) were below the 25th percentile. Significant low to medium positive associations were found between all the dependent variables, except for staying healthy. A negative association was identified between compassion satisfaction and burnout, as well as between compassion satisfaction and secondary trauma. High levels of compassion satisfaction and contentment, low levels of secondary trauma, and having a managerial position were predicted to be 40% of the RAW. Lower levels of maintaining perspective, secondary trauma, and being younger predicted 27% of burnout. Higher levels of finding your calling, living authentically, maintaining perspective, interacting cooperatively, being older, and not being a manager predicted 58% of compassion satisfaction. Lower levels of burnout, maintaining perspective, and being younger predicted 36% of secondary trauma. As the COVID-19 pandemic still challenges most societies, policymakers should consider ways to integrate mechanisms that will enhance social workers’ resilience at work. work-resilience social workers COVID-19 burnout compassion satisfaction secondary trauma This research received no external funding. ==== Body pmc1. Introduction Social work is a practice-based profession aimed at empowering individuals and communities through the enhancement of social change. Although implemented in communities worldwide, great diversity can be identified concerning the types of practices adopted, the varied organizations in which it operates, the target populations, the services provided, and the types of interventions that are commonly used (Arazi et al., 2020) [1]. Social workers are required to assist vulnerable populations routinely, but their importance is accentuated in the current COVID-19 crisis. The European Social Workers Union has stated that social workers are vital frontline responders in the campaign to contain COVID-19, as they facilitate the capacity of communities to protect themselves and others through physical distancing and social solidarity [2]. The social services workforce in general, and social workers in particular, provide crucial services during COVID-19 for diverse populations, including children, people with disabilities, the elderly, and more. During COVID-19, social workers provided services not only to individuals who were under their care prior to the pandemic but also to a large range of previously unknown populations who were substantially affected by its restrictions. These included the elderly, people with health risks, and wide groups that were negatively affected by the resulting economic crises [3,4,5]. Considering the social distancing regulations, social workers need to apply innovative methods and skills to provide the required services to clients and communities. Such supporting services encompass the provision of moral and instrumental support, mitigation, assistance to enable individuals to access services or advocate for their needs, therapy to enhance resilience and well-being, and more [6]. The work of social workers is characterized by numerous challenges and constraints in routine and, even more so, during adversities. This profession suffers from a lack of manpower, resources, and the budget required to manage many caseloads, which affect the ability of social workers to provide optimal care. Furthermore, in the past years, social workers have been exposed to negative media and social media coverage, which affect the social image of the profession, at times even leading to their being violently attacked [1,7,8]. Although there is an increasing need for social services during crises, throughout the COVID-19 pandemic, allocated budgets did not change, and manpower was not expanded, yet the work was even more demanding [9,10]. Social workers put themselves and their loved ones at risk of infection while they continue to work as frontline responders, meeting clients and organizations despite very limited access to personal protective equipment [11]. To meet this risk, some of the organizations temporarily ceased to provide services or, alternatively, used distant communication, but these solutions were found to impact the service effectiveness negatively [12]. Treatment methodologies were required to adapt to new concerns of the clients, such as anxiety concerning the virus, rising unemployment rates, heightened financial difficulties, and, most noted, loneliness. The social workers themselves face a joint changing of reality with their clients and the uncertainty of the stressful situation, which may complicate their capacity to provide a calm and reassuring atmosphere. In addition, social workers face ethical challenges, especially concerning equity in service provision, as well as the restricted ability to assist in situations of end-of-life, as they are unable to allow family members to be by the side of their dying loved ones [11,13]. Social workers have reported being unable to provide services as expected according to their ethical codes, agency policy, or regulations, most especially due to varied restrictions [6,14]. This may substantially impact their resilience at work and levels of burnout, partly resulting from secondary trauma. Secondary trauma is the negative impact of indirect exposure to trauma on caregivers and therapists, such as social workers, who provide aid to patients after traumatic events. Caregivers and therapists are at risk of developing symptoms and reactions, such as suffering from post-traumatic stress disorder (PTSD) [15,16,17]. This risk is similar to other frontline workers, such as physicians and nurses, who were found to have a higher risk for mental health problems compared to the general population due to their exposure to COVID-19 patients [18]. Approximately 41% of healthcare workers were found to have symptoms of secondary trauma [19]. Ben-Porat [20] (2013) found a medium level of secondary trauma among social workers, with similar levels among those who work in domestic violence prevention centers and/or in welfare departments. Dagan et al. [17] (2016) found that over 50% of social workers working in the domain of child protection indicated that they have a high to severe level of secondary trauma. In another study that examined levels of secondary trauma among 412 social workers, low to moderate levels of secondary trauma were identified [21]. All the aforementioned studies used the secondary traumatic stress scale questionnaire [22], which includes 17 items, measuring intrusion, avoidance, and arousal symptoms associated with indirect exposure to traumatic events. A study of a sample of 181 social workers conducted in the United States using the ProQOL scale during the COVID-19 pandemic found that approximately 50% of the respondents reported signs of secondary trauma at a medium level [23]. In a qualitative study conducted during the COVID-19 pandemic in California, in which 34 Master’s students of the Social Work Program were interviewed, most of the respondents expressed themes related to secondary trauma in their workplace [24]. Burnout is related to work conditions and a lack of coping resources. The symptoms of burnout can be presented as emotional (depression, compassion fatigue, and secondary trauma) or physical exhaustion (varied health conditions or illnesses). Burnout among social workers has been extensively studied, presenting high correlations between the social work profession and burnout syndromes, which are most frequently caused by it being a high-stress occupation [25,26,27]. Beyond the personal effects of burnout on social workers, it has also been shown that burnout negatively impacts professional functioning and leads to lower levels of quality of care for patients [28]. Furthermore, there are additional risk factors that contribute to the burnout of social workers derived from their low-income profession, which is characterized by a high level of workload, short deadlines, low appreciation, and limited resources, both for the social workers as service providers as well as for their organizations [29,30]. Conversely, enhancing psychological capital may serve as a protective measure for social workers against burnout and secondary trauma [31]. During the COVID-19 pandemic, the levels of stress and burnout increased for varied reasons, including work-related conditions [32]. Although many researchers have suggested that social workers are at a high risk of burnout [8,33], only a few studies have actually measured these levels during the COVID-19 pandemic. One such study, conducted in Spain using the Maslach Burnout Inventory [34], identified high levels of emotional exhaustion, depersonalization, and a reduced sense of personal accomplishment (all the constructs that demonstrate burnout) among social workers. Almost 70% of the social workers considered applying for psychotherapy due to COVID-19 [35]. In contrast, Holmes et al. [23] (2021) used the ProQOL scale and found that among 181 social workers screened, 63.71% expressed medium levels of burnout (M = 25.629, SD = 6.29, range 11–40). Although social workers are at risk of secondary trauma and burnout as a negative reaction to their work, they can also experience compassion satisfaction. Compassion satisfaction signifies positive feelings perceived by caregivers and service providers derived from helping people in need. Research has shown that compassion satisfaction is a protective occupation factor for varied professionals, such as social workers, therapists, and emergency responders. Compassion satisfaction can protect from stress, burnout, and secondary trauma [36,37,38]. In research conducted in Norway among child protection social workers, most of the workers were found to be moderately satisfied with their work [38]. Senreich et al. [39] (2020) assessed 6112 social workers from 13 different states in the United States and found that nearly 60% of them reported high levels of compassion satisfaction. They also found that the workplace environment’s wellness contributed to promoting compassion satisfaction. Similarly, during the COVID-19 pandemic, 99% of the social workers that were assessed reported a high level of compassion satisfaction [23]. Despite an extensive literature review, no additional studies on this subject were found. As delineated, emotional and physical work experience and conditions can lead to burnout and secondary trauma but, in contrast, may also result in compassion satisfaction [37,40]. A negative correlation was found between secondary trauma and burnout compared to compassion satisfaction [38,40,41]. For example, Staudt and Williams [42] (2019) found that members of a child advocacy interdisciplinary team expressed lower levels of secondary trauma and burnout and had higher levels of compassion satisfaction. Craig and Sprang [43] (2010), who assessed a sample of 532 clinical psychologists and social workers, also found a negative association between burnout and secondary trauma to compassion satisfaction and a positive relationship between burnout and secondary trauma. It has been suggested that resilience at work and compassion satisfaction are personal resources (the perception of the individuals that they are able to manage complex situations successfully) and, therefore, contribute toward lower levels of burnout and secondary trauma [44,45]. Two models have been proposed to explain the relationships between these four variables: the Job Demands-Resources Model (JD-R), which posits that work resources (such as participatory engagement in decision-making) contribute toward resilience, while work demands (such as high workloads) may lead to burnout [46]; and the Conservation of Resources (COR) model, which posits that individuals continuously strive to maintain their resources, but when the resources are nearing depletion, they lead to stress, secondary trauma, and burnout [47]. Resilience may be perceived as a personal resource or as a mediator for compassion fatigue and burnout [48]. Age and gender have both been found to impact workplace resilience, as well as levels of burnout [49,50]. Numerous studies have shown that males, as well as older personnel, report higher levels of workplace resilience compared to women and younger staff members [51,52]. The aim of the study was to assess the levels of resilience at work, burnout, secondary trauma, and compassion satisfaction among social workers in Israel during the COVID-19 pandemic and to identify associations between these variables. The hypotheses of the study were as follows: 1. Social workers will present low levels of resilience at work and compassion satisfaction during COVID-19; 2. Women will display lower levels of resilience at work and higher levels of burnout compared to men. 3. Resilience at work will be positively associated with compassion satisfaction and negatively associated with burnout and secondary trauma. 2. Materials and Methods 2.1. Study Design The first COVID-19-confirmed patients arrived in Israel in February 2020 from the infected ship, the Princess Diamond, and by March 2020, cases of community infectivity were identified. Since then, four waves of the pandemic occurred, resulting in the morbidity of 1,354,001 confirmed cases and 8230 deaths (MoH, 2021, updated 17 December 2021). The study was conducted quantitatively among social workers in Israel in the midst of the COVID-19 pandemic from May to June 2020, which was during the first wave of the COVID-19 pandemic. The respondents constitute a convenience sample recruited by a snowball approach through social media by inviting social workers to respond to the online questionnaire. The study was approved by the ethics committee of Tel Aviv University (#0001431-1 from 6 May 2020). All the participants signed an electronic informed consent form before joining the research and only then completed the questionnaire. 2.2. Study Population A convenience sample of 332 social workers participated in the study. In order to be included in the study, participants had to be certified social workers with at least one year of experience. They were recruited through social media (by posts published on Facebook and in social workers’ WhatsApp groups). 2.3. Study Tool The study tool was a structured questionnaire distributed online through Qualtrics software. The questionnaire included three parts: (1) Bio-demographic characteristics that included 9 items [age, gender, family, occupational status, type of fieldwork, income, contentment (self-report scale, 1–5), supervision (self-report concerning the social worker’s access to meetings with professional supervisors; dichotomic answer, yes\no), and management position (self-report dichotomic answer, yes\no)]; (2) Professional Quality of Life (ProQOL) scale in the employment realm [36]. This validated scale contains 30 items that incorporate two aspects: a positive scale (compassion satisfaction) and a negative scale called compassion fatigue (the negative scale has two categories—burnout and secondary trauma). Compassion fatigue includes feelings such as exhaustion, frustration, anger, and depression, which are typical of burnout. An example of the burnout category is: “I am not as productive at work because I am losing sleep over traumatic experience of people I help;” for the secondary trauma category: “I am preoccupied with more than one person [I help];” and an example for the compassion satisfaction category is: “I get satisfaction from being able to help people.” The ProQOL scale was used in Israel in a study with educational counselors [53]. The internal reliability of the general score was α = 0.78, burnout α = 0.79, secondary trauma α = 0.82, and compassion satisfaction α = 0.86. In another study conducted among palliative personnel, the internal reliability values were burnout α = 0.75, secondary trauma α = 0.81, and compassion satisfaction α = 0.88 [54].; and, finally, (3) Resilience at work (RAW)—a tool developed by Winwood et al. [55] (2013) to measure resilience at work for use in individual work-related performance and emotional distress contexts. The questionnaire has 7 sections: living authentically (3 items), finding your calling (4 items), maintaining perspective (3 items), managing stress (4 items), interaction cooperation (2 items), staying healthy (2 items), and building networks (2 items). The authors found that the questionnaire fits the study data, as evidenced by the fit statistics (i.e., goodness-of-fit index = 0.968; Tucker–Lewis index = 0.975; root mean square error of approximation = 0.038). 2.4. Data Analysis SPSS software (version 27, Chicago, IL, USA) was used to analyze the data. As all dependent variables did not have a normal distribution, we used the Spearman correlation coefficient to examine correlations between the dependent variables—resilience at work, burnout, secondary trauma, and compassion satisfaction. In order to identify the predictors of each of the dependent variables, we first conducted a log transformation using SPSS with conversion. The prediction of each dependent variable was investigated by entering the equation variables that were found to significantly correlate with it, which included the other dependent variables (those found significantly correlated with the specifically tested dependent variable) and the bio-demographic and professional characteristics (contentment, supervision, and managing role). Scores of 61 to 80 were considered average levels of the overall resilience at work [55]. Scores of 22 or less, between 23 to 41, and 42 or more for each of the three sub-scales of the ProQOL tool, were considered as low, average, and high scores, respectively [53]. p-values lower than 0.05 were considered to be statistically significant. We reported the level of the correlation coefficient based on the ranking proposed by Cohen [56] )1988(. 3. Results The respondents in the study ranged from 26 to 70, with a mean age of 41.92 (SD = 10.97). The sample included 291 women and 40 men. The mean years of experience was 13.82 (SD = 10.3), ranging from 1 to 47 years of experience. The mean score of contentment was 3.30 (SD = 0.954), ranging between 1 to 5. Additional characteristics of the study sample are presented in Table 1. 3.1. Levels of Resilience at Work, Burnout, Compassion Satisfaction, and Secondary Trauma The mean levels of resilience at work, burnout, compassion satisfaction, and secondary trauma were calculated. The first hypothesis was partially rejected. The overall mean level of the RAW among the social workers was found to be medium (M = 71, SD ± 8.9), as compared to the standardized scores with normative values [55]. When analyzing the varied components of the RAW, a diverse picture emerges. The mean scores for living authentically, managing stress, interacting cooperatively, and building networks were found to be higher than the normative values (+0.5–1, SD above mean). In contrast, the mean score for finding your calling was similar to the normative values of this scale (−0.5–1, SD of mean), while the mean scores of maintaining perspective and staying healthy were below the average of the normative score (−1–0.5, SD below mean). The subscales of maintaining perspective and staying healthy were in line with the hypothesis. The average compassion satisfaction score was very close to the 50th percentile (M = 48.25); the mean scores for burnout (M = 30.18) and secondary trauma (M = 26.27) were below the 25th percentile (43 and 42, respectively). The full results are presented in Table 2. The second hypothesis was partially verified. No significant differences were found in the average scores between genders concerning RAW, burnout, and compassion satisfaction, except for two sub-components of the RAW scale, as follows: the mean scores for females were significantly lower compared to those for males concerning finding your calling (M = 82.64, SD = 11.7 versus M = 86.87, SD = 8.3, respectively; Z = −2.235 and p = 0.025) and maintaining perspective (M = 35.28, SD = 14.28 versus M = 42.22, SD = 18.9, respectively; Z = −2.465 and p = 0.014). An opposite trend was found concerning the secondary trauma score, whereas the females’ scores were higher than those of the males’ (M = 26.76, SD = 7.54 versus M = 22.7, SD = 6.49, respectively; Z = −3.586 and p = 0.001). 3.2. Correlations between RAW, Burnout, Compassion Satisfaction, and Secondary Trauma The third hypothesis was verified. As expected, significant correlations were found among the varied components of resilience at work, except for staying healthy, which was significantly associated only with managing stress. More interesting to note are the significantly low to medium correlations between compassion satisfaction and all the other variables, except for staying healthy. A negative association was identified between compassion satisfaction and burnout, as well as between compassion satisfaction and secondary trauma. All the other associations between compassion satisfaction and the overall RAW score, as well as with its components (as stated, except for staying healthy), were positive. Burnout demonstrated significantly low to medium negative correlations for three (out of seven) components of the RAW scale, as well as with the overall RAW score. A significant positive association was found between burnout and secondary trauma. Secondary trauma demonstrated significantly low to medium negative correlations for all components of the RAW. Table 3 presents the correlations between the study variables. 3.3. Prediction of Resilience at Work, Burnout, Compassion Satisfaction, and Secondary Trauma A regression analysis was conducted to identify the factors that can predict the four study variables, including both demographic characteristics in each model as well as the varied dependent variables. The analysis revealed that high levels of compassion satisfaction, low levels of secondary trauma, high levels of contentment, and having a managerial position predicted 40.5% of the overall score of resilience at work (R2 = 0.405, with p < 0.001). Lower levels of maintaining perspective, secondary trauma, and being younger predicted 27% of burnout (R2 = 0.270, with p < 0.001). Higher levels of interacting cooperatively, finding your calling, living authentically, maintaining perspective, being older, and not being a manager predicted 58% of compassion satisfaction (R2 = 0.583, with p < 0.001). Being younger is negatively associated with secondary trauma. Higher levels of burnout, lower levels of being able to manage stress, and maintaining perspective predicted 35.9% of secondary trauma (R2 = 0.359, with p < 0.001). See Table 4, Table 5, Table 6 and Table 7. 4. Discussion The COVID-19 pandemic presented a great challenge to many professionals who needed to maintain direct contact with other individuals, amongst them social workers [2]. Similar to other frontline responders, social workers were required to provide services to vulnerable populations that were severely impacted by both the risk of contracting the virus as well as the protective measures that were adopted, such as social distancing, lockdowns, and more, which enhanced their perceived isolation and loneliness [6]. The expansion of the target populations that they had to serve, in conjunction with the personal and professional challenges caused by the pandemic, led to increased stress [11,14]. In contrast to the first hypothesis, which posited that the resilience at work among social workers during the COVID-19 pandemic would be low, the study findings displayed a medium level of resilience. Social work is a helping profession, providing the social workers, as well as other professionals such as medical personnel, psychologists, or other caretakers, with a sense of accomplishment and compassion satisfaction [57,58]. This enhances the RAW, solidarity among responders, and a heightened sense of involvement in an important and fulfilling position, which are all important work resources [47,48]. These positive perceptions are similar to those of other helping professions, such as the varied medical professions, which are also characterized by increased levels of stressful events [59]. This may explain the finding that, despite the stressors of the COVID-19 pandemic, the overall mean level of resilience at work among the social workers was found to be medium. The seemingly unchanged level of resilience at work may reflect the ongoing stressors that social workers continuously manage during their work (in both routine and during adversities), which paradoxically act as protective means [57,60]. Furthermore, the national and global collaboration to contain the COVID-19 pandemic may have served as psychological capital, raising hope, optimism, and resilience, which contributed to protecting social workers from the stressors of the pandemic [31]. The second hypothesis that posited a lower level of resilience at work during COVID-19 among women was verified. This tendency may be derived from contextual or cultural perceptions of the varied roles of females versus males or result from the overall tendency of men to report higher levels of resilience compared to women. The gender differences that were found concerning finding your calling and maintaining perspective, in which females had lower levels of resilience compared to males, have also been identified in other societies and professions. Conversely, women tended to have higher levels of secondary trauma compared to males. This tendency has also been observed in the other studies [61,62,63]. In line with the third hypothesis, resilience at work was found, during the COVID-19 pandemic, to be positively associated with compassion satisfaction and negatively associated with burnout and secondary trauma. These findings are in line with both the JD-R and the COR theories, which have delineated the association between personal as well as work resources and work demands [46,47]. The depletion of resources leads to a decrease in resilience and, subsequently, to an increase in negative reactions, such as secondary trauma, stress, and burnout [48]. The analysis revealed that high levels of compassion satisfaction, low levels of secondary trauma, high levels of contentment, and having a managerial position predicted 40% of the overall score of resilience at work. Contentment, compassion satisfaction, and perceived well-being were presented in the previous studies concerning other psychotherapeutic professions as associated with resilience [64,65]. As reported previously, concerning other adversities [38], a negative association was identified during the pandemic between compassion satisfaction and burnout, as well as between compassion satisfaction and secondary trauma. All the other associations between compassion satisfaction and the overall RAW score, as well as with its components (except for staying healthy), were positive. This is in line with the previous findings that illustrated that practitioners who presented positive perceptions of their profession presented higher levels of compassion satisfaction and lower levels of secondary trauma [63,64]. Not surprisingly, burnout demonstrated significantly low to medium negative correlations with the overall RAW score. Similarly, secondary trauma demonstrated a significantly low to medium negative correlation with all the components of the RAW, except for interacting cooperatively. Lower levels of resilience at work and higher levels of secondary trauma and burnout have been shown to decrease the quality of care that social workers provide to their clients and weaken the therapeutic results [28,60,64,66]. The self-care of social workers is, thus, crucial in order to prevent burnout and secondary trauma and to maintain resilience at work [57,63]. As demonstrated in the other studies, a significant positive association was found during the COVID-19 pandemic between burnout and secondary trauma [40,41]. As previously published, it is vital to assess the subjective well-being and levels of burnout of therapists, as these impact not only their work resilience but also the results of their therapeutic work [28]. In contrast to the previous findings, seniority (number of years in the workplace) in social work during the COVID-19 pandemic was not found to predict RAW and burnout [67,68]. This may be derived from the innovative nature of the COVID-19 pandemic. It is a novel experience for all social workers, regardless of their years of experience. Furthermore, we conjecture that, though senior social workers were found to have lower levels of burnout in routine compared to more junior ones, as COVID-19 posed a higher health risk to the older ones, this may have impacted their concerns for their own well-being and, thus, contributed to a higher perceived threat and burnout symptoms. There are two main theoretical implications of this study: The first is the centrality of the variable RAW and its impact on burnout, secondary trauma, and compassion satisfaction. To the best of our knowledge, the research concerning this issue is scarce. The second is that ongoing work demands, most especially loaded and stressful work conditions, may actually serve, during severe adversities, as a protective measure against a drastic decrease in resilience at work and compassion satisfaction. This conjecture should be further evaluated to understand whether it can be identified concerning other professions and/or other types of adversities. 5. Limitations and Future Research Three main limitations should be noted regarding the study. The study was conducted as cross-sectional research and, thus, it is not possible to make any causal inferences. The use of convenience sampling does not enable us to make generalizations of the results of the study to the overall social work population due to the possibility of bias and the under- or over-representation of specific groups of social workers. The third limitation is that data were collected at a one-time frame, which may reflect specific (and temporary) feelings and perceptions that each respondent felt, regardless of COVID-19. This might affect the exactness of the research conclusions. Further studies are warranted to build an evidence-based body of knowledge that will focus on the centrality of RAW and its impact on burnout, secondary trauma, and compassion satisfaction. We recommend that prospective longitudinal studies be conducted, which will facilitate follow-up over time, cluster sampling, and in-depth monitoring of the association between resilience at work and burnout among frontline workers. 6. Conclusions Social workers are vital frontline workers in both routine times and during adversities; they cater to the needs of the most vulnerable sects of the population. Considering their importance, it is imperative that social workers be part of the overall response to varied types of emergencies. Nonetheless, ensuring the effective participation of social workers in the overall response to any type of known or emerging threat necessitates empowering them to enhance their resilience at work and mitigate their levels of burnout. As the COVID-19 pandemic still challenges most societies and, furthermore, additional adversities (both human-made or those resulting from climate change or other natural events) are expected, policymakers should consider ways to integrate ongoing mechanisms that will enhance the resilience at work of all therapeutic professionals, including those of social workers. These measures include better work conditions, higher availability of resources, and extended budgets [9,10]. Considering the positive association found between resilience at work and compassion satisfaction, investing efforts in the well-being of social workers is expected to substantially enhance their capacity to contribute toward a holistic emergency response to varied crises. In order to generalize the findings of this study to varied frontline personnel, longitudinal studies should be conducted to identify commonalities and diversities concerning the well-being, workplace resilience, and burnout among different healthcare and welfare workers. Author Contributions Conceptualization, A.R., B.A. and M.F.; methodology, A.R., B.A. and N.R.; formal analysis, A.R. and N.R.; investigation, A.R.; resources, A.R. and B.A.; writing—original draft preparation, A.R. and B.A.; writing—review and editing, M.F. and N.R; supervision, M.F. and B.A. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the Tel Aviv University (#0001431-1 from 6 May 2020). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The data accumulated in the study are kept by the authors. Data are not published openly due to privacy issues, but the analyzed data are available from the authors upon request. Conflicts of Interest The authors declare no conflict of interest. ijerph-19-05500-t001_Table 1 Table 1 Characteristics of the study population (n = 332). Characteristic Percentage (n) Status of employment Salaried employee 93.1% (309) both salaried and self employed 2.4% (8) Supervision (ongoing) Yes 98.7% (228) No Management status Yes 39.5% (131) No Marital status Married 77.4 (257) Single 14.2 (47) Divorced 6.9 (23) Widowed 1.5 (5) Personal income below average 33.1 (110) close to average 53.9 (179) above average 6 (20) Household income below average 33.1 (110) close to average 45.2 (150) above average 10.5 (35) ijerph-19-05500-t002_Table 2 Table 2 Average scores of resilience at work and ProQOL scale; (n = 332).* Variable Name Mean Standard Deviation (SD) Minimum Score Maximum Score Resilience at Work Living authentically 86.31 8.35 38.89 100 Finding your calling 83.19 11.42 33.33 100 Maintaining perspective 36.12 15.03 5.56 72.22 Managing stress 69.32 16.78 16.67 100 Interacting cooperatively 77.96 13.3 25 100 Staying healthy 61.49 22.07 16.67 100 Building networks 79.92 17.7 16.67 100 Overall RAW 70.74 8.39 40.83 89.17 Professional Quality of Life scale Burnout 30.18 5.93 16 51 Compassion satisfaction 48.25 7.06 22 60 Secondary trauma 26.27 7.52 12 51 * Standard scores of 61 to 80 are considered as ‘average’ RAW scores [55]; standard scores of 23 to 41 are considered as ‘average’ ProQOL scores in each of the three sub-scales [53]. ijerph-19-05500-t003_Table 3 Table 3 Spearman coefficient correlation between all dependent variables (n = 332). Compassion Satisfaction Burnout Secondary Trauma Living Authentically Finding Your Calling Maintaining Perspective Managing Stress Interacting Cooperatively Staying Healthy Building Networks Compassion satisfaction Burnout −0.15 ** Secondary trauma −0.32 ** 0.37 ** Living authentically 0.54 ** 0.08 −0.20 ** Finding your calling 0.64 ** −0.16 ** −0.28 ** 0.44 ** Maintaining perspective 0.37 ** −0.38 ** −0.50 ** 0.20 ** 0.29 ** Managing stress 0.32 ** −0.24 ** −0.33 ** 0.28 ** 0.19 ** 0.41 ** Interacting cooperatively 0.26 ** −0.04 −0.07 0.19 ** 0.17 ** 0.11 * 0.20 ** Staying healthy 0.08 −0.04 −0.13 * 0.03 0.02 0.08 0.23 ** −0.07 Building networks 0.33 ** −0.05 −0.13 * 0.17 ** 0.42 ** 0.16 ** 0.22 ** 0.20 ** 0.16 ** RAW 0.57 ** −0.28 ** −0.45 ** 0.46 ** 0.56 ** 0.63 ** 0.75 ** 0.35 ** 0. 41 ** 0.53 ** * p ≤ 0.05, ** p ≤ 0.01 ijerph-19-05500-t004_Table 4 Table 4 Regression analysis; dependent variable: RAW, R 2 = 0.405, p < 0.001. Variable B SE B β Sig Age 0.000 0.000 −0.066 NS Gender 0.007 0.007 0.045 NS Contentment 0.009 0.003 0.152 0.002 Supervision 0.002 0.004 0.020 NS Management position 0.012 0.005 0.107 0.020 Income −0.001 0.003 −0.001 NS Professional status −0.006 0.006 −0.045 NS Compassion Satisfaction 0.356 0.038 0.469 0.001 Burnout −0.040 0.030 −0.065 NS Secondary trauma −0.099 0.022 −0.225 0.001 B = unstandardized B, SE B = coefficients Std Error, β = standardized coefficients Beta, Sig = Significant. ijerph-19-05500-t005_Table 5 Table 5 Regression analysis; dependent variable: burnout, R2 = 0.270, p < 0.001. Variable B SE B β Sig Age −0.001 0.000 −0.127 0.021 Contentment −0.008 0.005 −0.087 NS Supervision 0.012 0.007 0.077 NS Management position −0.004 0.009 −0.020 NS Income 0.005 0.006 −0.048 NS Compassion Satisfaction 0.026 0.081 0.021 NS Secondary trauma 0.150 0.041 0.211 0.001 Finding your calling −0.054 0.080 −0.402 NS maintaining perspective −0.078 0.024 −0.194 0.002 Managing stress −0.071 0.037 −0.103 NS Sig = Significant. ijerph-19-05500-t006_Table 6 Table 6 Regression analysis; dependent variable: compassion satisfaction, R2 = 0.583, p < 0.001. Variable B SE B β Sig Age 0.001 0.000 0.094 0.024 Contentment 0.004 0.003 0.049 NS Management position −0.013 0.006 −0.092 0.018 Income 0.002 0.004 0.022 NS Burnout −0.035 0.035 −0.042 NS Secondary trauma −0.018 0.026 −0.031 NS Living authentically 0.487 0.067 0.316 0.001 Finding your calling 0.381 0.047 0.368 0.001 maintaining perspective 0.038 0.015 0.117 0.012 Managing stress −0.009 0.024 −0.016 NS Interacting cooperatively 0.118 0.035 0.139 0.001 Building networks 0.005 0.023 0.009 NS Sig = Significant. ijerph-19-05500-t007_Table 7 Table 7 Regression analysis; dependent variable: secondary trauma, R2 = 0.359 p < 0.001. Variable B SE B β Sig Age −0.002 0.001 −0.192 0.041 Contentment −0.003 0.007 −0.022 NS Years of experience 0.017 0.010 0.001 NS Supervision 0.017 0.010 0.001 NS Management position 0.011 0.012 0.043 NS Compassion satisfaction −0.094 0.123 −0.055 NS Burnout 0.233 0.077 0.167 0.003 Living authentically −0.060 0.161 −0.023 NS Finding your calling −0.126 0.112 −0.072 NS maintaining perspective 0.183 0.032 −0.328 0.001 Managing stress −0.142 0.053 −0.145 0.008 Interacting cooperatively 0.058 0.080 0.040 NS Staying healthy −0.033 0.031 −0.053 NS Building networks 0.096 0.050 0.102 NS Sig = Significant. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Arazi T. Moran-Gilad L. Namer-Furstenberg R. The Professional Status of Social Workers: International Review Available online: https://brookdale.jdc.org.il/en/publication/the-professional-status-of-social-workers-international-review/ (accessed on 3 February 2022) 2. Truell R. News from Our Societies—IFSW: COVID-19: The Struggle, Success and Expansion of Social Work—Reflections on the Profession’s Global Response, 5 Months On Int. Soc. Work 2020 63 545 548 10.1177/0020872820936448 3. Brown N. Social Service Workers Mitigating the Impact of COVID-19 Available online: http://socialserviceworkforce.org/resources/blog/social-service-workers-mitigating-impact-COVID-19 (accessed on 3 February 2022) 4. Ben-Ezra M. Hamama-Raz Y. Social Workers during COVID-19: Do Coping Strategies Differentially Mediate the Relationship between Job Demand and Psychological Distress? Br. J. Soc. Work 2021 51 1551 1567 10.1093/bjsw/bcaa210 5. Maor M. Sulitzeanu-Kenan R. Chinitz D. When COVID-19, Constitutional Crisis, and Political Deadlock Meet: The Israeli Case from a Disproportionate Policy Perspective Policy Soc. 2020 39 442 457 10.1080/14494035.2020.1783792 35039730 6. Barsky A. Ethical Exceptions for Social Workers in Light of the COVID-19 Pandemic and Physical Distancing New Soc. Work. 2020 27 4 7. Gal J. About the crisis with the face forward Meidaos 2020 92 4 11 8. Peinado M. Anderson K.N. Reducing social worker burnout during COVID-19 Int. Soc. Work 2020 63 757 760 10.1177/0020872820962196 9. Shlomo S.B. Levin-Keini N. Once in a Hundred Years: Does COVID-19 Present an Opportunity to Restructure the Professional Image of the Social Worker in Israel? Br. J. Soc. Work 2021 51 1568 1584 10.1093/bjsw/bcab055 10. Schwartz-Tayri T.M. On the frontlines: The impact of the COVID-19 pandemic on social workers’ well-being Br. J. Soc. Work 2022 52 1724 1746 11. Abrams L. Dettlaff A. Voices from the frontlines: Social workers confront the covid19 pandemic Soc. Work 2020 65 302 305 10.1093/sw/swaa030 32642752 12. Danilovich M. Norrick C. Lessem R. Milstein L. Briggs N. Berman R. Responding to COVID-19: Lessons Learned from a Senior Living and Social Service Organization Geriatrics 2020 5 98 10.3390/geriatrics5040098 13. Raftery C. Lewis E. Cardona M. The crucial role of nurses and social workers in initiating end-of-life communication to reduce overtreatment in the midst of the COVID-19 pandemic Gerontology 2020 66 427 430 10.1159/000509103 32516766 14. Banks S. Cai T. de Jonge E. Shears J. Shum M. Sobocan A.M. Strom K. Truell R. U´riz M.J. Weinberg M. Practicing ethically during COVID-19: Social work challenges and responses Int. Soc. Work 2020 63 569 583 10.1177/0020872820949614 15. Figley C.R. Systemic traumatization: Secondary traumatic stress disorder in family therapists Integrating Family Therapy: Handbook of Family Psychology and Systems Theory Mikesell R.H. Lusterman D.D. McDaniel S.H. American Psychological Association Washington, DC, USA 1995 571 581 10.1037/10172-033 16. Vagni M. Maiorano T. Giostra V. and Pajardi D. Hardiness, stress and secondary trauma in Italian healthcare and emergency workers during the COVID-19 pandemic Sustainability 2020 12 5592 10.3390/su12145592 17. Dagan S.W. Ben-Porat A. Itzhaky H. Child protection workers dealing with child abuse: The contribution of personal, social and organizational resources to secondary traumatization Child Abus. Negl. 2016 51 203 211 10.1016/j.chiabu.2015.10.008 18. Arpacioglu S. Gurler M. Cakiroglu S. Secondary traumatization outcomes and associated factors among the health care workers exposed to the COVID-19 Int. J. Soc. Psychiatry 2021 67 84 10.1177/0020764020940742 32635786 19. Orrù G. Marzetti F. Conversano C. Vagheggini G. Miccoli M. Ciacchini R. Gemignani A. Secondary traumatic stress and burnout in healthcare workers during COVID-19 outbreak Int. J. Environ. Res. Public Health 2021 18 337 10.3390/ijerph18010337 20. Ben-Porat A. Predicting Factors of Secondary Traumatization: Comparison Between Domestic Violence Therapists and Welfare Bureau Therapists Soc. Welf. 2013 33 279 301 21. Cappiccie A. Sullivan D.J. Hyppolite M. Front-line social work jobs and secondary trauma: Impact on parenting J. Hum. Behav. Soc. Environ. 2020 30 892 904 10.1080/10911359.2020.1764897 22. Bride B.E. Robinson M.M. Yegidis B. Figley C.R. Development and validation of the secondary traumatic stress scale Res. Soc. Work Pract. 2004 14 27 35 10.1177/1049731503254106 23. Holmes M.R. Rentrope C.R. Korsch-Williams A. King J.A. Impact of COVID-19 Pandemic on Posttraumatic Stress, Grief, Burnout, and Secondary Trauma of Social Workers in the United States Clin. Soc. Work J. 2021 49 495 504 10.1007/s10615-021-00795-y 33678920 24. Parra P.A. Impact of COVID-19 on Future Social Workers in the Areas of Secondary Traumatic Stress, Role Conflict, Supervision, and Colleague Support Ph.D. Thesis California State University Stanislaus, CA, USA 2021 25. Sanchez Caravaca F.S. Carrión J.T. Pastor E.S. Burnout syndrome and work satisfaction in professionals of social work in prisons of Spain Rev. Esp. Sanid. Penit. 2018 20 40 46 30231150 26. Brown A.R. Walters J. Jones A. Akinsola O. Rural social work: Recruitment, job satisfaction, burnout, and turnover Contemp. Rural Soc. Work J. 2017 9 12 27. Stanley S. Mettilda G.B. Meenakshi A. Resilience as a Moderator of Stress and Burnout: A Study of Women Social Workers in India Int. Soc. Work 2021 64 40 58 10.1177/0020872818804298 28. Ostadhashemi L. Arshi M. Khalvati M. Eghlima M. Khankeh H.R. Social Workers’ Resilience: Preventing Burnout and Compassion Fatigue in Pediatric Oncology Int. J. Cancer Manag. 2019 12 e61376 10.5812/ijcm.61376 29. Ben-Porat A. Itzhaky H. Burnout among Social Workers: The Contribution of Personal and Environmental Resources J. Soc. Work 2015 15 606 620 10.1177/1468017314552158 30. Diaconescu M. Burnout, Secondary Trauma and Compassion Fatigue in Social Work Rev. De Asistenta Soc. 2015 15 57 63 31. Vîrgă D. Baciu E.L. Lazăr T.A. Lupșa D. Psychological capital protects social workers from burnout and secondary traumatic stress Sustainability 2020 12 2246 10.3390/su12062246 32. Coronavirus Disease (COVID-19) Available online: https://www.who.int/health-topics/coronavirus#tab=tab_1 (accessed on 6 February 2022) 33. Radut-Seliste D. Avoiding Burn-Out During the Coronavirus Pandemic: Self-Care Practices Among Worldwide Social Workers Working in the Social Sphere in Malkit Population Places—Predictions and Opportunities Cholakova M. Publishing House, Saint Cyril and Methodius Veliko Turnovo, Bulgaria 2021 36 44 34. Maslach C. Jackson S.E. Leiter M.P. Maslach Burnout Inventory. Manual 3rd ed. Consulting Psychologists Press Palo Alto, CA, USA 1996 35. Martínez-López J.Á. Lázaro-Pérez C. Gómez-Galán J. Predictors of burnout in social workers: The COVID-19 Pandemic as a scenario for analysis Int. J. Environ. Res. Public Health 2021 18 5416 10.3390/ijerph18105416 34069394 36. Stamm B.H. The Professional Quality of Life Scale: Compassion Satisfaction, Burnout, and Compassion Fatigue/Secondary Trauma Scales Sidran Press Derwood, MD, USA 2005 Available online: https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.470.313&rep=rep1&type=pdf (accessed on 3 February 2022) 37. Stamm B.H. The Concise ProQOL Manual 2nd ed. The ProQOL.org Pocatello, ID, USA 2010 Available online: https://proqol.org/proqol-manual (accessed on 6 February 2022) 38. Baugerud G.A. Vangbæk S. Melinder A. Secondary traumatic stress, burnout and compassion satisfaction among Norwegian child protection workers: Protective and risk factors Br. J. Soc. Work 2018 48 215 235 10.1093/bjsw/bcx002 39. Senreich E. Straussner S.L.A. Steen J. The work experiences of social workers: Factors impacting compassion satisfaction and workplace stress J. Soc. Serv. Res. 2020 46 93 109 10.1080/01488376.2018.1528491 40. Verheyden C. Van Holen F. West D. Vanderfaeillie J. Secondary traumatic stress, burnout and compassion satisfaction among Flemish foster care workers during the COVID-19 lockdown Dev. Child Welf. 2020 2 227 243 10.1177/2516103220987227 41. Bridger K.M. Binder J.F. Kellezi B. Secondary traumatic stress in foster carers: Risk factors and implications for intervention J. Child Fam. Stud. 2020 29 482 492 10.1007/s10826-019-01668-2 32431483 42. Staudt M. Williams-Hayes M. Secondary traumatic stress, burnout, and compassion satisfaction among child advocacy interdisciplinary team members Adv. Soc. Work 2019 19 416 429 10.18060/22957 43. Craig C.D. Sprang G. Compassion satisfaction, compassion fatigue, and burnout in a national sample of trauma treatment therapists Anxiety Stress Coping 2010 23 319 339 10.1080/10615800903085818 19590994 44. Luthans F. The need for and meaning of positive organizational behavior J. Organ. Behav. 2002 23 695 706 10.1002/job.165 45. Hobfoll S.E. Johnson R.J. Ennis N. Jackson A.P. Resource loss, resource gain, and emotional outcomes among inner city women J. Personal. Soc. Psychol. 2003 84 632 10.1037/0022-3514.84.3.632 46. Demerouti E. Bakker A.B. Nachreiner F. Schaufeli W.B. The job demands-resources model of burnout J. Appl. Psychol. 2001 86 499 10.1037/0021-9010.86.3.499 11419809 47. Hobfoll S.E. Social and psychological resources and adaptation Rev. Gen. Psychol. 2002 6 307 324 10.1037/1089-2680.6.4.307 48. Sandoval R.G. The Struggle is Real but the Hassle is Deeper: Personal Resources and the Job Demands-Resources Model Master’s Thesis California State University Sacramento, CA, USA 2018 49. Bose S. Pal D. Impact of Employee Demography, Family Responsibility and Perceived Family Support on Workplace Resilience Glob. Bus. Rev. 2020 21 1249 1262 10.1177/0972150919857016 50. Scott-Young C.M. Turner M. Holdsworth S. Male and female mental health differences in built environment undergraduates Constr. Manag. Econ. 2020 38 789 806 10.1080/01446193.2020.1748213 51. Foster K. Roche M. Giandinoto J.A. Furness T. Workplace stressors, psychological well-being, resilience, and caring behaviours of mental health nurses: A descriptive correlational study Int. J. Ment. Health Nurs. 2020 29 56 68 10.1111/inm.12610 31127973 52. Olusa A.O. Bolaji K. Job demand and turnover intention: A comparative analysis of male and female bank employee’s resilience Gend. Behav. 2020 18 16469 16482 53. Levkovich I. Ricon T. Understanding compassion fatigue, optimism and emotional distress among Israeli school counsellors Asia Pac. J. Couns. Psychother. 2020 11 159 180 10.1080/21507686.2020.1799829 54. Samson T. Shvartzman P. Association between level of exposure to death and dying and professional quality of life among palliative care workers Palliat. Supportive Care 2018 16 442 451 10.1017/S1478951517000487 28641599 55. Winwood P.C. Colon R. McEwen K. A practical measure of workplace resilience: Developing the resilience at work scale J. Occup. Environ. Med. 2013 55 1205 1212 10.1097/JOM.0b013e3182a2a60a 24064782 56. Cohen L.H. Measurement of life events L.H. Cohen ED Life Event and Psychological Functioning: Theoretical Issues Sage Newbury Park, CA, USA 1988 11 30 57. Brown M.E. Hazards of our helping profession: A practical self-care model for community practice Soc. Work 2020 65 38 44 10.1093/sw/swz047 58. Calhoun M. Lightfoot E. Okamoto K. Goodenough K. Zheng M. Contemporary perceptions of social work: Macro practice in the profession J. Community Pract. 2020 28 374 391 10.1080/10705422.2020.1841702 59. Cleary M. Kornhaber R. Thapa D.K. West S. Visentin D. The effectiveness of interventions to improve resilience among health professionals: A systematic review Nurse Educ. Today 2018 71 247 263 10.1016/j.nedt.2018.10.002 30342300 60. Pereira J.A. Barkham M. Kellett S. Saxon D. The role of practitioner resilience and mindfulness in effective practice: A practice-based feasibility study Adm. Policy Ment. Health Ment. Health Serv. Res. 2017 44 691 704 10.1007/s10488-016-0747-0 61. Turner M. Scott-Young C. Holdsworth S. Resilience and well-being: A multi-country exploration of construction management students Int. J. Constr. Manag. 2021 21 858 869 10.1080/15623599.2019.1588843 62. Luceño-Moreno L. Talavera-Velasco B. García-Albuerne Y. Martín-García J. Symptoms of posttraumatic stress, anxiety, depression, levels of resilience and burnout in Spanish health personnel during the COVID-19 pandemic Int. J. Environ. Res. Public Health 2020 17 5514 10.3390/ijerph17155514 32751624 63. Masson F. Enhancing resilience as a self-care strategy in professionals who are vicariously exposed to trauma: A case study of social workers employed by the South African Police Service J. Hum. Behav. Soc. Environ. 2019 29 57 75 10.1080/10911359.2018.1474159 64. Lakioti A. Stalikas A. Pezirkianidis C. The role of personal, professional, and psychological factors in therapists’ resilience Prof. Psychol. Res. Pract. 2020 51 560 570 10.1037/pro0000306 65. Kumar R. Resilience, psychological well-being, and coping strategies in medical students Indian J. Psychiatr. Nurs. 2019 16 92 10.4103/IOPN.IOPN_22_19 66. Nissen-Lie H.A. Havik O.E.A. Høglend A. Monsen J.T. Rønnestad M.H. The contribution of the quality of therapists’ personal lives to the development of the working alliance J. Couns. Psychol. 2013 60 483 495 10.1037/a0033643 23957765 67. Thompson T.S. Caring for Care Teams: Nurses Thriving and Innovating by Design Ph.D. Thesis Charleston Southern University North Charleston, SC, USA 2021 68. Gonçalves L. Sala R. Navarro J.B. Resilience, and occupational health of health care workers: A moderator analysis of organizational resilience and sociodemographic attributes Int. Arch. Occup. Environ. Health 2022 95 223 232 10.1007/s00420-021-01725-8 34076733

PMC009xxxxxx/PMC9099677.txt

==== Front Polymers (Basel) Polymers (Basel) polymers Polymers 2073-4360 MDPI 10.3390/polym14091857 polymers-14-01857 Article Experimental and Numerical Investigation on Fatigue Properties of Carbon Fiber Cross-Ply Laminates in Hygrothermal Environments Xu Mingrui 12* Zeng Benyin 12 An Ziqian 1 Xiong Xin 2 Cheng Xiaoquan 1 Campagnolo Alberto Academic Editor 1 School of Aeronautic Science and Engineering, Beihang University, Beijing 100083, China; laoben001@sina.com (B.Z.); anziqian@buaa.edu.cn (Z.A.); xiaoquan_cheng@buaa.edu.cn (X.C.) 2 Aviation Industry Corporation of China, Ltd. (AVIC), China Helicopter Research and Development Institute, Jingdezhen 333001, China; xiongx005@avic.com * Correspondence: xumingrui@buaa.edu.cn 30 4 2022 5 2022 14 9 185704 4 2022 28 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). The fatigue properties of composite materials are degraded seriously in hygrothermal environments, so taking into account their influence is very important when evaluating the fatigue life of composite structures. Tensile fatigue experiments of carbon fiber reinforced resin composite cross-ply laminates were conducted in room temperature/dry (RTD), cool temperature/dry (CTD) and elevated temperature/wet (ETW) conditions. The S-N curves and fatigue failure modes of the cross-ply laminates were obtained in three conditions. On this basis, a finite element model was established to discuss the influence of temperature and moisture content on the fatigue properties, as well as a method for determining environmental factors of fatigue life of cross-ply laminates was established. The results show that the saturation moisture absorption and temperature have a significant influence on the tensile fatigue properties of cross-ply laminates. The high-cycle fatigue property is weakened significantly by the saturation moisture absorption and high temperature, but the low-cycle fatigue properties were strengthened in cool temperature conditions. The delamination failure mode in ETW is the most severe, presenting with an obvious necking phenomenon. The influence of temperature has a greater effect than that of moisture content, but moisture absorption would play its affect obviously when temperature exceeds 40 °C. laminates tensile fatigue hygrothermal environments fatigue properties S–N curve This research received no external funding. ==== Body pmc1. Introduction Carbon fiber reinforced resin matrix composites (CFRP) have been widely used in aircraft structures due to their high specific strength and stiffness [1,2,3]. The composite structure of the aircraft not only needs to stand cyclic fatigue loads during the service period, but also may encounter harsh environmental conditions such as heat and humidity [4]. Due to them being sensitive to heat and humidity [5], an environment with elevated temperature and high moisture could seriously weaken the mechanical properties of the composite material structure, and that will make the fatigue problem more prominent. The durability of the composite structure in severe environment conditions has become a big challenge of their application in aerospace industries [6]. Some studies have been conducted on the fatigue properties of composite laminates in wet and thermal circumstances [7,8,9,10,11,12,13,14,15,16,17,18]. Mismatch in moisture expansion and thermal expansion coefficients between the fiber and matrix will induce wet stress and thermal stress, and then initiate micro-cracks and debonding in the interface of the fiber/matrix, reducing the structural bearing capacity [19,20]. Kawai et al. [21] investigated the effect of moisture absorption on the fatigue strength of plane fabric fiber quasi-isotropic composite laminates under constant stress amplitude and different stress ratios. The results indicate that the fatigue life of wet laminates is lower than that of dry laminates, and the fatigue strength of wet laminates is 11% lower than that of dry laminates at room temperature. Mcbagonluri et al. [22] compared and studied the tensile fatigue properties of vinyl glass fiber composites in a dry state, fresh water and salty water immersion. The S–N curve after moisture absorption is lower than that in the dry state, but the slope of the S–N curve in three environments is almost the same. Malpot et al. [23] investigated the effect of moisture absorption on the fatigue property of composites in three wet conditions (RH0, RH50% and RH100%) and found that when the cycle number was less than 104, the fatigue life decreased with the increase in relative humidity. However, when the number of cycles is greater than 104, the humidity condition hardly affects the fatigue life. Therefore, the effect of hygrothermal environment on high-cycle fatigue far exceeds that of low-cycle fatigue. Kawai et al. [24] investigated the fatigue property of three composites, AS4/PEEK, T800H/Polyimide and T800H/Epoxy, at high temperature (100 °C) with different lay-up angles. The AS4/PEEK composites were found to exhibit stronger fatigue strength except for the 0° lay-up, while the T800H/Polyimide composites had weaker fatigue strength. The fatigue strength of all the lay-up materials except for the 0° layup showed a significant decrease when the number of cycles reached 105. In terms of numerical simulation, Shokrieh et al. [25,26] combined stress analysis, failure analysis and material performance attenuation to propose a fatigue model suitable for carbon fiber composite unidirectional bands. Harper et al. [27] established a FEM mode by using cohesive element to simulate fatigue delamination. All the investigations above focus on the individual effect on the fatigue property of composites about temperature and moisture absorption, so there are few studies on the combined effects of moisture and temperature. That contributes to the insufficiency of experimental verification, a deficient depth of numerical simulation, as well as the lack of the in-depth discussion upon the coupling between moisture and temperature. We have previously investigated the fatigue property of carbon fiber cross-ply laminates under hygrothermal environmental conditions in the article [28] using the method of experiments and numerical simulations, and then established an analysis model about the environmental influence factor of fatigue life for laminates. The effect of the hygrothermal environment on the properties of laminates largely depends on the influence of the matrix and the matrix/fiber interface. The sensitivity of angle-ply laminates to that under the hygrothermal environment is much stronger than that of cross-ply laminates. Therefore, we continue this investigation to gain a deeper understanding of coupling effect of the temperature and moisture absorption as well as better to improve the analysis model. 2. Specimens and Tests 2.1. Specimen Design and Manufacture In this work, T300 level CF3052/3238A carbon fiber-reinforced epoxy composite laminates were investigated whose configuration is shown in Figure 1. The specimens were designed based on ASTM D3479M [29] and a typical configuration of a [(45/−45)]8 lay-up was selected based on the application in rotorcraft industry. Prepregs were supplied by Guangwei Composites Co., Ltd. (Weihai, China). Laminates were manufactured by the AVIC composite corporation then. The maximum temperature for curing of laminates was 120 °C and they were kept for 120 mins under that condition. A pressure of 0.45 MPa was applied for the whole curing cycle. The test category and number of specimens are presented in Table 1. Among them, there are 3 moisture absorption specimens, which were used to determine the saturated moisture absorption state and time of all specimens. There are three experiment environments, namely in the room temperature and dry condition (RTD), in the cool temperature and dry condition (CTD, −40 °C) as well as in the elevated temperature and wet condition (ETW, 55 °C). In order to determine the fatigue test load, the corresponding static tensile tests were conducted first. 2.2. Experimental Procedures Three kinds of experiments were conducted, involving moisture absorption, tensile and tensile–tensile fatigue tests. All the specimens of ETW were firstly put in the hot water to reach wet condition, and then tensile and tensile–tensile fatigue tests were performed in RTD, RTW and ETW. When testing, the thermal condition was provided by the conditioning chamber. 2.2.1. Moisture Absorption Moisture absorption experiments were designed based on HB7401 [30] and ASTM Standard D5229 [31]. The specimens in ETW were immersed in a deionized water tank to achieve the moisture adsorption equilibrium as the water temperature was set to be 70 °C which is lower than the glass transmission of the matrix. Weights of specimens were regularly measured by an analytical balance with the accuracy of 0.1 mg. The specimens were cleaned by the absorbent cloth before weighing, and the time of weighing was within 5 min. Moisture equilibrium was reached when the moisture content met the following equation [31]. (1) M(%)=Wi−W0W0×100% where M is the moisture absorption content of the test specimens, W0 and Wi are the mass of the test specimen before and after moisture absorption. 2.2.2. Tensile Fatigue Tests As shown in Figure 2a, the fatigue tests were conducted on the Instron 8801 servo-hydraulic testing system (Instron Inc., Shanghai, China). Parameters of the sinusoidal waveform and stress ratio R = 0.0526 were set. RTD specimens were directly exposed to the environment and CTD/ETW specimens were placed in the conditioning chamber during testing. Hang extensometers with a maximum measurement range of 2.5 m in the longitudinal and transverse directions, respectively, were used to measure strains. The standard test method [29] points out that the loading frequency should ensure the change of the surface temperature of the test specimens does not exceed 10 °C. Therefore, an infrared temperature tester was used to monitor the temperature, and the loading frequency is set to 1 Hz according to the monitoring results. Due to the phenomenon of heating and water loss in the fatigue test of composite materials, the test specimens were moisturized by the method shown in Figure 2b under the ETW environment. Four stress levels which represent the maximum fatigue stress were selected based on some certain percentages of static strength. The test was terminated once the specimen was broken or the dynamic stiffness of the specimen was reduced by 10%. 3. Numerical Study 3.1. Basic Property Degradation The wet and thermal environment seriously weaken the properties of the matrix and matrix/fiber interface, so its basic mechanical properties are deteriorated. The mechanical properties in RTD are shown in Table 2. The model established by Shan et al. [32] is used to describe the change law of basic mechanical properties in wet and thermal environments, in which the dimensionless T* considers the influence of temperature and moisture absorption on basic mechanical properties. T* can be described as (2) T*=Tgw−TTg0−T0 In which T is the current temperature, T0 is the starting temperature, Tgw is the glass transition temperature of resin when the current moisture absorption is C, and Tg0 is the glass transition temperature of resin in dry state. Tgw and Tg0 satisfy the following relationship, (3) Tgw=Tg0−gC In which, C is the current moisture absorption of resin and g is the constant. The resin moisture absorption C is calculated from the moisture absorption content M of laminates, resin density ρm, laminate density ρ and resin volume content Vm according to Equation (4), (4) C=ρMρmVm After the dimensionless T* is obtained, the strength and stiffness of laminates are attenuated according to Equation (5), (5) E11E110=E22E220=(T∗)aXTXT0=YTYT0=(T∗)bG12G120=(T∗)cS12S120=(T∗)d In which, E11 and E22 are the elastic modulus of laminates along longitudinal and transverse directions under the current environment, respectively, E110 and E220 are elastic modulus along two directions in RTD, respectively, XT and YT are the tensile strength along two directions under the current environment, respectively, XT0 and YT0 are the tensile strength along two directions in RTD, respectively, G12 and S12 are the in-plane shear modulus and strength of laminates under the current environment, respectively, G120 and S120 are the in-plane shear modulus and strength in RTD, respectively. Finally, a, b, c, and d are the degradation constants of material properties in wet and thermal environments. The parameters used in Equations (2)–(5) are shown in Table 2. Based on the test data in RTD, the calculated value of material properties in CTD and ETW are shown in Table 3. Compared with the simulation results, it was found that the theoretical value is in good agreement with the test value, and the correlation coefficients R2 is greater than 0.99, indicating that the prediction of mechanical property parameters for laminates is reliable. 3.2. Fatigue Failure Criterion and Property Degradation Model In the process of fatigue loading, the residual strength and stiffness will show a downward trend. The degradation law is described as follows [25,26], (6) XT(n,σ,R)XT=XC(n,σ,R)XC=YT(n,σ,R)YT=YC(n,σ,R)XC=1−[1−(σ1XT)](nN)1.3218E11(n,σ,R)E11=1−[1−(σ10.8857XT)1/1.0702](nN)0.5418S12(n,σ,R)S12=1−[1−(τ12S12)](nN)9.3459G12(n,σ,R)G12=1−[1−(τ124.7853S12)1/14.9393](nN)0.2623 where XT(n,σ,R), XC(n,σ,R), YT(n,σ,R), YC(n,σ,R) denote the residual strength along the transverse and longitudinal direction, respectively, as well as the subscript T and C is for tensile and compression. E11(n,σ,R), S12(n,σ,R), G12(n,σ,R) are the residual elastic modulus, in-plane shear strength and in-plane shear modulus respectively after n cycles, XT, XC, YT, YC, E11, S12, G1 are the corresponding strength and modulus, respectively, before fatigue testing, σ1 and τ12 are the normal stress and in-plane shear stress, respectively, and N is the fatigue life. Damage criteria of composites were selected according to Shokrieh [25,26] (Equations (7)–(11)). Longitudinal fiber tensile failure:(7) f1t2=(σ1XT(n,σ,R))2+(τ12S12(n,σ,R))2+(τ13S13(n,σ,R))2≥1 Longitudinal fiber compressive failure:(8) f1c2=(σ1XC(n,σ,R))≥1 Transverse fiber tensile failure:(9) f2t2=(σ2YT(n,σ,R))2+(τ12S12(n,σ,R))2+(τ23S23(n,σ,R))2≥1 Transverse fiber compressive failure:(10) f2c2=(σ2YC(n,σ,R))≥1 Fiber-matrix shear failure:(11) f122=(σ1XC(n,σ,R))2+(τ12S12(n,σ,R))2+(τ23S23(n,σ,R))2≥1 where σi and τij are the components of normal and shear stress. S13(n,σ,R) and S23(n,σ,R) are the ris1ual shear strength in the 1-3 and 2-3 planes after n cycles. The stress–strain relationship of a single-layer lamina can be given as, (12) [σ1σ2σ3τ23τ31τ12]=[C11C12C13000C21C22C23000C31C32C33000000C44000000C55000000C66][ε1ε2ε3γ23γ31γ12] or {σ}=[C]{ε} where εi and γij represent linear and shear strain, respectively, and [C] denotes the three-dimensional stiffness matrix. The damage quantity D is introduced to characterize the damage degree of the material. D = 0 means that the material is not damaged, and d = 1 means that the material is completely damaged. When damage occurs, the propagation was described by stiffness degradation as, (13) {σ}=[C^]{ε} where [C^] is the reduced stiffness matrix, C^ij is given by, C12^=(1-d1t)(1-d1c)(1-d2t)(1-d2c)C12, C12^=(1-d1t)(1-d1c)(1-d2t)(1-d2c)C12, C22^=(1-d2t)(1-d2c)C22, C13^=(1-d1t)(1-d1c)C13, C23^=(1-d2t)(1-d2c)C23, C33^=C33, C44^=(1-d12)C44, C55^=C55, and C66^=C66. dit and dic (i = 1, 2), respectively, express the damage degree of tensile and compression along 2 directions, and d12 is for the change of in-plane shear damage. The damage variable da is a continuous function of fa (a = 1t, 1c, 2t, 2c, 12), (14) d1t=1−1f1te(1−f1t)(XtεtxG1t)=1−1f1te(1−f1t)(XtεtxLcW1t)d1c=1−1f1ce(1−f1c)(XcεcxG1c)=1−1f1ce(1−f1c)(XcεcxLcW1c)d2t=1−1f2te(1−f2t)(YtεtyG2t)=1−1f2te(1−f2t)(YtεtyLcW2t)d2c=1−1f2ce(1−f2c)(YcεcyG2c)=1−1f2ce(1−f2c)(YcεcyLcW2c)d12=1−γ(γf−ε12)ε12(γf−γ)(γf=2GSS12) where Ga (a = 1t, 1c, 2t, 2c, 12) is the fracture energy density of materials under different damage forms. 3.3. Fatigue Finite Element Model As shown in Figure 3, Abaqus 6.14 is used to establish the finite element analysis model (FEA). The model is divided into three parts: a fixed support section, working section and loading section and laminates contains eight layers along the thickness direction. In order to improve the calculation accuracy, the grid element type is set to C3D8R, which contributes to 39,820 units in total. The load and constraint conditions are as follows: the load is applied through the in-plane shear force on the upper and lower surfaces of the loading section; translational constraints are applied to all nodes on the upper and lower surfaces of the fixed support section. The material property parameters in the FEA model are as shown in Table 4. A UMAT subroutine written in the Fortran language according to fatigue failure criterion and properties degradation model was used to carry out the progressive damage evolution of laminates, the flow chart of which is shown in Figure 4. 4. Results and Discussion 4.1. Tensile Tests The average tensile properties in three environments are shown in Table 5. It can be found that, compared with the elastic modulus and tensile strength in RTD, they increase by 12.67% and 12.84%, respectively, in CTD, and decrease by −8.54% and −37.11%, respectively, in ETW. Therefore, it can draw a conclusion that the increase in matrix brittleness at cool temperature increases the elastic modulus and tensile strength correspondingly, while the coupling effect of high temperature and high moisture seriously weakens them. 4.2. Fatigue Test 4.2.1. Hygrothermal Environment on the Fatigue Properties The fatigue test results in three environments are shown in Table 6. A total of 16 specimens were conducted in each environment, which are divided into four groups according to the stress level, with an average of four in each group. The life is listed in the column of fatigue life and the stress level refers to the relative value of the average tensile strength. The S–N curve, also known as the Wöhler curve, is obtained after a number of fatigue tests at different stress levels. In this paper, the two-parameter linear mode is applied [33], (15) S=A+B*lgN where A and B are material constants, S is max fatigue stress, N is the cycle number when composites are fractured. As shown in Figure 5, the fatigue S-N curves in RTD, CTD and ETW are fitted as Equations (16)–(18), respectively, (16) S=229.55−15.61lgN (17) S=306.32−34.03lgN (18) S=116.49−8.38lgN Figure 5 shows that, for the same cycle times, when the number of cycles n is greater than 14,715, the fatigue strength sequence from large to small is RTD, CTD, and ETW. When the number is less than 14,715, that sequence is CTD, RTD, and ETW. It means the cool temperature makes the high-cycle fatigue strength decrease, but it is beneficial to the low-cycle fatigue. In addition, the fatigue dispersion in RTD and ETW is comparable, and it is obviously reduced in CTD, which should be related to the embrittlement of the matrix at a cool temperature. The decreasing rate of fatigue strength from large to small is RTD > ETW > CTD which means that the fatigue strength decreases most rapidly at cool temperatures. Taking the fitting curve results when the fatigue life is 106 times for comparison, the fatigue strength in RTD, CTD and ETW are 135.9 MPa, 102.1 MPa and 66.2 MPa, respectively. Compared with in RTD, the fatigue strength in CTD and ETW is reduced by 24.88% and 51.28%, respectively. The coupling effect of high temperature and moisture on fatigue strength is significantly stronger than that of cool temperature alone. 4.2.2. Failure Mode Analysis The fatigue failure morphology in three environments is shown in Figure 6 and Figure 7. From Figure 6a and Figure 7a, it can be found that the fatigue failure is mainly matrix cracking and fiber fracture in RTD, accompanied by a certain 45° delamination propagation failure. At the same time, it can be seen from Figure 6b, Figure 7b,c and Figure 7b,c that failure modes such as fiber fracture, matrix cracking and delamination can be found in CTD and ETW as well. Delamination runs through the whole thickness direction and basically expands along the 0° direction in CTD and ETW. In addition, an obvious necking phenomenon can be seen in ETW which means the properties of matrix were weakened seriously. At last, it can be seen from Figure 7 that the delamination failure of ETW is the most severe in three environments. 4.3. Finite Element Results 4.3.1. Model Verification Table 7 presents the comparison of the fatigue life between the predicted values and the experimental values in three environments to verify the validity of the FEM model. It can be found that there is a larger error in the low-cycle section. However, in the high-cycle section, the simulation results are in good agreement with the experimental results, and the error is 13.2% at most. In general, the logarithmic life error is relatively small, which is less than 7%. The failure morphology obtained by finite element calculation is shown in Figure 8. From Figure 8a,b, it can be found that the fiber fracture damage occurs in a small area on the left and right of the specimen in RTD and CTD, and the damage distribution direction is at a certain angle with the edge of the specimen. From Figure 8c, it can be found that the specimen in ETW will have fiber failure in a wider range, which is similar to the failure mode shown in the test. In conclusion, the model can effectively calculate the fatigue life of specimens in different environments and reflect their failure modes. 4.3.2. Failure Process Analysis According to the test results, the final fatigue failure of the specimen is mainly fiber fracture. Therefore, the development process of fiber fracture damage is observed with the help of the finite element model, so as to analyze the fatigue damage mechanism of the specimen. The damage evolution process of fiber fracture in three environments is shown in Figure 9 with the stress being set to 60% of their respective static strength. From Figure 9a in RTD, after 190,000 cycles of loading there is a small area of damage in the middle of the specimen, and then the damage extends roughly along the 45° direction of the edge of the specimen. After 474,000 cycles, the damage extends from the inside to the edge, and finally to the whole thickness and width of the specimen. From Figure 9b in CTD, the results of the fatigue failure process of the specimen are similar to those in RTD. from Figure 9c in ETW, after the initial damage occurs the damage propagation range is relatively larger, and finally a large area of fiber fracture occurs in the middle of the specimen then the overall life decreases significantly. 4.3.3. Effect of Different Environments on Fatigue Life In order to explore the coupling effect of temperature and moisture on the fatigue life of angel-ply laminates, the RTD environmental fatigue life of 106 was taken as a fixed point, and the load at this point was kept unchanged at 135.5 MPa. The fatigue life of laminates at different temperatures (−40 °C, −20 °C, 0 °C, 20 °C, 40 °C and 55 °C) and different moisture absorption (0%, 0.5%, 1%, 1.5% and 1.8%) was calculated by the finite element model. Figure 10 shows the fatigue life obtained by finite element calculation for different temperature and moisture combinations. Figure 10a is for fatigue life, and Figure 10b is the normalization type of logarithmic life. Taking the moisture absorption as the horizontal axis and the temperature as the vertical axis, the calculated values under different temperature and moisture absorption are drawn in a palace diagram. Taking the number of life cycles down to one tenth of the value at room temperature in dry state as the critical value, the effect of the hygrothermal environment on fatigue life is divided into a strong influence area and weak influence area, marked with red and blue, respectively. From the calculation results, when the moisture absorption content reaches 1.0% or the temperature exceeds 40 °C, all results are in the strong influence area. Assuming that the effects of temperature and moisture are independent of each other, by fitting the above data, the environmental factors of fatigue life can be expressed as following, (19) F=logNlogN0={1−0.358(MM∞)0.725−0.171(T−T0T1−T0)1.539T≥20 °C1−0.230(MM∞)0.771−0.088(T−T0T1−T0)0.138T<20 °C where F is the environmental factors of fatigue life, N and N0 are the fatigue life in hygrothermal environment and in RTD, respectively, and M and M∞ are the current and maximum moisture absorption content, respectively. T is the current temperature, T0 is the reference temperature, and this formula takes 20 °C. T1 is the maximum temperature and 55 °C is taken in this formula. 5. Conclusions In this paper, the tests of moisture absorption, static tensile and tensile–tensile fatigue were conducted on carbon fiber CF3250/3238A angle-ply laminates. Fatigue S–N curves and failure damage modes in three environments were obtained. Based on the experimental study, the finite element analysis model of progressive damage was established, and the influence of temperature and humidity on fatigue performance was discussed. Finally, the determination method of the environmental factors of fatigue life was established. Some conclusions were drawn as follows:Compared with the RTD environment, the tensile strength in CTD and ETW increased by 12.84% and decreased by −37.11%, and the tensile modulus increased by 12.67% and decreased by −8.54%, respectively. Only cool temperatures have obvious negative effects on the fatigue life dispersion of the test specimens, and high temperature and high moisture have no effect on that. The decline rate of the S–N curve is the largest in CTD and the smallest in ETW. Cool temperature has a positive effect on low-cycle fatigue but has a negative effect on high-cycle fatigue. When N = 106, the fatigue limit in CTD and ETW is, respectively, decreased by 24.88% and 51.28% of that in RTD. The temperature plays a large role on the fatigue limit and the combination of high temperature and moisture has a much larger effect on that. The failure morphology in three environments is similar, including fiber fracture, matrix cracking and delamination damage. However, the severity of failure is different, the failure of specimens in ETW is the most significant as well as the least in CTD. The fatigue morphology in ETW shows the obvious necking phenomenon since the matrix was seriously weakened. Numerical analysis shows that the effect of temperature on fatigue property is significantly stronger than that of moisture absorption, and when the temperature exceeds 40 °C, the effects of moisture absorption are great. A method for determining the environmental factors of fatigue life of angle-ply composite materials is proposed, which can be used for the life prediction of that in different environmental conditions. Author Contributions Writing—original draft, M.X.; formal analysis, M.X. and Z.A.; methodology, X.C. and B.Z.; data curation, X.X. and Z.A.; validation, M.X. and Z.A.; investigation, M.X., Z.A., X.X.; resources, X.X.; writing—review and editing, M.X.; supervision, Z.A. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Data are contained within the article. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Static and tensile fatigue specimens. Figure 2 Test set for the tensile fatigue. (a) Test loading and measuring in RTD; (b) moisture retention of ETW test specimens. Figure 3 Finite element model. Figure 4 Flow chart of fatigue simulation. Figure 5 S-N curve of angel-ply laminates in three environments. Figure 6 Fatigue failure morphology in three environments. (a) RTD; (b) CTD; (c) ETW. Figure 7 Fatigue failure morphology along the direction of thickness in three environments. (a) RTD; (b) CTD; (c) ETW. Figure 8 The final fatigue failure morphology calculated by the finite element method. (a) RTD; (b) CTD; (c) ETW. Figure 9 Fatigue failure process calculated by the finite element method. (a) RTD; (b) CTD; (c) ETW. Figure 10 Fatigue life under different hygrothermal conditions. (a) Fatigue life; (b) Normalization of logarithmic life. polymers-14-01857-t001_Table 1 Table 1 Test types and specimen number. Test Category RTD CTD ETW Moisture absorption 3 Static tensile 3 3 3 Tension fatigue 16 16 16 polymers-14-01857-t002_Table 2 Table 2 Material parameters [28]. E110/GPa E220/GPa XT0/MPa YT0/MPa G120/GPa S120/MPa ρ ρm/g/cm3 Vm/% 54.3 54.3 680 680 3.26 116 1.42 1.23 45 polymers-14-01857-t003_Table 3 Table 3 Empirical constants [28]. Tg0/°C T0/°C g/°C/c a b c d 120 20 5 0.05 0.15 0.22 0.56 polymers-14-01857-t004_Table 4 Table 4 Elastic engineering constants of materials [28]. E1/GPa E2/GPa E3/GPa G12/GPa G13/GPa G23/GPa ν12 ν13 54.3 54.3 3.3 3.26 2.17 2.17 0.04 0.01 ν23 XT/MPa XC/MPa YT/MPa YC/MPa S12/MPa S13/MPa S23/MPa 0.01 680 614.29 680 614.29 115.98 73.5 73.5 polymers-14-01857-t005_Table 5 Table 5 Average tensile properties in three environments. Static Properties RTD CTD ETW Maximum load/kN 16.85 19.02 10.75 Elastic modulus/GPa 12.10 13.63 11.07 Tensile strength/MPa 231.86 261.63 145.82 polymers-14-01857-t006_Table 6 Table 6 Fatigue test results in three environments (stress ratio R = 0.0526). Environments Stress Level /% Max Stress /MPa Fatigue Life /Cycle RTD 60 139.1 645,427, 77,795, 280,167, 925,905 68 157.7 59,040, 15,576, 35,494, 33,051 70 162.3 21,348, 15,576, 47,731, 39,773 73 169.3 4213, 10,871, 11,142, 399,891 CTD 38 99.4 874,358, 1,207,366, 1,015,576, 840,721 55 143.9 79,826, 88,424, 95,197, 82,243 60 157 18,463, 29,169, 27,646, 20,312 63 164.8 7866, 17,618, 9382, 20,316 ETW 45 65.6 1,124,138, 122,504, 917,306, 895,576 50 72.9 213,125, 125,555, 123,153, 48,097 56 81.7 70,968, 107,692, 35,724, 3890 60 87.5 1608, 2988, 6730, 6038 polymers-14-01857-t007_Table 7 Table 7 Comparison between the predicted values and the experimental values. Stress Level Life Cycle Number/Cycle Logarithmic Life Number Test Simulation Error Simulation Test Error RTD-60% 482,324 546,000 13.2% 5.53 5.74 3.8% RTD-68% 35,790 36,500 2.0% 4.51 4.56 1.1% RTD-70% 31,107 17,500 −43.7% 4.45 4.24 −4.72% RTD-73% 16,500 6800 −58.8% 4.08 3.83 −6.1% CTD-38% 984,505 1,014,000 3.0% 5.99 6.01 0.3% CTD-55% 86,423 50,000 −42.1% 4.94 4.70 −4.9% CTD-60% 23,898 27,000 13.0% 4.37 4.43 1.4% CTD-63% 13,796 12,000 −13.0% 4.11 4.08 −0.7% ETW-45% 764,881 780,000 2.0% 5.76 5.89 2.3% ETW-50% 127,483 123,750 −3.0% 5.05 5.09 0.8% ETW-56% 54,569 16,000 −70.7% 4.51 4.20 −6.9% ETW-60% 4341 3400 −21.7% 3.57 3.53 −1.1% Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Zhang Q. Cheng X.Q. Zhang J. Wang S. Cheng Y. Zhang T. Experimental and numerical investigation of composite box joint under tensile load Compos. B Eng. 2016 107 75 83 10.1016/j.compositesb.2016.09.056 2. Liu S.F. Zhang J. Zhang Q. Guo X. Cheng X.Q. Bao J.W. An investigation of hygrothermal effects on adhesive materials and double lap shear joints of CFRP composite laminates Compos. B Eng. 2016 91 431 440 10.1016/j.compositesb.2016.01.051 3. Cheng X.Q. Zhang J. Bao J.W. Zeng B.Y. Cheng Y.J. Hu R.W. Low-velocity impact performance and effect factor analysis of scarf-repaired composite laminates Int. J. Impact Eng. 2018 111 85 93 10.1016/j.ijimpeng.2017.09.004 4. Jin H. Study on the fatigue properties of carbon fiber composite laminates under environmental conditions Sci. Technol. Innov. Her. 2014 1 76 77 5. Cheng X.Q. Yang M.M. Zhang J. Zhang Q. Zhang J.K. Thermal behavior and tensile properties of composite joints with bonded embedded metal plate under thermal circumstance Compos. B Eng. 2016 99 340 347 6. Sousa J.M. Correia J.R. Cabral-Fonseca S. Some permanent effects of hygrothermal and outdoor ageing on a structural polyurethane adhesive used in civil engineering applications Int. J. Adhes. Adhes. 2018 84 406 419 10.1016/j.ijadhadh.2018.04.010 7. Du Y. Ma Y.E. Fatigue performance of fiber reinforced polymer composites under hygrothermal environment—A review Acta Mater. Compos. Sin. 2021 19 1 15 8. Venkatesha B.K. Saravanan K. Anand B. Effect of moisture absorption on woven bamboo/glass fiber reinforced epoxy hybrid composites Mater. Today Proc. 2021 45 216 221 10.1016/j.matpr.2020.10.421 9. Bachchan A.A. Das P.P. Chaudhary V. Effect of moisture absorption on the properties of natural fiber reinforced polymer composites: A review Mater. Today Proc. 2021 49 3403 3408 10.1016/j.matpr.2021.02.812 10. Bao L.R. Yee A.F. Moisture diffusion and hygrothermal aging in bismaleimide matrix carbon fiber composites—Part I: Uni-weave composites Compos. Sci. Technol. 2002 62 2099 2110 10.1016/S0266-3538(02)00161-6 11. Wu F.Q. Yao W.X. Residual strength degradation model of fiber reinforced plastic J. Nanjing Univ. Aeronaut. Astronaut. 2008 40 517 520 12. Meziere Y. Bunsell A.R. Favrya Y. Teissedrea J.C. Dob A.T. Large strain cyclic fatigue testing of unidirectional carbon fiber reinforced epoxy resin Compos. Part A Appl. Sci. Manuf. 2005 36 1627 1636 10.1016/j.compositesa.2005.03.020 13. Karad S.K. Jones F.R. Attwood D. Moisture absorption by cyanate ester modified epoxy resin matrices. Part II. The reverse thermal effect Polymer 2002 43 5643 5649 10.1016/S0032-3861(02)00484-6 14. Guo F.-L. Guo P. Huang Y.L. Hu N. Fu S.-Y. Multiscale modeling of mechanical behaviors of carbon fiber reinforced epoxy composites subjected to hygrothermal aging Compos. Struct. 2021 256 113098 10.1016/j.compstruct.2020.113098 15. Guo F. Yan Y. Hong Y. Tian Z. Li J. Prediction and optimization design for thermal expansion coefficients of three dimensional directional braided composites Polym. Compos. 2019 40 2495 2509 10.1002/pc.25132 16. Ray B.C. Temperature effect during humid ageing on interfaces of glass an carbon fibers reinforced epoxy composites J. Colloid Interface Sci. 2006 298 111 117 10.1016/j.jcis.2005.12.023 16386268 17. Yu Z.G. Yang S.C. Song B.F. Comparison of wet and hot aging resistance of T700 and T300 carbon fiber strengthened epoxy resin composites Mater. Mech. Eng. 2009 33 48 51 18. Sha M. Xiong X. Xu M.R. Zeng B.Y. Cheng X.Q. Effect of hygrothermal environment on fatigue properties of composite materials Hi Tech Fiber Appl. 2017 42 37 43 19. Chen H.R. Xi Z.C. He X.D. Non-linear analysis of thermal deformations of composite laminates under transient thermal loading J. Dalian Univ. Technol. 1994 34 280 286 20. Sethi S. Ray B.C. Environmental effects on fiber reinforced polymeric composites: Evolving reasons and remarks on interfacial strength and stability Adv. Colloid Interface Sci. 2015 217 43 67 10.1016/j.cis.2014.12.005 25578406 21. Kawai M. Yagihashi Y. Hoshi H. Iwahori Y. Anisomorphic constant fatigue life diagrams for quasi-isotropic woven fabric carbon/epoxy laminates under different hygrothermal environments Adv. Compos. Mater. 2013 22 79 98 10.1080/09243046.2013.777172 22. Mcbagonluri F. Garcia K. Hayes M. Verghese K.N.E. Lesko J.J. The Characterization of fatigue and combined environment on durability performance of glass/vinyl ester composite for a new method for the processing of multifunction applications Int. J. Fatigue 2000 22 53 64 10.1016/S0142-1123(99)00100-0 23. Malpot A. Touchard S. Bergamo An investigation of the influence of moisture on fatigue damage mechanisms in a woven glass-fibre-reinforced PA66 composite using acoustic emission and infrared thermography Compos. Part B Eng. 2017 130 11 20 10.1016/j.compositesb.2017.07.017 24. Kawai M. Yajima S. Hachinohe A. High-temperature off-axis fatigue behavior of unidirectional carbon-fiber-reinforced composites with different resin matrices Compos. Sci. Technol. 2001 61 1285 1302 10.1016/S0266-3538(01)00027-6 25. Shokrieh M.M. Lessard L.B. Progressive fatigue damage modeling of composite materials, Part I: Modeling J. Compos. Mater. 2000 34 1056 1080 10.1177/002199830003401301 26. Shokrieh M.M. Lessard L.B. Progressive fatigue damage modeling of composite materials, Part II: Material characterization and model verification J. Compos. Mater. 2000 34 1081 1116 10.1177/002199830003401302 27. Harper P.W. Hallett S.R. A fatigue degradation law for cohesive interface elements-Development and application to composite materials Int. J. Fatigue 2010 32 1774 1787 10.1016/j.ijfatigue.2010.04.006 28. Zhang X.L. Meng Q.C. Xu M.R. Zeng B.Y. Cheng X.Q. Sun W. Tensile fatigue properties of carbon fiber reinforced composite orthogonal laminates after moisture absorption J. Mater. Eng. 2021 49 169 177 29. D3479/D3479M-2012 Standard Test Method for Tension-Tension Fatigue of Polymer Matrix Composite Materials American Society of Testing Materials International West Conshehoken, PA, USA 2012 1 6 30. HB7401-1996 People’s Republic of China Aviation Industry Standard. Moisture Absorption Test Method for Resin Matrix Composite Laminates under Hygrothermal Environment Aviation Industry Corporation of China Beijing, China 1996 31. ASTM D5229/D5229M-2014 Standard Test Method for Moisture Absorption Properties and Equilibrium Conditioning of Polymer Matrix Composite Materials American Society of Testing Materials International West Conshehoken, PA, USA 2014 1 6 32. Shan M. Zhao L. Hong H. Liu F. Zhang J. A progressive fatigue damage model for composite structures in hygrothermal environments Int. J. Fatigue 2018 111 299 307 10.1016/j.ijfatigue.2018.02.019 33. Spindel J. Haibach E. Some considerations in the statistical determination of the slope of SN curves Statistical Analysis of Fatigue Data American Society of Testing Materials International West Conshehoken, PA, USA 1981

PMC009xxxxxx/PMC9099678.txt

==== Front Materials (Basel) Materials (Basel) materials Materials 1996-1944 MDPI 10.3390/ma15093400 materials-15-03400 Article Manufacturing of Aluminum Matrix Composites Reinforced with Carbon Fiber Fabrics by High Pressure Die Casting https://orcid.org/0000-0002-2523-7309 Bedmar Javier Torres Belén https://orcid.org/0000-0003-0837-3437 Rams Joaquín * Xiao Bolv Academic Editor Department of Applied Mathematics, Materials Science and Engineering and Electronics Technology, ESCET, Rey Juan Carlos University, Mostoles, 28933 Madrid, Spain; javier.bedmar@urjc.es (J.B.); belen.torres@urjc.es (B.T.) * Correspondence: joaquin.rams@urjc.es 09 5 2022 5 2022 15 9 340006 4 2022 06 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Aluminum matrix composites reinforced with carbon fiber have been manufactured for the first time by infiltrating an A413 aluminum alloy in carbon fiber woven using high-pressure die casting (HPDC). Composites were manufactured with unidirectional carbon fibers and with 2 × 2 twill carbon wovens. The HPDC allowed full wetting of the carbon fibers and the infiltration of the aluminum alloy in the fibers meshes using aluminum at 680 °C. There was no discontinuity at the carbon fiber-matrix interface, and porosity was kept below 0.1%. There was no degradation of the carbon fibers by their reaction with molten aluminum, and a refinement of the microstructure in the vicinity of the carbon fibers was observed due to the heat dissipation effect of the carbon fiber during manufacturing. The mechanical properties of the composite materials showed a 10% increase in Young’s modulus, a 10% increase in yield strength, and a 25% increase in tensile strength, which are caused by the load transfer from the alloy to the carbon fibers. There was also a 70% increase in elongation for the unidirectionally reinforced samples because of the finer microstructure and the load transfer to the fibers, allowing the formation of larger voids in the matrix before breaking. The comparison with different mechanical models proves that there was an effective load transference from the matrix to the fibers. metal matrix composites aluminum carbon fibers carbon wovens high-pressure die casting infiltration Agencia Estatal de InvestigaciónRTI 2018-096391-B-C31 PID2021-124341OB-C21 PID2021-123891OB-I00 Comunidad de Madrid Project ADITIMAT-CMS2018/NMT-4411 This research was funded by Agencia Estatal de Investigación (RTI 2018-096391-B-C31, PID2021-124341OB-C21, PID2021-123891OB-I00) and Comunidad de Madrid Project ADITIMAT-CM (S2018/NMT-4411). ==== Body pmc1. Introduction Metal matrix composites (MMCs) reinforced with carbon fibers (Cf) are interesting materials as they combine the different properties of the metal and the reinforcement used, providing better toughness and high yield and tensile resistances [1]. Aluminum matrix composites (AMCs), other works specify more advantages such as lightness and high corrosion resistance [2,3]. There are several processing techniques to fabricate this kind of composite. Spark Plasma Sintering (SPS) has been used to fabricate composites with good bonding between matrix and reinforcement [4] and with high mechanical properties, as shown by Tan et al. in their study [5]. However, most of the methods used for the manufacture of these types of composites are in a liquid phase, i.e., the molten aluminum is poured on the fibers [6], because it is a fast processing route and it allows the attainment of near-net-shape pieces [7,8] However, Ramos-Masana and Colominas [9] claimed that the carbon fibers do not wet by the molten aluminum at the temperatures used for casting aluminum composites (700–800 °C) as the wetting angle is higher than 90° between Cf and aluminum because of the low reactivity of carbon fibers with the molten aluminum. Sarina et al. [10] confirmed that this effect remains even after long-time exposures of the Cf to the molten aluminum at higher temperatures. In the case of high modulus carbon fibers, the lack of wettability is enhanced because the basal planes of the graphite structure are placed parallel to the surface; this provides a low energy surface that further reduces the reactivity and hence the wettability of the carbon fibers [11]. The most common routes followed to get the infiltration of the fibers have been the increase of processing temperature, the coating of the fibers, the hot-pressing technique, and the use of high infiltration pressure. Increasing the temperature of the aluminum alloy has been used by several authors to increase the wettability of different ceramic reinforcements: SiC particles [12], BC4 [13], and Cf [14]. However, in Cf, when the temperature exceeds 900 °C, aluminum reacts with the fibers and increases their wettability, but the reaction product formed is Al4C3, which is brittle and tends to hydrolyze in humid environments, expanding and degrading the composites. To improve wettability, coatings have been deposited on the carbon fibers. The coatings inhibit the contact between the molten aluminum and the fibers thus that the wetting system is the molten aluminum coating, which is favored in most cases. Some authors have used the electroless technique to deposit different metals on long and short carbon fibers, such as Ni [15,16] and copper [17]. Refractory materials have been also used because of their inertness, and also pyrolytic carbon around the fibers has been suggested [18] but the results obtained have not fully solved the wettability problems in dense fiber zones. Finally, the incorporation of Mg to the alloy used also improves the wettability of the systems because it tended to situate at the Cf–Al interfaces [19]. Another way to avoid the wettability problem is using the hot-pressing technique on fibers coated by electroless. Nour-Eldin et al. obtained good bonding between matrix and reinforcement in short times of production with this method [20]. These composites showed a good distribution of the reinforcement with increased hardness and toughness [21] and increased wear resistance [22]. Another method is the continuous infiltration of wire composites to fabricate double composites. In this technique, wires of reinforcements are impregnated in the molten metal in a continuous process, and, after that, composites are obtained by a casting process. In this case, Kientzl and Dobránszky claimed that this kind of composite has improved mechanical properties [23]. In addition, Kientzl et al. claimed that the infiltration of the composite wire was successfully achieved [24]. In addition, from the manufacturing perspective, another way to solve the low wetting of the carbon fibers is to apply an external pressure to force the wetting of carbon fibers by the liquid aluminum. In the case of carbon fiber preforms, the pressure threshold that must be surpassed to infiltrate the fibers preforms has been established in the 35–106 MPa range [25], and pressures even higher are needed to get the full infiltration of the carbon fibers meshes of wovens. The lack of spontaneous wetting of the fibers also implies that there would be a tendency to form defects in the material, particularly in zones with a high density of fibers. The pressure required for the infiltration can be applied by many different routes. The most used ones are squeeze casting, gas pressure-assisted infiltration, and centrifugal infiltration. During squeeze casting, pressure is mechanically applied to the alloy used after pouring it on the carbon fibers [26]. In some cases, this can be made on preforms made of short carbon fibers [27]. However, this method usually gives rise to the formation of pores and voids, and subsequent processing such as rolling reduced their influence on the mechanical properties of the composites [28]. The effect of the pressure used has also been analyzed in different systems. Sukumaran et al. [29] established that a pressure of 100 MPa was required for the microstructural refinement and to get very low porosity in Al 2124-10%SiCp composites. This indicates that once the infiltration threshold has been surpassed, it is not needed to apply much higher pressures to avoid the formation of defects. Finally, the centrifugal infiltration phenomenon has been used by authors like Nishida et al. [30] on fibrous preforms and metallic coated short Cf preforms, showing that it is a valid route, although it may cause gradients in the composition of the alloys, both in the location of the reinforcement and in the distribution of alloying elements in the matrix, what occurs in several kinds of AMCs [31]. On the other hand, aluminum matrix composites have been fabricated as coatings to improve the surface properties of metals like aluminum. In this case, thermal spray is a method that has achieved aluminum reinforced with SiC coatings with improved hardness, low porosity, and good bonding between matrix and reinforcement [32]. This wettability and low porosity can be enhanced by applicating sol-gel coatings on the SiC particles before the thermal spray [33]. On the other hand, one of the most used casting methods for aluminum pieces is High-Pressure Die Casting (HPDC) [34]. In this technique, molten aluminum is injected into a metallic die, and before the cooling and solidification of the aluminum alloy, high pressure is suddenly applied by a piston driven by compressed nitrogen. The alloy solidifies quickly due to the heat evacuated by the metal mold, and a fine microstructure is obtained [35]. In addition, the applied pressure reduces the size and number of the defects, which in most cases are trapped gas bubbles from the lubricant used for the demolding of the piece [36]. This process allows obtaining aluminum pieces for very different applications with a very competitive cost and with high production rates [37]. Therefore, HPDC seems to be optimal for the infiltration of carbon fibers with no need for preforms since the molten aluminum and the reinforcement adapt to the mold. However, it has not been used in composite manufacturing until now because of the difficulties of placing the fibers, the need for highly complex machinery, and the low reinforcement rates that there seems to be possibly obtained. There are many differences between the previous methods used and the one proposed in this work. The most important is the combination of pressure and time used. Pressure is much higher in the HPDC than in the other methods, except for squeeze casting. However, in HPDC the die is cold; thus, the process is much faster than in squeeze casting, resulting in much different structures and properties. In this work, aluminum matrix composites reinforced with long carbon fibers have been manufactured by infiltrating continuous carbon fibers in a metallic die. The design of the die prevented the movement of the fibers and allowed the infiltration of the carbon fiber meshes. The microstructure and the presence of defects, as well as the distribution of fibers in the composite, have been studied. Microhardness measurements and tensile tests have been carried out on the manufactured samples, and several micromechanical models have been used to show the load transference from the matrix to the fibers. 2. Materials and Methods The aluminum alloy used, supplied by Aluminio La Estrella, was the AA413 alloy (EN AC-44100), which is an alloy specifically designed for the die casting process. Its composition was (wt.%): 12.8 Si, 0.65 Fe, 0.55 Mn, 0.20 Ti, 0.15 Cu, 0.15 Zn, 0.10 Mg, 0.10 Ni, 0.10 Pb, 0.05 Sn, and Al-balance. The mechanical properties provided by the manufacturer were 150–170 MPa tensile strength; 1–2% elongation at break; and 75 GPa of Young’s modulus. The carbon fibers used were AS4 and were supplied by Hexcel. They had an average diameter of 7.1 µm and were supplied as 2 × 2 twill wovens. Before the composite manufacturing, the fibers were heat-treated in an oven for 24 h at 175 °C to remove their sizing. In the case of the unidirectional reinforced composites, fibers were obtained by removing the fibers of the warp of the woven. This allowed having the same type of composites and effective reinforcement in the main direction of the samples, while in the transverse direction, there were substantial differences. The properties of the carbon fibers provided by the supplier were 4480 MPa tensile strength; 1.8% elongation at break; 231 GPa longitudinal Young’s modulus; and 13 GPa of transversal elastic modulus. The procedure used to fabricate the samples is shown in Figure 1 and is based on using a high-pressure horizontal opening die casting machine. Different HPDC conditions were tested, and several evolutions of the mold were used before having the definitive manufacturing process. The die used had a dog bone shape with a length of 55.0 mm, a width of 10.0 mm, and a thickness of 2.0 mm in the central zone, the total length of the specimen was 116 mm. The carbon fibers were cut with a length of 150 mm and a width of 20 mm, and they were placed in the die to be used as preforms for the pressure infiltration process. The die had a system of cavities and bumps that allowed clamping of the fiber before closing the die and which kept it fixed and stretched during casting. To ensure the filling of the die, the mold included different thoughts, and the inlet was dimensioned to avoid the formation of defects caused by the solidification shrinkage characteristic of the aluminum-silicon alloys. The die was also preheated to 280 °C. The molten aluminum was injected at 680 °C into the die with a piston. After the filling of the mold with liquid aluminum, a 200 bar pressure was applied to the aluminum through the piston by the expansion of compressed nitrogen. The external pressure was maintained until infiltrated aluminum solidified in the die. Finally, the composite material was extracted using 2 pushing elements that were placed at the wider zones of the samples manufactured. Two types of carbon fiber preforms were used: unidirectional carbon fibers (UD); and a twill fabric with fibers with 0° and 90° orientation. Previously, different volume fractions were used to fabricate the composites by adding layers of carbon fiber. There was a limited infiltration of aluminum between the Cf layers causing defects, and the number and size of pores also increased. Therefore, only 1 layer of reinforcement was used to fabricate the composites. Table 1 summarizes the type of samples manufactured and the final proportion of carbon fibers in the samples. After manufacturing, the samples were tested by X-rays. The microstructure was evaluated with an optical microscope (OM) equipped with a Leica ProPlus software to measure the porosity of the specimens and a scanning electron microscope (SEM) with the help of image analysis software. For it, samples were cut, embedded in conductive resin, and mechanically polished (up to 3 µm). The grain size was determined from the obtained micrographs by optical microscopy, following the Equation (1):(1) d=CnL·M where d is the average diameter; C is a constant whose value is 1.5; nL is the number of grains per unit of length, and M is the magnification used. Vickers hardness testing was carried out using a Microhardness Tester (SHIMADZU HMV-2TE) with applied loads of 980.7 mN (HV0.1) and 9.807 N (HV1) for 15 s in different zones of the cross-sections of the samples as well as to identify the hardening effect of the carbon fibers used. The number of measures per zone was 15. Tensile tests were performed in a ZWICK/ROEL TYPE 8594.60 testing machine equipped with a contact extensometer. Elastic limit, tensile strength, and Young’s modulus of each kind of sample were evaluated. At least 10 samples of each kind of material were evaluated. Finally, the fracture surface of each sample was analyzed by scanning electron microscopy (SEM) using a Hitachi S-3400N microscope. 3. Results 3.1. Samples Manufactured The samples were manufactured following the procedure indicated in the experimental section. Figure 2a shows one of the manufactured samples with the dog bone shape and with the throughs for the excess aluminum and the aluminum inlet. The samples did not show external defects, and no carbon fibers were present on the surface. This indicates that the carbon fibers were embedded in the aluminum alloy and that the fabric was not pushed away by the aluminum injected. To evaluate the internal characteristics, the samples were evaluated by X-ray imaging. The X-ray test is not sensitive to carbon fibers, but voids or pores in the aluminum can be easily detected. The images obtained (Figure 2b) showed that there were no relevant or visible defects in the interior. Therefore, the aluminum filled the die and infiltrated the carbon fiber meshes. 3.2. Microstructure of Composites The distribution of the carbon fibers in the composite was evaluated on the cross-section of the central zone of the samples. Figure 3 shows the OM images of the cross-section of the composites. The different orientations of the carbon fibers in the woven-reinforced composite can be observed in Figure 3a,b. The circular sections correspond to fibers that were perpendicular to the observed surface, while elliptical ones correspond to fibers that were oblique to the surface. The microstructure shows that the fibers maintained the woven order during casting. The order was also kept in the unidirectional composite (Figure 3c,d) and all the fibers maintained their original orientation. The density of fibers in the images is higher for the woven (Figure 3a,b) than for the unidirectional composites (Figure 3c,d), as corresponds with the higher reinforcement rate used in the woven (8% vs. 4%). The porosity measured was below 0.5%, which is a particularly good value even for die casting of aluminum pieces. Some small pores (below 30 µm) were observed in the zones where fibers agglomerated, but they were not predominant in the samples. No significant differences in terms of porosity or other defects were observed associated with the type of fabric used. The carbon fibers were distributed all over the sample, but they tended to be located in the central zone of the sample. This was caused by the method developed for holding the carbon fibers during the manufacturing process and is also related to the use of preforms of carbon fibers that were much thinner than the hollow in the die. The fibers in the woven were more aggregated than in the unidirectional composite. In the woven, the carbon fibers were intertwined in two directions, forming a 2D structure that reduces the mobility of the individual fibers during the casting process. On the other hand, the unidirectional fibers were looser and were more easily moved by the flow of molten aluminum during the injection of the metal through the casting process. SEM images in Figure 4 show that aluminum surrounded the carbon fibers and that there were no voids or lack of wetting of the carbon fibers, showing that the Cf were well integrated into the composites. The microstructures of the samples obtained by optical microscopy are shown in Figure 5. In all cases, they were mainly constituted by eutectic morphologies, although some dendritic α-Al zones could also be observed in some zones. This is caused by the high Si content of the AA413 aluminum alloy used. In the microstructure of the alloy shown in Figure 5a, three different zones can be distinguished. Zone 1 corresponds to the zones of the alloy that are in the vicinity of the die; Zone 2 corresponds to those parts of the alloy that are farther from the mold but that are not highly reinforced with carbon fibers; and Zone 3 to those zones in which carbon fibers were preferentially located. In Zone 1, the grain size of the alloy was 3.4 ± 0.2 µm (Figure 5b). This small value is characteristic of the HPDC manufacturing processes as the cold metallic die causes the rapid solidification and cooling of the aluminum alloy. In Zone 2 a dendritic structure with a grain size of 22.0 ± 0.1 µm was observed (Figure 5c). The bigger grain size indicates that the alloy cooled at a lower rate than at the surface of the samples. It also indicates that in these zones, there were no solids that could act as nucleation points for the formation of the solid. In Zone 3, i.e., the zones closer to the carbon fibers, the grain size was 9.0 ± 0.2 µm (Figure 5d,e). This size is smaller than in Zone 2 because the carbon fiber allows a faster solidification of the alloy due to its high thermal conductivity; also, they act as nucleation zones, as has been claimed by different authors with different reinforcements [38]. In the case of the unreinforced alloy, the grain size in the vicinity of the metal die was similar to that of the composite materials (3.6 ± 0.3 µm, as mentioned before). At the center of the samples, the grain size was 24.3 ± 0.1 µm (Figure 5d), which is greater than that observed in the composites, even in zones far from the carbon fibers. This indicates that the carbon fibers affect the cooling of the whole material, not only of the zones closer to them. Apart from the eutectic and the dendritic α-Al zones, other precipitates appeared in the microstructure as bright zones (arrowed in Figure 5f) with diameters of about 5 µm. These precipitates correspond to the α-AlFeMnSi phase and are the result of the presence of Fe and Mn in the alloy used [39]. 3.3. Hardness of Composites The hardness of the composites was evaluated in the cross-section of the samples with a 0.1 and 1 kg loads. The hardness was HV1 112 ± 2 and HV0.1 111 ± 4 for the alloy; HV1 141 ± 2 and HV0.1 135 ± 5 for the woven reinforced composite; and HV1 117 ± 1 and HV0.1 124 ± 4 for the composite with unidirectional fibers. The presence of the carbon fibers increased the microhardness by 22% for the 8 vol.% Cf reinforced woven and by 12% for 4 vol.% Cf reinforced with unidirectional fibers, while the HV1 hardness was 4% greater in the unidirectional composite and 20% in the fiber woven reinforced composite. These results show that the reinforcement has an impact on the hardness of the composite that is greater as the reinforcement increases and that wovens increase hardness in all load ranges, presumably because of its bidirectional effect. The changes observed indicate that the carbon fibers had a hardening effect. Apart from the higher hardness of the reinforcement, the reduction in grain size associated with the presence of fibers could have also increased the hardness of the aluminum matrix. This effect was observed by Gajalakshmi et al. [40] in aluminum matrix composites reinforced with fibers coated with copper and nickel. The hardness varied across the cross-section of the samples (Figure 6). In the unreinforced alloy, the hardness was ~5% higher at the edges than in the central part of the samples. In the composites, the hardness was similar in the center and at the borders, with a variation of ±1% on both sides of the center. The zones with the carbon fibers showed the highest hardness values, i.e., the central zones of the composites, due to the higher hardness of the fibers and their reinforcing effect. In addition, the zones with smaller grain sizes showed high hardness values, i.e., the edges of the samples. Small grain size increases the strength and hardness since grain boundaries act as blocking dislocation movement, as claimed by Inoue et al. [41]. 3.4. Tensile Properties of the Composites The different samples were tensile tested. Table 2 resumes the results from the tests, and Figure 7 represents the values for a better comparison of the results. The yield strength of the die-casting aluminum was 177 MPa, the tensile strength was 216 MPa, and the elongation at break was nearly 1%. The strength values were above those indicated by the supplier, and the elongation at break was comparable. The values provided by the supplier should be valid for casting very different shapes and dimensions; thus, they suppose a lower limit for the material properties. In addition, the slender shape of the samples manufactured and the small grain size observed contributed to increasing the yield and tensile strength of the samples. Obtaining values that were similar to or above the references validates the manufacturing method used, as has been indicated by authors like Mahaviradhan et al. [42]. The yield strength of the composites was 199 ± 3 MPa for the woven-reinforced composite and 188 ± 6 MPa for the unidirectional reinforced. Therefore, the increase in the yield was 12% for the woven and 7% for the unidirectionally reinforced composite compared to the alloy. The tensile strength of the composites was ~275 MPa, which was 25% higher than the unreinforced. Finally, there was also an increase in the values of Young’s modulus. The values measured for the composites were ~84 GPa for the composites, while for the unreinforced alloy was 77 GPa. The increase in Young’s modulus indicates that the load was being transferred from the alloy to the fibers, which is also supported by the increase in strength. The increase in the yield suggests that the fibers, along with smaller grain size, would limit the sliding of atomic planes and pin the defects of the alloys. The higher yield of the woven composite is explained by its higher carbon fiber content. Elongation at break also provided some differences between the alloy and the composites. The elongation of the alloy was 0.84%, which is common in HPDC processes where defects and heterogeneity of the microstructure limit the ductility of the samples. In the composites, the values increased to 1.02% and 1.68% for the woven and unidirectional reinforcements, respectively. The differences indicate that the fibers withheld part of the load during the tests, delaying the failure of the matrix. Both composites show higher elongations at break than the alloy, but there was no contribution from part of the fibers in the woven. It seems that those that were perpendicular to the tensile test direction acted more as defects and strain accumulating zones than as reinforcement. Finally, the standard deviation of the tensile tests for the composites was similar to or lower than for the unreinforced alloy. The deviation arises from premature test failure, generally associated with the presence of porosity and defects. Therefore, a smaller deviation indicates that the samples were reproducible, and the quality of the materials was high. 4. Discussion 4.1. Wettability and Pressure Infiltration The infiltration of aluminum alloys in carbon fibers finds different challenges. The first one is the penetration of the molten aluminum within the fiber meshes because the lack of wettability of the carbon fibers under 900 °C avoids the spontaneous filling of the interfiber spaces. Apart from increasing the temperature, other alternatives have been explored to infiltrate the aluminum: modifying the composition of aluminum alloys by adding elements such as Mg as suggested by Landry et al. [43]; coating the carbon fibers with metals such as Ni [14] or Cu [44], but this results in the modification of the composition of the alloy or using higher pressures. Constantin et al. [45] used a gas pressure-assisted method that allowed applying pressures from 1.2 to 5 MPa, which allowed filling carbon preforms made of short carbon fibers. Squeeze casting processes used pressures from 30 MPa to 50 MPa but to avoid the solidification of the aluminum alloy, the mold and fibers were at 750 °C, and the slow cooling of the composite gave rise to thick microstructures. In addition, the productivity of the technique was very limited [26]. In our case, the infiltration pressure was applied by an HPDC system, in which the pressure was applied to the aluminum alloy by a piston that is firstly mechanically moved and finally impulsed by a gas. There were many differences between the previous methods used and the one proposed in this work. In HPDC infiltration, the pressure used was 200 bar, which allowed the infiltration of aluminum at 680 °C in a few seconds. In addition, unlike the other systems, a cold metallic die was used, which fastly cooled the molten aluminum, resulting in improved properties of the composite matrix, Figure 7. 4.2. Microstructure of the Samples The HPDC process developed was capable of integrating the carbon fibers in the aluminum matrix. The high magnification images of the carbon fibers in the aluminum alloy (Figure 4) showed that there were no gaps between the matrix and reinforcement. This evidences that the pressure used was much higher than the minimum one required to overcome the lack of wetting of the carbon fibers with aluminum at 680 °C even for the short duration of the process. Cf maintained their circular shape after the casting process (Figure 4), indicating that the carbon fibers were neither consumed nor degraded by their reaction with the molten aluminum. In addition, SEM images did not detect the presence of Al4C3. Lee et al. [46] observed the formation of Al4C3 and the degradation of the Cf in low-pressure infiltration at high temperature, and Yang and Scott [47] and de Sanctis et al. [48] observed the same reaction in squeeze casting. Aluminum carbide is brittle and reduces the mechanical properties of the composites, as Zhang et al. [49] reported. In addition, it is highly hygroscopic and, in the presence of humidity, expands and breaks, causing a strong degradation of the composite. We have not observed either Al4C3 or the degradation of the carbon fibers, and this explains the excellent mechanical properties obtained for the composites. In HPDC the process temperature was kept low, slightly above the aluminum melting temperature, and the interaction time between the molten aluminum and the carbon fibers was very short. The low interaction time was the second characteristic of the processing route used. The die was filled with aluminum in less than 0.5 s, and then high pressure was applied and maintained. The solidification of the aluminum took only a few seconds more. Simulations carried out in different conditions indicate that, even without carbon fibers, solidification times would be less than 4 s for the entire sample. These times were further reduced by the presence of carbon fibers, which also remove heat from the molten aluminum alloy. The short interaction time, together with the low temperature used in this process, prevented the formation of detrimental Al4C3. Wielage and Dorner [50] stated that composites with Al4C3 have higher corrosion than those without this phase. In the microstructure, some phases constituted by Al, Fe, Mn, and Si were observed (arrowed zone in Figure 4b). The alloy used contained 0.65 wt.% Fe that was used to reduce the adherence of the samples to the mold, but it could promote the formation of acicular β-AlFeSi precipitates. These precipitates were not observed in the microstructure of the alloy because of the presence in the alloy of 0.55 wt.% Mn, which transformed the β-AlFeSi precipitates into the compact α-AlFeMnSi ones which were less detrimental to the properties of the material, as was explained by Lu et al. [51]. The composition of the alloy included Mg. The precipitates that Mg usually forms in 4xx alloys were Mg2Si and helped to increase the strength of the alloys. However, the alloy contained 0.1 wt.%, which makes its proportion small. In any case, their size is usually in the nanometer range; thus, special techniques would be necessary to determine their presence. One important feature that requires analysis is the heterogeneous distribution of the carbon fibers in the composite. Fibers were preferentially located at the central zone of the transversal section of the samples because the cavity of the die was designed to be filled with aluminum on the fixed and mobile platen to avoid the presence of carbon fibers at the surface of the samples. This distribution would have important effects on the corrosion and mechanical properties of the materials. The fibers were not present at the surface; thus, corrosion galvanic couples between carbon and the aluminum matrix were not formed. This solves the corrosion risks related to these composites, which are the cause of avoiding the combination of carbon fiber composites with aluminum bolts or pieces in the aeronautical and automotive industry. This is one of the causes of giving preference to the location of fibers inside the composite rather than forcing a homogenous distribution in the transversal section of the samples. On the other hand, the preferential location of carbon fibers at the central zone of the cross-section can be a challenge because the properties of the composite are not homogenous across its section. These differences may cause the apparition of some added stress within the samples. In addition, the behavior of the composite would be affected by the uneven distribution of the carbon fibers. To assess how critical this aspect could be, it is worth comparing it to carbon fiber-reinforced polymers (CFRPs). CFRPs are manufactured by stacking plies that are typically less than 0.2 mm thick in different orientations. The tensile strength of a layer in the direction of its fibers is very high. The resistance of a layer with fibers perpendicular to the applied load would be only that of the resin, which is very low [52]. However, the combination of layers that have such different properties has been accepted and used in CFRP laminates. In our system, the resistance parallel to the carbon fibers is the combination of the resistance of the carbon fibers added to that of the aluminum alloy, which has values of strength above those of the epoxy matrices used in many CFRPs. In the transversal direction, the resistance would be that of the aluminum alloy, which is still much higher than that of many polymers used in CFRPs. Therefore, the properties of the aluminum composites manufactured could be easily modeled, and the heterogeneous distribution of the carbon fibers should not result in a detrimental behavior of the composites for structural applications. In addition, the aluminum matrix composites reinforced with unidirectional fibers should show different properties between the parallel and the transversal directions of the material, i.e., mechanical anisotropy, as the orientation of the carbon fibers determines the behavior of the composite. The anisotropy should be lower when using wovens as the reinforcement improves the properties in two perpendicular directions equivalently. 4.3. Mechanical Properties 4.3.1. Modeling the Mechanical Properties The composites developed have a higher Young’s modulus, yield, and maximum resistance than the alloy, indicating that load was transferred between the two phases of the composite, i.e., the aluminum matrix and the carbon fibers. The Young’s modulus is similar in all aluminum alloys as it is not affected by the presence of different alloying elements or different tempering treatments. Therefore, the reinforcing effect of carbon fibers caused the observed change. We have used the rule of mixtures (ROM) and the Halpin–Tsai (HT) models to evaluate the effective reinforcement rate of the composites. The ROM model considers that the two phases present in the composite, i.e., carbon fibers and aluminum matrix, behave with a perfect transference of load from one another. In this model, the final Young’s modulus value only depends on Young’s modulus of each phase and the corresponding volume fractions. The general equation for our system could be the one shown in Equation (2). (2) Ec=Efvf+Emvm   with   vf+vm=1 where Ec, Ef, and Em refer to the Young’s modulus of the composite, the fiber, and the matrix, respectively; vf and vm refer to the volume fraction of the fiber and the matrix, respectively. This equation can be directly applied to the unidirectional reinforced composite as Young’s modulus of aluminum is isotropic, and carbon fibers are aligned in the direction of the tensile test. In the case of the woven composite, the fibers that are in the longitudinal direction of the tensile tests (0°) have a different contribution from the ones placed in the transversal one (90°). In this case, although a laminated model could provide more precise results, the ROM also allows us to consider that we have three different phases as Equation (3) shows. (3) Ec=EfLvfL+EfTvfT+Emvm   with   vfL+vfT+vm=1 where EfL and EfT refer to the Young’s Modulus of the 0° fibers and 90° fibers, respectively; while vfL and vfT refer to the volume fraction of the fiber along the 0° and 90°, respectively. The values measured for Young’s modulus were 85 GPa and 84 GPa for the unidirectional and bidirectionally reinforced composites, respectively. The Young’s modulus of the carbon fibers was 231 GPa in the fiber direction and 13 GPa in the transversal one, and its maximum resistance was 4480 MPa, as shown by Herráez et al. [53]. The ROM according to Equation (1) indicates that for 4% vol. of Cf Young’s modulus should be 83.2 GPa and in the case of the bidirectional reinforcement, the expected Young’s modulus was 89.3 GPa (Table 3). These results approximate to values measured; thus, it indicates that there was an adequate transference of load from the matrix to the fibers. Islam and Begum [54] explained in their work the Halpin–Tsai model, which incorporates the effect of the interactions between fibers and matrix, and the form factor, i.e., the length (L) to diameter (D) rate, of the reinforcing fibers. Therefore, we assume that we have an ideal solid with no pores and that there is no interphase between matrix and reinforcement. Equation (4) can be directly applied to unidirectional composites and in it. (4) Ec=Em1+ξηvf1−ηvf   with   η=Ef/Em−1Ef/Em+ξ   and   ξ=2LD In our system, the form factor of the fibers was ~30,000; thus, the model approximates better than 0.5% to the ROM model, and the expected values nearly reproduce the ones from ROM, indicating an expected value for the unidirectional composite of 83.5 GPa. In the case of the transversal fibers, two different approaches may be applied: (i) the transversal carbon fibers act as defects in the matrix; thus, it is like considering the presence of voids with a volume proportion corresponding to that of the transversal fibers; (ii) the transversal fibers act as very short fibers with a length similar to its diameter. The model is not valid for this second type of approximation, as the shape of the fibers is not adequate; thus, the first considered case has been included in Table 3. The value proposed by the model is 83.5 GPa for the unidirectional composite and 80.2 GPa for the bidirectionally reinforced composites. Both values are close to the experimental results, but we have not observed the reduction in Young’s modulus expected for the bidirectional reinforcement. The models used fit the experimental results of the Young’s Modulus thus it can be determined that an effective reinforcement was obtained by the incorporation of unidirectional carbon fibers in the composite and that the presence of transversal fibers did not degrade Young’s modulus of the composite. The combined results of a 7–12% increase in yield strength, 25% in tensile strength, and 10% in Young’s modulus indicate that the carbon fibers had an effect on the mechanical properties of the composite and that the load applied to the samples was partially transferred to the carbon fibers. Otherwise, the carbon fibers would have acted as defects in the composites and would have reduced the performance of the composites in the tests. One relevant consideration is that, although the woven doubles the reinforcement rate of the alloy, Young’s modulus and the tensile strength are the same. This can be explained by the relative direction of carbon fibers with the direction of the stress applied to the samples during the tests. Only the carbon fibers in the test direction were capable of withholding load and, therefore, increasing Young’s modulus of the composite. The fibers that were perpendicular to the test direction contributed with their transversal Young’s modulus, which is substantially lower than the longitudinal one. Therefore, they could act as defects and lower the properties of the composite, but this effect has not been observed. In particular, the models used to calculate Young’s modulus in composites provide effective reinforcement rates of ~4.6%, which are very similar to the volume fraction of reinforcement aligned in the test direction. In addition, considering the tensile resistance of the alloy and the carbon fibers, the maximum expected resistance for the composite would be 303 MPa, while the values obtained were ~276 MPa. Considering the most basic models for fiber-reinforced composites, it can be indicated that the effective reinforcement rate was in the range of 1.6–2.6%. This value is a minimum reference as the ripples of the fibers in the woven reduced the effectivity of the load transfer. Again, this indicates that most of the fibers in the load direction were effectively contributing to the stiffness of the composites. The transference of load between matrix and reinforcement modified the hardness of the samples. Hardness has usually been related to the elastic yield of the alloys, and a proportional behavior has been observed. The woven reinforced composite showed a hardness increase of 22%, while the increase was 12% for the unidirectional composite. These values are similar to the observed increases in yield strength, which was 12% for the aluminum reinforced with woven and 6% for the unidirectionally reinforced one. The hardness increase was not greater because of the orientation of the fibers. In the woven, half of the fibers were perpendicular to the cross-section where hardness was measured, but the other half was placed in the surface plane. Therefore, the effect of the fibers perpendicular to the indent was limited. In addition to the mechanical contribution of the fibers, the grain size distribution also had a strong effect on hardness. Therefore, the carbon fibers had a twofold contribution. On the one hand, Sree Manu et al. [55] claimed that they increase hardness as a result of their mechanical properties, particularly their high modulus and strength, and they act as a barrier to matrix dislocation. On the other hand, carbon fibers also form finer microstructures, like in other composites made by different techniques, as Aynalem [56] indicates, because their high thermal conductivity leads to the fast cooling and solidification of the alloy, while they act as pinning zones for the grains nucleation, resulting in a synergic increase of hardness in the alloy. 4.3.2. Fractographic Analysis Figure 8 shows the fracture surface of the different samples as observed by SEM. In all cases, the morphology corresponds to samples without casting defects. In the case of the alloy, cracks started by the clustering of microvoids (Figure 8a–c with orange arrows in the dimples). The fracture surface showed the characteristic dimples of the ductile fracture of many metallic materials, but due to the fine microstructure, their size was very small. In the case of the samples reinforced with carbon fiber woven (Figure 8d–f), a more complex fracture mechanism was present. Cracks started at zones of the alloy that were not highly reinforced. Then cracks coalesced at the fibers that were approximately perpendicular to the tensile test direction. Moreover, the final fracture took place at zones with high content of carbon fibers. Due to this, the end of long fibers was visible protruding from the surface of the fractured samples (Figure 8e). In all cases, fibers were free of aluminum, and their shape was not modified by the HPDC process (Figure 8e). The unidirectional reinforced composite fracture image (Figure 8g) shows that the cracks did not start with any defect in the structure. The coalescence of microvoids in the central zone of the samples was the cause of the fracture of the samples. Observing them with a higher magnification (Figure 8h,i) allowed determining the presence of many fibers sticking out of the surface of the fracture surface, as well as many holes left by the fibers that remained on the other side of the sample. This can be observed with more detail in Figure 8g in which some fibers protrude on the surface while some perfectly circular holes can be observed. In addition, in Figure 8i, it can be appreciated that the voids formed in the aluminum matrix were bigger than those observed in the unreinforced alloy. Therefore, the carbon fibers played an important role in the fracture of the composites. On the one hand, they helped to increase the elongation at the break of the samples, particularly for the unidirectionally reinforced composite. In this case, the direction of the fibers was along the applied strain, which favored the higher elongation of the samples. According to the models shown, the 4 vol.% of Cf retained 30% of the load applied, thus the rest of the alloy evolved with less strain. The higher elongation observed is related to forming of larger voids in the matrix before breaking, favored by the fine microstructure. In the case of the woven-reinforced alloys, the fibers that were in the 0° direction had the same effect as in the UD composite, supporting the load and helping for the deformation of the samples, but the ones oriented at 90° behaved differently. As they could not slide or retain load, they acted as barriers to deformation of the composite and as defects that helped fracture the samples. This causes the elongation at break in this configuration to be lower than in the UD composite. Despite this analysis, if the load had been in another direction, the woven-reinforced material would have had similar behavior, while the UD composite would have behaved similarly to that of the matrix or even worse. 5. Conclusions The main conclusions derived from this work are the following:Aluminum matrix composites consisting of AA413 aluminum alloy reinforced with long carbon fibers in the shape of unidirectional and woven fabrics were manufactured by HPDC by incorporating the fabrics in the metallic die; Fibers were not degraded by the interaction with aluminum and most of the fibers remained in the zones where they were positioned before the entrance of the molten aluminum; The pressure applied by the HPDC process allowed the filling of the zones within the woven meshes and fibers, and the composites were practically free from defects. The carbon fibers limited the grain growth in the composites; thus, finer microstructures were obtained. As a result of this, hardness increased in the composites and more in the woven-reinforced composite because of its higher volumetric reinforcement fraction; The carbon fiber reinforcement used provided a 10% increase in Young’s modulus, a 6% with UD Cf and 12% with woven in yield strength, 28% increase in tensile strength, and the increase in elongation at break was 86% for the UD and 21% for the woven. The properties observed have been explained by the rule of mixtures models or by the Halpin–Tsai one, with a general deviation of less than 7%. Fractographic tests showed that the presence of fibers acted differently depending on the orientation. Fibers along the tensile test direction (0°) favored the ductile behavior of the alloy, and after the break, fibers arose from the fracture surface. In the woven-reinforced composite, the 90° fibers did not help to increase the strength of the composite and limited the deformation of the sample, therefore acting as defects. The material with the highest yield strength and tensile strength was the aluminum reinforced with woven with 8 vol.% Cf, while the material with the highest Young’s Modulus and elongation at break was the aluminum reinforced with 4 vol.% of unidirectional carbon fibers. 6. Patents Marino Martín, P.L.; Arias Martín, R.; Carrero Hinojal, A.; Torres Barreiro, B.; Rams Ramos, J.; Ureña Fernández, A.; Sánchez Martínez, M.; López Galisteo, A.J.; Rodrigo Herrero, P.; Bedmar Sanz, J.; Mercado Sapia, C. Fabricación de materiales compuestos reforzados con fibra de carbono mediante inyección de una aleación de aluminio de alta presión. Spain Patent ES2802282, 18 May 2021. Author Contributions Conceptualization, B.T. and J.R.; methodology, B.T. and J.R.; validation, B.T. and J.R.; formal analysis, J.B., B.T. and J.R.; investigation, J.B., B.T. and J.R.; resources, B.T. and J.R.; data curation, J.B.; writing—original draft preparation, J.B.; writing—review and editing, B.T. and J.R.; visualization, B.T. and J.R.; supervision, B.T. and J.R.; project administration, B.T. and J.R.; funding acquisition, B.T. and J.R. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Scheme of the high-pressure die casting process. Figure 2 (a) As-manufactured composite sample, and (b) X-ray image of the sample. Figure 3 Optical micrographs of the AMCs: (a,b) samples manufactured with woven, and (c,d) samples manufactured with unidirectional fiber. Figure 4 SEM images of the unidirectional carbon fibers in the Al matrix: (a) microstructure of aluminum reinforced with unidirectional carbon fiber; and (b) magnification of (a). Figure 5 Microstructure of the sample obtained by optical microscopy: (a) general view of the cross-section of a sample reinforced with woven: zone 1 corresponds to the microstructure close to the die; zone 2 to the microstructure far from the die and the fibers and zone 3 to the microstructure close to the fiber; (b) zones in contact with the mold; (c) zones far from the mold and the fibers with arrowed pore; (d) samples with reinforcement in zones near the fibers; (e) magnification of (d,f) zones with arrowed α-AlFeMnSi phases. Figure 6 Vickers’s microhardness values measured on the cross-sections of the alloy and AMCs samples. Figure 7 Results of the tensile tests: (a) yield strength; (b) Young’s modulus; (c) tensile strength; and (d) elongation at break. Figure 8 SEM fracture surfaces: (a) AA413 aluminum alloy and (b,c) magnification with dimples arrowed in orange; (d) aluminum reinforced with woven and (e,f) magnification of the extracted fibers; (g) aluminum reinforced with 0° unidirectional fiber and (h,i) magnification of different zones. materials-15-03400-t001_Table 1 Table 1 Samples manufactured. Sample Reinforcement Alloy None 0% Woven reinforced 0/90° 2 × 2 twill 8 vol.% 0° UD reinforced 0° UD 4 vol.% materials-15-03400-t002_Table 2 Table 2 Results of the mechanical properties. Sample Yield Strength (MPa) Tensile Strength (MPa) Elongation at Break (%) Young’s Modulus (GPa) Alloy 177 ± 8 216 ± 15 0.84 ± 0.18 77 ± 5 Woven reinforced 199 ± 3 276 ± 10 1.02 ± 0.12 84 ± 4 0° UD reinforced 188 ± 6 273 ± 19 1.68 ± 0.17 85 ± 5 materials-15-03400-t003_Table 3 Table 3 Results of the mechanical models. ROM (GPa) HT (GPa) Al—Cf u Al—Cf twill Al—Cf u Al—Cf twill 83.2 89.3 83.5 80.2 u: unidirectional reinforced composites; twill: bidirectional reinforced composite, woven. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Reddy P.V. Kumar G.S. Krishnudu D.M. Rao H.R. Mechanical and Wear Performances of Aluminium-Based Metal Matrix Composites: A Review J. Bio-Tribo-Corrosion 2020 6 83 10.1007/s40735-020-00379-2 2. Akinwamide S.O. Akinribide O.J. Olubambi P.A. Microstructural Evolution, Mechanical and Nanoindentation Studies of Stir Cast Binary and Ternary Aluminium Based Composites J. Alloy. Compd. 2021 850 156586 10.1016/j.jallcom.2020.156586 3. Akinwamide S.O. Abe B.T. Akinribide O.J. Obadele B.A. Olubambi P.A. Characterization of Microstructure, Mechanical Properties and Corrosion Response of Aluminium-Based Composites Fabricated via Casting—A Review Int. J. Adv. Manuf. Technol. 2020 109 975 991 10.1007/s00170-020-05703-1 4. Ding H. Bao X. Jamili-Shirvan Z. Jin J. Deng L. Yao K. Gong P. Wang X. Enhancing Strength-Ductility Synergy in an Ex Situ Zr-Based Metallic Glass Composite via Nanocrystal Formation within High-Entropy Alloy Particles Mater. Des. 2021 210 110108 10.1016/j.matdes.2021.110108 5. Tan Z. Wang L. Xue Y. Zhang P. Cao T. Cheng X. High-Entropy Alloy Particle Reinforced Al-Based Amorphous Alloy Composite with Ultrahigh Strength Prepared by Spark Plasma Sintering Mater. Des. 2016 109 219 226 10.1016/j.matdes.2016.07.086 6. Chak V. Chattopadhyay H. Dora T.L. A Review on Fabrication Methods, Reinforcements and Mechanical Properties of Aluminum Matrix Composites J. Manuf. Process. 2020 56 1059 1074 10.1016/j.jmapro.2020.05.042 7. Arunkumar S. Sundaram M.S. Suketh K.M. Vigneshwara S. Materials Today: Proceedings A Review on Aluminium Matrix Composite with Various Reinforcement Particles and Their Behaviour Mater. Today Proc. 2020 33 484 490 10.1016/j.matpr.2020.05.053 8. Srinivasan R. Pridhar T. Kirubakaran R. Ramesh A. Materials Today: Proceedings Prediction of Wear Strength of Squeeze Cast Aluminium Hybrid Metal Matrix Composites Using Response Surface Methodology Mater. Today Proc. 2020 27 1806 1811 10.1016/j.matpr.2020.03.779 9. Ramos-masana A. Colominas C. Surface & Coatings Technology Evaluation of DC-MS and HiPIMS TiB 2 and TaN Coatings as Diffusion Barriers against Molten Aluminum: An Insight into the Wetting Mechanism Surf. Coat. Technol. 2019 375 171 181 10.1016/j.surfcoat.2019.06.031 10. Sarina B.A.O. Kai T. Kvithyld A. Engh T. Tangstad M. Wetting of Pure Aluminium on Graphite, SiC and Al 2 O 3 in Aluminium Filtration. Trans Nonferrous Met. Soc. China 2012 22 1930 1938 10.1016/S1003-6326(11)61410-6 11. Fu Y. Li H. Cao W. Enhancing the Interfacial Properties of High-Modulus Carbon Fiber Reinforced Polymer Matrix Composites via Electrochemical Surface Oxidation and Grafting. Compos Part A Appl. Sci. Manuf. 2020 130 105719 10.1016/j.compositesa.2019.105719 12. Tong H. Qiu F. Zuo R. Shen P. Cong X. Liu J. Yang H. Jiang Q. The Effect and Mechanism of Alloying Elements on Al/SiC Interfacial Reaction in Al Melt Appl. Surf. Sci. 2020 501 144265 10.1016/j.apsusc.2019.144265 13. Bhujanga D.P. Manohara H.R. Processing and Evaluation of Mechanical Properties and Dry Sliding Wear Behavior of AA6061-B4C Composites Mater. Today Proc. 2018 5 19773 19782 10.1016/j.matpr.2018.06.340 14. Rams J. Ureña A. Escalera M.D. Sánchez M. Electroless Nickel Coated Short Carbon Fibres in Aluminium Matrix Composites Compos. Part A Appl. Sci. Manuf. 2007 38 566 575 10.1016/j.compositesa.2006.02.010 15. Alten A. Erzi E. Gürsoy Ö. Hapçı Ağaoğlu G. Dispinar D. Orhan G. Production and Mechanical Characterization of Ni-Coated Carbon Fibers Reinforced Al-6063 Alloy Matrix Composites J. Alloys Compd. 2019 787 543 550 10.1016/j.jallcom.2019.02.043 16. Zhang J. Liu S. Liu J. Lu Y. Liu Y. Li T. Electroless Nickel Plating and Spontaneous Infiltration Behavior of Woven Carbon Fibers Mater. Des. 2020 186 108301 10.1016/j.matdes.2019.108301 17. Bhav Singh B. Balasubramanian M. Processing and Properties of Copper-Coated Carbon Fibre Reinforced Aluminium Alloy Composites J. Mater. Process. Technol. 2009 209 2104 2110 10.1016/j.jmatprotec.2008.05.002 18. Lancin M. Marhic C. TEM Study of Carbon Fibre Reinforced Aluminium Matrix Composites: Influence of Brittle Phases and Interface on Mechanical Properties J. Eur. Ceram. Soc. 2000 20 1493 1503 10.1016/S0955-2219(00)00021-2 19. Ahlatci H. Koçer T. Candan E. Çimenoğlu H. Wear Behaviour of Al/(Al2O3p+SiCp) Hybrid Composites Tribol. Int. 2006 39 213 220 10.1016/j.triboint.2005.01.029 20. Nour-Eldin M. Elkady O. Yehia H.M. Timeless Powder Hot Compaction of Nickel-Reinforced Al/(Al2O3-Graphene Nanosheet) Composite for Light Applications Using Hydrazine Reduction Method J. Mater. Eng. Perform. 2022 1 16 10.1007/s11665-022-06749-w 21. Yehia H.M. Allam S. Hot Pressing of Al-10 Wt% Cu-10 Wt% Ni/x (Al2O3–Ag) Nanocomposites at Different Heating Temperatures Met. Mater. Int. 2021 27 500 513 10.1007/s12540-020-00824-4 22. Elkady O.A. Yehia H.M. Ibrahim A.A. Elhabak A.M. Elsayed E.M. Mahdy A.A. Direct Observation of Induced Graphene and SiC Strengthening in Al–Ni Alloy via the Hot Pressing Technique Crystals 2021 11 1142 10.3390/cryst11091142 23. Kientzl I. Dobránszky J. Production and Examination of Double Composites Mater. Sci. Forum 2007 537–538 191 198 10.4028/www.scientific.net/MSF.537-538.191 24. Kientzl I. Orbulov I.N. Dobránszky J. Németh Á. Mechanical Behaviour Al-Matrix Composite Wires in Double Composite Structures Advances in Science and Technology Lublin University of Technology Lublin, Poland 2006 147 152 10.4028/www.scientific.net/AST.50.147 25. Venkatesan S. Xavior M.A. Analysis of Mechanical Properties of Aluminum Alloy Metal Matrix Composite by Squeeze Casting—A Review Mater. Today Proc. 2018 5 11175 11184 10.1016/j.matpr.2018.01.141 26. Hajjari E. Divandari M. Arabi H. Effect of Applied Pressure and Nickel Coating on Microstructural Development in Continuous Carbon Fiber-Reinforced Aluminum Composites Fabricated by Squeeze Casting Mater. Manuf. Process. 2011 26 599 603 10.1080/10426910903447311 27. Baghi M. Niroumand B. Emadi R. Fabrication and Characterization of Squeeze Cast A413-CSF Composites J. Alloys Compd. 2017 710 29 36 10.1016/j.jallcom.2017.03.136 28. Ju C.P. Chen K.I. Lin J.H.C. Process, Microstructure and Properties of Squeeze-Cast Short-Carbon-Fibre-Reinforced Aluminium-Matrix Composites J. Mater. Sci. 1994 29 5127 5134 10.1007/BF01151107 29. Sukumaran K. Ravikumar K.K. Pillai S.G.K. Rajan T.P.D. Ravi M. Pillai R.M. Pai B.C. Studies on Squeeze Casting of Al 2124 Alloy and 2124-10% SiCp Metal Matrix Composite Mater. Sci. Eng. A 2008 490 235 241 10.1016/j.msea.2008.01.054 30. Nishida Y. Shirayanagi I. Sakai Y. Infiltration of Fibrous Preform by Molten Aluminum in a Centrifugal Force Field Metall. Mater. Trans. A 1996 27 4163 4169 10.1007/BF02595664 31. Zhiming Y. Tengteng W. Jinxu L. Lin Z. Shukui L. Interface Optimization and Mechanical Properties of Cu-Coated Carbon Fiber Cloth/Titanium Alloy Composite Rare Met. Mater. Eng. 2017 46 869 875 10.1016/S1875-5372(17)30114-5 32. Torres B. Taltavull C. López A.J. Campo M. Rams J. Al/SiCp and Al11Si/SiCp Coatings on AZ91 Magnesium Alloy by HVOF Surf. Coat. Technol. 2015 261 130 140 10.1016/j.surfcoat.2014.11.045 33. Torres B. Campo M. Ureña A. Rams J. Thermal Spray Coatings of Highly Reinforced Aluminium Matrix Composites with Sol–Gel Silica Coated SiC Particles Surf. Coat. Technol. 2007 201 7552 7559 10.1016/j.surfcoat.2007.02.023 34. Das P. Bhuniya B. Samanta S.K. Dutta P. Studies on Die Filling of A356 Al Alloy and Development of a Steering Knuckle Component Using Rheo Pressure Die Casting System J. Mater. Process. Technol. 2019 271 293 311 10.1016/j.jmatprotec.2019.04.014 35. Gu C. Lu Y. Cinkilic E. Miao J. Klarner A. Yan X. Luo A.A. Predicting Grain Structure in High Pressure Die Casting of Aluminum Alloys: A Coupled Cellular Automaton and Process Model Comput. Mater. Sci. 2019 161 64 75 10.1016/j.commatsci.2019.01.029 36. Tian C. Law J. Van der Touw J. Murray M. Yao J.-Y. Graham D. John D.S. Effect of Melt Cleanliness on the Formation of Porosity Defects in Automotive Aluminium High Pressure Die Castings J. Mater. Process. Technol. 2002 122 82 93 10.1016/S0924-0136(01)01229-8 37. Qi M. Kang Y. Li J. Wulabieke Z. Xu Y. Li Y. Liu A. Chen J. Microstructures Refinement and Mechanical Properties Enhancement of Aluminum and Magnesium Alloys by Combining Distributary-Confluence Channel Process for Semisolid Slurry Preparation with High Pressure Die-Casting J. Mater. Process. Technol. 2020 285 116800 10.1016/j.jmatprotec.2020.116800 38. Samal P. Vundavilli P.R. Meher A. Mahapatra M.M. Recent Progress in Aluminum Metal Matrix Composites: A Review on Processing, Mechanical and Wear Properties J. Manuf. Process. 2020 59 131 152 10.1016/j.jmapro.2020.09.010 39. Zhang X. Wang D. Nagaumi H. Zhou Y. Yu W. Chong X. Li X. Zhang H. Morphology, Thermal Stability, Electronic Structure and Mechanical Properties of α-AlFeMnSi Phases with Varying Mn/Fe Atomic Ratios: Experimental Studies and DFT Calculations J. Alloy. Compd. 2022 901 163523 10.1016/j.jallcom.2021.163523 40. Gajalakshmi K. Senthilkumar N. Mohan B. Anbuchezhiyan G. An Investigation on Microstructure and Mechanical Behaviour of Copper-Nickel Coated Carbon Fibre Reinforced Aluminium Composites Mater. Res. Express 2020 7 115701 10.1088/2053-1591/abcb4c 41. Inoue T. Horita Z. Somekawa H. Ogawa K. Effect of Initial Grain Sizes on Hardness Variation and Strain Distribution of Pure Aluminum Severely Deformed by Compression Tests Acta Mater. 2008 56 6291 6303 10.1016/j.actamat.2008.08.042 42. Mahaviradhan N. Sivaganesan S. Padma Sravya N. Parthiban A. Experimental Investigation on Mechanical Properties of Carbon Fiber Reinforced Aluminum Metal Matrix Composite Mater. Today Proc. 2020 39 743 747 10.1016/j.matpr.2020.09.443 43. Landry K. Kalogeropoulou S. Eustathopoulos N. Wettability of Carbon by Aluminum and Aluminum Alloys Mater. Sci. Eng. A 1998 254 99 111 10.1016/S0921-5093(98)00759-X 44. Ureña A. Rams J. Escalera M.D. Sánchez M. Effect of Copper Electroless Coatings on the Interaction between a Molten Al–Si–Mg Alloy and Coated Short Carbon Fibres. Compos Part A Appl. Sci. Manuf. 2007 38 1947 1956 10.1016/j.compositesa.2007.02.005 45. Constantin H. Harper L. Kennedy A.R. Pressure-Assisted Infiltration of Molten Metals into Non-Rigid, Porous Carbon Fibre Structures J. Mater. Process. Technol. 2018 255 66 75 10.1016/j.jmatprotec.2017.11.059 46. Lee M. Choi Y. Sugio K. Matsugi K. Sasaki G. Effect of Aluminum Carbide on Thermal Conductivity of the Unidirectional CF/Al Composites Fabricated by Low Pressure Infiltration Process Compos. Sci. Technol. 2014 97 1 5 10.1016/j.compscitech.2014.03.022 47. Yang M. Scott V.D. Carbide Formation in a Carbon Fibre Reinforced Aluminium Composite Carbon N.Y. 1991 29 877 879 10.1016/0008-6223(91)90162-C 48. De Sanctis M. Pelletier S. Bienvenue Y. On the Formation of Interfacial Carbides in a Carbon Fibre-Reinforced Aluminium Composite Carbon N.Y. 1994 32 925 930 10.1016/0008-6223(94)90050-7 49. Zhang J. Liu S. Lu Y. Jiang L. Zhang Y. Li T. Semisolid-Rolling and Annealing Process of Woven Carbon Fibers Reinforced Al-Matrix Composites J. Mater. Sci. Technol. 2017 33 623 629 10.1016/j.jmst.2017.01.002 50. Wielage B. Dorner A. Corrosion Studies on Aluminium Reinforced with Uncoated and Coated Carbon Fibres Compos. Sci. Technol. 1999 59 1239 1245 10.1016/S0266-3538(98)00163-8 51. Deng Y. Mao Z. Yang N. Niu X. Lu X. Collaborative Optimization of Density and Surface Roughness of 316L Stainless Steel in Selective Laser Melting Materials 2020 13 1601 10.3390/ma13071601 52. Jeoung S.K. Hwang Y.J. Lee H.W. Son S.K. Kim H.S. Ha J.U. Study on Properties of CFRP Fabricated by VA-RTM Process AIP Publishing LLC Melville, NY, USA 2016 120016 10.1063/1.4942331 53. Herráez M. Mora D. Naya F. Lopes C.S. González C. Llorca J. Transverse Cracking of Cross-Ply Laminates: A Computational Micromechanics Perspective Compos. Sci. Technol. 2015 110 196 204 10.1016/j.compscitech.2015.02.008 54. Islam M.A. Begum K. Prediction Models for the Elastic Modulus of Fiber-Reinforced Polymer Composites: An Analysis J. Sci. Res. 2011 3 225 238 10.3329/jsr.v3i2.6881 55. Manu K.S. Raag L.A. Rajan T. Pai B. Petley V. Verma S.N. Self-Lubricating Bidirectional Carbon Fiber Reinforced Smart Aluminum Composites by Squeeze Infiltration Process J. Mater. Sci. Technol. 2019 35 2559 2569 10.1016/j.jmst.2019.04.034 56. Aynalem G.F. Processing Methods and Mechanical Properties of Aluminium Matrix Composites Adv. Mater. Sci. Eng. 2020 2020 1 19 10.1155/2020/3765791

PMC009xxxxxx/PMC9099679.txt

==== Front J Clin Med J Clin Med jcm Journal of Clinical Medicine 2077-0383 MDPI 10.3390/jcm11092486 jcm-11-02486 Review Asian Race and Primary Open-Angle Glaucoma: Where Do We Stand? https://orcid.org/0000-0003-2247-6942 Belamkar Aditya 1 Harris Alon 2 Oddone Francesco 3 Verticchio Vercellin Alice 2 Fabczak-Kubicka Anna 24 Siesky Brent 2* Pinazo-Duran Maria Dolores Academic Editor 1 Department of Ophthalmology, Indiana University School of Medicine, Indianapolis, IN 46202, USA; abelamka@iu.edu 2 Department of Ophthalmology, Icahn School of Medicine at Mount Sinai, New York, NY 10029, USA; alon.harris@mssm.edu (A.H.); alice.verticchio@mssm.edu (A.V.V.); anna.fabczak-kubicka@mssm.edu (A.F.-K.) 3 IRCCS-Fondazione Bietti, 00198 Rome, Italy; oddonef@gmail.com 4 New York Eye & Ear Infirmary of Mount Sinai, New York, NY 10003, USA * Correspondence: brent.siesky@mssm.edu; Tel.: +1-212-241-2831 28 4 2022 5 2022 11 9 248623 3 2022 24 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Primary open-angle glaucoma (POAG) is an optic neuropathy characterized by irreversible retinal ganglion cell damage and visual field loss. The global POAG prevalence is estimated to be 3.05%, and near term is expected to significantly rise, especially within aging Asian populations. Primary angle-closure glaucoma disproportionately affects Asians, with up to four times greater prevalence of normal-tension glaucoma reported compared with high-tension glaucoma. Estimates for overall POAG prevalence in Asian populations vary, with Chinese and Indian populations representing the majority of future cases. Structural characteristics associated with glaucoma progression including the optic nerve head, retina, and cornea are distinct in Asians, serving as intermediates between African and European descent populations. Patterns in IOP suggest some similarities between races, with a significant inverse relationship between age and IOP only in Asian populations. Genetic differences have been suggested to play a role in these differences, however, a clear genetic pattern is yet to be established. POAG pathogenesis differs between Asians and other ethnicities, and it may differ within the broad classification of the Asian race. Greater awareness and further research are needed to improve treatment plans and outcomes for the increasingly high prevalence of normal tension glaucoma within aging Asian populations. glaucoma primary open angle glaucoma normal tension glaucoma Asian population race intraocular pressure optic nerve head cornea retina genetics National Institutes of HealthR01EY030851 National Science FoundationNSF DMS (1853222/2021192) NYEE FoundationNYEE Foundation grant Research to Prevent BlindnessChallenge Grant award Alon Harris is supported by NIH grant (R01EY030851), NSF DMS (1853222/2021192), NYEE Foundation grant, and in part by a Challenge Grant award from Research to Prevent Blindness, NY. ==== Body pmc1. Introduction Glaucoma describes a family of optic neuropathies characterized by retinal ganglion cell loss, optic nerve head alterations, and the retinal nerve fiber layer (RNFL) thinning with associated progressive visual field loss, beginning in the peripheries. The global prevalence of the disease has been estimated to be 3.54% in populations aged 40 to 80 and 3.05% for primary open-angle glaucoma specifically (POAG) [1]. The disease is expected to afflict over 100 million individuals by 2040 [1]. Currently, intraocular pressure (IOP) is the only modifiable risk factor, with strong evidence establishing the reduction of IOP, even for those with clinically normal pressures, to significantly decrease the risk of progression [2,3]. Several risk factors for glaucoma have been identified including elevated IOP, family history, older age, and, of particular importance, race [1,3]. In fact, it has been estimated that persons of African descent may have a 2.8 times higher prevalence of POAG than those of European descent [1]. Additionally, normal-tension glaucoma (NTG), a particular type of POAG characterized by IOP within normal limits, and primary angle-closure closure glaucoma (PACG), have been established to be more prevalent in Asian populations [1]. More specifically, China and India have been predicted to have the most and second most total glaucoma cases, second and third most POAG cases, and most and second most PACG cases in the world, respectively [4]. These racial differences in risk and potential mechanistic pathways necessitate an improved understanding of glaucoma pathology to more effectively treat individuals of Asian descent. To date, the most important data available on the epidemiology of primary forms of glaucoma (including POAG with high IOP, NTG, and PACG) in Asians are from a diverse set of studies [5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23]; however, findings have not been well synthesized to properly evaluate the specifics of the disease within the Asian population as well as its ethnic subgroups; therefore, in this review, we will summarize the current literature on primary forms of glaucoma with an emphasis on understanding NTG and its impact in aging Asian populations. 2. Materials and Methods PubMed and Embase searches were conducted for all pertinent articles and abstracts published between 1 January 2000 and 31 December 2021. Key words utilized in varying combinations include glaucoma, primary open-angle glaucoma, open-angle glaucoma, Asian, incidence, prevalence, intraocular pressure, ocular blood flow, ocular vasculature, optic nerve head, retina, retinal nerve fiber layer, cornea, central corneal thickness, and genetics. Only articles written in English were considered. Articles were screened for relevance and analyzed for inclusion in the paper by the authors. The authors first sorted through abstracts for relevance to the below outline. Abstracts that mentioned or described these topics were included for further full article review. Additionally, citations of screened articles were also reviewed and considered for inclusion based on relevance. 3. Results 3.1. Incidence and Prevalence Asian populations, specifically of East Asian ethnic origin, have been commonly found to have the highest prevalence of and be a major risk factor for PACG [5,6,7,8]. Additionally, the pattern of disease has been described differently in Asian populations as compared to European populations, potentially due to anterior chamber and angle anatomy [8]. In comparison, studies done on POAG have been conducted in a wide variety of Asian populations, finding prevalence rates ranging from as low as 1.0% to as high as 6.5% (Table 1) [9,10,11,12,13,14,15,16,17,18,19,20,21,22,23]. Additionally, studies have estimated that approximately 70% of all POAG cases in Asians were of the normal-tension variety [13,24]. In comparison to other races, Asian populations have been found to have a relatively lower risk of POAG per decade in age, along with subtle regional differences in risk [22,25]. This risk parameter was found to be the highest in Hispanics followed by white populations; black populations had a similar risk score as Asian populations, but also had the highest prevalence at any age group [22,25]. Additionally, within Asian populations, POAG prevalence has been estimated to increase significantly in South and Central Asia, increasing from 17.06 million in 2016 to 32.09 million individuals afflicted by 2040 [11]. This increase may potentially be attributed to an aging population base or an increased access to primary care and glaucoma screening; however, further study is necessary to develop an appropriate conclusion regarding these epidemiological changes and the factors influencing them [26]. Asian populations may not have the highest prevalence percentage or increase in prevalence as other racial demographics, but their rapidly increasing population positions them to be a heavily affected group in the future. In addition to primary forms of glaucoma, secondary forms of the disease, including exfoliative and pigmentary glaucoma, exist and need further investigation in the Asian populations. Moreover, another disease highly prevalent among Asians is myopic maculopathy, a leading cause of irreversible blindness that shares a common risk factor, represented by axial elongation, with glaucoma [27]. Myopic optic neuropathy is expected to increase in the future and mostly in Asian populations, therefore, investigating the natural course of the glaucomatous disease in high myopic eyes, specifically in the Asian race, is suggested. 3.2. Risk Factors Risk factors for glaucoma in Asian populations vary widely within the literature, but similarly to other racial demographics, they include: older age, family history of glaucoma, elevated IOP, thinner central corneal thickness (CCT), and myopia [28,29,30,31,32,33]. The Beijing Eye Study found glaucoma progression to be associated with smaller optic disc rim area in multiple regression analysis, but not with optic disc size, mean blood pressure, ocular perfusion pressure, retinal vessel diameter, retinal microvascular abnormalities, refractive error, and prevalence of dyslipidemia [33]. Additionally, systemic diabetes and hypertension have been implicated to increase the risk for glaucoma in Asians, which is similar to Western populations, but the role of these conditions in glaucomatous pathology is still undefined [33,34,35]. Obstructive sleep apnea and hypopnea syndrome have also been associated with an increased risk of glaucoma in Asian and Caucasian populations [36]. In addition to systemic conditions, serum levels of vitamin C and D, uric acid, and ferritin have been associated with glaucoma [37,38,39]. More specifically, lower serums levels of vitamin C and increased levels of uric acid have been associated with NTG [37]. Similarly, reduced vitamin D levels were associated, in a reverse J-shaped manner, with a significantly elevated risk of POAG [38]. Additionally, IOP and vertical and horizontal cup-to-disc ratios were significantly related to vitamin D level, potentially providing an explanation for this association [38]. In terms of social risk factors, education level has not been demonstrated to be associated with IOP or POAG in an Asian population [40]. Overall, risk factors for glaucoma may not vary between racial demographics, but further research is necessary to better understand their role in glaucoma pathology. 3.3. Association of Ocular Structures and Glaucoma 3.3.1. Optic Nerve Head Racial variations in the ONH and their association with glaucoma have been proposed, however, a consensus is yet to be established [41,42,43,44]. The Tanjong Pagar study considered morphometric characteristics of the ONH in a Singaporean Chinese population, and found that these parameters were incredibly similar to those described previously in other populations [41]; however, another study looking at white, Asian, African, Hispanic, and Filipino-Americans found mean optic disc sizes of 2.15 mm, 2.38 mm, 2.55 mm, 2.57 mm, and 2.48 mm, respectively, with white-Americans having significantly smaller optic discs than the other groups [42]. Mean data also suggested that Asian populations may present as a distinct intermediate between white and black patients, but this varies within subgroups [42]. Additionally, a study conducted in European Caucasian and East Asian children found that East Asian children had significantly larger mean cup-to-disc ratios [43]. Finally, significantly greater disc area, cup-to-disc area, vertical cup-to-disc area, and cup volume have been observed in Chinese populations compared with Caucasian populations [44]. Additionally, Malay patients with NTG have been found to have larger optic disc and cup areas than patients with POAG, which may further complicate its clinical utilization [45]. 3.3.2. Retina Differences in retinal measurements have also been noted between Asian and Caucasian populations and within Asian ethnic groups. East Asian children have been found to have a thicker RNFL compared with European Caucasian children [43]. Additionally, a study comparing Caucasian and Chinese subjects found a significantly greater thickness in all peripapillary RNFL parameters, except for the nasal quadrant, in Chinese subjects after adjusting for age, sex, axial length, IOP, disc area, and ganglion cell complex thickness [44]; however, Caucasian subjects were found to have significantly greater ganglion cell complex thicknesses [44]. These findings indicate potential racial differences in retinal parameters, but further comparative studies are needed to appropriately draw conclusions with other racial demographics. It is important to note that studies have found similar trends in other ethnic groups. In Indian eyes, a negative correlation between age and average RNFL thickness was identified without significant differences in RNFL thickness between males and females [46]. Additionally, Indian patients with POAG were found to have significantly thinner RNFL parameters than OHT patients whereas OHT patients were found to have significantly thinner RNFL parameters than normal patients [47]. These trends have generally been identified in other ethnicities, which may thus describe commonalities in glaucoma pathogenesis in spite of population-based variations. Looking specifically at Asian ethnic groups, the Singapore Epidemiology of Eye Diseases Study examined normal eyes of Chinese, Malay, and Indian adults. The study found the average ganglion cell-inner plexiform layer thickness to be 82.6 ± 6.1, 81.5 ± 6.8, and 78.0 ± 6.9 μm in Chinese, Malay, and Indian participants, respectively [48]. This parameter was found to be significantly thinner in Indians compared with Chinese and Malays [48]. Additionally, the study found the average RNFL thickness to be 95.7 ± 9.6, 94.9 ± 10.6, and 87.3 ± 10.6 μm in Chinese, Malay, and Indian participants, respectively [49]. This parameter was also found to be significantly thinner in Indians compared with Chinese and Malays [49]. No statistically significant difference was observed between Chinese and Malay participants [49]. This recent, large-scale study proposes anatomical differences in key glaucomatous parameters between various Asian ethnic groups. It is essential to better identify racial differences to improve upon our understanding of glaucoma pathology as well as to develop more individualized treatment options. 3.3.3. Cornea Although corneal measurements, particularly CCT, have been studied in a range of racial demographics, conclusions regarding Asians are still conflicted. It has been established that patients of African descent have thinner corneas whereas those of European and Latin American descent have relatively thicker corneas [50,51,52,53]. In comparison, Asians are known to fall between these extremes; however, their exact position as intermediates is yet to be established, with some studies identifying Asians as more similar to their European descent counterparts and others with their African descent counterparts [50,51,52,53]. Additionally, differences in ethnic groups have been found within the broad classification of Asian race. South and Southeast Asians, Filipinos, and Pacific Islanders have been found to have 6 to 13 µm thicker corneas than Chinese, Japanese, and Koreans [54]. Studies have identified Chinese participants to have the thinnest corneas within the Asian cohort [50,55], but a definitive conclusion has yet to be established due to similar findings in Japanese participants [53]. Additionally, ethnic variations within rural Chinese populations have been identified. A large-scale study examining 6504 adults categorized as ethnic Bai, Yi, or Han over the age of 50 years showed that ethnicity contributed significantly to the presence of thinner corneas compared with other factors including age, gender, body mass index, blood pressure, and other anterior ocular structural parameters [56]. More specifically, those of Han ethnicity were found to have the thinnest corneas [56]. Although the role of CCT in glaucoma pathogenesis is not clear, the parameter has been found to explain, at least partially, the effect of older age on increased risk of glaucoma in those of African and Latin American descent, but not in Asians [54]. The Singapore Malay Eye Study found thinner CCT to be associated with a smaller rim area and greater cup-to-disc area in POAG patients but not normal subjects, which may suggest a potential relationship between CCT and glaucomatous pathogenesis [57]. Moreover, the role of CCT in NTG compared to high-pressure POAG has yet to be understood. A Korean study found CCT to be thinner in NTG patients compared with POAG patients and control subjects [58]; however, a Chinese study found no significant difference between the CCT of NTG patients and healthy age- and gender-matched subjects [59]. Although ethnic differences may exist, further study is necessary to better understand this relationship. Greater CCT has been proposed, though, it is associated with higher IOP, younger age, male sex, non-hypertensive status, and diabetes and hyperglycemia [60,61,62,63]. 3.3.4. Ocular Vasculature Retinal vascular geometry has been widely studied in Asian populations. Overall, decreased retinal arteriolar and venular tortuosity or straighter retinal vessels, narrower retinal venular branching angle, narrower retinal arteriolar and venular caliber, and decreased retinal vascular fractal dimension have been associated with an increased risk for glaucoma and progressive alterations in structural parameters including increased cup-to-disc ratio, reduced RNFL and ganglion cell-inner plexiform layer thickness, and thinner neuroretinal rim [64,65,66,67,68,69,70,71]. Interestingly, the Handan Eye Study found that both POAG and PACG patients had narrower retinal arteries and veins [71]. These findings in Asian populations have been studied in Caucasian population-based studies with conflicting results. The Blue Mountain Eye Study found that POAG eyes were much more likely to have retinal arteriolar narrowing than normal eyes [72,73]; however, the Beaver Dam Eye Study found no association between retinal vascular caliber and increased prevalence of glaucoma or larger cup-to-disc ratio [72,74]. In comparison, the Montrachet study found that decreased retinal vessel calibers were associated with decreased RNFL thickness in healthy elderly eyes [75]. Importantly, the Montrachet study also established the diagnostic ability of spectral-domain optical coherence tomography in discriminating between glaucoma patients from healthy controls, a technique which needs to also be applied to Asians and other populations to establish its diagnostic role in a clinician’s arsenal [76]. Another European study found that narrowing of both arterial and venous retinal vessels was associated with POAG [77]. The large-scale European Eye Epidemiology study noted peripapillary RNFL thickness to be associated with systemic vascular and neurovascular disease [78]. These findings may describe common mechanisms of the vascular hypothesis of glaucomatous progression; however, further research in diverse populations is required to appropriately describe the pathogenesis in both of these racial groups. Research on retinal vascular geometry within specific Asian ethnic groups is much more limited. One study found, through multiple linear regression modeling, that healthy Indian participants had the largest arteriolar and venular calibers and Chinese participants had the smallest vessel calibers, with Malay participants falling between these groups [79]. In addition, Chinese participants were identified to have the largest arteriolar and venular tortuosity and venular fractal dimension [79]. Both of these parameters have been previously associated with a reduced risk of glaucoma or structural progression in broader Asian studies, which may be an important consideration in understanding disease pathogenesis as it differs between ethnic groups. New developments in optical coherence tomography angiography have allowed for the collection of quantitative data on optic disc and peripapillary nerve fiber layer plexus vessel density that could serve as normative clinical references, as per a recent large-scale study conducted in Chinese adults [80]. Retinal vasculature and blood pressure differences do exist between ethnic and racial demographics and may be important in understanding the pathogenesis of glaucoma in these diverse groups. 3.3.5. Other Ocular Structures Other ocular structures have been briefly studied in Asian populations. A cross-sectional and meta-analysis study of healthy and POAG Chinese subjects found that there was no significant difference in choroidal thickness between the groups after adjusting for IOP, age, and axial length, indicating the choroid may not play a significant role in glaucoma pathogenesis, at least in this population [81]. A Korean study found CCT and anterior scleral thickness to be correlated only in NTG patients and anterior scleral thickness to be thinner in NTG patients compared with POAG patients and control subjects [58]. A study comparing iris structural measurements in American Caucasians and Chinese and mainland Chinese subjects found that Chinese subjects had thicker irises and greater iris area under dark conditions that Caucasian subjects [82]. The group also found that Chinese subjects had smaller angle recess area and trabecular-iris space area than Caucasian subjects but greater dark-to-light changes in angle opening distance and trabecular-iris space area [83]. Another study found that the lamina cribrosa thickness to be reduced in the glaucomatous group compared with the normal group [84]. This finding supports the mechanistic hypothesis postulating the role of pressures in damaging posterior ocular structures and influencing visual function. Furthermore, in healthy Asian eyes, a greater lamina cribrosa depth was found to be associated with age, the female gender, Indian race, axial length, retinal nerve fiber layer thickness, choroidal thickness, vertical cup-to-disc ratio, and disc size [85]. Although these measurements may not be directly related to the pathogenesis of glaucoma, they are still important for a better understanding of the disease and racial differences in the eye as a whole. 3.4. Intraocular Pressure Due to the relatively higher incidence of NTG in Asian populations, it is thought that these patients may have a lower IOP than other racial groups more commonly afflicted with high-pressure POAG. This, at least on the surface, goes against the popular mechanistic hypothesis regarding glaucomatous progression and thus must be evaluated carefully; however, a definite conclusion regarding racial variations in IOP is difficult to establish [51,52]. A study comparing aqueous humor dynamics between Chinese and Caucasian adults found that Caucasians had lower IOP, a slower aqueous flow rate, and a faster uveoscleral outflow rate [86]. Ethnic differences may also exist in IOP, with East Asians having been suggested to have the lowest IOP [55]. The Singapore Epidemiology of Eye Diseases Study found the mean IOP to be 14.3 ± 3.1, 15.3 ± 3.7, and 15.8 ± 2.9 mmHg in Chinese, Malay, and Indian participants, respectively, with multivariate regression analysis suggesting Chinese participants to have significantly lower IOP [55]. Additionally, the prevalence of study participants with elevated IOP, defined as 21 mm Hg or greater, was found to be 2.6%, 6.2%, and 4% in Chinese, Malay, and Indians [55]. It has been suggested that this variation may be highly heritable by a large-scale Korean study. Additive genetics was found to estimate 36% of the total variance in the IOP phenotype, whereas a unique environment explained the remaining 64% [87]. Additionally, a child’s risk of having high IOP was almost 10 times greater if they have parents with high IOP [87]. Another study conducted in Korean and Mongolian populations found higher heritability estimates of approximately 50% [88]. Although IOP may display significant individual variation, it is important to consider racial and ethnic variations when utilizing target IOPs as treatment goals due to potential genetic predispositions. Similar to other racial demographics, elevated IOP is known to increase the risk of glaucomatous progression and has been associated with a number of ocular and systemic risk factors including the female sex, thicker central corneal thickness, high myopia, high body mass index, high blood pressure, diabetes, and hyperlipidemia [88,89,90,91,92,93,94,95]; however, rather surprisingly, older age has not been associated with increased IOP in Asian populations. Instead, IOP has repeatedly been found to significantly decrease with age in this racial demographic [88,89,92,96,97,98]. A positive association between age and IOP has been previously established in white and black populations by the Beaver Dam Eye Study and Barbados Eye Study [99,100]. This racial discrepancy in IOP is incredibly significant for clinicians to consider as IOP reduction is currently the only therapy for preventing or limiting disease progression. Asian glaucoma patients may therefore display artificial reductions in IOP as they age, which is problematic as epidemiologic studies have indicated an increased risk of glaucoma with age in this population, complicating disease management as current medical therapies are focused on IOP reduction [25]. New treatment modalities may thus be vital to adequately manage glaucomatous progression in Asian populations. Theories for this discrepancy have been proposed, but further research is absolutely necessary to better understand these varied patterns to effectively treat patients of different racial backgrounds and to better understand this family of pathomechanisms classified as POAG. 3.5. Blood and Perfusion Pressures Although hypertension and elevated blood pressure parameters have frequently been associated with glaucoma in those of European and African descent, their role in Asian populations is not as widely established, potentially due to the relatively lower incidence of these cardiovascular conditions in these populations [101]; however, the study of these vascular factors is incredibly important as it may provide a potential explanation, through the vascular hypothesis, for the differences in glaucomatous progression noted between Asian populations and other racial groups. The Beijing Eye Study found no significant association between arterial hypertension, blood pressure parameters, ocular perfusion pressure (OPP), and POAG progression or prevalence [33,102]; however, other studies have identified an elevated prevalence of glaucoma, particularly NTG, in hypertensive patients [103,104]. In fact, a study conducted in Chinese subjects with systemic hypertension found that blood pressure was negatively correlated with a range of RNFL thickness parameters and positively correlated with mean IOP [104]. Interestingly, the Singapore Malay Eye Study found low diastolic blood pressure, mean OPP, and diastolic OPP to be independent risk factors for POAG [105]. Additionally, the Singapore Epidemiology of Eye Diseases Study found that both low and high levels of systolic OPP, but not mid-range levels, were associated with an increased risk for POAG [106]. Studies comparing the role of hypertensive status and OPP between racial demographics were not identified. A Korean study of healthy participants found little difference between pulsatile ocular blood flow between Koreans and Caucasians [107]. Studies do indicate that ocular and systemic blood pressures may be important in the pathogenesis of POAG in Asian populations as well. 3.6. Genetics The heritability of glaucoma has been widely studied in Asian populations, and many genes have been implicated to play a role in its pathogenesis. Several of the structural parameters involved in glaucomatous progression have been found to be highly heritable, with one Chinese twin study identifying approximately 80% of phenotypic variations in the optic disc to be determined genetically [108]. The study found that the correlation coefficients of heritability for disc area, cup disc, and cup-to-disc ratio were 0.79, 0.83, and 0.80 in monozygotic twin pairs and 0.30, 0.37, and 0.35 in dizygotic twin pairs, respectively [108]. Additionally, the literature suggests that less than one-tenth of POAG cases in the general population may be caused by specific genetic mutations, and most are instead explained by polygenic alterations [109]. Racial differences in genetic polymorphisms have been identified, however, the significance of these in explaining population-based differences in disease presentation has yet to be fully understood [110,111]. In fact, one multi-ethnic genome-wide meta-analysis including 34,179 cases and 349,321 controls identified 127 significant POAG loci across Europeans, Africans, and Asians [110]. It also found moderately high cross-ancestry concordance of loci involved in POAG, but also many racial-specific loci [110]. Important genes noted across ethnicities include SVEP1, RERE, VCAM1, ZNF638, CLIC5, SLC2A12, YAP1, MXRA5, and SMAD6 [110]. Another study found that European, American, and South Asian populations may share similar genetic heatmap patterns for single nucleotide polymorphisms of risk alleles for POAG, whereas African, East Asian, and Korean populations each have a distinct pattern [111]. In addition, similar findings regarding the heritability and polygenic nature of PACG have been identified, however, very different gene loci have been implicated in this disease compared with POAG [112,113]. The most common POAG genes studied in Asian populations include populations include MYOC, OPTN, CYP1B1, CAV1-CAV2, TGFBR3, ATOH7, CDKN2B/CDKN2B-AS1, SIX6, MMP, LOXL-1, TP53, TNF, APOE, TLR4, NFT4, WDR36, IL-1, and VAV2-VAV3, but numerous other loci have been noted (Table 2). Overall, genetic differences between populations are expected, and they do exist, but their role in elucidating racial differences has yet to be uncovered. 4. Discussion The risk of glaucoma within aging Asian populations has never been higher or in more need of targeted research. Although African and European descent populations have been compared and contrasted to some degree, Asian populations and their specific differential glaucoma risk and disease profiles are not as well understood outside of a large number of NTG patients expected within these populations. Asians are a rapidly growing population and will account for a majority of glaucoma cases in the near future. Both PACG and NTG are known to be more prevalent in this group, but it is not yet definitively known why this may be the case. Structural parameters including the optic disc, retina, and cornea have shown both racial and ethnic differences across populations. It is not yet known whether these variations may predispose populations to glaucoma or if they are secondary to other population-based variations. Regardless, these structural differences are clinically noteworthy and may impact diagnostic standards in different racial and ethnic demographics. Patterns in IOP and relevant parameters have also shown important differences. Aqueous humor dynamics have been indicated to differ between Asian and Caucasian populations. More significantly, in Asian populations, older age has been associated with decreased IOP, which conflicts with previous findings of a positive association in African and European descent populations. With medical therapies for glaucoma limited to IOP reduction, lowering IOP may not be as effective in aging Asian populations given their tendency to present with NTG and a decreasing IOP with age. Novel modalities of treatment may thus be necessary to effectively manage glaucoma in this population. Systemic blood pressure and OPP have both been specifically identified to be risk factors for glaucoma in Asians, as well as other, populations. Additionally, retinal vascular geometry has been studied in detail, with narrower retinal vessel calibers, decreased tortuosity, and decreased fractal dimension having been associated with glaucoma and parameters of progression. It is still unclear whether these findings differ from other racial demographics, as ocular blood flow parameters have been studied in greater detail in white and black populations. Further studies focused on ocular blood flow in Asian populations are necessary, as well as studies on vascular geometry in other populations. Finally, distinct genetic patterns have been identified between races; however, the extent of these similarities and differences has yet to be elucidated along with associations of genetic polymorphisms and definitive pathomechanisms. Further studies are needed to specifically investigate glaucoma genes, which are commonly shared both in Asian and other races, and genes unique to Asian populations. It is important to note that other factors may interact with specific genetic patterns and influence the disease pathophysiology, specifically in Asian ethnicities, including environmental conditions and dietary habits. Asians make up a very large portion of the global population and will soon make up the majority of glaucoma-afflicted population. With a significant impact expected both within the communities and health care systems, studies are needed to investigate mechanisms driving the different types of glaucoma within the Asian population. Focus should be placed not only on POAG forms with both high and low IOP, but also within PACG and secondary forms (such as exfoliative glaucoma, pigmentary glaucoma, and myopic optic neuropathy). Looking forward, increased awareness for clinicians and patients are important to help mitigate the outsized POAG and NTG burden within aging Asian populations, and further targeted research is needed to understand specific risk factors and pathophysiologic mechanisms driving different glaucoma types (primary and secondary) in the Asian race. A greater understanding of ocular structure, including cup-to-disc biomarkers and their differences to Western populations along with a better understanding of IOP and vascular dynamics is needed, with a particular emphasis on accounting for both genetical and environmental factors. Author Contributions Conceptualization, B.S.; methodology, A.B.; data curation, A.B.; writing—original draft preparation, A.B., A.V.V., F.O. and B.S; writing—review and editing, A.V.V., A.H., F.O., A.F.-K. and B.S.; supervision, A.H. and B.S. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement No new data were created or analyzed in this study. Data sharing is not applicable to this article. Conflicts of Interest Alon Harris would like to disclose that he received remuneration from AdOM, Qlaris, Luseed, and Cipla for serving as a consultant, and he serves on the board of AdOM, Qlaris, and Phileas Pharma. Alon Harris holds an ownership interest in AdOM, Luseed, Oxymap, Qlaris, Phileas Pharma, SlitLed, and QuLent. All relationships listed above are pursuant to Icahn School of Medicine’s policy on outside activities. The contribution of the author Francesco Oddone was supported by Fondazione Roma and by the Italian Ministry of Health. None of the other authors listed have any financial disclosures. The authors declare no conflict of interest. jcm-11-02486-t001_Table 1 Table 1 Prevalence and Incidence of Glaucoma in Asian Populations. POAG: Primary open-angle glaucoma; NTG: normal-tension glaucoma. Author Population Prevalence/Incidence of Glaucoma Prevalence/Incidence of POAG Prevalence/Incidence of NTG Stein et al., 2011 [9] Asian Americans 6.52% 0.73% Baskaran et al., 2015 [10] Chinese 4.0% 1.7% Chan et al., 2016 [11] Asians 3.54% 2.34% Guo et al., 2021 [12] Laos 1.54% 0.62% Zhao et al., 2018 [13] Chinese 2.0% 1.0% He et al., 2015 [14] Chinese 2.85% Sales et al., 2014 [15] Filipino 11.9% 6.8% Rauf et al., 2013 [16] Indian 1.0% Narayanaswamy et al., 2013 [17] Indian 1.95% 1.25% Rosman et al., 2012 [18] Singapore Malay 4.6% 3.2% Liang et al., 2011 [19] Chinese 1.0% Shen et al., 2008 [20] Singapore Malay 3.4% 2.5% He et al., 2006 [21] Chinese 2.1% Rudnicka et al., 2006 [22] Asian 1.4% Koh et al., 2021 [23] Indian 1.68% 1.37% jcm-11-02486-t002_Table 2 Table 2 Genetic Findings in Adult Asian Populations. NTG: normal-tension glaucoma; PACG: primary angle-closure glaucoma; PCG: primary congenital glaucoma; POAG: primary open-angle glaucoma; HTG: high-tension glaucoma. Race/Ethnicity Genetic Change Associated Type of Glaucoma MYOC Fan et al., 2020 [114] Chinese c.622G > T, p.D208Y POAG Lei et al., 2019 [115] Chinese c.1309T > C, p.Y437H POAG Yang et al., 2015 [116] Chinese c.761C < G, p.P254R POAG Guo et al., 2015 [117] Asian rs12035719, rs2075648 No association with POAG Jin et al., 2015 [118] Han Chinese rs183532 PACG Cai et al., 2012 [119] Chinese Uygur c.1151A > G, p.D384G POAG Chen et al., 2011 [120] Chinese c.1099G > A, p.G367R POAG Qu et al., 2010 [121] Chinese c.1084G>- POAG Jia et al., 2009 [122] Northern Chinese p.Val53Ala POAG Xie et al., 2008 [123] Chinese c.38C > T, p.Pro13Leu; c.1009C > del, p.Gln337Stop POAG Megkegale et al., 2008 [124] Japanese c.297G > C, p.Gln297His; c.363G > A, p.Ala363Thr POAG Kumar et al., 2007 [125] Indian p.Gln48His POAG Funayama et al., 2006 [126] Japanese c.227G > A, p.Arg76Lys No association with POAG Chakrabarti et al., 2005 [127] Indian c.144G > T, p.Gln48His POAG, PCG Ishikawa et al., 2004 [128] Japanese c.1105T > C, p.Phe369Leu; c.1079T > A, p.Ile360Asn; c.1087G > A, p.Ala363Thr; c.1342A > C, p.Thr448Pro POAG Kanagavalli et al., 2003 [129] Indian c.1099G > A, p.Gly367Arg; c.1130C > T, p.Thr377Met POAG Mukhopadhyay et al., 2002 [130] Indian c.144G > T, p.Gln48His; c.1109C > T, p.Pro370Leu POAG OPTN Cheng et al., 2010 [131] Asian T34T POAG Xiao et al., 2009 [132] Chinese c.1274A > G, p.Lys322Glu POAG Kumar et al., 2007 [125] Indian c.915C > G, p.Thr202Arg POAG Ayala-Luge et al., 2007 [133] Asian M98K NTG Sripriya et al., 2006 [134] Indian M98K; IVS7 + 24G > A POAG Funayama et al., 2004 [135] Japanese c.412G > A, p.Thr34Thr; c.603T→A, p.Met98Lys POAG; NTG Leung et al., 2003 [136] Chinese E103D; H486R; V148V; IVS13 + 21C > G POAG CYP1B1 Gong et al., 2015 [137] Chinese p.P93S; p.R259C; p.A295T; p.L475P POAG Dong et al., 2012 [138] Asian rs180040, rs1056836, rs10012, rs1056827, rs1056837, rs2567206 No association with POAG Chen et al., 2011 [120] Chinese g.17120037A > G; g.17120090C > G; g.17120026T > C POAG Bhattacharjee et al., 2008 [139] Indian c.1666G, Leu432Val POAG Kumar et al., 2007 [125] Indian c.578C > T, p.Pro193Leu; c.685G > A, p.Glu229Lys; c.875T > A, p.Met292Lys; c.1103G > A, p.Arg368His POAG CAV1-CAV2 Huang et al., 2019 [140] Chinese rs548030386 Intraocular pressure Kim et al., 2015 [141] Korean minor allele G of rs17588172 HTG Huang et al., 2014 [142] Asian rs4236601[A] POAG Kato et al., 2013 [143] Japanese minor allele G of rs1052990 NTG TGFBR3 Chai et al., 2020 [144] Indian, Malay, Chinese rs1192415 Optic disc parameters Li et al., 2015 [145] Asian rs1192415 POAG Khor et al., 2011 [146] Indian, Malay rs1192415 Optic disc parameters ATOH7 Chai et al., 2020 [144] Indian, Malay, Chinese rs1900004 Optic disc parameters Mabuchi et al., 2015 [147] Japanese rs1900004 POAG Chen et al., 2012 [148] Chinese rs61854782, rs3858145 NTG, HTG Khor et al., 2011 [146] Indian, Malay rs7916697 Optic disc parameters CDKN2B/CDKN2B-AS1 Chai et al., 2020 [144] Indian, Malay, Chinese rs1360589 Optic disc parameters Hu and He, 2017 [149] Asian rs1063192 POAG Mabuchi et al., 2015 [147] Japanese rs1063192 POAG Nakano et al., 2012 [150] Japanese rs523096:A; rs518394:C; rs564398:A; rs7865618:A POAG, NTG Osman et al., 2012 [151] Japanese rs1063192 POAG Takamoto et al., 2012 [152] Japanese rs523096 NTG SIX6 Chai et al., 2020 [144] Indian, Malay, Chinese rs33912345 Optic disc parameters Lu et al., 2019 [153] Asian rs10483727, rs33912345, rs12436579 POAG Rong et al., 2019 [154] Chinese, Japanese rs10483727, rs33912345, rs12436579 POAG Mabuchi et al., 2015 [147] Japanese rs10483727 POAG Osman et al., 2012 [151] Japanese rs10483727 POAG MMP Zhao et al., 2020 [155] Chinese rs2250889; rs3918242; rs17576 PACG; POAG; both He et al., 2017 [156] Asian rs1799750 POAG, PACG LOXL-1 Wu et al., 2015 [157] Asian rs1048661, C allele of rs2165241 POAG Sun et al., 2014 [158] Asian rs2165241, rs1048661, rs3825942 No association with POAG TP53 Zhang and Wang, 2019 [159] Chinese rs4938723, rs1042522 POAG Gupta et al., 2018 [160] Indian P72R PACG; No association with POAG Guo et al., 2012 [161] Asian Arg72Pro, intron 3 16-bp insertion POAG Fan et al., 2010 [162] Chinese R72P; rs1042522 NTG TNF Passan et al., 2019 [163] North Indian c.-308G > A, c.-863C > A POAG Wang et al., 2012 [164] Chinese (-863)A allele POAG Fan et al., 2010 [162] Chinese -308G > A; rs1800629 HTG APOE Guo et al., 2015 [117] Asian rs405509, rs769446, rs449647 No association with POAG Wang et al., 2014 [165] Asian ε4ε4 genotype POAG Jia et al., 2009 [122] Northern Chinese −491A > T, −427T > C, −219T > G, c.526C > T for ε2, c.388T > C for ε4 No association with POAG Lam et al., 2006 [166] Chinese −219T > G; −427T > C NTG and HTG; NTG TLR4 Takano et al., 2012 [167] Japanese rs10759930, rs1927914, rs1927911, rs12377632, rs2149356, rs7037117 POAG, NTG Chen et al., 2012 [168] Southern Chinese rs7037117 POAG Suh et al., 2011 [169] Korean rs10759930, rs1927914, rs1927911, rs12377632, rs2149356, rs11536889, rs7037117, rs7045953 No association with NTG NTF4 Chen et al., 2012 [170] Chinese c.453G > A, p.Pro151Pro; c.470G > C, p.Gly157Ala; c.545C > T, p.Ala182Val POAG Rao et al., 2010 [171] Indian c.263C > T, p.A88V; c.453G > A, p.P151P; c.790T > G, 3′UTR; c.811G > A, 3′UTR No association with POAG, PACG Vithana et al., 2010 [172] Chinese c.338T > C, p.Leu113Ser POAG WDR36 Lee et al., 2010 [173] Mongolian No specific polymorphism studied Heritability of intraocular pressure Jia et al., 2009 [122] Northern Chinese IVS5 + 30C > T No association with POAG Fan et al., 2009 [174] Chinese p.I713V HTG Miyazawa et al., 2007 [175] Japanese p.S664L; p.I264V; c.1965-30A > G HTG IL-1 How et al., 2007 [176] Chinese IL1α (c.-889C > T); IL1β (c.3953C < T); IL1β (c.-511C < T) No association with POAG, PACG Wang et al., 2007 [177] Chinese IL1β c.-511; c.+3953 No association with NTG Wang et al., 2007 [178] Chinese IL1α c.-889C > T No association with NTG Wang et al., 2006 [179] Chinese IL1α c.-889C > T POAG VAV2-VAV3 Shi et al., 2013 [180] Japanese rs2156323, rs2801219 No association with POAG, NTG Rao et al., 2010 [171] Indian rs2156323, rs2801219 No association with POAG, PACG Chromosome 2p16.3 Meng et al., 2015 [181] Chinese rs1533428, rs12994401, rs10202118 POAG Chen et al., 2012 [168] Southern Chinese rs1533428 POAG ABCA1 Huang et al., 2019 [140] Chinese rs2472494 Intraocular pressure Chen et al., 2014 [168] Southern Chinese rs2487032 POAG PMM2 Chen et al., 2014 [182] Southern Chinese rs3785176 POAG GLIS3 Li et al., 2020 [183] Han Chinese rs736893 POAG Huang et al., 2019 [140] Chinese rs7047871 Intraocular pressure RAMP2 Gong et al., 2019 [184] Han Chinese p.Glu39Asp; p.Glu54Lys; p.Phe103Ser; p.Asn113Lysfs*10; p.Glu143Lys; p.Ser171Arg POAG ABCC5 Tang et al., 2017 [185] Chinese rs939336, rs1132776, rs983667 PACG Nongpiur et al., 2014 [186] Asian rs1401999 PACG HTR3D Tang et al., 2017 [185] Chinese rs12493550 PACG hOGG1 Zeng et al., 2017 [187] Han Chinese p.Ser326Cys PACG APE1 Zeng et al., 2017 [187] Han Chinese p.Asp148Glu PACG XRCC1 Zeng et al., 2017 [187] Han Chinese p.Arg399Gln PACG Yousaf et al., 2011 [188] Pakistani c.1316G > A (rs25487) POAG XPD Yousaf et al., 2011 [188] Pakistani c.2298A > C (rs13181) POAG MFRP Wang et al., 2018 [189] Northern Chinese rs2510143, rs36015759, rs3814762 No association with PACG Shi et al., 2013 [190] Han Chinese rs3814762 PACG ZNRF3 Wang et al., 2018 [189] Northern Chinese rs7290117, rs2179129, rs4823006, rs3178915 No association with PACG HGF Wang et al., 2018 [189] Northern Chinese rs5745718, rs12536657, rs12540393, rs17427817, rs3735520 No association with PACG CAT Gong et al., 2018 [191] Chinese rs769217 POAG GJA1 Huang et al., 2015 [192] Chinese c.791_792delAA, p.K264Ifs*43 POAG SOD2 Zhou et al., 2015 [193] Chinese rs6917589, rs5746136 POAG CD2 Liu et al., 2014 [194] Han Chinese p.Gln596Trp POAG GSTM1/GSTT1 Lu et al., 2013 [195] East Asian Null genotype POAG Huang et al., 2013 [196] Asian Null genotype POAG HSP70 Shi et al., 2013 [190] Han Chinese rs1043618 PACG OPA1 Guo et al., 2012 [197] Asian rs166850, rs10451941 No association with NTG Woo et al., 2004 [198] Korean IVS8 + 4C > T; c.32T > C No association with NTG SLC1A3 Yasumura et al., 2011 [199] Japanese rs13173144, rs1366632, rs1428967, rs930072, rs2301066 No association with NTG HLA Suzuki et al., 2010 [200] Japanese 27 HLA-DRB1 alleles, 14 HLA-DQB1 alleles No association with NTG GLC1F Murakami et al., 2010 [201] Japanese 163 allele of D7S1277i NTG OLFM2 Funayama et al., 2006 [126] Japanese p.Arg144Gln POAG EDNRA Kim et al., 2006 [202] Korean c.*1222C > T NTG Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Tham Y.-C. Li X. Wong T.Y. Quigley H.A. Aung T. Cheng C.-Y. Global Prevalence of Glaucoma and Projections of Glaucoma Burden through 2040: A systematic review and meta-analysis Ophthalmology 2014 121 2081 2090 10.1016/j.ophtha.2014.05.013 24974815 2. Kass M.A. Heuer D.K. Higginbotham E.J. Johnson C. Keltner J.L. Miller J.P. Parrish R.K. Wilson M.R. Gordon M.O. The Ocular Hypertension Treatment Study: A randomized trial determines that topical ocular hypotensive medication delays or prevents the onset of primary open-angle glaucoma Arch. Ophthalmol. 2002 120 701 713 10.1001/archopht.120.6.701 12049574 3. Leske M.C. Heijl A. Hussein M. Bengtsson B. Hyman L. Komaroff E. Factors for Glaucoma Progression and the Effect of Treatment: The early manifest glaucoma trial Arch. Ophthalmol. 2003 121 48 56 10.1001/archopht.121.1.48 12523884 4. Quigley H.A. Broman A.T. The number of people with glaucoma worldwide in 2010 and 2020 Br. J. Ophthalmol. 2006 90 262 267 10.1136/bjo.2005.081224 16488940 5. Zhang N. Wang J. Chen B. Li Y. Jiang B. Prevalence of Primary Angle Closure Glaucoma in the Last 20 Years: A Meta-Analysis and Systematic Review Front. Med. 2021 7 624179 10.3389/fmed.2020.624179 33537335 6. Jonas J.B. Aung T. Bourne R.R. Bron A.M. Ritch R. Panda-Jonas S. Glaucoma Lancet 2017 390 2183 2193 10.1016/S0140-6736(17)31469-1 28577860 7. Weinreb R.N. Leung C. Crowston J.G. Medeiros F.A. Friedman D.S. Wiggs J.L. Martin K.R. Primary open-angle glaucoma Nat. Rev. Dis. Primers 2016 2 16067 10.1038/nrdp.2016.67 27654570 8. He M. Foster P.J. Johnson G.J. Khaw P.T. Angle-closure glaucoma in East Asian and European people. Different diseases? Eye 2006 20 3 12 10.1038/sj.eye.6701797 15688051 9. Stein J.D. Kim D.S. Niziol L.M. Talwar N. Nan B. Musch D. Richards J.E. Differences in Rates of Glaucoma among Asian Americans and Other Racial Groups, and among Various Asian Ethnic Groups Ophthalmology 2011 118 1031 1037 10.1016/j.ophtha.2010.10.024 21310489 10. Baskaran M. Foo R.C. Cheng C.-Y. Narayanaswamy A.K. Zheng Y.-F. Wu R. Saw S.-M. Foster P. Wong T.-Y. Aung T. The Prevalence and Types of Glaucoma in an Urban Chinese Population: The Singapore Chinese Eye Study JAMA Ophthalmol. 2015 133 874 880 10.1001/jamaophthalmol.2015.1110 25974263 11. Chan E.W. Li X. Tham Y.-C. Liao J. Wong T.Y. Aung T. Cheng C.-Y. Glaucoma in Asia: Regional prevalence variations and future projections Br. J. Ophthalmol. 2016 100 78 85 10.1136/bjophthalmol-2014-306102 26112871 12. Guo B. Tan Y. Nygaard S. Carrillo C. Nouansavanh K.O. Souksamone K. Casson R.J. Prevalence of glaucoma in the Lao People’s Democratic Republic: The Vientiane Eye Study Br. J. Ophthalmol. 2021 Online ahead of print 10.1136/bjophthalmol-2020-317658 13. Zhao J. Solano M.M. Oldenburg C.E. Liu T. Wang Y. Wang N. Lin S.C. Prevalence of Normal-Tension Glaucoma in the Chinese Population: A Systematic Review and Meta-Analysis Am. J. Ophthalmol. 2018 199 101 110 10.1016/j.ajo.2018.10.017 30352196 14. He J. Zou H. Lee R.K. Tong X. Tang W. Zhang Y. Zhao R. Ge L. Prevalence and risk factors of primary open-angle glaucoma in a city of Eastern China: A population-based study in Pudong New District, Shanghai BMC Ophthalmol. 2015 15 134 10.1186/s12886-015-0124-x 26462564 15. Sales C. Lee R. Agadzi A.K. Hee M.R. Singh K. Lin S.C. Open-angle Glaucoma in Filipino and White Americans: A comparative study J. Glaucoma 2014 23 246 253 10.1097/IJG.0b013e318279b3e2 23221903 16. Rauf A. Malik R. Bunce C. Wormald R. The British Asian Community Eye Study: Outline of results on the prevalence of eye disease in British Asians with origins from the Indian subcontinent Indian J. Ophthalmol. 2013 61 53 58 10.4103/0301-4738.107191 23412521 17. Narayanaswamy A. Baskaran M. Zheng Y. Lavanya R. Wu R. Wong W.-L. Saw S.-M. Cheng C.-Y. Wong T.-Y. Aung T. The Prevalence and Types of Glaucoma in an Urban Indian Population: The Singapore Indian Eye Study Investig. Ophthalmol. Vis. Sci. 2013 54 4621 4627 10.1167/iovs.13-11950 23745009 18. Rosman M. Zheng Y. Lamoureux E. Saw S.M. Aung T. Tay W.T. Wang J.J. Mitchell P. Tai E.S. Wong T.Y. Review of key findings from the Singapore Malay Eye Study (SiMES-1) Singap. Med. J. 2012 53 82 87 19. Liang Y.B. Friedman D.S. Zhou Q. Yang X. Sun L.P. Guo L.X. Tao Q.S. Chang D.S. Wang N.L. Prevalence of Primary Open Angle Glaucoma in a Rural Adult Chinese Population: The Handan Eye Study Investig. Ophthalmol. Vis. Sci. 2011 52 8250 8257 10.1167/iovs.11-7472 21896871 20. Shen S.Y. Wong T.Y. Foster P. Loo J.-L. Rosman M. Loon S.-C. Wong W.L. Saw S.-M. Aung T. The Prevalence and Types of Glaucoma in Malay People: The Singapore Malay Eye Study Investig. Ophthalmol. Vis. Sci. 2008 49 3846 3851 10.1167/iovs.08-1759 18441307 21. He M. Foster P. Ge J. Huang W. Zheng Y. Friedman D. Lee P.S. Khaw P.T. Prevalence and Clinical Characteristics of Glaucoma in Adult Chinese: A Population-Based Study in Liwan District, Guangzhou Investig. Ophthalmol. Vis. Sci. 2006 47 2782 2788 10.1167/iovs.06-0051 16799014 22. Rudnicka A.R. Mt-Isa S. Owen C. Cook D. Ashby D. Variations in Primary Open-Angle Glaucoma Prevalence by Age, Gender, and Race: A Bayesian Meta-Analysis Investig. Ophthalmol. Vis. Sci. 2006 47 4254 4261 10.1167/iovs.06-0299 17003413 23. Koh V. Tham Y.-C. Tan N.Y. Yu M. Thakur S. Teo Z.L. Mani B. Wong T.-Y. Aung T. Cheng C.-Y. Six-Year Incidence and Risk Factors of Primary Glaucoma in the Singapore Indian Eye Study Ophthalmol. Glaucoma 2021 4 201 208 10.1016/j.ogla.2020.09.004 32947057 24. Pekmezci M. Vo B. Lim A.K. Hirabayashi D.R. Tanaka G.H. Weinreb R.N. Lin S.C. The Characteristics of Glaucoma in Japanese Americans Arch. Ophthalmol. 2009 127 167 10.1001/archophthalmol.2008.593 19204234 25. Kapetanakis V.V. Chan M.P.Y. Foster P. Cook D. Owen C. Rudnicka A. Global variations and time trends in the prevalence of primary open angle glaucoma (POAG): A systematic review and meta-analysis Br. J. Ophthalmol. 2016 100 86 93 10.1136/bjophthalmol-2015-307223 26286821 26. Stein J.D. Khawaja A.P. Weizer J.S. Glaucoma in Adults—Screening, Diagnosis, and Management: A Review JAMA 2021 325 164 174 10.1001/jama.2020.21899 33433580 27. Song Y. Wang W. Lin F. Chen S. Jin L. Li F. Gao K. Cheng W. Xiong J. Zhou R. Natural history of glaucomatous optic neuropathy in highly myopic Chinese: Study protocol for a registry cohort study BMJ Open 2020 10 e039183 10.1136/bmjopen-2020-039183 28. Chiam N. Baskaran M. Li Z. Perera S. Goh D. Husain R. Khor C.C. Cheng C. Aung T. Vithana E.N. Social, health and ocular factors associated with primary open-angle glaucoma amongst Chinese Singaporeans Clin. Exp. Ophthalmol. 2018 46 25 34 10.1111/ceo.13008 28621876 29. Doss E.L. Doss L. Han Y. Huang S. Porco T. Pekmezci M. Lin S. Risk Factors for Glaucoma Suspicion in Healthy Young Asian and Caucasian Americans J. Ophthalmol. 2014 2014 726760 10.1155/2014/726760 25143829 30. Perera S.A. Wong T.Y. Tay W.-T. Foster P.J. Saw S.-M. Aung T. Refractive Error, Axial Dimensions, and Primary Open-Angle Glaucoma: The Singapore Malay Eye Study Arch. Ophthalmol. 2010 128 900 905 10.1001/archophthalmol.2010.125 20625053 31. Suzuki Y. Iwase A. Araie M. Yamamoto T. Abe H. Shirato S. Kuwayama Y. Mishima H.K. Shimizu H. Tomita G. Risk Factors for Open-Angle Glaucoma in a Japanese Population: The Tajimi Study Ophthalmology 2006 113 1613 1617 10.1016/j.ophtha.2006.03.059 16828504 32. Tham Y.-C. Aung T. Fan Q. Saw S.-M. Siantar R.G. Wong T.Y. Cheng C.-Y. Joint Effects of Intraocular Pressure and Myopia on Risk of Primary Open-Angle Glaucoma: The Singapore Epidemiology of Eye Diseases Study Sci. Rep. 2016 6 19320 10.1038/srep19320 26758554 33. Wang Y.X. Hu L.N. Yang H. Jonas J.B. Xu L. Frequency and associated factors of structural progression of open-angle glaucoma in the Beijing Eye Study: Table 1 Br. J. Ophthalmol. 2012 96 811 815 10.1136/bjophthalmol-2011-301224 22408234 34. Luo X.-Y. Tan N.Y.Q. Chee M.-L. Shi Y. Tham Y.-C. Wong T.Y. Wang J.J. Cheng C.-Y. Direct and Indirect Associations Between Diabetes and Intraocular Pressure: The Singapore Epidemiology of Eye Diseases Study Investig. Ophthalmol. Vis. Sci. 2018 59 2205 2211 10.1167/iovs.17-23013 29801155 35. Bae H.W. Lee N. Lee H.S. Hong S. Seong G.J. Kim C.Y. Systemic Hypertension as a Risk Factor for Open-Angle Glaucoma: A Meta-Analysis of Population-Based Studies PLoS ONE 2014 9 e108226 10.1371/journal.pone.0108226 25254373 36. Wu X. Liu H. Obstructive sleep apnea/hypopnea syndrome increases glaucoma risk: Evidence from a meta-analysis Int. J. Clin. Exp. Med. 2015 8 297 303 25785000 37. Yuki K. Murat D. Kimura I. Ohtake Y. Tsubota K. Reduced-serum vitamin C and increased uric acid levels in normal-tension glaucoma Graefe’s Arch. Clin. Exp. Ophthalmol. 2010 248 243 248 10.1007/s00417-009-1183-6 19763599 38. Yoo T.K. Oh E. Hong S. Is vitamin D status associated with open-angle glaucoma? A cross-sectional study from South Korea Public Health Nutr. 2014 17 833 843 10.1017/S1368980013003492 24476947 39. Lin S.-C. Wang S.Y. Yoo C. Singh K. Lin S.C. Association Between Serum Ferritin and Glaucoma in the South Korean Population JAMA Ophthalmol. 2014 132 1414 1420 10.1001/jamaophthalmol.2014.2876 25171442 40. Xu L. Wang Y.X. Jonas J.B. Level of education associated with ophthalmic diseases. The Beijing Eye Study Graefe’s Arch. Clin. Exp. Ophthalmol. 2010 248 49 57 10.1007/s00417-009-1204-5 19821117 41. Bourne R.R.A. Foster P. Bunce C. Peto T. Hitchings R.A. Khaw P.T. Seah S.K.L. Garway-Heath D. The morphology of the optic nerve head in the Singaporean Chinese population (the Tanjong Pagar study): Part 1—Optic nerve head morphology Br. J. Ophthalmol. 2008 92 303 309 10.1136/bjo.2007.123273 18303151 42. Seider M.I. Lee R. Wang D. Pekmezci M. Porco T.C. Lin S.C. Optic Disk Size Variability Between African, Asian, White, Hispanic, and Filipino Americans Using Heidelberg Retinal Tomography J. Glaucoma 2009 18 595 600 10.1097/IJG.0b013e3181996f05 19826388 43. Samarawickrama C. Wang J.J. Huynh S.C. Pai A. Burlutsky G. Rose K. Mitchell P. Ethnic differences in optic nerve head and retinal nerve fibre layer thickness parameters in children Br. J. Ophthalmol. 2010 94 871 876 10.1136/bjo.2009.158279 19822916 44. Chansangpetch S. Huang G. Coh P. Oldenburg C. Amoozgar B. He M. Lin S.C. Differences in Optic Nerve Head, Retinal Nerve Fiber Layer, and Ganglion Cell Complex Parameters Between Caucasian and Chinese Subjects J. Glaucoma 2018 27 350 356 10.1097/IJG.0000000000000889 29394205 45. Shatriah I. Adlina A.-R. Alisa-Victoria K. Liza-Sharmini A.T. Ahmad M.S. Optic disc topography in Malay patients with normal-tension glaucoma and primary open-angle glaucoma Clin. Ophthalmol. 2014 8 2533 2539 10.2147/OPTH.S71136 25540578 46. Malik A. Singh M. Arya S. Sood S. Ichhpujani P. Retinal nerve fiber layer thickness in Indian eyes with optical coherence tomography Nepal. J. Ophthalmol. 2012 4 59 63 10.3126/nepjoph.v4i1.5852 22343998 47. Gyatsho J. Kaushik S. Gupta A. Pandav S.S. Ram J. Retinal Nerve Fiber Layer Thickness in Normal, Ocular Hypertensive, and Glaucomatous Indian Eyes: An optical coherence tomography study J. Glaucoma 2008 17 122 127 10.1097/IJG.0b013e31814b9817 18344758 48. Tham Y.-C. Chee M.L. Dai W. Lim Z.W. Majithia S. Siantar R. Thakur S. Rim T. Cheung C.Y. Sabanayagam C. Profiles of Ganglion Cell-Inner Plexiform Layer Thickness in a Multi-Ethnic Asian Population: The Singapore Epidemiology of Eye Diseases Study Ophthalmology 2020 127 1064 1076 10.1016/j.ophtha.2020.01.055 32197910 49. Ho H. Tham Y.-C. Chee M.L. Shi Y. Tan N.Y. Wong K.-H. Majithia S. Cheung C.Y. Aung T. Wong T.Y. Retinal Nerve Fiber Layer Thickness in a Multiethnic Normal Asian Population: The Singapore Epidemiology of Eye Diseases Study Ophthalmology 2019 126 702 711 10.1016/j.ophtha.2018.11.031 30529130 50. Badr M. Solano M.M. Amoozgar B. Nguyen A. Porco T. Lin S. Central Corneal Thickness Variances Among Different Asian Ethnicities in Glaucoma and Nonglaucoma Patients J. Glaucoma 2019 28 223 230 10.1097/IJG.0000000000001180 30624387 51. Shimmyo M. Ross A.J. Moy A. Mostafavi R. Intraocular pressure, Goldmann applanation tension, corneal thickness, and corneal curvature in Caucasians, Asians, Hispanics, and African Americans Am. J. Ophthalmol. 2003 136 603 613 10.1016/S0002-9394(03)00424-0 14516799 52. Knight O.J. Girkin C.A. Budenz D.L. Durbin M.K. Feuer W.J. The Cirrus OCT Normative Database Study Group Effect of Race, Age, and Axial Length on Optic Nerve Head Parameters and Retinal Nerve Fiber Layer Thickness Measured by Cirrus HD-OCT Arch. Ophthalmol. 2012 130 312 318 10.1001/archopthalmol.2011.1576 22411660 53. Aghaian E. Choe J.E. Lin S. Stamper R.L. Central corneal thickness of Caucasians, Chinese, Hispanics, Filipinos, African Americans, and Japanese in a glaucoma clinic Ophthalmology 2004 111 2211 2219 10.1016/j.ophtha.2004.06.013 15582076 54. Wang S.Y. Melles R. Lin S.C. The Impact of Central Corneal Thickness on the Risk for Glaucoma in a Large Multiethnic Population J. Glaucoma 2014 23 606 612 10.1097/IJG.0000000000000088 25055208 55. Chua J. Tham Y.C. Liao J. Zheng Y. Aung T. Wong T.Y. Cheng C.-Y. Ethnic Differences of Intraocular Pressure and Central Corneal Thickness: The Singapore Epidemiology of Eye Diseases study Ophthalmology 2014 121 2013 2022 10.1016/j.ophtha.2014.04.041 24950592 56. Pan C.-W. Li J. Zhong H. Shen W. Niu Z. Yuan Y. Chen Q. Ethnic Variations in Central Corneal Thickness in a Rural Population in China: The Yunnan Minority Eye Studies PLoS ONE 2015 10 e0135913 10.1371/journal.pone.0135913 26274328 57. Wu R.-Y. Zheng Y.-F. Wong T.-Y. Cheung C. Loon S.-C. Chauhan B.C. Aung T. Relationship of Central Corneal Thickness with Optic Disc Parameters: The Singapore Malay Eye Study Investig. Ophthalmol. Vis. Sci. 2011 52 1320 1324 10.1167/iovs.10-6038 21087973 58. Yoo C. Eom Y.S. Suh Y.-W. Kim Y.Y. Central Corneal Thickness and Anterior Scleral Thickness in Korean Patients with Open-angle Glaucoma: An anterior segment optical coherence tomography study J. Glaucoma 2011 20 95 99 10.1097/IJG.0b013e3181dde051 20577104 59. Chen H.-C. Ho J.-D. Chang S.H.-L. Chen Y.-H.S. Wu S.-C. Central corneal thickness of normal-tension glaucoma and non-glaucoma populations in ethnic Chinese Chang. Gung Med. J. 2004 27 50 55 15074890 60. Su D.H. Wong T.Y. Foster P. Tay W.-T. Saw S.-M. Aung T. Central Corneal Thickness and its Associations with Ocular and Systemic Factors: The Singapore Malay Eye Study Am. J. Ophthalmol. 2009 147 709 716.e1 10.1016/j.ajo.2008.10.013 19152872 61. Hwang Y.H. Kim T.-W. Sohn Y.H. Central Corneal Thickness in a Korean Population: The Namil Study Investig. Ophthalmol. Vis. Sci. 2012 53 6851 6855 10.1167/iovs.12-10173 22952116 62. Day A.C. Machin D. Aung T. Gazzard G. Husain R. Chew P.T.K. Khaw P.T. Seah S.K.L. Foster P.J. Central Corneal Thickness and Glaucoma in East Asian People Investig. Ophthalmol. Vis. Sci. 2011 52 8407 8412 10.1167/iovs.11-7927 21917940 63. Luo X.-Y. Dai W. Chee M.-L. Tao Y. Chua J. Tan N.Y.Q. Tham Y.-C. Aung T. Wong T.Y. Cheng C.-Y. Association of Diabetes with Central Corneal Thickness Among a Multiethnic Asian Population JAMA Netw. Open 2019 2 e186647 10.1001/jamanetworkopen.2018.6647 30646192 64. Wu R. Cheung C.Y.-L. Saw S.M. Mitchell P. Aung T. Wong T.Y. Retinal Vascular Geometry and Glaucoma: The Singapore Malay Eye Study Ophthalmology 2013 120 77 83 10.1016/j.ophtha.2012.07.063 23009894 65. Amerasinghe N. Aung T. Cheung N. Fong C.W. Wang J.J. Mitchell P. Saw S.-M. Wong T.Y. Evidence of Retinal Vascular Narrowing in Glaucomatous Eyes in an Asian Population Investig. Ophthalmol. Vis. Sci. 2008 49 5397 5402 10.1167/iovs.08-2142 18719076 66. Tham Y.-C. Cheng C.-Y. Zheng Y. Aung T. Wong T.Y. Cheung C. Relationship between Retinal Vascular Geometry with Retinal Nerve Fiber Layer and Ganglion Cell-Inner Plexiform Layer in Nonglaucomatous Eyes Investig. Ophthalmol. Vis. Sci. 2013 54 7309 7316 10.1167/iovs.13-12796 24135760 67. Tham Y.-C. Siantar R.G. Cheung C. Tan S.-P. Koh V.T. Aung T. Wong T.Y. Cheng C.-Y. Inter-Relationships Between Retinal Vascular Caliber, Retinal Nerve Fiber Layer Thickness, and Glaucoma: A Mediation Analysis Approach Investig. Ophthalmol. Vis. Sci. 2016 57 3803 3809 10.1167/iovs.15-18683 27421069 68. Wang S. Xu L. Wang Y. Wang Y. Jonas J.B. Retinal vessel diameter in normal and glaucomatous eyes: The Beijing eye study Clin. Exp. Ophthalmol. 2007 35 800 807 10.1111/j.1442-9071.2007.01627.x 18173406 69. Zheng Y. Cheung N. Aung T. Mitchell P. He M. Wong T.Y. Relationship of Retinal Vascular Caliber with Retinal Nerve Fiber Layer Thickness: The Singapore Malay Eye Study Investig. Ophthalmol. Vis. Sci. 2009 50 4091 4096 10.1167/iovs.09-3444 19443726 70. Koh V. Cheung C. Zheng Y. Wong T.Y. Wong W. Aung T. Relationship of Retinal Vascular Tortuosity with the Neuroretinal Rim: The Singapore Malay Eye Study Investig. Ophthalmol. Vis. Sci. 2010 51 3736 3741 10.1167/iovs.09-5008 20207964 71. Gao J. Liang Y. Wang F. Shen R. Wong T. Peng Y. Friedman D. Wang N. Retinal Vessels Change in Primary Angle-Closure Glaucoma: The Handan Eye Study Sci. Rep. 2015 5 9585 10.1038/srep09585 25928115 72. Chan K.K.W. Tang F. Tham C.Y.C. Young A.L. Cheung C.Y. Retinal vasculature in glaucoma: A review BMJ Open Ophthalmol. 2017 1 e000032 10.1136/bmjophth-2016-000032 29354699 73. Mitchell P. Leung H. Wang J.J. Rochtchina E. Lee A.J. Wong T.Y. Klein R. Retinal vessel diameter and open-angle glaucoma: The Blue Mountains Eye Study Ophthalmology 2005 112 245 250 10.1016/j.ophtha.2004.08.015 15691558 74. Klein R. Klein B.E. Tomany S.C. Wong T.Y. The relation of retinal microvascular characteristics to age-related eye disease: The Beaver Dam eye study Am. J. Ophthalmol. 2004 137 435 444 10.1016/j.ajo.2003.10.020 15013865 75. Arnould L. Guillemin M. Seydou A. Gabrielle P.-H. Bourredjem A. Kawasaki R. Binquet C. Bron A.M. Creuzot-Garcher C. Association between the retinal vascular network and retinal nerve fiber layer in the elderly: The Montrachet study PLoS ONE 2020 15 e0241055 10.1371/journal.pone.0241055 33085730 76. Arnould L. De Lazzer A. Seydou A. Binquet C. Bron A.M. Creuzot-Garcher C. Diagnostic ability of spectral-domain optical coherence tomography peripapillary retinal nerve fiber layer thickness to discriminate glaucoma patients from controls in an elderly population (The MONTRACHET study) Acta Ophthalmol. 2020 98 e1009 e1016 10.1111/aos.14448 32333503 77. Chiquet C. Gavard O. Arnould L. Mautuit T. Macgillivray T.J. Bron A.M. Semecas R. Trucco E. Florent A. Retinal vessel phenotype in patients with primary open-angle glaucoma Acta Ophthalmol. 2020 98 e88 e93 10.1111/aos.14192 31359603 78. Mauschitz M.M. Bonnemaijer P.W. Diers K. Rauscher F.G. Elze T. Engel C. Loeffler M. Colijn J.M. Ikram M.A. Vingerling J.R. Systemic and Ocular Determinants of Peripapillary Retinal Nerve Fiber Layer Thickness Measurements in the European Eye Epidemiology (E3) Population Ophthalmology 2018 125 1526 1536 10.1016/j.ophtha.2018.03.026 29716786 79. Li X. Wong W.L. Cheung C. Cheng C.-Y. Ikram M.K. Li J. Chia K.S. Wong T.Y. Racial Differences in Retinal Vessel Geometric Characteristics: A Multiethnic Study in Healthy Asians Investig. Ophthalmol. Vis. Sci. 2013 54 3650 3656 10.1167/iovs.12-11126 23652487 80. Zhu M.M. Choy B.N.K. You Q.S. Chan J.C.H. Ng A.L.K. Shih K. Cheung J.J.C. Wong J.K.W. Shum J. Ni M.Y. Optic disc and peripapillary vessel density measured with optical coherence tomography angiography and its associations in Chinese adults: A large population-based study Br. J. Ophthalmol. 2021 10.1136/bjophthalmol-2020-318199 81. Wang W. Zhang X. Choroidal Thickness and Primary Open-Angle Glaucoma: A Cross-Sectional Study and Meta-Analysis Investig. Ophthalmol. Vis. Sci. 2014 55 6007 6014 10.1167/iovs.14-14996 25168904 82. Wang D. He M. Wu L. Yaplee S. Singh K. Lin S. Differences in iris structural measurements among American Caucasians, American Chinese and mainland Chinese Clin. Exp. Ophthalmol. 2012 40 162 169 10.1111/j.1442-9071.2011.02658.x 21794048 83. Wang D. Chiu C. He M. Wu L. Kao A. Lin S. Differences in Baseline Dark and the Dark-to-Light Changes in Anterior Chamber Angle Parameters in Whites and Ethnic Chinese Investig. Ophthalmol. Vis. Sci. 2011 52 9404 9410 10.1167/iovs.11-8470 22076986 84. Ren R. Wang N. Li B. Li L. Gao F. Xu X. Jonas J.B. Lamina Cribrosa and Peripapillary Sclera Histomorphometry in Normal and Advanced Glaucomatous Chinese Eyes with Various Axial Length Investig. Ophthalmol. Vis. Sci. 2009 50 2175 2184 10.1167/iovs.07-1429 19387083 85. Tun T.A. Wang X. Baskaran M. Nongpiur M.E. Tham Y.C. Nguyen D.Q. Strouthidis N.G. Aung T. Cheng C.-Y. Boote C. Determinants of lamina cribrosa depth in healthy Asian eyes: The Singapore Epidemiology Eye Study Br. J. Ophthalmol. 2021 105 367 373 10.1136/bjophthalmol-2020-315840 32434775 86. Fan S. Guo T. Chen B. Xiong J. Havens S. Gulati V. Toris C.B. Differences in ocular biometrics and aqueous humour dynamics between Chinese and Caucasian adults Br. J. Ophthalmol. 2019 103 1845 1849 10.1136/bjophthalmol-2018-313132 30894342 87. Kim N.R. Park H.J. Suh Y.J. Chin H.S. Kim C.Y. Heritabilities of Intraocular Pressure in the Population of Korea: The Korean National Health and Nutrition Examination Survey 2008–2009 JAMA Ophthalmol. 2014 132 278 285 10.1001/jamaophthalmol.2013.7612 24435617 88. Lee M.K. Cho S.-I. Kim H. Song Y.-M. Lee K. Kim J.-I. Kim D.-M. Chung T.-Y. Kim Y.S. Seo J.-S. Epidemiologic Characteristics of Intraocular Pressure in the Korean and Mongolian Populations: The Healthy Twin and the GENDISCAN Study Ophthalmology 2012 119 450 457 10.1016/j.ophtha.2011.09.016 22244945 89. Zhou Q. Liang Y. Wong T.Y. Yang X.H. Lian L. Zhu D. Sun L.P. Wang N.L. Friedman D. Intraocular Pressure and its Relationship to Ocular and Systemic Factors in a Healthy Chinese Rural Population: The Handan Eye Study Ophthalmic Epidemiol. 2012 19 278 284 10.3109/09286586.2012.708084 22978528 90. Wang D. Huang W. Li Y. Zheng Y. Foster P. Congdon N. He M. Intraocular Pressure, Central Corneal Thickness, and Glaucoma in Chinese Adults: The Liwan Eye Study Am. J. Ophthalmol. 2011 152 454 462.e1 10.1016/j.ajo.2011.03.005 21679915 91. Jonas J.B. Nangia V. Matin A. Sinha A. Kulkarni M. Bhojwani K. Intraocular Pressure and Associated Factors: The central India eye and medical study J. Glaucoma 2011 20 405 409 10.1097/IJG.0b013e3181f7af9b 21897285 92. Kawase K. Tomidokoro A. Araie M. Iwase A. Yamamoto T. Tajimi Study Group Japan Glaucoma Society Ocular and systemic factors related to intraocular pressure in Japanese adults: The Tajimi study Br. J. Ophthalmol. 2008 92 1175 1179 10.1136/bjo.2007.128819 18669541 93. Foster P.J. Machin D. Wong T.-Y. Ng T.-P. Kirwan J.F. Johnson G.J. Khaw P.T. Seah S.K.L. Determinants of intraocular pressure and its association with glaucomatous optic neuropathy in Chinese Singaporeans: The Tanjong Pagar Study Investig. Ophthalmol. Vis. Sci. 2003 44 3885 3891 10.1167/iovs.03-0012 12939305 94. Kim Y.-H. Jung S.W. Nam G.-E. Han K.D. Bok A.R. Baek S.J. Cho K.-H. Choi Y.S. Kim S.M. Ju S.-Y. High intraocular pressure is associated with cardiometabolic risk factors in South Korean men: Korean National Health and Nutrition Examination Survey, 2008–2010 Eye 2014 28 672 679 10.1038/eye.2014.43 24603415 95. Tham Y.-C. Liao J. Vithana E.N. Khor C.C. Teo Y.-Y. Tai E.S. Wong T.Y. Aung T. Cheng C.-Y. Aggregate Effects of Intraocular Pressure and Cup-to-Disc Ratio Genetic Variants on Glaucoma in a Multiethnic Asian Population Ophthalmology 2015 122 1149 1157 10.1016/j.ophtha.2015.01.024 25745874 96. Baek S.U. Kee C. Suh W. Longitudinal analysis of age-related changes in intraocular pressure in South Korea Eye 2015 29 625 629 10.1038/eye.2015.11 25697455 97. Chua J. Chee M.L. Chin C. Tham Y.C. Tan N. Lim S.H. Aung T. Cheng C.-Y. Wong T.Y. Schmetterer L. Inter-relationship between ageing, body mass index, diabetes, systemic blood pressure and intraocular pressure in Asians: 6-year longitudinal study Br. J. Ophthalmol. 2019 103 196 202 10.1136/bjophthalmol-2018-311897 29632002 98. Nomura H. Ando F. Niino N. Shimokata H. Miyake Y. The relationship between age and intraocular pressure in a Japanese population: The influence of central corneal thickness Curr. Eye Res. 2002 24 81 85 10.1076/ceyr.24.2.81.8161 12187477 99. Klein B.E. Klein R. Linton K.L. Intraocular pressure in an American community. The Beaver Dam Eye Study Investig. Ophthalmol. Vis. Sci. 1992 33 2224 2228 1607232 100. Wu S.-Y. Leske M.C. Associations with Intraocular Pressure in the Barbados Eye Study Arch. Ophthalmol. 1997 115 1572 1576 10.1001/archopht.1997.01100160742012 9400792 101. Huck A. Harris A. Siesky B. Kim N. Muchnik M. Kanakamedala P. Amireskandari A. Tobe L.A. Vascular considerations in glaucoma patients of African and European descent Acta Ophthalmol. 2014 92 e336 e340 10.1111/aos.12354 24460758 102. Xu L. Wang Y.X. Jonas J.B. Ocular perfusion pressure and glaucoma: The Beijing Eye Study Eye 2009 23 734 736 10.1038/eye.2008.342 18989341 103. Gangwani R.A. Chan J. Lee J.W. Kwong A. Lai J.S.M. Detection of Glaucoma in a Cohort of Chinese Subjects with Systemic Hypertension J. Ophthalmol. 2013 2013 463710 10.1155/2013/463710 23401742 104. Gangwani R.A. Lee J.W. Mo H. Sum R. Kwong A.S. Wang J.H. Tsui W.W. Chan J. Lai J.S. The Correlation of Retinal Nerve Fiber Layer Thickness with Blood Pressure in a Chinese Hypertensive Population Medicine 2015 94 e947 10.1097/MD.0000000000000947 26061324 105. Zheng Y. Wong T.Y. Mitchell P. Friedman D.S. He M. Aung T. Distribution of Ocular Perfusion Pressure and Its Relationship with Open-Angle Glaucoma: The Singapore Malay Eye Study Investig. Ophthalmol. Vis. Sci. 2010 51 3399 3404 10.1167/iovs.09-4867 20164462 106. Tham Y.-C. Lim S.-H. Gupta P. Aung T. Wong T.Y. Cheng C.-Y. Inter-relationship between ocular perfusion pressure, blood pressure, intraocular pressure profiles and primary open-angle glaucoma: The Singapore Epidemiology of Eye Diseases study Br. J. Ophthalmol. 2018 102 1402 1406 10.1136/bjophthalmol-2017-311359 29331952 107. Kim S.K. Cho B.J. Hong S. Kang S.Y. Kim J.S. Kim C.Y. Seong G.J. Pulsatile Ocular Blood Flow in Healthy Koreans Korean J. Ophthalmol. 2008 22 6 9 10.3341/kjo.2008.22.1.6 18323699 108. He M. Liu B. Huang W. Zhang J. Yin Q. Zheng Y. Wang D. Ge J. Heritability of Optic Disc and Cup Measured by the Heidelberg Retinal Tomography in Chinese: The Guangzhou Twin Eye Study Investig. Ophthalmol. Vis. Sci. 2008 49 1350 1355 10.1167/iovs.07-1146 18385049 109. Abu-Amero K. Kondkar A.A. Chalam K.V. An Updated Review on the Genetics of Primary Open Angle Glaucoma Int. J. Mol. Sci. 2015 16 28886 28911 10.3390/ijms161226135 26690118 110. Gharahkhani P. Jorgenson E. Hysi P. Khawaja A.P. Pendergrass S. Han X. Ong J.S. Hewitt A.W. Segrè A.V. Rouhana J.M. Genome-wide meta-analysis identifies 127 open-angle glaucoma loci with consistent effect across ancestries Nat. Commun. 2021 12 1258 10.1038/s41467-020-20851-4 33627673 111. Shin H.-T. Yoon B.W. Seo J.H. Analysis of risk allele frequencies of single nucleotide polymorphisms related to open-angle glaucoma in different ethnic groups BMC Med. Genom. 2021 14 80 10.1186/s12920-021-00921-2 112. Wang J. Yusufu M. Khor C.C. Aung T. Wang N. The genetics of angle closure glaucoma Exp. Eye Res. 2019 189 107835 10.1016/j.exer.2019.107835 31634478 113. Shi H. Chen Y. Lu H. Zhu R. Zhang J. He M. Guan H. In-depth analysis of eight susceptibility loci of primary angle closure glaucoma in Han Chinese Exp. Eye Res. 2021 202 108350 10.1016/j.exer.2020.108350 33227294 114. Fan W. Li W. Duan C. Zhang W. Guo Y. Chen F. Characterization of a novel mutation in the MYOC gene in a Chinese family with primary open-angle glaucoma Mol. Med. Rep. 2020 22 3263 3270 10.3892/mmr.2020.11441 32945492 115. Lei L. Li S. Liu X. Zhang C. The clinical feature of myocilin Y437H mutation in a Chinese family with primary open-angle glaucoma Br. J. Ophthalmol. 2019 103 1524 1529 10.1136/bjophthalmol-2018-313069 30612094 116. Yang Y. Shi Y. Huang X. Li X. Ye Z. Shuai P. Qu C. Chen R. Xu J. Yang Z. Identification of a novel MYOC mutation in a Chinese family with primary open-angle glaucoma Gene 2015 571 188 193 10.1016/j.gene.2015.06.042 26095806 117. Guo H. Li M. Wang Z. Liu Q. Wu X. Association of MYOC and APOE promoter polymorphisms and primary open-angle glaucoma: A meta-analysis Int. J. Clin. Exp. Med. 2015 8 2052 2064 25932136 118. Jin X. Wang D.-J. Qu L.-H. Hou B.-K. Gong Y. Xu W.-W. Haplotype Analysis of Association of the MYOC Gene with Primary Angle-Closure Glaucoma in a Han Chinese Population Genet. Test. Mol. Biomark. 2015 19 3 8 10.1089/gtmb.2014.0130 119. Cai S.P. Muhemaiti P. Yin Y. Cheng H. Di Ya A. Keyimu M. Cao X. Fan N. Jiang L. Yan N. A novel MYOC heterozygous mutation identified in a Chinese Uygur pedigree with primary open-angle glaucoma Mol. Vis. 2012 18 1944 1951 22876119 120. Chen J. Cai S.-P. Yu W. Yan N. Tang L. Chen X. Liu X. Sequence analysis of MYOC and CYP1B1 in a Chinese pedigree of primary open-angle glaucoma Mol. Vis. 2011 17 1431 1435 21655360 121. Qu X. Zhou X. Zhou K. Xie X. Tian Y. New mutation in the MYOC gene and its association with primary open-angle glaucoma in a Chinese family Mol. Biol. Rep. 2010 37 255 261 10.1007/s11033-009-9667-3 19688280 122. Jia L.-Y. Tam P.O.-S. Chiang S.W.-Y. Ding N. Chen L.J. Yam G.H.-F. Pang C.-P. Wang N.-L. Multiple gene polymorphisms analysis revealed a different profile of genetic polymorphisms of primary open-angle glaucoma in northern Chinese Mol. Vis. 2009 15 89 98 19145250 123. Xie X. Zhou X. Qu X. Wen J. Tian Y. Zheng F. Two novel myocilin mutations in a Chinese family with primary open-angle glaucoma Mol. Vis. 2008 14 1666 1672 18776955 124. Mengkegale M. Fuse N. Miyazawa A. Takahashi K. Seimiya M. Yasui T. Tamai M. Nakazawa T. Nishida K. Presence of myocilin sequence variants in Japanese patients with open-angle glaucoma Mol. Vis. 2008 14 413 417 18334962 125. Kumar A. Basavaraj M.G. Gupta S.K. Qamar I. Ali A.M. Bajaj V. Ramesh T. Prakash D.R. Shetty J.S. Dorairaj S.K. Role of CYP1B1, MYOC, OPTN and OPTC genes in adult-onset primary open-angle glaucoma: Predominance of CYP1B1 mutations in Indian patients Mol. Vis. 2007 13 667 676 17563717 126. Funayama T. Mashima Y. Ohtake Y. Ishikawa K. Fuse N. Yasuda N. Fukuchi T. Murakami A. Hotta Y. Shimada N. SNPs and Interaction Analyses of Noelin 2, Myocilin, and Optineurin Genes in Japanese Patients with Open-Angle Glaucoma Investig. Ophthalmol. Vis. Sci. 2006 47 5368 5375 10.1167/iovs.06-0196 17122126 127. Chakrabarti S. Kaur K. Komatireddy S. Acharya M. Devi K.R. Mukhopadhyay A. Mandal A.K. Hasnain S. Chandrasekhar G. Thomas R. Gln48His is the prevalent myocilin mutation in primary open angle and primary congenital glaucoma phenotypes in India Mol. Vis. 2005 11 111 113 15723004 128. Ishikawa K. Funayama T. Ohtake Y. Tanino T. Kurosaka D. Suzuki K. Ideta H. Fujimaki T. Tanihara H. Asaoka R. Novel MYOC Gene Mutation, Phe369Leu, in Japanese Patients with Primary Open-Angle Glaucoma Detected by Denaturing High-Performance Liquid Chromatography J. Glaucoma 2004 13 466 471 10.1097/0.ijg.0000138204.6d 15534471 129. Kanagavalli J. Krishnadas S.R. Pandaranayaka E. Krishnaswamy S. Sundaresan P. Evaluation and understanding of myocilin mutations in Indian primary open angle glaucoma patients Mol. Vis. 2003 9 606 614 14627955 130. Mukhopadhyay A. Acharya M. Mukherjee S. Ray J. Choudhury S. Khan M. Ray K. Mutations in MYOC gene of Indian primary open angle glaucoma patients Mol. Vis. 2002 8 442 448 12447164 131. Cheng J.-W. Li P. Wei R.-L. Meta-analysis of association between optineurin gene and primary open-angle glaucoma Med. Sci. Monit. 2010 16 CR369-77 20671613 132. Xiao Z. Meng Q. Tsai J.C. Yuan H. Xu N. Li Y. A novel optineurin genetic mutation associated with open-angle glaucoma in a Chinese family Mol. Vis. 2009 15 1649 1654 19710941 133. Ayala-Lugo R.M. Pawar H. Reed D.M. Lichter P.R. Moroi S.E. Page M. Eadie J. Azocar V. Maul E. Ntim-Amponsah C. Variation in optineurin (OPTN) allele frequencies between and within populations Mol. Vis. 2007 13 151 163 17293779 134. Sripriya S. Nirmala J. George R. Hemamalini A. Baskaran M. Prema R. Ramesh S.V. Karthiyayini T. Amali J. Job S. OPTN gene: Profile of patients with glaucoma from India Mol. Vis. 2006 12 816 820 16885925 135. Funayama T. Ishikawa K. Ohtake Y. Tanino T. Kurosaka D. Kimura I. Suzuki K. Ideta H. Nakamoto K. Yasuda N. Variants in Optineurin Gene and Their Association with Tumor Necrosis Factor-α Polymorphisms in Japanese Patients with Glaucoma Investig. Ophthalmol. Vis. Sci. 2004 45 4359 4367 10.1167/iovs.03-1403 15557444 136. Leung Y.F. Fan B.J. Lam D.S.C. Lee W.S. Tam P.O.S. Chua J.K.H. Tham C.Y.C. Lai J.S.M. Fan D.S.P. Pang C.P. Different optineurin mutation pattern in primary open-angle glaucoma Investig. Ophthalmol. Vis. Sci. 2003 44 3880 3884 10.1167/iovs.02-0693 12939304 137. Gong B. Qu C. Li X. Shi Y. Lin Y. Zhou Y. Shuai P. Yang Y. Liu X. Zhang D. Mutation spectrum ofCYP1B1in Chinese patients with primary open-angle glaucoma Br. J. Ophthalmol. 2015 99 425 430 10.1136/bjophthalmol-2014-306054 25527694 138. Dong S. Yang J. Yu W. Kota P. Xia X. Xu H. No association of genetic polymorphisms in CYP1B1 with primary open-angle glaucoma: A meta- and gene-based analysis Mol. Vis. 2012 18 786 796 22509109 139. Bhattacharjee A. Banerjee D. Mookherjee S. Acharya M. Banerjee A. Ray A. Sen A. Indian Genome Variation Consortium Ray K. Leu432Val polymorphism in CYP1B1 as a susceptible factor towards predisposition to primary open-angle glaucoma Mol. Vis. 2008 14 841 850 18483560 140. Huang L. Chen Y. Lin Y. Tam P.O.S. Cheng Y. Shi Y. Gong B. Lu F. Yang J. Wang H. Genome-wide analysis identified 17 new loci influencing intraocular pressure in Chinese population Sci. China Life Sci. 2019 62 153 164 10.1007/s11427-018-9430-2 30591961 141. Kim S. Kim K. Heo D.W. Kim J.-S. Park C.K. Kim C.-S. Kang C. Expression-associated polymorphisms of CAV1-CAV2 affect intraocular pressure and high-tension glaucoma risk Mol. Vis. 2015 21 548 554 26015768 142. Huang W. Wang W. Zhou M. Zhang X. Association of single-nucleotide polymorphism rs4236601 near caveolin 1 and 2 with primary open-angle glaucoma: A meta-analysis Clin. Exp. Ophthalmol. 2014 42 515 521 10.1111/ceo.12201 24034151 143. Kato T. Meguro A. Nomura E. Uemoto R. Nomura N. Ota M. Kashiwagi K. Mabuchi F. Iijima H. Kawase K. Association study of genetic variants on chromosome 7q31 with susceptibility to normal tension glaucoma in a Japanese population Eye 2013 27 979 983 10.1038/eye.2013.123 23743525 144. Chai X. Low K.Y. Tham Y.C. Chee M.L. Thakur S. Zhang L. Tan N.Y. Khor C.C. Aung T. Wong T.Y. Association of Glaucoma Risk Genes with Retinal Nerve Fiber Layer in a Multi-ethnic Asian Population: The Singapore Epidemiology of Eye Diseases Study Investig. Ophthalmol. Vis. Sci. 2020 61 37 10.1167/iovs.61.10.37 32821913 145. Yu C.C. Allingham R.R. Nakano M. Li-Jia C. Chen Y. Ikeda Y. Mani B. Chen L.-J. Kee C. Garway-Heath D. A common variant near TGFBR3 is associated with primary open angle glaucoma Hum. Mol. Genet. 2015 24 3880 3892 10.1093/hmg/ddv128 25861811 146. Khor C.C. Ramdas W.D. Vithana E.N. Cornes B.K. Sim X. Tay W.-T. Saw S.-M. Zheng Y. Lavanya R. Wu R. Genome-wide association studies in Asians confirm the involvement of ATOH7 and TGFBR3, and further identify CARD10 as a novel locus influencing optic disc area Hum. Mol. Genet. 2011 20 1864 1872 10.1093/hmg/ddr060 21307088 147. Mabuchi F. Sakurada Y. Kashiwagi K. Yamagata Z. Iijima H. Tsukahara S. Involvement of Genetic Variants Associated with Primary Open-Angle Glaucoma in Pathogenic Mechanisms and Family History of Glaucoma Am. J. Ophthalmol. 2015 159 437 444.e2 10.1016/j.ajo.2014.11.023 25461262 148. Chen J.-H. Wang D. Huang C. Zheng Y. Chen H. Pang C.-P. Zhang M. Interactive Effects ofATOH7andRFTN1in Association with Adult-Onset Primary Open-Angle Glaucoma Investig. Ophthalmol. Vis. Sci. 2012 53 779 785 10.1167/iovs.11-8277 22222511 149. Hu Z. He C. CDKN2B gene rs1063192 polymorphism decreases the risk of glaucoma Oncotarget 2017 8 21167 21176 10.18632/oncotarget.15504 28416752 150. Nakano M. Ikeda Y. Tokuda Y. Fuwa M. Omi N. Ueno M. Imai K. Adachi H. Kageyama M. Mori K. Common Variants in CDKN2B-AS1 Associated with Optic-Nerve Vulnerability of Glaucoma Identified by Genome-Wide Association Studies in Japanese PLoS ONE 2012 7 e33389 10.1371/journal.pone.0033389 22428042 151. Osman W. Low S.-K. Takahashi A. Kubo M. Nakamura Y. A genome-wide association study in the Japanese population confirms 9p21 and 14q23 as susceptibility loci for primary open angle glaucoma Hum. Mol. Genet. 2012 21 2836 2842 10.1093/hmg/dds103 22419738 152. Takamoto M. Kaburaki T. Mabuchi A. Araie M. Amano S. Aihara M. Tomidokoro A. Iwase A. Mabuchi F. Kashiwagi K. Common Variants on Chromosome 9p21 Are Associated with Normal Tension Glaucoma PLoS ONE 2012 7 e40107 10.1371/journal.pone.0040107 22792221 153. Lu S.Y. He Z.Z. Xu J.X. Yang C. Chen L.J. Gong B. Association of Polymorphisms at the SIX1-SIX6 Locus with Primary Open-Angle Glaucoma Investig. Ophthalmol. Vis. Sci. 2019 60 2914 2924 10.1167/iovs.18-26489 31284308 154. Rong S.S. Lu S.Y. Matsushita K. Huang C. Leung C.K. Kawashima R. Usui S. Tam P.O. Young A.L. Tsujikawa M. Association of the SIX6 locus with primary open angle glaucoma in southern Chinese and Japanese Exp. Eye Res. 2019 180 129 136 10.1016/j.exer.2018.12.014 30586556 155. Zhao F. Fan Z. Huang X. Role of matrix metalloproteinase-9 gene polymorphisms in glaucoma: A hospital-based study in Chinese patients J. Clin. Lab. Anal. 2020 34 e23105 10.1002/jcla.23105 31713905 156. He M. Wang W. Han X. Huang W. Matrix metalloproteinase-1 rs1799750 polymorphism and glaucoma: A meta-analysis Ophthalmic Genet. 2017 38 211 216 10.1080/13816810.2016.1193877 27428613 157. Wu M. Zhu X.-Y. Ye J. Associations of polymorphisms of LOXL1 gene with primary open-angle glaucoma: A meta-analysis based on 5293 subjects Mol. Vis. 2015 21 165 172 25750511 158. Sun W. Sheng Y. Weng Y. Xu C.-X. Williams S.E. Liu Y.-T. Hauser M.A. Allingham R.R. Jin M.-J. Chen G.-D. Lack of association between lysyl oxidase-like 1 polymorphisms and primary open angle glaucoma: A meta-analysis Int. J. Ophthalmol. 2014 7 550 556 10.3980/j.issn.2222-3959.2014.03.29 24967207 159. Zhang J. Wang L. Association between rs4938723 polymorphism and the risk of primary open-angle glaucoma (POAG) in a Chinese population J. Cell. Biochem. 2019 120 12875 12886 10.1002/jcb.28559 30861198 160. Gupta S. Chatterjee S. Chandra A. Maurya O.P.S. Mishra R.N. Mukherjee A. Mutsuddi M. TP53 codon 72 polymorphism and the risk of glaucoma in a north Indian cohort: A genetic association study Ophthalmic Genet. 2018 39 228 235 10.1080/13816810.2017.1413661 29261364 161. Guo Y. Zhang H. Chen X. Yang X. Cheng W. Zhao K. Association of TP53 Polymorphisms with Primary Open-Angle Glaucoma: A Meta-Analysis Investig. Ophthalmol. Vis. Sci. 2012 53 3756 3763 10.1167/iovs.12-9818 22562509 162. Fan B. Liu K. Wang D.Y. Tham C.Y.C. Tam P.O.S. Lam D.S.C. Pang C.P. Association of Polymorphisms of Tumor Necrosis Factor and Tumor Protein p53 with Primary Open-Angle Glaucoma Investig. Ophthalmol. Vis. Sci. 2010 51 4110 4116 10.1167/iovs.09-4974 20357201 163. Passan S. Goyal S. Bhat M.A. Singh D. Vanita V. Association of TNF-α gene alterations (c.-238G > A, c.-308G > A, c.-857C > T, c.-863C > A) with primary glaucoma in north Indian cohort Gene 2019 709 25 35 10.1016/j.gene.2019.05.035 31132515 164. Wang C.-Y. Shen Y.-C. Wei L.-C. Lin K.-H. Feng S.-C. Yang Y.-Y. Chiu C.-H. Tsai H.-Y. Polymorphism in the TNF-α(−863) locus associated with reduced risk of primary open angle glaucoma Mol. Vis. 2012 18 779 785 22509108 165. Wang Y. Zhou Y.-F. Zhao B.-Y. Gu Z.-Y. Li S.-L. Apolipoprotein E gene ε4ε4 is associated with elevated risk of primary open angle glaucoma in Asians: A meta-analysis BMC Med. Genet. 2014 15 60 10.1186/1471-2350-15-60 24885013 166. Lam C.Y. Fan B.J. Wang D.Y. Tam P.O.S. Tham C.C.Y. Leung D.Y.L. Fan D.S.P. Lam D.S.C. Pang C.P. Association of Apolipoprotein E Polymorphisms with Normal Tension Glaucoma in a Chinese Population J. Glaucoma 2006 15 218 222 10.1097/01.ijg.0000212217.19804.a7 16778644 167. Takano Y. Shi D. Shimizu A. Funayama T. Mashima Y. Yasuda N. Fukuchi T. Abe H. Ideta H. Zheng X. Association of Toll-like Receptor 4 Gene Polymorphisms in Japanese Subjects with Primary Open-Angle, Normal-Tension, and Exfoliation Glaucoma Am. J. Ophthalmol. 2012 154 825 832.e1 10.1016/j.ajo.2012.03.050 22831837 168. Chen L.J. Tam P.O. Leung D.Y. Fan A.H. Zhang M. Tham C.C. Chiang S.W. Fan B.J. Wang N. Pang C.P. SNP rs1533428 at 2p16.3 as a marker for late-onset primary open-angle glaucoma Mol. Vis. 2012 18 1629 1639 22773901 169. Suh W. Kim S. Ki C.-S. Kee C. Toll-like Receptor 4 gene polymorphisms do not associate with normal tension glaucoma in a Korean population Mol. Vis. 2011 17 2343 2348 21921986 170. Chen L.J. Ng T.K. Fan A.H. Leung D.Y. Zhang M. Wang N. Zheng Y. Liang X.Y. Chiang S.W. Tam P.O. Evaluation of NTF4 as a causative gene for primary open-angle glaucoma Mol. Vis. 2012 18 1763 1772 22815630 171. Rao K.N. Kaur I. Parikh R.S. Mandal A.K. Chandrasekhar G. Thomas R. Chakrabarti S. Variations inNTF4, VAV2, andVAV3Genes Are Not Involved with Primary Open-Angle and Primary Angle-Closure Glaucomas in an Indian Population Investig. Ophthalmol. Vis. Sci. 2010 51 4937 4941 10.1167/iovs.10-5553 20463313 172. Vithana E.N. Nongpiur M.E. Venkataraman D. Chan S.H. Mavinahalli J. Aung T. Identification of a novel mutation in the NTF4 gene that causes primary open-angle glaucoma in a Chinese population Mol. Vis. 2010 16 1640 1645 20806036 173. Lee M.K. Woo S.J. Kim J.-I. Cho S.-I. Kim H. Sung J. Seo J.-S. Kim D.M. Replication of a Glaucoma Candidate Gene on 5q22.1 for Intraocular Pressure in Mongolian Populations: The GENDISCAN Project Investig. Ophthalmol. Vis. Sci. 2010 51 1335 1340 10.1167/iovs.09-3979 19875670 174. Fan B.J. Wang D.Y. Cheng C.-Y. Ko W.C. Lam S.C. Pang C.P. Different WDR36 mutation pattern in Chinese patients with primary open-angle glaucoma Mol. Vis. 2009 15 646 653 19347049 175. Miyazawa A. Fuse N. Mengkegale M. Ryu M. Seimiya M. Wada Y. Nishida K. Association between primary open-angle glaucoma and WDR36 DNA sequence variants in Japanese Mol. Vis. 2007 13 1912 1919 17960130 176. How A.C.S. Aung T. Chew X. Yong V.H.K. Lim M.C.C. Lee K.Y.C. Toh J.-Y. Li Y. Liu J. Vithana E.N. Lack of Association between Interleukin-1 Gene Cluster Polymorphisms and Glaucoma in Chinese Subjects Investig. Ophthalmol. Vis. Sci. 2007 48 2123 2126 10.1167/iovs.06-1213 17460270 177. Wang C.Y. Shen Y.C. Su C.H. Lo F.Y. Lee S.H. Tsai H.Y. Fan S.S. Investigation of the association between interleukin-1beta polymorphism and normal tension glaucoma Mol. Vis. 2007 13 719 723 17563722 178. Wang C.-Y. Shen Y.-C. Lo F.-Y. Su C.-H. Lee S.-H. Tsai H.-Y. Fan S.-S. Normal Tension Glaucoma is not Associated with the Interleukin -1α (−889) Genetic Polymorphism J. Glaucoma 2007 16 230 233 10.1097/IJG.0b013e3180300818 17473736 179. Wang C.-Y. Shen Y.-C. Lo F.-Y. Su C.-H. Lee S.-H. Lin K.-H. Tsai H.-Y. Kuo N.-W. Fan S.-S. Polymorphism in the IL-1alpha (-889) locus associated with elevated risk of primary open angle glaucoma Mol. Vis. 2006 12 1380 1385 17149369 180. Shi D. Takano Y. Nakazawa T. Mengkegale M. Yokokura S. Nishida K. Fuse N. Molecular genetic analysis of primary open-angle glaucoma, normal tension glaucoma, and developmental glaucoma for the VAV2 and VAV3 gene variants in Japanese subjects Biochem. Biophys. Res. Commun. 2013 432 509 512 10.1016/j.bbrc.2013.02.003 23402756 181. Meng N. Ma J. Xia L. Zhang Y. Shi C. Association between SNPs (rs1533428, rs12994401, rs10202118) on chromosome 2p16.3 and primary open angle glaucoma Curr. Eye Res. 2015 40 839 846 10.3109/02713683.2014.958172 25271776 182. Chen Y. Lin Y. Vithana E.N. Jia L. Zuo X. Wong T.Y. Chen L.J. Zhu X. Tam P.O.S. Gong B. Common variants near ABCA1 and in PMM2 are associated with primary open-angle glaucoma Nat. Genet. 2014 46 1115 1119 10.1038/ng.3078 25173107 183. Li K. Yang C. Wan X. Xu J. Luo Q. Cheng Y. Peng J. Gong B. Jiang L. Liu Y. Evaluation of the association between five genetic variants and primary open-angle glaucoma in a Han Chinese population Ophthalmic Genet. 2020 41 252 256 10.1080/13816810.2020.1747089 32281515 184. Gong B. Zhang H. Huang L. Chen Y. Shi Y. Tam P.O.-S. Zhu X. Huang Y. Lei B. Sundaresan P. Mutant RAMP2 causes primary open-angle glaucoma via the CRLR-cAMP axis Genet. Med. 2019 21 2345 2354 10.1038/s41436-019-0507-0 31000793 185. Tang F.Y. Ma L. Tam P.O.S. Pang C.P. Tham C.C. Chen L.J. Genetic Association of the PARL-ABCC5-HTR3D-HTR3C Locus with Primary Angle-Closure Glaucoma in Chinese Investig. Ophthalmol. Vis. Sci. 2017 58 4384 4389 10.1167/iovs.17-22304 28813580 186. Nongpiur M.E. Khor C.C. Jia H. Cornes B.K. Chen L.-J. Qiao C. Nair K.S. Cheng C.-Y. Xu L. George R. ABCC5, a Gene That Influences the Anterior Chamber Depth, Is Associated with Primary Angle Closure Glaucoma PLoS Genet. 2014 10 e1004089 10.1371/journal.pgen.1004089 24603532 187. Zeng K. Zhong B. Fang M. Shen X.-L. Huang L.-N. Common polymorphisms of the hOGG1, APE1 and XRCC1 genes correlate with the susceptibility and clinicopathological features of primary angle-closure glaucoma Biosci. Rep. 2017 37 BSR20160644 10.1042/BSR20160644 28396513 188. Yousaf S. Khan M.I. Micheal S. Akhtar F. Ali S.H.B. Riaz M. Ali M. Lall P. Waheed N.K. Hollander A.I.D. XRCC1 and XPD DNA repair gene polymorphisms: A potential risk factor for glaucoma in the Pakistani population Mol. Vis. 2011 17 1153 1163 21617750 189. Wang S. Zhuang W. Ma J. Xu M. Piao S. Hao J. Zhang W. Chi H. Xue Z. Ha S. Association of Genes implicated in primary angle-closure Glaucoma and the ocular biometric parameters of anterior chamber depth and axial length in a northern Chinese population BMC Ophthalmol. 2018 18 271 10.1186/s12886-018-0934-8 30348125 190. Shi H. Zhu R. Hu N. Shi J. Zhang J. Jiang L. Jiang H. Guan H. Association of frizzled-related protein (MFRP) and heat shock protein 70 (HSP70) single nucleotide polymorphisms with primary angle closure in a Han Chinese population: Jiangsu Eye Study Mol. Vis. 2013 19 128 134 23378726 191. Gong B. Shi Y. Qu C. Ye Z. Yin Y. Tan C. Shuai P. Li J. Guo X. Cheng Y. Association of catalase polymorphisms with primary open-angle glaucoma in a Chinese population Ophthalmic Genet. 2018 39 35 40 10.1080/13816810.2017.1342132 28829657 192. Huang X. Wang N. Xiao X. Li S. Zhang Q. A novel truncation mutation in GJA1 associated with open angle glaucoma and microcornea in a large Chinese family Eye 2015 29 972 977 10.1038/eye.2015.74 25976645 193. Zhou Y. Shuai P. Li X. Liu X. Wang J. Yang Y. Hao F. Lin H. Zhang D. Gong B. Association of SOD2 Polymorphisms with Primary Open Angle Glaucoma in a Chinese Population Ophthalmic Genet. 2015 36 43 49 10.3109/13816810.2014.985844 25417767 194. Liu T. Xie L. Ye J. He X. Family-based analysis identified CD2 as a susceptibility gene for primary open angle glaucoma in Chinese Han population J. Cell. Mol. Med. 2014 18 600 609 10.1111/jcmm.12201 24597656 195. Lu Y. Shi Y. Yin J. Huang Z. Are glutathione S-transferase polymorphisms (GSTM1, GSTT1) associated with primary open angle glaucoma? A meta-analysis Gene 2013 527 311 315 10.1016/j.gene.2013.06.031 23827458 196. Huang W. Wang W. Zhou M. Chen S. Zhang X. Association of glutathione S-transferase polymorphisms (GSTM1 and GSTT1) with primary open-angle glaucoma: An evidence-based meta-analysis Gene 2013 526 80 86 10.1016/j.gene.2013.05.032 23747403 197. Guo Y. Chen X. Zhang H. Li N. Yang X. Cheng W. Zhao K. Association of OPA1 Polymorphisms with NTG and HTG: A Meta-Analysis PLoS ONE 2012 7 e42387 10.1371/journal.pone.0042387 22879959 198. Woo S.J. Kim D.M. Kim J.Y. Park S.S. Ko H.S. Yoo T. Investigation of the Association between OPA1 Polymorphisms and Normal-Tension Glaucoma in Korea J. Glaucoma 2004 13 492 495 10.1097/01.ijg.0000137870.25779.40 15534475 199. Yasumura R. Meguro A. Ota M. Nomura E. Uemoto R. Kashiwagi K. Mabuchi F. Iijima H. Kawase K. Yamamoto T. Investigation of the association between SLC1A3 gene polymorphisms and normal tension glaucoma Mol. Vis. 2011 17 792 796 21528001 200. Suzuki M. Meguro A. Ota M. Nomura E. Kato T. Nomura N. Kashiwagi K. Mabuchi F. Iijima H. Kawase K. Genotyping HLA-DRB1 and HLA-DQB1 alleles in Japanese patients with normal tension glaucoma Mol. Vis. 2010 16 1874 1879 21031025 201. Murakami K. Meguro A. Ota M. Shiota T. Nomura N. Kashiwagi K. Mabuchi F. Iijima H. Kawase K. Yamamoto T. Analysis of microsatellite polymorphisms within the GLC1F locus in Japanese patients with normal tension glaucoma Mol. Vis. 2010 16 462 466 20309402 202. Kim S.H. Kim J.Y. Kim D.M. Ko H.S. Kim S.Y. Yoo T. Hwang S.S. Park S.S. Investigations on the association between normal tension glaucoma and single nucleotide polymorphisms of the endothelin-1 and endothelin receptor genes Mol. Vis. 2006 12 1016 1021 16971893

PMC009xxxxxx/PMC9099680.txt

==== Front Nutrients Nutrients nutrients Nutrients 2072-6643 MDPI 10.3390/nu14091923 nutrients-14-01923 Review Carbohydrate Maldigestion and Intolerance https://orcid.org/0000-0002-1489-504X Fernández-Bañares Fernando 12 Mela David J. Academic Editor 1 Department of Gastroenterology, Hospital Universitary MútuaTerrassa, 08221 Terrassa, Spain; ffbanares@mutuaterrassa.es 2 Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas, Instituto de Salud Carlos III, 28029 Madrid, Spain 04 5 2022 5 2022 14 9 192311 4 2022 29 4 2022 © 2022 by the author. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). This review summarizes dietary carbohydrate intolerance conditions and recent advances on the possible role of carbohydrate maldigestion and dietary outcomes in patients with functional bowel disease. When malabsorbed carbohydrates reach the colon, they are fermented by colonic bacteria, with the production of short-chain fatty acids and gas lowering colonic pH. The appearance of diarrhoea or symptoms of flatulence depends in part on the balance between the production and elimination of these fermentation products. Different studies have shown that there are no differences in the frequency of sugar malabsorption between patients with irritable bowel disease (IBS) and healthy controls; however, the severity of symptoms after a sugar challenge is higher in patients than in controls. A diet low in ‘Fermentable, Oligo-Di- and Monosaccharides and Polyols’ (FODMAPs) is an effective treatment for global symptoms and abdominal pain in IBS, but its implementation should be supervised by a trained dietitian. A ‘bottom-up’ approach to the low-FODMAP diet has been suggested to avoid an alteration of gut microbiota and nutritional status. Two approaches have been suggested in this regard: starting with only certain subgroups of the low-FODMAP diet based on dietary history or with a gluten-free diet. sugar malabsorption lactose fructose sorbitol sucrose FODMAP irritable bowel syndrome hydrogen breath test No funding was received. ==== Body pmc1. Introduction In the average Western diet, dietary starch provides around 60% of ingested absorbable carbohydrates; the disaccharides sucrose and lactose and the monosaccharide fructose are consumed to a lesser extent. Absorbable dietary carbohydrates must be broken down to their monosaccharide components before being transported across the surface membrane of the absorptive enterocytes. The process of digestion is completed by the action of disaccharide-specific hydrolases bound to the brush-border membrane of enterocytes lining the villi of the small intestine. The monosaccharide products of saccharidase hydrolysis (glucose and galactose) and the monosaccharides in the human diet (glucose and fructose) are ultimately transported into the enterocyte via specific membrane transport systems [1]. While, under normal conditions, most of an ingested carbohydrate load is completely absorbed before reaching the colon, several conditions can result in the impairment of absorption in the small intestine. Carbohydrate malabsorption can provoke an osmotically driven influx of fluid into the small bowel, leading to intestinal distension and rapid propulsion into the colon [2]. In addition, unabsorbed carbohydrates are rapidly fermented by colonic microbiota, generating gas, lactate, and short-chain fatty acids that have effects on gastrointestinal function (Figure 1) [3]. All of these factors can produce diarrhoea, gas, bloating, flatulence, and abdominal pain, which are symptoms that patients with carbohydrate intolerance usually report. Other factors involving the development of symptoms are: 1. the quantity and quality of the ingested carbohydrate load; 2. the rate of gastric emptying; 3. the response of the small intestine to an osmotic load; 4. gastrointestinal motility; 5. the metabolic capacity of the colonic microbiota; and 6. the compensatory capacity of the colon to reabsorb water and short-chain fatty acids [4]. This review summarizes carbohydrate intolerance conditions and recent clinical and basic science reports on the possible role of carbohydrate malabsorption and dietary outcomes in patients with functional bowel disease. Table 1 describes the mechanism of absorption of dietary carbohydrates and the enzyme drugs available to facilitate digestion. 2. Lactose Malabsorption and Intolerance Recently, the different aspects of lactose malabsorption (LM) and its relationship to the development of clinical intolerance have been excellently reviewed [5,6], and the main points are summarized below. LM is typically caused by lactase downregulation after infancy due to lactase non-persistence, which, in Caucasians, is mediated by the LCT−13910:C/C genotype [7]. A recent meta-analysis estimated the global prevalence of LM at 68%, with higher rates reported for genotyping data than hydrogen breath test (HBT) data [8]. Thus, lactase non-persistence cannot be considered a disease. Both lactase persistence and non-persistence are common phenotypes in healthy humans. LM refers to an inefficient digestion of lactose in the small intestine. There are primary and secondary causes (such as viral gastroenteritis, giardiasis, and celiac disease). Lactose intolerance is the occurrence of gastrointestinal symptoms such as bloating, borborygmi, flatulence, abdominal pain, and diarrhoea in patients with LM after ingestion of lactose. The difference between LM and lactose intolerance is important since a lactose-restricted diet is only indicated in patients with both malabsorption and intolerance [9]. Symptoms of lactose intolerance characteristically do not arise until there is less than 50% of lactase activity. However, since adults with lactase deficiency often maintain between 10% and 30% of intestinal lactase activity, symptoms only develop when they eat enough lactose to overcome the compensatory mechanisms of the colon. There appears to be a dose-dependent effect. A study of a Chinese population with proven lactase deficiency showed few symptoms at lower doses (10 g of lactose) and more symptoms at higher doses (20 g and 40 g) with an increase in a linear fashion [10]. When malabsorbed lactose reaches the colon, it is fermented by colonic bacteria, with the production of short-chain fatty acids (mainly acetate, propionate, and butyrate) and gas (hydrogen, carbon dioxide, and in some subjects, methane). The appearance of diarrhoea or symptoms of flatulence depends in part on the balance between the production and elimination of these fermentation products [2,11,12]. The short-chain fatty acids are rapidly absorbed by the colonic mucosa, favouring the concomitant absorption of water and electrolytes, and are an important energy fuel for the colonocyte and the organism. Diarrhoea occurs only under certain circumstances: (a) if the arrival speed of lactose to the colon exceeds the rate of bacterial fermentation of this sugar, which causes an osmotic overload in the colon and the appearance of diarrhoea; (b) if the capacity of bacterial fermentation in the colon is decreased (for instance, with the bacterial shift after the use of broad-spectrum antibiotics), which results in decreased production of short-chain fatty acids and, therefore, a lower capacity to absorb water and electrolytes; and (c) if there is a decrease in the absorption of short-chain fatty acids, as occurs in inflammatory diseases of the colon. Furthermore, gases produced by fermentation are consumed by the same bacteria or are absorbed and pass into the circulatory bloodstream [13]. For all these reasons, small amounts of lactose can be malabsorbed without inducing symptoms of intolerance. HBT is the test of choice to assess LM and symptoms of intolerance [14,15,16]. Lactose HBT measures hydrogen produced by intestinal bacteria in the end-expiratory air after an oral challenge with a standard dose of lactose. In clinical practice, an intermediate lactose dose of 20–25 g in a 10% water solution is recommended [10,14,15,16]. As lactose intolerance is of greater clinical interest than LM, it has been debated whether it would be more relevant to use either ‘report of symptoms during a positive HBT’ or ‘disappearing of symptoms by a long-term lactose restricted diet after a positive HBT’ as a reference standard in diagnosis [9,17]. It is important to document that there is a relationship between carbohydrate intake and the occurrence of symptoms, and to consider that the symptoms are caused by the test carbohydrate [15]. Often, lactose malabsorbers do not develop symptoms after a challenge, and malabsorption without symptoms is not a major determinant for the outcome of the diet. In this sense, it has been suggested that the occurrence of symptoms during a lactose HBT strongly suggests a favourable response but does not help in predicting whether symptoms would subside or be reduced [18]. Conversely, their absence during the test was not associated with an acceptable negative predictive value. In another study, there were no significant differences in clinical improvement after a sugar-restricted diet between patients with either the presence or absence of clinical symptoms during HBT [19]. Thus, it is not as yet sufficiently clear whether the absence of symptoms after a positive HBT always gives an indication as to the role of sugar in the genesis of the patient’s symptoms. At present, therefore, lactose intolerance should be confirmed by a sustained, significant symptomatic relief of intestinal complaints after a lactose-free diet [9,15]. Reduction of lactose intake rather than exclusion is the key to the treatment of lactose intolerance [20]. In fact, while the intake of a large dose of lactose, of the order of 50 g of lactose (corresponding to one litre of milk), causes symptoms in most of the population with LM, most patients with lactose intolerance can ingest 12 g of lactose (equivalent to 200–250 mL of milk) with no symptoms, mainly when taken with other foods [21]. That is, the consumption of lactose with other foods likely slows gastric emptying and small intestinal transit, allowing lactose more time to be hydrolysed and absorbed. The role of lactase supplementation or the intake of the probiotics that produce lactase in the gut has been reviewed, and a positive effect was confirmed, though the effect was modest, and the quality of these studies was poor [5]. 3. Sucrase-Isomaltase Deficiency Sucrose, or saccharose, is commercially known as cane sugar or regular table sugar and consists of one glucose and one fructose molecule. The bond between these two molecules is broken by the membrane-bound enzyme sucrase-isomaltase. The same enzyme also hydrolyses the glucose molecules in the short oligosaccharides and starch [22]. Congenital sucrase-isomaltase deficiency (CSID) is a rare autosomal recessive condition with mutations of the sucrase-isomaltase gene on chromosome 3q25-26 [23]. Sucrase activity in intestinal villi is practically non-existent. The prevalence of CSID varies but has been described as 5–10% in Greenland, 3–7% in Canada, and 3% in Alaska. The prevalence in North America and Europe varies between 1/500 and 1/2000 [23]. Acquired forms of sucrase-isomaltase deficiency may be secondary to other chronic gastrointestinal conditions associated with intestinal villous atrophy, such as enteric infection, celiac disease, Crohn’s disease, and other enteropathies affecting the small intestine. The symptoms usually appear in childhood and do not manifest until sucrose is included in the diet, which usually results from the introduction of fruit in the diet. It can also manifest at birth if a child is fed with a milk formula containing sucrose. In some individuals, it appears in adulthood with symptoms suggestive of IBS. In these cases, a careful anamnesis of the patient and their parents usually reveals a lifelong history of intestinal symptoms. Symptoms in childhood and in adulthood are similar, but the consequences are more serious in children. Severe watery diarrhoea may occur after the ingestion of small amounts of sucrose, which can be accompanied by abdominal pain, bloating, growth retardation, and the rejection of sugary foods. Isomaltose as such is not consumed in the diet. However, this oligosaccharide is released in the hydrolysis of starch. Although some patients may have mild symptoms after the ingestion of starch, most tolerate it well, mainly due to the low osmotic power of the molecule of undigested isomaltose. Recent studies on patients with clinical symptoms suggestive of IBS, abdominal pain, diarrhoea, or bloating have shown the presence of sucrase-isomaltase deficiency after assessing disaccharidase activity in small bowel samples. Functional sucrase-isomaltose genetic variants appear to be more common in patients with symptoms suggestive of IBS than in controls and can lead to similar symptoms of maldigestion as those described in classical CSID [24,25,26]. This has the potential to identify groups among patients with IBS for individualized management. It seems to be a heterogeneous disorder, the severity of which is likely related to the biochemical phenotypes of the sucrase-isomaltase mutants, as well as the environment and diet of patients [27]. These studies suggest that the screening for sucrase-isomaltase mutations in patients with IBS may prove helpful when considering either a restrictive diet or enzyme replacement therapy as an appropriate treatment. However, genetic testing for sucrase-isomaltase mutations in individuals presenting with IBS-like symptoms is not routinely done in practice since it can be extremely costly. A confirmatory CSID diagnosis can be performed by a disaccharidase assay using a small bowel tissue biopsy, genetic testing, or sucrose breath testing. The disaccharidase assay shows the levels of various enzymes such as sucrase-isomaltase, lactase, and maltase. Results are consistent with CSID if the amount of sucrose broken down by sucrase-isomaltase is lower than expected [28]. There are two available sucrose breath tests. The carbon-13 (13C) sucrose breath test involves the challenge with 13C-sucrose, and it is regarded to be the most direct and definitive measure for detecting CSID, with a sensitivity and specificity of 100% [29]. Additionally, the sucrose HBT may also be useful in diagnosing CSID [30], though many paediatric patients experience severe symptoms, passage of watery stools, bloating abdomen, and cramps from the 2 g/kg sucrose load. In contrast, this symptomatic response is not observed with the 13C-sucrose test because the load of sucrose ingested is only 0.02 g [29]. Genetic testing allows an unequivocal diagnosis in children with CSID. At least 80% of patients with CSID have one of four common mutations [31]. There are no comparative studies of these tests in adults with IBS-like symptoms who are carriers of functional sucrase-isomaltose genetic variants. A British Society of Gastroenterology review on the management of IBS stated that, at present, there is not enough evidence to consider routine testing for sucrase-isomaltase deficiency [32]. Most patients require a dietetic manipulation that, in general, must be stricter in childhood than in adult life. The degree of sucrose restriction necessary is different for each patient, who, by a trial-and-error method, becomes an expert in manipulating the diet to be symptom-free. It has been suggested that the use of sacrosidase (Sucraid®), an enzyme produced by Saccharomyces cerevisiae that hydrolyses sucrose, is effective as a treatment for sucrase-isomaltase deficiency. Double-blind studies have revealed that this enzyme, administered along with food, significantly prevents symptoms of intolerance in patients on a sucrose-containing diet as compared with placebo [33]. 4. Fructose Malabsorption Fructose is commonly obtained from sugar beets, sugar cane, and maize and is the sweetest of all natural sugars [34]. Fructose can be present as a monosaccharide or as disaccharide sucrose in a one-to-one molecular ratio with glucose. In recent decades, the solubility and sweetness of fructose have been exploited by the food industry in the form of sweetener blends derived from the enzymatic conversion of starches, with the increasing popularity of the use of high-fructose corn syrup (HFCS), a mixture of glucose and fructose in monosaccharide form, which may be of concern if it contains more fructose than glucose. In addition, use of fructose enhances the flavour and physical appeal of many foods and beverages. Thus, fructose is used in place of sucrose and other carbohydrates to reduce the caloric content of dietetic products while conserving high-quality sweetening profiles [35]. Between 1970 and 2004, the share of HFCS as a percentage of total sweetener used in the United States increased from half a percentage point to 42% [36]. The product is found in many beverages, including nearly all non-diet soda brands, as well as breakfast cereals, salad dressings, cheese spreads, yogurts, jams, peanut butter, and other foods. HFCS is an American definition; in Europe, the term isoglucose or glucose–fructose syrup refers to a liquid sweetener composed of mainly glucose and fructose in varying compositions, which has a 20–30% fructose content compared to 42% (HFCS 42) and 55% (HFCS 55) in the United States [37]. The use of glucose–fructose syrup in soft drinks is limited in the European Union because manufacturers do not have a sufficient supply containing at least 42% fructose. As a result, soft drinks are primarily sweetened with sucrose, and the use of isoglucose as a replacement for sucrose in foods and beverages is not as widespread in Europe as in the United States. Data from Europe showed that daily free-fructose intake in adults was a mean of 17 g/d in Finland in 2007 [38], and a median of 15 g/d in the Netherlands in 2015 [39], far less than the 55 g/d reported for US adults in 2008 [40]. Fructose is mainly absorbed in the proximal small intestine. The absorption of monosaccharides is mainly mediated by the Na+-glucose cotransporter SGLT1 and the facilitative transporters GLUT2 and GLUT5 (reviewed in 1) (Figure 2). In brief, SGLT1 and GLUT2 are important for the absorption of glucose and galactose, while the GLUT5 transporter is related to fructose absorption. SGLT1 and GLUT5 are continuously localized in the apical brush-border membrane of enterocytes, whereas GLUT2 is localized in the basolateral membrane at low luminal glucose concentrations or in both the apical brush-border membrane and the basolateral membrane at high luminal glucose concentrations. Fructose monosaccharide is transported across the apical intestinal brush-border membrane via a Na+-independent facilitated diffusion mechanism via the GLUT5 transporter. GLUT5 expression is induced by fructose, which exerts a fast and strong upregulation of GLUT5 mRNA expression, leading to an increase in GLUT5 protein and activity levels. Approximately 90% of the fructose entering the enterocyte is metabolized, increasing the intracellular pool of glucose. The remaining fructose subsequently exits the enterocytes to enter the blood via the GLUT2 and GLUT5 transporters present at the basolateral membrane. GLUT2 may be recruited transiently in the apical brush-border membrane in response to high luminal glucose concentrations to assist in the absorption of excess luminal fructose. GLUT2 is a high-capacity, low-affinity glucose/galactose transporter that can co-transport fructose in a one-to-one ratio. GLUT2 is unable to transport fructose without the presence of glucose, although the mechanism for this is currently unknown. Fructose malabsorption may be due to the insufficient uptake of fructose into enterocytes relative to the amount of luminal fructose. In addition, it may be caused by insufficient intracellular digestion of fructose, resulting in high intracellular fructose concentrations, which may contribute to decreased fructose uptake. In humans, fructose absorption capacity in the small intestine is much lower than glucose absorption capacity; it is very small after birth and increases later in response to dietary fructose. However, in combination with glucose, the capacity for fructose absorption increases due to the additional fructose uptake associated with Na+-glucose cotransport. Thus, fructose is well-absorbed in the presence of equimolar amounts of glucose in the proximal small intestine, while free fructose is absorbed slowly along the small intestine. Therefore, glucose co-ingestion significantly increases fructose absorption: glucose stimulates fructose absorption in a dose-dependent manner, and malabsorption occurs when fructose is present in excess of glucose. Unabsorbed fructose then passes into the colon and is fermented in the same manner as lactose in patients who have lactase non-persistence. HBTs have been used to assess fructose malabsorption. However, there is insufficient information regarding the frequency of incomplete absorption of fructose in the healthy population. There are large individual variations in the ability to absorb fructose in healthy adults. Various studies have shown that the absorption capacity of fructose as a monosaccharide in healthy subjects ranges from less than 5 g to more than 50 g [41]. Thirty-seven to ninety percent of healthy subjects present malabsorption of a 50 g fructose load [41,42,43]. In addition, not only the dose, but also the solution concentration, affects fructose absorption. Incomplete absorption after a 50-g fructose load was observed in 37.5% (10% solution) and 71% (20% solution) of healthy people [42]. In a double-blind study in healthy subjects, increasing doses of fructose, 15 g, 25 g, and 50 g (at 10% solutions) were malabsorbed by 0%, 10%, and 80% of subjects. After a 50 g fructose load, 55% of subjects experienced symptoms of intolerance, which did not occur at lower doses [43]. In light of these results, it was claimed that healthy subjects have the capacity to absorb up to 25 g fructose. However, in a review of six studies, it was shown that 10% to 55% of healthy subjects did not absorb a 25 g fructose load, with either no symptoms or mild symptoms [44]. In addition, the ingestion of fructose as sucrose (50 to 100 g) did not result in appreciable malabsorption in these cases [41,42,45]. Neither is it clear whether fructose administration alone in fructose HBT reasonably reflects the conditions under which most dietary fructose is ingested. When fructose is ingested as part of a solid food, which delays gastric emptying and small intestinal transit, it does not reach the colon for at least 180 min; but when it is administered as an aqueous solution, it undergoes a fast and exponential gastric emptying, passing rapidly through the small intestine to reach the colon within 30 min [46]. In addition, the concomitant intake of sorbitol interferes with fructose absorption, whereas the concomitant ingestion of glucose in food enhances fructose absorption (via GLUT2). The European guidelines on the indications, performance, and clinical impact of hydrogen and methane breath tests recommend that the dose of fructose in adults for the diagnosis of fructose malabsorption and intolerance should be 20–25 g [15]. However, the clinical utility of fructose HBT is debated. In fact, the Rome Consensus Conference on ‘Methodology and indications of H2-breath testing in gastrointestinal diseases’ stated that a fructose breath test is not recommended in clinical practice [14]. Moreover, the British Society of Gastroenterology guidelines for the investigation of chronic diarrhoea stated that, at present, fructose breath testing cannot be said to inform the diagnosis and treatment of fructose intolerance [47]. In a randomized control trial, it was observed that, despite the challenge of a relatively high dose of fructose (50 g) administered to reduce the risk for false-negative breath tests, 64.5% of patients with a normal fructose breath test improved with a fructose-restricted diet [48]. Thus, it is not clear that fructose HBT is an informative test regarding the response to a fructose-restricted diet since patients may improve on the diet despite a normal test and vice versa. Testing for fructose intolerance may replace testing for fructose malabsorption in the future as a way to obtain relevant clinical information. In another prospective randomized trial, patients with IBS were randomized for 12 weeks with/without a fructose-reduced diet in addition to a standard IBS diet [49]. In this study, the criteria for the diagnosis of fructose intolerance were based on improvement of self-reported symptoms while on a fructose-restricted diet and exacerbation of symptoms following a fructose challenge test. Using valid outcome measures to assess the fulfilment of these criteria, 56% of patients with IBS were considered to have fructose intolerance. Thus, there is the need to use standardized, validated symptom scales to obtain a reliable assessment of intestinal symptoms. In this sense, the ‘Carbohydrate Perception Questionnaire’, in both the paediatric and adult versions, seems to be a valid instrument for the assessment of symptoms developed after carbohydrate ingestion, with excellent psychometric properties [50]. The objective of a fructose-restricted diet is to restrict the intake of foods rich in fructose to a level that does not trigger intestinal symptoms. As mentioned, the main determinants of fructose malabsorption are the amount of fructose in excess of glucose and the intake of foods containing both fructose and sorbitol, since sorbitol interferes with fructose absorption. Patients are recommended to eliminate all foods with either an excess of free fructose or sorbitol (and other polyols) from the diet [19]. Xylose isomerase has been proposed as a potential treatment for fructose intolerance. The ability of xylose isomerase to convert between glucose and fructose has led to the suggestion of its use as a treatment for fructose intolerance. A double-blind, placebo-controlled study showed that oral administration of xylose isomerase was associated with a significant reduction in breath hydrogen after fructose ingestion, as well as a significant improvement in nausea and abdominal pain [51]. Further research is needed to assess the long-term health effects of xylose isomerase and to determine which patients may be most suitable for treatment. 5. Sorbitol Intolerance Sugar alcohols, a class of low-molecular weight polyols, can occur naturally or be obtained by the hydrogenation of sugars. The most common are sorbitol, mannitol, maltitol, isomalt, lactitol, and xylitol. Sorbitol (D-glucitol) is the most frequently consumed sugar alcohol. Small amounts of sorbitol are present in some fruits of the Rosaceae family (apples, pears, cherries, apricots, peaches, and prunes). Dietary intake in the UK National Diet and Nutrition Survey data showed that the average daily intake of polyols was 3.5 g, with the 95th percentile at 10.4 g [52]. Most sorbitol intake comes from added sources [53]. Sorbitol contents in certain sugar-free sweet foods (e.g., sugar-free chewing gum, candy, mints, jam, diet drinks, chocolate) may be considerable. Sorbitol is also used as an additive for purposes other than sweetening in foods because of its unique combination of functional properties, including its role as a humectant, thickener, stabilizer, plasticizer, and emulsifier. Sorbitol is poorly absorbed by the small intestine. It is well-known that, at high doses and concentrations, sorbitol is a laxative. Test solutions containing 10 g and 20 g resulted in 90% and 100%, respectively, of healthy volunteers showing malabsorption [54]. After a 5 g dose administered at concentrations of 2%, 4%, 8%, and 16%, malabsorption was manifested in 10%, 12%, 22%, and 43% of healthy volunteers [54]. Simultaneous ingestion of sorbitol and fructose seems to increase the malabsorption of the latter [55,56]. Thus, polyol restriction has been incorporated as a part of a FODMAP (acronym for Fermentable, Oligo-Di- and Monosaccharides and Polyols) diet [57] (see below). It should be emphasized that, as with fructose HBT, sorbitol HBT should not be recommended in clinical practice for either adults or children [14]. 6. Sugar Malabsorption and Functional Bowel Disease Different studies have shown that there are no differences in the frequency of sugar malabsorption between patients with IBS and healthy controls, although the severity of symptoms after a sugar challenge is greater in patients than in controls [10,45]. In a single-blind randomized controlled study in patients with diarrhoea-predominant IBS and healthy subjects [45], sugar malabsorption was assessed by HBT after an oral load of various solutions containing lactose (50 g), fructose (25 g), sorbitol (5 g), fructose plus sorbitol (25 + 5 g), and sucrose (50 g). The frequency of sugar malabsorption was high in both patients and healthy controls, with malabsorption of at least one sugar solution in more than 90% of the subjects, but all subjects absorbed the sucrose solution. However, the symptoms score after both lactose and fructose plus sorbitol malabsorption was significantly higher in patients than in control subjects. In addition, more severe symptoms were observed in the IBS-D group after both lactose and fructose–sorbitol malabsorption than after the sucrose load administered as a control solution. Significantly more symptoms, although of mild intensity, were also observed after the sucrose load in patients with IBS than in healthy subjects. Finally, the administration of 10 g lactulose, a nonabsorbable carbohydrate, induced more symptoms and H2 production in patients with IBS than in healthy controls. In this study, 40–50% of both patients with IBS and healthy controls malabsorbed the 25 g fructose load, and the severity of symptoms was no different between patients and controls. In another double-blind randomized study [10], patients with IBS-D and controls were given HBT to detect malabsorption and intolerance following the administration of 10, 20, and 40 g lactose. All participants had the lactase genotype C/C-13910, which is associated with lactase non-persistence. Malabsorption of 40 g lactose was observed in 93% of controls and 92% of IBS-D patients. Fewer controls than patients with IBS-D were intolerant to 10 g lactose, 20 g lactose, and 40 g lactose. The frequency of positive tests for lactose malabsorption and intolerance in both controls and IBS-D patients increased with the lactose dose, and the breath hydrogen excretion (peak and AUC H2 excretion) was associated with the severity of abdominal symptoms. Therefore, the frequency of sugar intolerance after malabsorption seems to be higher in patients with IBS-D than in controls. The presence of a functional bowel disease increased the likelihood that an individual would report abdominal pain, bloating, and diarrhoea. In fact, symptoms experienced during breath testing, but not malabsorption, seem to correlate with previous functional bowel disease symptoms [58]. In that study, adequate symptom relief with dietary adaptation was achieved in >80% of intolerant patients, irrespective of malabsorption, which is a response rate similar to that in comparable studies [19,59,60], leading the way to a more comprehensive diet that reduces all sources of poorly digested, rapidly fermentable carbohydrates. 7. Carbohydrate-Reduced Diets in Functional Bowel Disease More recently, there has been renewed interest in evaluating the role of FODMAPs in patients with IBS and carbohydrate-related symptoms [57]. FODMAPs are fermentable short-chain carbohydrates found in a variety of fruits, vegetables, pulses, dairy products, artificial sweeteners, and wheat. Evidence of a relationship between dietary FODMAPs and intestinal symptoms comes from a double-blind, cross-over study challenging patients with IBS with increasing doses of either glucose, fructose, fructans, or a mixture of the latter two for up to two weeks [61]. Fructose or fructans were significantly more likely than glucose to induce symptom recurrence. A further cross-over study found that patients developed fewer symptoms after a low-FODMAP diet compared with a typical Australian diet [62]. Afterwards, a randomized controlled trial found no differences between a low-FODMAP diet and an empiric IBS diet (NICE guidelines) based on healthy eating patterns, low fat content, and the avoidance of high-fibre food and resistant starch [63]. Current British Society of Gastroenterology guidelines on the management of IBS recommend that a diet low in FODMAPs, which is considered to be a second-line dietary therapy, is an effective treatment for global symptoms and abdominal pain in IBS, although its implementation should be supervised by a trained dietitian, and fermentable oligosaccharides, disaccharides and monosaccharides, and polyols should be reintroduced according to tolerance (recommendation: weak, quality of evidence very low) [32]. The low-FODMAP diet is designed as a three-phase approach, whereby patients restrict all FODMAP subgroups for a 6–8 week period, followed by a reintroduction phase for the re-challenge of subgroups to test tolerance, followed by a personalised long-term maintenance phase with periodic re-challenge of poorly tolerated foods [64], in a ‘top-down’ approach. In addition, a ‘bottom-up’ approach has been suggested, starting with the restriction of a few specific foods or FODMAP subgroups based on baseline diet history and patient-reported triggers [46] (Figure 3). This ‘bottom-up’ approach has been promoted to avoid prolonged dietary restrictions on a low-FODMAP diet, potentially preventing the disruption of the gut microbiota and micronutrient status. In addition, it has been suggested that a gluten-free diet may be the easiest way of achieving fructan reduction [65], since fructans are a key component to be reduced in a long-term adapted low-FODMAP diet, as demonstrated in a prospective study [66]. In this sense, it has been proposed that a gluten-free diet may be administered as a ‘bottom-up’ approach in the FODMAP diet for patients with functional bowel disease [67,68,69]. In addition, patients have regarded a gluten-free diet as more acceptable than a low-FODMAP diet [70]. In fact, only 40% of patients show a good adherence to the low-FODMAP diet [71]. Twenty to fifty percent of IBS patients do not experience a reduction in gastrointestinal symptoms when following NICE guidelines and/or a low-FODMAP diet [64]. Other dietary approaches have been suggested. In a recent study that randomized patients with IBS to a 4-week starch- and sucrose-reduced diet or to a habitual diet (control group) [72], the intervention group presented a significantly lower total IBS-SSS, ‘abdominal pain’, ‘bloating/flatulence’, and ‘intestinal symptoms influence on daily life’ scores compared to controls, and a 74% response rate. Although promising, this dietary treatment needs to be further evaluated and compared to establish dietary treatments before it can be routinely used in a clinical setting. 8. Mechanisms of Carbohydrate-Reduced Diet Improvement in Functional Bowel Disease The mechanism by which dietary changes may affect symptoms in IBS patients was explored in a cross-over trial in which patients with IBS and bloating were recruited alongside a parallel cohort of healthy volunteers without functional gastrointestinal symptoms who followed the same trial regimen [73]. Subjects were given 40 g of carbohydrate (glucose, fructose, and inulin in random order) in a 500-mL solution. Levels of breath hydrogen were measured, and intestinal content was assessed by MRI before and at various time points after the consumption of each drink. IBS patients and healthy subjects had similar physiological responses following fructose or inulin ingestion. These results indicate that colonic hypersensitivity to distension, rather than excessive gas production, produces the carbohydrate-related symptoms in IBS patients. Zhu et al. [74] reported on lactose responsiveness in a Chinese population with a high prevalence of lactose maldigestion. They included IBS patients and healthy controls who underwent a 20 g lactose HBT, with assessments of hydrogen gas production and lactose intolerance symptoms. Lactose intolerance was more frequent in IBS than in healthy controls, especially bloating and borborygmus. Rectal hypersensitivity assessed by barostat was associated with a higher odds ratio of bloating than hydrogen production, suggesting that visceral hypersensitivity plays an important role in carbohydrate intolerance in IBS. Therefore, osmotically active unabsorbed monosaccharides and disaccharides distend the small bowel with fluid and, subsequently, the colon, where they produce a gas increase and, in those subjects with visceral hypersensitivity, induce more severe gastrointestinal symptoms. Furthermore, the rapid colonic fermentation of unabsorbed carbohydrates generates gas and produces short-chain fatty acids, which lower colonic pH and trigger bowel symptoms [75]. When delivered as liquid drinks, they speed gastric emptying, and the increase in small-bowel water content also accelerates intestinal transit, reducing small-bowel absorption, which may make symptoms more severe. Evidence that there are differences in visceral hypersensitivity in subsets of IBS patients suggests that the same magnitude of stimulus will produce different degrees of symptom response in patients depending on their sensory threshold [32,76]. In the case of carbohydrate malabsorbers without IBS, symptom generation may be mainly triggered by rapid colonic fermentation. A role for barrier dysfunction and inflammation produced by a high-FODMAP diet has been proposed. In animal models, there is an association between fructose or high-FODMAP intake, increased intestinal permeability, barrier dysfunction, and inflammation [77,78]. Thus, changing carbohydrate intake may induce barrier dysfunction, which may, at least in part, be driven by the colonic microbiota. In addition, in humans, changes in histamine levels suggest that increased FODMAP intake may influence immune activation [79]. Moreover, reductions in pro-inflammatory cytokines following a low-FODMAP diet support the immune modulation hypothesis [80]. 9. Final Outlook The development of a well-accepted, practical, and cost-effective carbohydrate intolerance test capable of predicting the outcome of dietary management is one of the major clinical challenges in the field of functional bowel disease. In this regard, more studies are needed to explore the role of fructose intolerance testing, with or without concomitant measures of H2 and CH4 excretion, using standardized and validated symptom scales. The role of screening for sucrase-isomaltase mutations for guiding the development of a rational treatment with sucrose-reduced diets in a subgroup of IBS patients appears promising, but also requires further study. The ‘bottom-up’ approach of a low-FODMAP diet should also be further evaluated. It is likely that many patients may benefit from this approach, thereby avoiding a long-term, highly carbohydrate-restrictive diet and preventing an alteration of gut microbiota and nutritional status. Two approaches have been suggested in this regard, starting with only some subgroups of the low-FODMAP diet based on dietary history or with a gluten-free diet. Further studies should provide high-quality evidence to document the clinical response and the long-term effects of these strategies. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The author declares no conflict of interest. Figure 1 Short-chain fatty acid and gas (H2, CO2, CH4) production by the colonic microbiota fermentation of dietary unabsorbed carbohydrates. Thirty to fifty grams of soluble dietary fibre, resistant starch, oligosaccharides (fructo-oligosaccharides, galacto-oligosaccharides, raffinose, and stachyose), disaccharides (lactose, sucrose), and monosaccharides (fructose) enter the colon each day and become available for colonic fermentation by the microbiota. Acetate (C2), propionate (C3), and butyrate (C4) are the main short-chain fatty acids that play important roles in gastrointestinal function. Figure 2 Enterocyte monosaccharide transporters involved in D-glucose, D-galactose, and D-fructose absorption in the small intestine. GLUT2, which is only observed in the apical brush-border membrane at high D-glucose concentrations in intestinal lumen, is indicated in red (see reference [1]). Red dots between enterocytes represent the tight junctions. Figure 3 ‘Top-down’ and ‘bottom-up’ approach to the low-FODMAP diet (adapted with permission from reference [46]; Copyright 2019, John Wiley & Sons, Inc). Reducing a few specific foods/subgroups in the ‘bottom-up’ approach implies an adequate diet history, for example, that in a patient who consumes chewing gum and/or candies, these are withdrawn; if he/she consumes large quantities of fruits and fruit juices, they may benefit from the restriction of only excess fructose and polyols to start with; or if he/she consumes large amounts of wheat, onion, artichokes, and pulses, they may be more likely to benefit from a restriction of fructans. Furthermore, in the ‘bottom-up’ approach, a gluten-free diet has been claimed to be the easiest way to reduce fructan intake (see text). nutrients-14-01923-t001_Table 1 Table 1 Absorption mechanisms of dietary carbohydrates and specific drugs available to facilitate digestion. Carbohydrate Type Absorption Mechanisms Available Specific Drugs Monosaccharides Fructose Absorption of excess fructose occurs in the small intestine: rapidly via GLUT-2, the sodium-dependent active transport mechanism in conjunction with glucose; slowly via GLUT-5, a specific transporter for fructose using carrier-mediated facilitated diffusion. Thus, fructose is well-absorbed in the presence of equimolar glucose in the proximal small intestine, whereas free fructose is absorbed slowly along the length of the small intestine. Xylose-isomerase Disaccharides Lactose Sucrose Unabsorbed in small intestine if lactase is absent. Unabsorbed in small intestine in case of sucrase-isomaltase deficiency Lactase Sacrosidase Oligosaccharides FOS * GOS Humans do not possess small intestinal hydrolases to hydrolyse oligosaccharides, and they are unabsorbed. None Polyols Sorbitol Mannitol Maltitol Isomalt Lactitol Xylitol Sugar alcohols are poorly absorbed along the length of the small intestine by slow passive diffusion. None * FOS: Fructo-oligosaccharides; GOS: Galacto-oligosaccharides. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Koepsell H. Glucose transporters in the small intestine in health and disease Pflug. Arch. Eur. J. Physiol. 2020 472 1207 1248 10.1007/s00424-020-02439-5 32829466 2. Christopher N.L. Bayless T.M. Role of the Small Bowel and Colon in Lactose-Induced Diarrhea Gastroenterology 1971 60 845 852 10.1016/S0016-5085(71)80083-5 5581327 3. He T. Venema K. Priebe M.G. Welling G.W. Brummer R.J. Vonk R.J. The role of colonic metabolism in lactose intolerance Eur. J. Clin. Investig. 2008 38 541 547 10.1111/j.1365-2362.2008.01966.x 18573099 4. Caspary W.F. Diarrhoea associated with carbohydrate malabsorption Clin. Gastroenterol. 1986 15 631 655 10.1016/S0300-5089(21)00742-2 3742842 5. Misselwitz B. Butter M. Verbeke K. Fox M.R. Update on lactose malabsorption and intolerance: Pathogenesis, diagnosis and clinical management Gut 2019 68 2080 2091 10.1136/gutjnl-2019-318404 31427404 6. Jansson-Knodell C.L. Krajicek E.J. Savaiano D.A. Shin A.S. Lactose Intolerance: A Concise Review to Skim the Surface Mayo Clin. Proc. 2020 95 1499 1505 10.1016/j.mayocp.2020.04.036 32622451 7. Enattah N.S. Sahi T. Savilahti E. Terwilliger J.D. Peltonen L. Järvelä I. Identification of a variant associated with adult-type hypolactasia Nat. Genet. 2002 30 233 237 10.1038/ng826 11788828 8. Storhaug C.L. Fosse S.K. Fadnes L.T. Country, regional, and global estimates for lactose malabsorption in adults: A systematic review and meta-analysis Lancet Gastroenterol. Hepatol. 2017 2 738 746 10.1016/S2468-1253(17)30154-1 28690131 9. Fernández-Bañares F. Reliability of symptom analysis during carbohydrate hydrogen-breath tests Curr. Opin. Clin. Nutr. Metab. Care 2012 15 494 498 10.1097/MCO.0b013e328356689a 22878243 10. Yang J. Deng Y. Chu H. Cong Y. Zhao J. Pohl D. Misselwitz B. Fried M. Dai N. Fox M. Prevalence and presentation of lactose intolerance and effects on dairy product intake in healthy subjects and patients with irritable bowel syndrome Clin. Gastroenterol. Hepatol. 2013 11 262 268 10.1016/j.cgh.2012.11.034 23246646 11. He T. Priebe M.G. Harmsen H.J. Stellaard F. Sun X. Welling G.W. Vonk R.J. Colonic fermentation may play a role in lactose intolerance in humans J. Nutr. 2006 136 58 63 10.1093/jn/136.1.58 16365059 12. Binder H.J. Role of colonic short-chain fatty acid transport in diarrhea Annu. Rev. Physiol. 2010 72 297 313 10.1146/annurev-physiol-021909-135817 20148677 13. Gibson G.R. Cummings J.H. Macfarlane G.T. Allison C. Segal I. Vorster H.H. Walker A.R. Alternative pathways for hydrogen disposal during fermentation in the human colon Gut 1990 31 679 683 10.1136/gut.31.6.679 2379871 14. Parodi A. Colecchia A. Orsola-Malpighi P.S. Festi D. Perri F. Methodology and indications of H2-breath testing in gastrointestinal diseases: The Rome Consensus Conference Aliment. Pharmacol. Ther. 2009 29 1 49 19344474 15. Hammer H.F. Fox M.R. Keller J. Salvatore S. Basilisco G. Hammer J. Lopetuso L. Benninga M. Borrelli O. Dumitrascu D. European guideline on indications, performance, and clinical impact of hydrogen and methane breath tests in adult and pediatric patients: European Association for Gastroenterology, Endoscopy and Nutrition, European Society of Neurogastroenterology and Motility, and European Society for Paediatric Gastroenterology Hepatology and Nutrition consensus United Eur. Gastroenterol. J. 2022 10 15 40 16. Rezaie A. Buresi M. Lembo A. Lin H. McCallum R. Rao S. Schmulson M. Valdovinos M. Zakko S. Pimentel M. Hydrogen and methane-based breath testing in gastrointestinal disorders: The North American Consensus Am. J. Gastroenterol. 2017 112 775 784 10.1038/ajg.2017.46 28323273 17. Jellema P. Schellevis F.G. Van Der Windt D.A.W.M. Kneepkens C.M.F. Van Der Horst H.E. Lactose malabsorption and intolerance: A systematic review on the diagnostic value of gastrointestinal symptoms and self-reported milk intolerance QJM Med. 2010 103 555 572 10.1093/qjmed/hcq082 18. Vernia P. Ricciardi M.R. Frandina C. Bilotta T. Frieri G. Lactose malabsorption and irritable bowel syndrome. Effect of a long-term lactose-free diet Ital. J. Gastroenterol. 1995 27 117 121 7548919 19. Fernández-Bañares F. Rosinach M. Esteve M. Forné M. Espinós J.C. Viver J.M. Sugar malabsorption in functional abdominal bloating: A pilot study on the long-term effect of dietary treatment Clin. Nutr. 2006 25 824 831 10.1016/j.clnu.2005.11.010 16410032 20. Alkalay M.J. Nutrition in Patients with Lactose Malabsorption, Celiac Disease, and Related Disorders Nutrients 2021 14 2 10.3390/nu14010002 35010876 21. Shaukat A. Levitt M.D. Taylor B.C. MacDonald R. Shamliyan T.A. Kane R.L. Wilt T.J. Systematic review: Effective management strategies for lactose intolerance Ann. Intern. Med. 2010 152 797 803 10.7326/0003-4819-152-12-201006150-00241 20404262 22. Ohlsson B. Theories behind the effect of starch- and sucrose-reduced diets on gastrointestinal symptoms in irritable bowel syndrome (Review) Mol. Med. Rep. 2021 24 732 10.3892/mmr.2021.12372 34414452 23. Treem W.R. Clinical aspects and treatment of congenital sucrase-isomaltase deficiency J. Pediatr. Gastroenterol. Nutr. 2012 55 7 13 10.1097/01.mpg.0000421401.57633.90 23103658 24. Kim S.B. Calmet F.H. Garrido J. Garcia-Buitrago M.T. Moshiree B. Sucrase-isomaltase deficiency as a potential masquerade in irritable bowel syndrome Dig. Dis. Sci. 2000 65 534 540 10.1007/s10620-019-05780-7 25. Henström M. Diekmann L. Bonfiglio F. Hadizadeh F. Kuech E.M. von Köckritz-Blickwede M. Thingholm L.B. Zheng T. Assadi G. Dierks C. Functional variants in the sucrase-isomaltase gene associate with increased risk of irritable bowel syndrome Gut 2018 67 263 270 10.1136/gutjnl-2016-312456 27872184 26. Garcia-Etxebarria K. Zheng T. Bonfiglio F. Bujanda L. Dlugosz A. Lindberg G. Schmidt P.T. Karling P. Ohlsson B. Simren M. Increased Prevalence of Rare Sucrase-isomaltase Pathogenic Variants in Irritable Bowel Syndrome Patients Clin. Gastroenterol. Hepatol. 2018 16 1673 1676 10.1016/j.cgh.2018.01.047 29408290 27. Husein D.M. Rizk S. Naim H.Y. Differential Effects of Sucrase-Isomaltase Mutants on Its Trafficking and Function in Irritable Bowel Syndrome: Similarities to Congenital Sucrase-Isomaltase Deficiency Nutrients 2020 13 9 10.3390/nu13010009 28. Burke M. Carbohydrate Intolerance and Disaccharidase Measurement—A Mini-Review Clin. Biochem. Rev. 2019 40 167 174 10.33176/AACB-19-00025 31857738 29. Robayo-Torres C.C. Opekun A.R. Quezada-Calvillo R. Xavier V. Smith E.B. Navarrete M. Baker S.S. Nichols B.L. 13C-breath tests for sucrose digestion in congenital sucrase isomaltase-deficient and sacrosidase-supplemented patients J. Pediatr. Gastroenterol. Nutr. 2009 48 412 418 10.1097/MPG.0b013e318180cd09 19330928 30. Davidson G.P. Robb T.A. Value of breath hydrogen analysis in management of diarrheal illness in childhood: Comparison with duodenal biopsy J. Pediatr. Gastroenterol. Nutr. 1985 4 381 387 10.1097/00005176-198506000-00010 4020571 31. Uhrich S. Wu Z. Huang J.-Y. Scott C.R. Four mutations in the SI gene are responsible for the majority of clinical symptoms of CSID J. Pediatr. Gastroenterol. Nutr. 2012 55 S34 S35 10.1097/01.mpg.0000421408.65257.b5 23103650 32. Vasant D.H. Paine P.A. Black C.J. Houghton L.A. Everitt H.A. Corsetti M. Agrawal A. Aziz I. Farmer A.D. Eugenicos M.P. British Society of Gastroenterology guidelines on the management of irritable bowel syndrome Gut 2021 70 1214 1240 10.1136/gutjnl-2021-324598 33903147 33. Treem W.R. McAdams L. Stanford L. Kastoff G. Justinich C. Hyams J. Sacrosidase therapy for congenital sucrase-isomaltase deficiency J. Pediatr. Gastroenterol. Nutr. 1999 28 137 142 10.1097/00005176-199902000-00008 9932843 34. Moskowitz H.R. Ratio scales of sugar sweetness Percept. Psychophys. 1970 7 315 320 10.3758/BF03210175 35. Hanover L.M. White J.S. Manufacturing, composition, and applications of fructose Am. J. Clin. Nutr. 1993 58 724S 732S 10.1093/ajcn/58.5.724S 8213603 36. Duffey K.J. Popkin B.M. High-fructose corn syrup: Is this what’s for dinner? Am. J. Clin. Nutr. 2008 88 1722S 1732S 10.3945/ajcn.2008.25825C 19064537 37. Glucose-Fructose Syrup: An Ingredient Worth Knowing Starch EU: Brussels, Belgium, June 2017 Available online: https://www.starch.eu/wp-content/uploads/2017/06/5247.055_starch_eu-fiche-glucose-fructose-webC-1.pdf (accessed on 8 March 2022) 38. Montonen J. Järvinen R. Knekt P. Heliovaara M. Reunanen A. Consumption of sweetened beverages and intakes of fructose and glucose predict type 2 diabetes occurrence J. Nutr. 2007 137 1447 1454 10.1093/jn/137.6.1447 17513405 39. Sluik D. Engelen A.I. Feskens E.J. Fructose consumption in the Netherlands: The Dutch National Food Consumption Survey 2007–2010 Eur. J. Clin. Nutr. 2015 69 475 481 10.1038/ejcn.2014.267 25514900 40. Vos M.B. Kimmons J.E. Dietary fructose consumption among US children and adults: The third National Health and Nutrition Examination Survey Medscape J. Med. 2008 10 160 18769702 41. Rumessen J.J. Gudmand-Hoyer E. Absorption capacity of fructose in healthy adults. Comparison with sucrose and its constituent monosaccharides Gut 1986 27 1161 1168 10.1136/gut.27.10.1161 3781328 42. Ravich W.J. Bayless T.M. Thomas M. Fructose: Incomplete intestinal absorption in humans Gastroenterology 1983 84 26 29 10.1016/S0016-5085(83)80162-0 6847852 43. Rao S.S. Attaluri A. Anderson L. Stumbo P. Ability of the normal human small intestine to absorb fructose: Evaluation by breath testing Clin. Gastroenterol. Hepatol. 2007 5 959 963 10.1016/j.cgh.2007.04.008 17625977 44. Jones H.F. Butler R.N. Brooks D.A. Intestinal fructose transport and malabsorption in humans Am. J. Physiol. Gastrointest. Liver Physiol. 2011 300 G202 G206 10.1152/ajpgi.00457.2010 21148401 45. Fernandez-Banares F. Esteve-Pardo M. De Leon R. Humbert P. Cabre E. Llovet J.M. Gassull M.A. Sugar malabsorption in functional bowel disease: Clinical implications Am. J. Gastroenterol. 1993 88 2044 2050 8249972 46. Wang X.J. Camilleri M. Vanner S. Tuck C. Review article: Biological mechanisms for symptom causation by individual FODMAP subgroups-the case for a more personalised approach to dietary restriction Aliment. Pharmacol. Ther. 2019 50 517 529 10.1111/apt.15419 31309595 47. Arasaradnam R.P. Brown S. Forbes A. Fox M.R. Hungin P. Kelman L. Major G. O’Connor M. Sanders D.S. Sinha R. Guidelines for the investigation of chronic diarrhoea in adults: British Society of Gastroenterology, 3rd edition Gut 2018 67 1380 1399 10.1136/gutjnl-2017-315909 29653941 48. Berg L.K. Fagerli E. Martinussen M. Myhre A.O. Florholmen J. Goll R. Effect of fructose-reduced diet in patients with irritable bowel syndrome, and its correlation to a standard fructose breath test Scand. J. Gastroenterol. 2013 48 936 943 10.3109/00365521.2013.812139 23834159 49. Berg L.K. Fagerli E. Myhre A.O. Florholmen J. Goll R. Self-reported dietary fructose intolerance in irritable bowel syndrome: Proposed diagnostic criteria World J. Gastroenterol. 2015 21 5677 5684 10.3748/wjg.v21.i18.5677 25987795 50. Hammer J. Sonyi M. Engeßer K.M. Riedl G. Luong S. Hammer H.F. Carbohydrate-induced gastrointestinal symptoms: Development and validation of a test-specific symptom questionnaire for an adult population, the adult Carbohydrate Perception Questionnaire Eur. J. Gastroenterol. Hepatol. 2021 32 171 177 10.1097/MEG.0000000000001880 32796357 51. Komericki P. Akkilic-Materna M. Strimitzer T. Weyermair K. Hammer H.F. Aberer W. Oral xylose isomerase decreases breath hydrogen excretion and improves gastrointestinal symptoms in fructose malabsorption: A double-blind, placebo-controlled study Aliment. Pharmacol. Ther. 2012 36 980 987 10.1111/apt.12057 23002720 52. Tennant D.R. Potential intakes of total polyols based on UK usage survey data Food Addit. Contam. Part A 2014 31 574 586 10.1080/19440049.2014.886132 53. Dills W.L. Jr. Sugar alcohols as bulk sweeteners Ann. Rev. Nutr. 1989 9 161 186 10.1146/annurev.nu.09.070189.001113 2669868 54. Corazza G.R. Strocchi A. Rossi R. Sirola D. Gasbarrini G. Sorbitol malabsorption in normal volunteers and in patients with celiac disease Gut 1998 29 44 48 10.1136/gut.29.1.44 55. Rumessen J.J. Gudmand-Hoyer E. Malabsorption of fructose-sorbitol mixtures. Interactions causing abdominal distress Scand. J. Gastroenterol. 1987 22 431 436 10.3109/00365528708991486 3602922 56. Rumessen J.J. Gudmand-Hoyer E. Functional bowel disease: Malabsorption and abdominal distress after ingestion of fructose, sorbitol, and fructose-sorbitol mixtures Gastroenterology 1988 95 694 700 10.1016/S0016-5085(88)80016-7 3396816 57. Gibson P.R. Halmos E.P. Muir J.G. Review article: FODMAPS, prebiotics and gut health-the FODMAP hypothesis revisited Aliment. Pharmacol. Ther. 2020 52 233 246 10.1111/apt.15818 32562590 58. Wilder-Smith C.H. Materna A. Wermelinger C. Schuler J. Fructose and lactose intolerance and malabsorption testing: The relationship with symptoms in functional gastrointestinal disorders Aliment. Pharmacol. Ther. 2013 37 1074 1083 10.1111/apt.12306 23574302 59. Shepherd S. Gibson P. Fructose malabsorption and symptoms of Irritable Bowel Syndrome: Guidelines for effective dietary management J. Am. Diet. Assoc. 2006 106 1631 1639 10.1016/j.jada.2006.07.010 17000196 60. Choi Y.K. Kraft N. Zimmerman B. Jackson M. Rao S.S. Fructose intolerance in IBS and utility of fructose-restricted diet J. Clin. Gastroenterol. 2008 42 233 238 10.1097/MCG.0b013e31802cbc2f 18223504 61. Shepherd S.J. Parker F.C. Muir J.G. Gibson P.R. Dietary triggers of abdominal symptoms in patients with irritable bowel syndrome: Randomized placebo-controlled evidence Clin. Gastroenterol. Hepatol. 2008 6 765 771 10.1016/j.cgh.2008.02.058 18456565 62. Halmos E.P. Power V.A. Shepherd S.J. Gibson P.R. Muir J.G. A diet low in FODMAPs reduces symptoms of irritable bowel syndrome Gastroenterology 2014 146 67 75 10.1053/j.gastro.2013.09.046 24076059 63. Böhn L. Störsrud S. Liljebo T. Collin L. Lindfors P. Törnblom H. Simrén M. Diet low in FODMAPs reduces symptoms of irritable bowel syndrome as well as traditional dietary advice: A randomized controlled trial Gastroenterology 2015 149 1399 1407 10.1053/j.gastro.2015.07.054 26255043 64. Mitchell H. Porter J. Gibson P.R. Barrett J. Garg M. Review article: Implementation of a diet low in FODMAPs for patients with irritable bowel syndrome-directions for future research Aliment. Pharmacol. Ther. 2019 49 124 139 10.1111/apt.15079 30589971 65. Rej A. Aziz I. Sanders D.S. A gluten-free diet: The express route to fructan reduction Am. J. Gastroenterol. 2019 114 1553 10.14309/ajg.0000000000000349 66. O’keeffe M. Jansen C. Martin L. Williams M. Seamark L. Staudacher H.M. Irving P.M. Whelan K. Lomer M.C. Long-term impact of the low-FODMAP diet on gastrointestinal symptoms, dietary intake, patient acceptability, and health care utilization in irritable bowel syndrome Neurogastroenterol. Motil. 2018 30 e13154 10.1111/nmo.13154 67. Aziz I. Trott N. Briggs R. North J.R. Hadjivassiliou M. Sanders D.S. Efficacy of a gluten-free diet in subjects with irritable bowel syndrome-diarrhea unaware of their HLA-DQ2/8 genotype Clin. Gastroenterol. Hepatol. 2016 14 696 703 10.1016/j.cgh.2015.12.031 26748221 68. Barmeyer C. Schumann M. Meyer T. Zielinski C. Zuberbier T. Siegmund B. Schulzke J.D. Daum S. Ullrich R. Long-term response to gluten-free diet as evidence for non-celiac wheat sen-sitivity in one third of patients with diarrhea-dominant and mixed-type irritable bowel syndrome Int. J. Colorectal Dis. 2017 32 29 39 10.1007/s00384-016-2663-x 27695975 69. Fernández-Bañares F. Arau B. Raga A. Aceituno M. Tristán E. Carrasco A. Ruiz L. Martín-Cardona A. Ruiz-Ramírez P. Esteve M. Long-Term Effect of a Gluten-Free Diet on Diarrhoea- or Bloating-Predominant Functional Bowel Disease: Role of the ’Low-Grade Coeliac Score’ and the ’Coeliac Lymphogram’ in the Response Rate to the Diet Nutrients 2021 13 1812 10.3390/nu13061812 34073569 70. Paduano D. Cingolani A. Tanda E. Usai P. Effect of Three Diets (Low-FODMAP, Gluten-free and Balanced) on Irritable Bowel Syndrome Symptoms and Health-Related Quality of Life Nutrients 2019 11 1566 10.3390/nu11071566 71. Tuck C.J. Reed D.E. Muir J.G. Vanner S.J. Implementation of the low FODMAP diet in functional gastrointestinal symptoms: A real-world experience Neurogastroenterol. Motil. 2020 32 e13730 10.1111/nmo.13730 31571351 72. Nilholm C. Larsson E. Sonestedt E. Roth B. Ohlsson B. Assessment of a 4-Week Starch- and Sucrose-Reduced Diet and Its Effects on Gastrointestinal Symptoms and Inflammatory Parameters among Patients with Irritable Bowel Syndrome Nutrients 2021 13 416 10.3390/nu13020416 33525489 73. Major G. Pritchard S. Murray K. Alappadan J.P. Hoad C.L. Marciani L. Gowland P. Spiller R. Colon Hypersensitivity to Distension, Rather Than Excessive Gas Production, Produces Carbohydrate-Related Symptoms in Individuals with Irritable Bowel Syndrome Gastroenterology 2017 152 124 133 10.1053/j.gastro.2016.09.062 27746233 74. Zhu Y. Zheng X. Cong Y. Chu H. Fried M. Dai N. Fox M. Bloating and distention in irritable bowel syndrome: The role of gas production and visceral sensation after lactose ingestion in a population with lactase deficiency Am. J. Gastroenterol. 2013 108 1516 1525 10.1038/ajg.2013.198 23917444 75. Spiller R. Impact of diet on symptoms of the irritable bowel syndrome Nutrients 2021 13 575 10.3390/nu13020575 33572262 76. Simrén M. Törnblom H. Palsson O.S. van Tilburg M.A. Van Oudenhove L. Tack J. Whitehead W.E. Visceral hypersensitivity is associated with GI symptom severity in functional GI disorders: Consistent findings from five different patient cohorts Gut 2018 67 255 262 10.1136/gutjnl-2016-312361 28104632 77. Volynets V. Louis S. Pretz D. Lang L. Ostaff M.J. Wehkamp J. Bischoff S.C. Intestinal barrier function and the gut microbiome are differentially affected in mice fed a Western-style diet or drinking water supplemented with fructose J. Nutr. 2017 147 770 780 10.3945/jn.116.242859 28356436 78. Zhou S.Y. Gillilland M. Wu X. Leelasinjaroen P. Zhang G. Zhou H. Ye B. Lu Y. Owyang C. FODMAP diet modulates visceral nociception by lipopolysaccharide-mediated intestinal inflammation and barrier dysfunction J. Clin. Investig. 2018 128 267 280 10.1172/JCI92390 29202473 79. McIntosh K. Reed D.E. Schneider T. Dang F. Keshteli A.H. De Palma G. Madsen K. Bercik P. Vanner S. FODMAPs alter symptoms and the metabolome of patients with IBS: A randomised controlled trial Gut 2016 66 1241 1251 10.1136/gutjnl-2015-311339 26976734 80. Hustoft T.N. Hausken T. Ystad S.O. Valeur J. Brokstad K. Hatlebakk J.G. Lied G.A. Effects of varying dietary content of fermentable short-chain carbohydrates on symptoms, fecal microenvironment, and cytokine profiles in patients with irritable bowel syndrome Neurogastroenterol. Motil. 2017 29 e12969 10.1111/nmo.12969

PMC009xxxxxx/PMC9099681.txt

==== Front Polymers (Basel) Polymers (Basel) polymers Polymers 2073-4360 MDPI 10.3390/polym14091730 polymers-14-01730 Review Microencapsulation of Essential Oils: A Review https://orcid.org/0000-0002-0539-6433 Sousa Vânia Isabel https://orcid.org/0000-0003-4594-6361 Parente Joana Filipa Marques Juliana Filipa https://orcid.org/0000-0002-5774-7733 Forte Marta Adriana https://orcid.org/0000-0001-5757-0096 Tavares Carlos José * Cazan Cristina Academic Editor Physics Center of Minho and Porto Universities (CF-UM-PT), Campus of Azurém, University of Minho, 4804-533 Guimarães, Portugal; vaniafernandesousa@gmail.com (V.I.S.); joanacp_17@hotmail.com (J.F.P.); juliana.g.marques@hotmail.com (J.F.M.); martadrianaff@gmail.com (M.A.F.) * Correspondence: ctavares@fisica.uminho.pt 23 4 2022 5 2022 14 9 173018 3 2022 18 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Essential oils (EOs) are complex mixtures of volatile compounds extracted from different parts of plants by different methods. There is a large diversity of these natural substances with varying properties that lead to their common use in several areas. The agrochemical, pharmaceutical, medical, food, and textile industry, as well as cosmetic and hygiene applications are some of the areas where EOs are widely included. To overcome the limitation of EOs being highly volatile and reactive, microencapsulation has become one of the preferred methods to retain and control these compounds. This review explores the techniques for extracting essential oils from aromatic plant matter. Microencapsulation strategies and the available technologies are also reviewed, along with an in-depth overview of the current research and application of microencapsulated EOs. essential oils extraction techniques microencapsulation controlled release microcapsules pharmacology ==== Body pmc1. Introduction Essential oils (EOs) are liquid products present in plants and can be defined as complex natural mixtures of volatile secondary lipophilic metabolites that give plants and spices their essence and colour [1,2]. These compounds can be obtained by hydrodistillation, solvent extraction, and supercritical CO2 extraction, among other methods that can also be used for essential oil extraction [1]. These oils can be extracted from different parts of the plant, such as the flowers, leaves, stems, roots, fruits, and bark, and have different biological and pharmaceutical properties [3]. Due to their versatile nature, the oils can be utilised for several purposes, from contact toxicant and fumigant to attractive or repellent applications [4]. However, there are many factors that affect the chemical composition of essential oils, including genetic variation, type or variety of plants, plant nutrition, fertilizer applications, geographical location of the plant, climate, seasonal variations, stress during growth or maturation, as well as post-harvest drying and storage. EOs that have antimicrobial properties are alternatives to the use of antibiotics and chemical additives [5]. As they have been used worldwide in many industries, their prices differ due to the supply of raw materials, issues related to harvesting, climate factors, and extraction yields. Some of these EOs also have antioxidant properties, with studies reporting that EOs from celery, citronella, cloves, oregano, parsley, tarragon, and thyme seeds were able to inhibit 50% of the 2,2-diphenylpicliryl-hydrazil (DPPH) radical elimination activity [6]. The application of EOs as antioxidants has been evaluated in different types of foods, and research is currently being conducted to optimise the process [7]. Essential oils have gained renewed interest in several areas over the years. Its use was expanded to the medical field due to its biocidal activities (bactericides, viricides, and fungicides) and medicinal properties [8]. The use of natural compounds has become popular in the food industry, with EOs being used as preservatives and food additives due to their antioxidant and antimicrobial properties and pleasant flavour. EOs are included in the composition of many dosage forms in pharmaceutical products. Studies have been carried out on the many biological activities of essential oils (Figure 1) and their components, and particular interests have also been established to elucidate their modes of action, allowing for the improved and targeted intervention in new drugs [9]. EOs are unstable and highly susceptible to changes caused by external factors, such as light, temperature, oxygen, and humidity [10]. The high volatility and reactivity of these compounds represent challenges for the application of essential oils in several industries [11]. To overcome these limitations, the microencapsulation technique is often used to maintain the functional and biological characteristics of these compounds and to control their release [12]. Microencapsulation is a technology based on the coating of solid, liquid, or gaseous particles through an encapsulating agent that acts as a barrier, completely isolating the core material from the external environment [13]. Most microcapsules have a diameter within the range 1–1000 µm [14]. The shell material can be a film of a natural, semi-synthetic, or synthetic polymer and its choice has a key role in the stability of core material [15]. Arabic gum, agar, alginate, proteins, and dextrins are some of the materials used as encapsulating agents in the microencapsulation process [16]. Due to the enormous interest of the scientific community and the industry in the microencapsulation of active substances, several microencapsulation methods have been developed over time. Encapsulation processes are usually divided into three main categories: physicochemical, mechanical, and chemical processes [17]. In this review article, some of these methods, which are used in EO microencapsulation, are described and an outlook of scientific works developed in this area is approached. 2. Essential Oils Essential oils are defined, according to the European Pharmacopoeia, as an “odorous product, usually of complex composition, obtained from a botanically defined plant raw material by steam distillation, dry distillation, or a suitable mechanical process without heating. EOs are usually separated from the aqueous phase by a physical process that does not significantly affect their composition” [18]. EOs are extracted from aromatic plant materials such as oily aromatic liquids, and they can be biosynthesised in different plant organs as secondary metabolites, such as flowers, herbs, buds, leaves, fruits, branches, bark, zest, seeds, wood, rhizomes, and roots. Essential oils are complex mixtures of highly volatile aromatic compounds named after the plant from which they are derived. Within the different species of plants, only 10% contain EOs and are called aromatic plants. These natural products exert the function of protecting the plants, guaranteeing the growth of the plant and the propagation of species. Essential oil provides the essence, odour, or flavour of the plant and some of the functions that it performs in plants can also be made in living organisms [19]. EOs are generally liquid at room temperature and are hydrophobic (immiscible with water) and lipophilic (miscible with other oils and organic solvents) substances [20]. In general, essential oils are a mixture of compounds with their own physicochemical characteristics that, when combined, give the oil a particular odour. The different aroma of oils is fundamentally due to variations in the volatility and relative concentration of its constituents [21]. 2.1. Chemistry of Essential Oils The chemical composition of EOs can be complex due to the number of different components, which can have promising chemical and biological properties [22]. Essential oils are complex mixtures that can contain over 300 different compounds. Most EOs are characterised by two or three main components in reasonably high concentrations (20–70%) compared to other components present in small amounts [8]. The organic constituents have a low molecular weight, and their vapour pressure (at atmospheric pressure and at room temperature) is high enough for them to be partially in vapour state [23]. Chemically, EOs mainly belong to two classes of compounds: terpenes and phenylpropanoids (Table 1). The terpene family is predominant, and phenylpropanoids, when they appear, are responsible for the characteristic odour and taste [24]. 2.1.1. Terpenoids Terpenes, also called terpenoids, constitute the largest class of natural products with several structurally diversified known compounds [25]. Their structures contain carbon skeletons and are formed by isoprene units, being classified according to the number of these units that compose their structure. They can be classified as hemiterpenes (1 isoprene unit; 5 carbons), monoterpenes (2 isoprene units; 10 carbons), sesquiterpenes (3 isoprene units; 15 carbons), diterpenes (4 isoprene units; 20 carbons), triterpenes (6 isoprene units; 30 carbons), and tetraterpenes (8 isoprene units; 40 carbons), among others. Monoterpenes and sesquiterpenes are mostly found in volatile essential oils. Terpenes can present aromatic, aliphatic, and cyclic structures and can contain oxygen atoms, being called terpenoids (Figure 2) [26]. 2.1.2. Phenylpropanoids Phenylpropanoids are natural substances commonly found in plants and consist of a six-carbon aromatic ring joined to a three-carbon side chain. This side chain contains a double bond and the aromatic ring may be substituted. These compounds are biosynthesised from shikimic acid, which forms the basic units of cinnamic and p-coumaric acids. These units, through enzymatic reductions, produce propenylbenzenes and/or allylbenzenes and, through oxidations with side chain degradation, generate aromatic aldehydes [27,28]. polymers-14-01730-t001_Table 1 Table 1 Composition of compounds found in essential oils [29]. Essential Oil Compounds Classes Constituents Terpenes Monoterpene (−)-Camphene, p-cymene, (+)-limonene, β-ocimene α-phellandrene, α-pinene, α-terpinene, terpinoleneorange, Sesquiterpene (−)-β-isabolene, α-cadinene, β-caryophyllene, α-copaene, β-elemene, α-farnesene, α-humulene, α-zingiberene Phenylpropanoids (E)-Anethole, cinnamaldehyde, cinnamic acid cinnamic alcohol, eugenol, methyleugenol, myristicin 2.2. Extraction Methods Aromatic herbs or parts thereof, such as leaves, flowers, bark, seeds, and fruits, are subjected to extraction processes after being collected at specific stages of maturity and stored under controlled conditions (light, temperature, and humidity). Extraction techniques are essentially divided into classical and conventional methods and innovative methods. Classical methods are based on the distillation of water by heating, to extract the EOs from the plant matter. Hydrodistillation, steam distillation, hydrodiffusion, organic solvent extraction, and cold pressing are some of these methods. New extraction technologies have been developed in order to overcome some of the disadvantages of conventional methods. Methods such as ultrasound-assisted extraction and microwave-assisted extraction use energy sources that make the process more environmentally friendly. On the other hand, methods such as supercritical fluid extraction and subcritical liquid extraction allow the non-polar components from the material to be extracted [30]. 2.2.1. Hydrodistillation Hydrodistillation is the oldest and simplest method for extracting OEs. This method is characterised by direct contact between the solvent and the plant material, that is, the raw material is submerged in boiling water (Figure 3) [31]. In this procedure, the cell walls are broken, and the oil is evaporated together with the water, and then condensed into a mixture of water vapour and volatile compounds of vegetable raw material. However, these two phases (volatile compounds and water) are immiscible, rendering possible an additional separation according to the difference in density [32]. This technique is inexpensive, but, at the same time, it is not selective because of the waste of large amounts of the compound in the solvent (part of the extract can be lost in the aqueous phase) and can provide low yields [33,34]. Despite being the oldest method, hydrodistillation is still used today for extracting oils from different matrices. Essential oils from Rosmarinus officinalis L. [35], Ziziphora clinopodioides L. [36], Citrus latifolia Tanaka [37], and Zingiber officinale [38] are some of the medicinal plants where EOs can be extracted by hydrodistillation. 2.2.2. Steam Distillation Steam distillation is one of the preferred methods of extracting EOs. The extraction procedure is based on the same principles as hydrodistillation. The difference essentially lies in the absence of contact between the substrate to be extracted and the water, which causes a reduction in the extraction time. The sample is placed in a column where the bottom part is connected to a flask with water under heating (Figure 4). The top part is connected to a condenser, where the steam produced passes through the sample, taking essential oils to the condenser. This process causes the condensation of the water–oil mixture, and this mixture can be separated by liquid–liquid extraction [39]. This method is applied commercially and on a large scale in the extraction of essential oils from hops [40] and in the extraction of several EOs such as lavender [41] and patchouli essential oil [42]. 2.2.3. Organic Solvent Extraction Some essential oils (such as rose and jasmine) have low thermal stability and are unable to withstand high temperatures. In these cases, organic solvents that have a low boiling temperature, are chemically inert, and have low cost can be used. In organic solvent extraction, the sample is placed in contact with the organic solvent (which can be hexane, benzene, toluene, or petroleum ether, among others) for a period that allows the transfer of the soluble content of the sample. The extracted matrix is concentrated by evaporating the solvent present in the liquid phase. This method allows the sample to be permanently in contact with a quantity of fresh solvent and, at the end of the process, it is not necessary to carry out filtration, as long as there are high yields [34]. Solvent extraction is the most-used conventional method in the cosmetic industry [43,44,45]. Figure 5 represents the extraction of organic solvents through a Soxhlet extraction [32,46]. 2.2.4. Cold Pressing Essential oils are mechanically removed by cold pressing, where the oil glands are broken and volatile oils are released. In this process, an aqueous emulsion is formed, where the oil present can be obtained through centrifugation, decantation, or fractional distillation [13]. The cold pressing method is essentially used to extract oils from citrus fruits [47,48,49]. 2.2.5. Supercritical Fluid Extraction (SCFE) Supercritical fluid extraction is an efficient, environmentally friendly, and clean technique for isolating EOs. In this technique, supercritical fluids are used as extraction agents due to the supercritical state of fluids, conferring excellent characteristics for the extraction process, such as low viscosity, high density (close to that of a liquid), and high diffusivity (high penetration power). Several substances can be used as supercritical solvents, such as water, carbon dioxide (CO2), methane, ethylene, and ethane. However, CO2 is the most-used solvent due to its critical point being easily reached (low temperature and pressure, 31.2 °C and 72.9 atm, respectively), low toxicity and reactivity, low cost, and non-flammability. After selecting the ideal temperature and pressure for extraction, supercritical fluid passes through the sample and the oils are dissolved and extracted. Subsequently, the extraction solution is maintained at a pressure below the critical point and as the pressure decreases, the supercritical fluid passes to the gaseous state and loses its solvating capacity, being recycled [50]. This method is increasingly used commercially, being applied in the extraction of EOs from the leaves of laurel [51], rosemary [52], sage [53], flowering plants [54], and horseradish tree [55]. 2.2.6. Microwave-Assisted Extraction (MAE) Due to the need to use more ecological and energy-efficient extraction methods, microwave-assisted extraction has become an alternative to conventional methods. The sample is placed in a microwave reactor without any solvent, where the electromagnetic energy that is converted into heat increases the internal temperature of sample cells due to the evaporation of the moisture present. The internal pressure increases, the glands rupture, and the essential oil is released [56]. Several EOs were extracted from plant matrices through this technique, such as orange [57], laurel [58], lemon [59], mint [60], rosemary [61], and basil [62]. 2.2.7. Ultrasound-Assisted Extraction (UAE) Ultrasound energy allows the intensification of EO extraction [63]. Therefore, it is usually combined with other extraction techniques in order to accelerate the extraction process and increase the speed of mass transfer. The sample is submerged in a solvent while being subjected to ultrasound. This method, through rapid solvent movements, induces a mechanical vibration of the walls and membranes of the sample that causes the release of essential oils. In some areas, it is already considered a large-scale application method, such as in the medical and food industry, where it is used to increase the quality of the extracted substrate, reduce working time, and increase the yield [64]. In general, alternative methods have emerged to overcome some of the disadvantages and limitations of conventional methods. Traditional methods have long extraction times (4 to 6 h), high energy consumption, and use solvents that increase environmental pollution. Furthermore, they can cause chemical changes to the EOs that are thermally unstable, causing a decrease in the quality of the extracted oils and changes in the chemical nature of compounds. The ‘greener’ alternatives are more sustainable and economical due to reduced water and energy consumption and reduced CO2 emissions. However, these methods are not easily accessible, and the initial investment is higher. Therefore, currently, the hydrodistillation method continues to be the most-used extraction technique in laboratory due to its accessibility, simplicity, and lower cost [46]. Table 2 summarises the advantages and disadvantages of the different EO extraction methods. 2.3. Essential Oils Application 2.3.1. Essential Oils in Plants EOs are stored in specific parts of plants, acting in extraordinarily different ways. Some aromatic plants have been widely explored due to their properties, such as bay laurel (Laurus nobilis). This plant is an aromatic tree, and laurel oil is extracted from the dry leaves and branches, appearing as a greenish yellow liquid with a powerful medicinal odour. In addition to being used in cuisine, the laurel tree leaves are used in medicine for having antioxidant [65], antibacterial [65,66], and antifungal [22] properties. According to the literature, laurel has also been proven to be an insect repellent [67,68]. However, it can cause dermatitis in some individuals, and due to the possible narcotic properties attributed to methyleugenol, this oil should be used in moderation. Cymbopogon nardus, commonly known as citronella, is an aromatic and perennial herb. Citronella oil can also be produced from Java or Maha Pengiri citronella (C. winterianus) [69]. Citronella leaves are used for their aromatic and medicinal value in many cultures, such as in the treatment of fever, intestinal parasites, and digestive and menstrual problems, as well as for use as an insect stimulant and repellent [69,70,71,72]. Citronella is also used in traditional Chinese medicine for rheumatic pain, and it has antifungal [73], antioxidant, and antibacterial [74] properties. It is non-toxic and non-irritating, but it can cause dermatitis in some people [69]. Regarding the medical properties of hops (Humulus lupulus), these are better known for treatments associated with nerves, insomnia, nervous tension, neuralgia, and for sexual neurosis in both sexes [5]. It has antibacterial [75,76], antifungal [76], anti-cancer [76,77], and repellent [78,79] properties. In China, it is used for pulmonary tuberculosis and cystitis treatment. It can also be used to make beer. It is non-toxic and non-irritating, but it can cause sensitivity in some individuals, and people with depression should avoid this oil [76]. Lemon balm (Melissa officinalis) is a herbaceous perennial from the mint family and it has antibacterial, antifungal [80], sedative, antipyretic, antispasmodic, anti-hypertensive, anti-Alzheimer, and antiseptic properties [81]. In addition to the treatment of several gastrointestinal, liver, and nervous system disorders, it has also been reported that lemon balm is useful in the treatment of asthma, bronchitis, coughs, and several pains [82]. Furthermore, this plant is notably marked by its antimicrobial applications in different medicines, exemplified by its use in insect bites (wasps and bees) and poisonous or infectious bites [81,83]. Azadirachta indica, better known as neem, is an ancient tree that has been used for centuries for the most varied purposes. The plant provides a great number of secondary metabolites with biological activity. The plant has gained great importance in several areas, such as agriculture, livestock, and medicine [84]. It is used as an insecticide [85], antiviral [86], antibacterial [87], and antimicrobial [88], among others. Neem oil is very effective for acne, psoriasis, and eczema treatments, but it can also be applied as a support in the treatment of topical fungal or viral conditions, such as nail fungus, athlete’s foot, warts, or wounds. The natural antihistamines contained in neem oil are effective in relieving the itching and burning caused by, for example, bee, mosquito, and spider bites. The main constituent of neem is azadiractin, found in the leaves, fruits, and seeds. Mentha pulegium, better known as pennyroyal or mint (Brazil), is one of the best-known species of the genus Mentha. Pennyroyal extracts are good insect repellents [89,90,91,92]. There are several studies that show that these extracts also have other properties, such as antimicrobial [92,93,94], antioxidant [92,93,95], antibacterial [95,96], and anti-tumour [96] uses. It is still current in the British Herbal Pharmacopoeia, indicated for flatulent dyspepsia, intestinal colic, common cold, delayed menstruation, skin rashes, and gout [69]. Illicium verum, popularly known as star anise, is a plant considered a spice for medicinal and culinary use. The extraction of Illicium verum has carminative, stomach, stimulating, and diuretic properties and is used as a pharmaceutical supplement [97]. The extracted shikimic acid is one of the main ingredients of the antiviral drug Tamiflu® (oseltamivir) that is used to treat symptoms caused by avian influenza [98]. It has also been reported to have antimicrobial properties [99] and antioxidant properties [100], as well as significant anti-cancer potential [101]. There are studies in which star anise has been used as an insect repellent, such as for the Indian flour moth (P. interpunctella larvae) [102,103]. The main constituent of star anise is trans-anethole (Table 3) (80–90%) and when used in large doses, it is narcotic and slows down circulation, which can lead to brain disorders [69]. Valerian is a perennial flowering plant with many chemical constituents, with more than 150 constituents identified in its essential oil [104]. Regarding therapeutic indications, it is advisable for people with nervous agitation, mild anxiety, and difficulties in sleeping [105,106]. There are also studies that demonstrate its antibacterial [107] and antimicrobial [108] properties. It is non-toxic, non-irritating, and can cause sensitisation. Table 3 describes the main components of the EOs present in plants, as well as their chemical structures and some of the biological properties. 2.3.2. Pharmacological and Medical Applications Essential oils have a wide range of biological properties, and there has been a growing interest in clinical applications (Table 4). Some of the properties include the chemo-preventive effects of cancer [109], antifungal [110], antiviral [111], antimicrobial, analgesic, anti-inflammatory [112], and antiparasitic activities [113]. An extensive range of EOs have antibacterial activity against Gram-positive and Gram-negative bacteria, along with antifungal properties. These compounds have been studied and have shown very promising results in salmonella, staphylococci, and other oral pathogens, and can be an alternative to antibiotics providing they are properly studied for these effects [115,116]. EOs that have shown antibacterial potential are basil [117], manuka oil (more potent among eucalyptus oil, rosmarinus, lavandula, and tea tree) [118], melaleuca oil [119], and essential leaf oils P. undulatum and Hedychium gardnerianum. With regard to antifungal activity, melaleuca oil showed positive results for all of its constituents, especially against dermatophytes and filamentous fungi [120]. In a reported study, germinated conidia of Aspergillus niger were more susceptible to non-germinated ones, with EOs of Melaleuca ericifolia, Melaleuca armillaris, Melaleuca leucadendron, and Melaleuca styphelioides exhibiting good activity against this fungus [120]. These same oils were evaluated for their antiviral activity in African green monkey kidney cells through the plaque reduction assay in the herpes simplex virus type 1 [121]. Other plants that have good antifungal activity are M. piperita, Brassica nigra, Angelica archangelica, Cymbopogon nardus, Skimmia laureola, Artemisia sieberi, and Cuminum cyminum [122,123,124,125,126,127] Regarding antioxidant activity, the essential oil of the seeds of Nigella sativa L. showed considerable activity in the elimination of hydroxyl radicals. The essential oil of M. armillaris has marked antioxidant potential, changing the parameters of superoxide dismutase, and improving the concentrations of vitamin E and vitamin C [128]. However, there have been promising insect repellency/toxicity results from the essential oils of Nepeta parnassica, in Culex pipiens molestus [129]. Geranial, neral, geraniol, nerol, and trans-anethole are well established to stimulate the estrogenic response, and citrus (a combination of geraniol, nerol, and eugenol) is effective in replacing [3H] 17β-estradiol at the oestrogen receptors in recombinant yeast cells [130,131]. Cancer is a growing health problem worldwide and is the second leading cause of death. Essential oil constituents play an important role in cancer prevention and treatment as alternatives to synthetic drugs. Mechanisms such antioxidant, antimutagenic, and antiproliferative properties, enhancement of immune function and surveillance, enzyme induction and enhancing detoxification, modulation of multidrug resistance, and the synergistic mechanism of EO constituents are accountable for their chemo-preventive properties [132]. It has been reported that mitochondrial damage and apoptosis/necrosis in the yeast Saccharomyces cerevisiae were reduced by essential oils [133]. Recently, some studies demonstrated that certain EOs exhibited antimutagenicity towards mutations caused by UV light [8]. Jaganathan et al. reported that the active constituent eugenol from Syzygium aromaticum (cloves), nutmeg, basil, cinnamon, and bay leaves showed antiproliferative activity against several cancer cell lines in animal models [134]. In addition to medicinal and pharmacological applications, essential oils are used in perfumes, cosmetics, hygiene products, disinfectants, repellents, candles, phytochemicals, preservatives, and food additives. 2.3.3. Food Applications In food, cosmetics, and personal care products, EOs are used as a natural aroma due to their chemical properties. In the food industry, EOs are being used as a food preservative because one of the main concerns is the preservation of food to prolong its useful life, ensuring safety and quality [11]. An expiration date is defined as the period of time during which the food product will remain safe. This ensures the maintenance of sensory, chemical, physical, microbiological, and functional characteristics. For example, spices can be encapsulated to extend their shelf life, maintain their properties, and inhibit reactions with other compounds [135]. Cinnamaldehyde, the aromatic agent present in cinnamon, has antimicrobial properties, and when encapsulated can slow the growth of yeasts in bakery products. Thus, the use of cinnamon in encapsulated form allows the product to be flavoured without interfering with the leavening process [136]. As the unpleasant taste and instability limit the application of EOs, the use of these encapsulated compounds can allow their application for several purposes. One of them is the intensification of the flavour of food products, where capsules can be used that release the product quickly when introduced into the mouth [20]. The packaging has the function of delaying deterioration, maintaining the quality and safety of packaged foods. For the packaging material to be satisfactory, it must be inert and scratch resistant, and must not allow molecular transfer to or from the packaging materials. Active packaging technologies extend the shelf life and control the quality of food products, decreasing microbial, biochemical, and enzymatic reactions through different strategies, such as adding chemical additives/preservatives, removing oxygen, controlling humidity and/or temperature, or a combination of these [137]. Oregano oil contains a high amount of carvacrol and is considered one of the most active plant extracts against pathogens due to its antimicrobial activity. Therefore, it has been used to preserve a variety of foods such as pizza, fresh beef [138], and cheddar cheese [139]. For the same purpose, limonene is reported for the preservation of strawberries [140], rosemary in chicken breast cuts [141], and cinnamon in pastries [142]. 2.3.4. Cosmetic and Cleaning Applications In the detergent and cosmetics industry, microcapsules of essential oils are used in many products such as perfumes, creams, and deodorants where the controlled release of EOs is essential, increasing the duration of fragrance and the properties of the EOs [45]. Aroma ingredients such as patchouli (Pogostemoncablin), citronella (Cymbopogon winterianus), sandalwood (Santalum álbum), bergamot (Citrusaurantium), rosemary (Rosmarinus officinalus), mint (Mentha piperita), and vetiver (Chrysopogon zizanioides) are frequently used [4]. Regarding the EOs from flowers, Lavandula officinalis, rose, jasmine, tuberose, narcissus, and gardenia are those most commonly exploited for cosmetic applications [143]. Products such as detergents, soaps, shampoos, and softeners are largely produced using these natural compounds. Over the years, EOs have also been used against nosocomial infections, as a cleaning liquid for disinfecting equipment and medical surfaces [9], or as an aerosol in operating rooms and waiting rooms to limit contamination [10]. 2.3.5. Agrochemical Applications The loss of quality of agricultural products is caused by the presence of insect pests. The presence of these pests leads to reduced quality, low yield, and economic losses. Furthermore, human and animal health is compromised due to the production of carcinogenic secondary metabolites. To overcome this problem, chemical insecticides were used to excess. Despite being highly efficient, their overuse caused physiological resistance in several insect species and irreversible damage to the environment. Essential oils have emerged as a natural plant alternative to protect agricultural products from pests [144]. The use of EOs has intensified, mainly in gardens and homes, for pest control (Table 5), being important due to their toxic (pesticide) effect. EOs can be inhaled, ingested, or absorbed through the skin of insects. Monoterpenoids are an important group of chemical compounds in essential oils that interfere with the octopaminergic system of insects, which represent a target for insect control. As vertebrates do not have octopamine receptors, most chemicals in EOs are relatively safe to use. The special regulatory status together with the availability of essential oils has made the commercialisation of EO-based pesticides possible. Microencapsulation technology is used to produce these natural pesticides in order to mimic chemical compartmentalisation in plants, by protecting essential oils from degradation [145]. 2.3.6. Textile Applications Essential oils are used in medical and technical fabrics. The technique used in industrial processes is encapsulation, which is used to give finishes and properties to textiles that were not possible or economical. The main application for encapsulation is durable fragrances and skin softeners, and other applications may include insect repellents, dye, vitamins, microbial agents, and phase-change materials, and medical applications, such as antibiotics, hormones, and other medications. The functionalisation of textiles with EOs with anti-mosquito repellent properties is a revolutionary way to protect human beings from insect bites and, thus, from many diseases such as malaria and dengue [146]. Plants, whose OEs have been reported to have repellent properties, include citronella, cedar, geranium, pine, cinnamon, basil, thyme, garlic, and mint. Khanna et al. performed the synthesis of a modified cyclodextrin host (β-CD CA) for inclusion complexation with the essential oils of cedarwood, clove, eucalyptus, peppermint, lavender, and jasmine for the assessment of repellent efficacy against Anopheles Stephensi in cotton. It was concluded that jasmine EO is the weakest against mosquitoes, as it worked as an attractant simulating flower nectar. Eucalyptus and clove are the feeding deterrents. On the other hand, lavender and peppermint are potential mosquito repellents, and cedarwood is an effective mosquito killer [147]. Soroh et al. reported that textiles treated with the Litsea and lemon EO microemulsion showed potential mosquito-repellent properties [148]. In general, citronella remains the most promising as an insect repellent and, therefore, is the most-incorporated EO in tissue functionalisation for this purpose. Specos et al. demonstrated citronella essential oil’s mosquito-repellent action, especially against Aedes aegypti [148]. Microcapsules with citronella are commonly incorporated into matrices such as cotton and polyester [149]. Another report determined that bio-based citronella oil has a better insect repellent effect than synthetic agents. Sariisik et al. concluded that, after washing, the insect repellent activity of the printing and coating method was increased, and the fabrics still showed repellency after five washing cycles [150]. 3. Microencapsulation Microencapsulation is the protection of small solid, liquid, or gaseous particles through a coating system (1–1000 mm) [151]. The encapsulated material is called the core and the material that forms the coating of particle is the wall or encapsulating agent [152]. Wall material can be a natural, synthetic, or semi-synthetic polymeric coating. In this technology, microparticles are formed, which can be classified in relation to their size and morphology, according to the encapsulating agent and microencapsulation method used [153]. Microparticles can be distinguished according to their form: they are classified as a reservoir-type system, ‘microcapsules’, when the core (encapsulated material) is concentrated in the central region, coated by a continuous wall material (encapsulating agent); or a monolithic system, ‘microspheres’, when the active agent (core) is dispersed in a matrix system (Figure 6). In general, the main difference is that in microspheres, part of the encapsulated material is exposed on the surface of the microparticle [154]. The physicochemical characteristics of the microcapsule are defined by the encapsulating agent and the active agent. The wall material must form a cohesive film that bonds with the encapsulated material [155]. Several materials can be used for the coating, with proteins, carbohydrates, and lipids being frequently used. Furthermore, the materials must be chemically compatible and the encapsulating agent chemically inert, so as not to react with the core [156]. Microencapsulation technologies achieve several objectives (Figure 7) and they are particularly used to protect the core active agent’s sensitivity to oxygen, light, and moisture, or to prevent interaction with other compounds. However, the most important reason for encapsulating an active agent is to obtain a controlled release [157]. The process of defining a microencapsulation system is mainly dependent on the purpose of the microcapsules. Characteristics such as shape, size, permeability, biodegradability, or biocompatibility are defined depending on the application of this material. Other physical and mechanical properties of the microcapsule, such as strength and flexibility, must also be defined [158]. One of the great advantages of microencapsulation is the mechanism of the controlled, sustained, or targeted release of the active agent. This release can occur at a certain defined time or not, through a mechanism of diffusion through or rupture of the wall. The release can be activated through temperature variations, solubility, pH changes, or even the biodegradability of the wall material [159]. Depending on the nature of the interaction of the encapsulating and encapsulated material, microencapsulation methods can be distinguished as chemical, physicochemical, and mechanical (Figure 8) [160]. In general, a microencapsulation method must be fast, easy, reproducible, and easily scalable for industry. The most-used microencapsulation methods are spray drying and coacervation, and these approaches will be mentioned in more detail below. 3.1. Emulsification Emulsification is a fundamental step in oil microencapsulation, being used in a wide variety of food and pharmaceutical products. It is applied for the encapsulation of bioactive substances in aqueous solutions, which can be used directly in liquid or dried (spray-or freeze-drying) to form powders. An emulsion consists of at least two immiscible liquids, with one of the liquids being dispersed as small spherical drops in the other. As can be seen in Figure 9, there are four systems, consisting of:Oil-in-water emulsion (O/W); Water-in-oil emulsion (W/O); Oil-in-water-in-oil emulsion (O/W/O); Water-in-oil-in-water emulsion (W/O/W). In these systems, the droplet diameters can vary from 0.1 to 100 μm [161] and have been extensively revised by scientists [162]. The O/W emulsion consists of small oil droplets that are dispersed in an aqueous medium, being the droplets wrapped in a thin interfacial layer. Its advantages are the ease of preparation and low cost, with some disadvantages such as physical instability and limited control [162]. Through modifications of the emulsifiers, features can be added, such as the use of Maillard reaction products. These products can increase encapsulation efficiency and are able to protect the microencapsulation oil and other oils from oxidation [163]. A straightforward method for obtaining small droplets with a stratum size distribution is the evaporation/extraction of the emulsifying substance. This method is used in the preparation of biodegradable and non-biodegradable polymeric microparticles and in the microencapsulation of a wide variety of liquid and solid materials [164]. However, it is an expensive method with a low encapsulation efficiency, leading to residual solvent amounts [165]. 3.2. Coacervation Coacervation is one of the most widely used microencapsulation techniques. The technique is based on oppositely charged polyelectrolyte polymers that interact and form a wall covering the active agent. The coacervation process can be classified as simple and complex if one or two (or more) polymers are used, respectively. Generally, this technique is defined by the separation of two liquid phases in a colloidal solution, where one phase is rich in polymer (coacervated phase) and the other phase does not contain polymer (equilibrium phase) [46]. Complex coacervation involves the interaction of two oppositely charged colloids, where the neutralisation of charges induces a phase separation. A polysaccharide and a protein are usually used as the different polymers. Wall material systems that are most widely investigated include gelatin/gum arabic, gelatin/alginate, gelatin/glutaraldehyde, gelatin/chitosan and gelatin/carboxymethyl cellulose [166]. In the process of the microencapsulation of hydrophobic materials (Figure 10), the emulsification of the encapsulated agent in an aqueous solution containing two different polymers occurs, usually at a temperature and pH above the gel and isoelectric point of the protein. Then, the separation into two liquid phases (polymer-rich phase and aqueous phase) follows, which results from the electrostatic interaction of the polymers. Subsequently, a microcapsule wall is formed as the deposition of the polymer-rich phase occurs around the hydrophobic particles of the active agent, due to controlled cooling below the gelation temperature. Finally, the microcapsule walls harden through the addition of a crosslinking agent [167]. Simple coacervation has advantages over complex coacervation in terms of the associated cost, as cheap inorganic salts are used to induce the separation phase, while expensive hydrocolloids are applied in the complex method. Furthermore, complex coacervation is more sensitive to small variations in pH. However, compared to other microencapsulation methods, complex coacervation is a simple, scalable, inexpensive, reproducible, and solvent-free method, enabling its industrial use [166]. 3.3. In Situ Polymerisation In situ polymerisation (Figure 11) is based on the formation of a wall through the addition of a reagent inside or outside the core material [168], becoming one of the most-used methods in the preparation of microcapsules and functional fibres. Polymerisation takes place in the continuous phase and not on both sides of the interface between the core material and the continuous phase. Microcapsule formation occurs through an oil emulsion in a solution of melamine–formaldehyde resin and a sonication process to emulsify the oil in the aqueous phase. Then, resin is added under agitation and the pH is adjusted, with the formation of shells, thus promoting the reaction of the melamine with the formaldehyde at the interface of the oil droplets. This type of microcapsule is used in fragrances, insect repellents, food packaging, and footwear. The microcapsules result in smooth surface morphologies and are able to preserve the encapsulated scented oils for a sufficient period of time. They also have good thermal and controlled release properties [168,169]. Using a polymer as a microcapsule wrapper is considered a good addition due to its high strength and stability [170]. On the other hand, using a copolymer to prepare microcapsules with a low molecular weight of formaldehyde–melamine avoids the toxicity of formaldehyde [171]. In situ polymerisation is a method of rapid and easy expansion [172] and, at the same time, provides high encapsulation efficiency. However, the polymerisation reaction is difficult to control [173] and requires a large amount of solvent, making the monomers non-biodegradable and/or non-biocompatible [174]. 3.4. Spray Drying Spray drying is the most-used technology in the microencapsulation of essential oils. It is mainly used on an industrial scale, as it allows simple, reproducible, continuous, and low-cost production. Being used more frequently in the food industry, this process is also utilised in the cosmetics, pesticides, and pharmaceutical industries [175]. This technique allows encapsulated and powdered Eos to be obtained due to the ability to dry them in just one operation. In this process, the atomisation of emulsions occurs in a drying chamber with relatively high temperatures, where the evaporation of the solvent takes place and, consequently, microcapsules are formed [176]. The spray-drying technique involves four steps (Figure 12), where the preparation of dispersion first occurs, i.e., the wall materials are dissolved in water with agitation and controlled temperature. Still in the same step, the addition of the EOs follows and, if necessary, the emulsifier can be added. Afterwards, the dispersion is homogenised to be injected into the equipment through an atomising nozzle, where small droplets are formed. In the third step, emulsion atomisation occurs, where the formed droplets enter the drying chamber with a flow of hot air present. Finally, the dehydration of the atomised microparticles is done through the evaporation of the solvent, which dries the microparticles, which can them be recovered in the form of powder in a collector or filter [177]. The main limitations of this technique are related to the wall material, which must have good water solubility, and to the number of encapsulating agents available. In addition, some materials may be sensitive to the high temperatures introduced in the atomisation process. In addition, the production of microcapsules in fine powder form can cause agglomeration and an additional process may be required [166]. 3.5. Freeze Drying Freeze drying, also known as lyophilisation, is a simple process (Figure 13) that is used to dehydrate most materials sensitive to heat and aromas such as oils. Sublimation is the major principle involved in this drying process, where water passes directly from a solid state to a vapour state without passing through the liquid state. Before starting this process, the oil is dissolved in water and frozen [178]. Afterwards, the pressure is reduced and heat is added to allow the frozen water to sublimate the material directly from the solid phase to the gas phase. Freeze-dried materials appear to have the maximum retention of volatile compounds compared to spray drying, and this technique is used to microencapsulate some oils, with high yields [179]. This method helps to better preserve the EO content in many types of herbs and spices compared with other preservation techniques [180]. Lyophilisation is simple and easy to operate, showing that lyophilised samples are more resistant to oxidation and less efficient in microencapsulation [181]. The process disadvantages include high energy use, long processing time, and high production costs [182]. 3.6. Supercritical Fluid (SCF) Technology Many pharmaceutical, cosmetic, and food industries use supercritical fluid technology (Figure 14) to form the microcapsules of essential oils due to their inherent advantages. The use of a wide variety of materials that produce controlled particle sizes and morphologies, the easy solvent removal, the non-degradation of the product, and being a non-toxic method are some of the many advantages of SCF technology. The methods used for supercritical fluids are the precipitation of gas anti-solvent, particles of saturated gas solutions, the extraction of fluid emulsions, and the rapid expansion of supercritical solutions [183,184]. The supercritical solvent impregnation process has proven to be successful in a wide variety of substances (essential oils, fragrances, active pharmaceutical compounds, and dyes) and matrices (wood, polymers, cotton, and contact lenses). An alternative to spray drying (that degrades oils at high temperatures) is impregnation with supercritical solvent, as it is an ecological process where supercritical carbon dioxide is used as a green solvent. 3.7. Coaxial Electrospray System The food, cosmetic, and pharmaceutical industries use a new technology to encapsulate oils, called coaxial electrospraying (Figure 15) [185,186]. This system is used in two phases, with external and internal solutions being sprayed coaxially and simultaneously through two feed channels separated by a nozzle. In the electrospray process, the Taylor cone is composed of a core-shell structure that is formed at the top of the spray nozzle, ending up with the polymeric solution encapsulating the internal liquid. This method is distinguished by its ease and efficiency, and the maximum speed of the core material. The coaxial electrospray system provides a uniform size distribution, a high encapsulation efficiency, and an effective protection of bioactivity. However, the encapsulation efficiency and the stability of the microcapsules are affected by the wall materials [186]. Furthermore, controlling the process in coaxial electrospraying is difficult to some extent [187]. In experimental terms, the reported work on coaxial electrospray is based on individual laboratory experiments, consisting of specific combinations of materials and empirical process parameters. The fabrication of polymeric microparticles and nanoparticles is hampered by the lack of standard protocols. Regarding the collection of particles, the methodology cannot facilitate the hardening of the shell or maintain the morphology of particle, or even prevent its aggregation. On the theoretical side, many existing process models are empirical or semi-quantitatively empirical. The simulated results are not enough for the quantitative control of the process, as numerical simulations, such as computational fluid dynamics modelling, have been used to simulate the formation of the liquid cone and atomisation in a single axial electrospray process [188]. In summary, more experimental and theoretical study is needed to better understand the physical nature of coaxial electrospray and to provide quantitative guidance for process control. 3.8. Fluidized Bed Coating Fluidised bed coating is one of the most efficient coating methods, in which the ingredients can be mixed, granulated, and dried in the same container. Consequently, the handling and processing time of the material is reduced. This approach was recently used to encapsulate fish oil by spraying and coating it (Figure 16) [189]. Fluidised bed coating is carried out by suspending the solid particles of the core material by an air stream under controlled temperature and humidity and then sprayed, building, over time, a thin layer on the surface of the suspended particles. This material must have an acceptable viscosity for atomisation, and the pumping should be able to form an appropriate film and be thermally stable [190]. There are several methods used in fluidised bed coating, including top spray, bottom spray, and tangential spray methods. In the top spray system, the coating solution is sprayed in the opposite direction with air in the fluid bed. The opposite flows lead to an increase in the efficiency of encapsulation and the prevention of agglomerates formation, achieving microcapsules with a size between 2 and 100 μm. The bottom spray, known as the Wurster system, uses a coating chamber that has a cylindrical steel nozzle (used to spray the coating material) and a cribriform bottom plate, coating small particles (100 μm). This multilayer coating procedure helps to reduce particle defects, although it is a time-consuming process. On the other hand, tangential spray consists of a coating chamber with a rotating bottom of the same diameter as the chamber. During the process, the drum is raised to create a space between the edge of the chamber and the drum. A tangential nozzle is placed above the rotating drum, where the coating material is released. Then, the particles move through the space into the spray zone and are finally encapsulated [191]. During this process, there are three mechanical forces, namely, centrifugal force, lifting force, and gravity. The particles to be coated must be spherical and dense, and must have a narrow size distribution and perfect fluidity, with the non-spherical particles having the largest possible surface area and requiring more coating material. This technique has a low operating cost and a high thermal efficiency process, allowing total temperature control. However, it can be time consuming, which becomes a disadvantage [173]. 3.9. Layer-by-Layer Self-Assembly Layer-by-layer (LbL) is a relatively simple and promising technique for the encapsulation, stabilisation, storage, and release of several active compounds [192]. This method consists of alternating the adsorption of oppositely charged wall materials through many intermolecular interactions onto a charged substrate (Figure 17). The microcapsules have good chemical and mechanical stability through a formation mechanism constituted by irreversible electrostatic interactions that allow the adsorption of successive layers of polyelectrolytes [193]. The adsorption of the layers is normally carried out by immersing the suspension in alternate solutions of cationic and anionic polymers, with washing processes being carried out after the deposition of each layer [194]. This technique has significant advantages over other microencapsulation methods, because it allows the control of the permeability, morphology, composition, size, and wall thickness of the microcapsules by adjusting the number of layers and experimental conditions [195]. Controlling these parameters allows a better adaptation of the microcapsule to its functionality in the target application. However, most LbL systems have some restrictions in terms of biocompatibility [196]. 4. Microencapsulation of Essential Oils Microencapsulation is an alternative that can be utilised to overcome several limitations in the application of essential oils. This application is profoundly affected by the high volatility and chemically unstable nature of EOs [198]. In addition, EOs are compounds that can be easily degraded due to interactions with other chemical components and exposure to several factors such as light, temperature, and oxygen [166]. Essential oils can be “trapped” in microcapsules, which act as micro-reservoirs, ensuring excellent protection [199]. The encapsulation process, where small particles are enclosed in solid carriers to increase their protection, has the ability to reduce evaporation, promote easier handling, and control the release of essential oils during storage and application [199]. Furthermore, through microencapsulation, it is possible to change the appearance of EOs (which behave like a powder), without changing their structure and properties [177]. In EO microencapsulation, the first step is often to emulsify or disperse the essential oils in an aqueous solution of a wall material, which also acts as an emulsifier. This process happens because the EOs exist in liquid form at room temperature. Then, the resulting microcapsules must be dried under controlled conditions, so that the loss of the encapsulated material by volatilisation is reduced [177]. One of the areas that has also aroused interest in the microencapsulation of EOs is in the agrochemical industry. Yang et al. prepared and characterised microcapsules based on polyurea, containing essential oils as an active agent for possible applications in the controlled release of agrochemical compounds [200]. The microcapsules were synthesised by O/W emulsion interfacial polymerisation and the synthetic conditions that showed the best results were used to encapsulate four essential oils (lemongrass, lavender, sage, and thyme), capable of interfering with the seed germination and root elongation of some plants. In cases of pest control, biological pesticides must be more effective than synthetic pesticides. Bagle et al. reported success in encapsulating neem oil, an effective biological insecticide, in phenol formaldehyde (PF) microcapsules [201]. The microcapsules were obtained using an in situ polymerisation process in an O/W emulsion and their size was determined using a particle size analyser. Controlled release was monitored by measuring optical observations in the UV range. Figure 18 shows scanning electron microscopy (SEM) micrographs of PF microcapsules containing neem oil. It was possible to visualise that the PF microcapsules were spherical and globular, with diameters between 30 and 50 µm at 400–500 rpm. The microcapsules’ surface was considered quite smooth and can be useful regarding the protection and sustained release of the neem oil inside. The chemical constitution of synthesised microcapsules was confirmed by Fourier-transform infrared spectroscopy (FTIR), and it was found to be a good thermal stability of MCs needed for the long-term preservation of the core, and it was concluded that neem oil can be better preserved in PF microcapsules. The controlled release behaviour of PF microcapsules containing neem oil was studied and the experimental data are shown in Figure 19. A release of about 30% was observed after 6 h, confirmed by the decrease in absorbance over time. Like neem oil, other essential oils also have insecticidal properties, such as Rosmarinus officinalis and Zataria multiflora (Lamiaceae), that can be used as pesticides for stored-product pests. In the study carried out by Ahsaei et al., these oils were encapsulated in octenyl succinic anhydride (OSA) starch to test their insecticidal activity against Tribolium confusum. The microcapsules were obtained using an O/W emulsion and dried using the spray drying technique [202]. The solid formulations were characterised by particle size, encapsulation efficiency, and water activity. The release rate under storage conditions was measured over a period of 40 days, and the insecticidal activity against T. confusum was determined using specific bioassays. It was concluded that the encapsulation efficiency depends directly on the surfactant-to-oil ratio. Regarding the morphology of microcapsules loaded with OEs, SEM micrographs reveal the presence of oval and spherical microcapsules with irregular surfaces. The microcapsules appear to be devoid of cracks or fractures, which is an advantageous feature for protecting the oil. The results also showed an optimised release of pesticides from controlled release formulations, which maximises their biological activity for a longer time. The food sector is probably the sector where the microencapsulation of essential oils is most explored, with the encapsulation of flavours being one of the great interests of this industry. Flavours are necessary for some foods, to promote consumer satisfaction and the consumption of those products. Nevertheless, the flavour stability in foods has been a challenge for this sector in order to achieve quality and acceptability. For the encapsulation of a flavour, Fernandes et al. evaluated, by spray drying, the effects of the partial or total substitution of arabic gum with modified starch, maltodextrin, and inulin in the encapsulation of rosemary essential oil [203]. Regarding the characterisation of microcapsules, moisture content, wettability and solubility, density and apparent density, and oil retention was determined. From SEM observations (Figure 20), the authors found that there was no evidence of cracking in the particles using any of the encapsulating formulations, ensuring low gas permeability and thus better protecting the EO of rosemary. Differences were observed in the surface of each type of particle, showing that the particles have a spherical shape. It was concluded that the total substitution of arabic gum with modified starch or a mixture of modified starch and maltodextrin did not affect the efficiency of the encapsulation, increasing the possibility of developing new formulations of encapsulants. With the addition of inulin, the oil retention of particles decreased. However, the combination of modified starch and inulin was shown to be a viable substitute for arabic gum in foods. A group of researchers compared the release properties of three different microcapsules, namely gelatin microcapsules loaded with holy basil essential oil (HBEO) (designated as UC), UC coated with aluminium carboxymethylcellulose (CC), and UC coated with aluminium compound carboxymethyl cellulose–beeswax (CB) [204]. To be applied as a feed additive, the HBEO was encapsulated in order to be a potential alternative to antibiotic growth promoters (AGP). However, its benefits depend on the available amount in the gastrointestinal tract. The SEM technique was used to characterise the internal and external factors of the microcapsule surface morphology. According to Figure 21, UC microcapsules (Figure 21a) are almost spherical in shape and after coating, the CC (Figure 21b) and CB (Figure 21c) microcapsules are more spherical. Upon magnification of these micrographs, it was possible to verify that UC microcapsules have a spongy structure (Figure 21d) and that CC (Figure 21e) and CB (Figure 21) microcapsules are denser. When cut transversely, UC microcapsules seem to have a gelatinous morphology (Figure 21g), whereas the CC (Figure 21h) and CB microcapsules (Figure 21i) reveal a thicker and more compact outer coating layer with a honeycomb structure. This method of encapsulation demonstrated an effective process for improving HBEO efficacy for pathogen reduction in the distal region of the intestine. Regarding food safety, the use of antimicrobial packaging materials offers the potential to retard the growth rate of spoilage microorganisms. The physical and antimicrobial properties of nanofibres manufactured for active packaging systems were studied by Munhuweyi et al. [205]. Microcapsules and active nanofibres derived from the precipitation of β-cyclodextrin (β-CD) with essential oils of cinnamon and oregano were developed and their antifungal activity in vitro against Botrytis sp. was examined. To induce microencapsulation, the solutions were subjected to co-precipitation. It was verified that cinnamon microcapsules have greater antimicrobial efficacy when compared to oregano. As food preservatives, this microencapsulation system could have promising applications in the development of active packaging systems. Using the thermogravimetric analysis (TGA) technique, the initial weight loss for simple β-CD occurred at ~100 °C and the greatest weight loss at ~330 °C (Figure 22a). The degradation temperature of β-CD in the CIN/β-CD and OREG/β-CD complexes decreased from ~330 °C to ~270 °C (Figure 22b,c). Comparing the TGA curves, there is a difference between them, demonstrating the presence of chemical and guest molecule interaction in the complex. Using the simple coacervation method, Leimann et al. encapsulated lemongrass, which is known for its broad spectrum antimicrobial activity [206]. Poly(vinyl alcohol) crosslinked with glutaraldehyde was used as the wall-forming polymer. The influence of the agitation rate and the fraction of oil volume on the microcapsule size distribution was evaluated. Sodium dodecyl sulphate (SDS) and poly(vinyl pyrrolidone) were tested to prevent the agglomeration of microcapsules during the process. The microcapsules did not show agglomeration when 0.03% by weight of SDS was used. The composition and antimicrobial properties of the encapsulated oil were determined, demonstrating that the microencapsulation process did not deteriorate the encapsulated essential oil. Cyclodextrins (CDs) are important supramolecular microcapsule hosts in foods and other fields, and the essential oil of Laurus nobilis (LEO) has natural antioxidant properties in food due to its main constituents being terpenic alcohols and phenols. For these reasons, Li et al. isolated LEO by microwave-assisted hydrodistillation [207]. The authors prepared chitosan (CS) microcapsules loaded with citrus essential oils (CEOs: D-limonene, linalool, a-terpinene, myrcene, and a-pinene) using six different emulsifiers (Tween 20, Tween 40, Tween 60, Tween 60/Tween 20/Span 80 1:1, Tween 20/sodium dodecyl benzene sulfonate (SDBS) 1:1, Span 80) through an emulsion gelation technique [208]. After preparing β-cyclodextrin (β-CD) microcapsules and their derivatives, several affecting factors were examined in detail. Figure 23 shows the total antioxidant activity of LEO. LEO caused Mo (VI) to be deoxidised to become Mo (V) through a mechanism of total antioxidant activity. Mo (V) exhibits maximum absorption at 695 nm and has a stronger antioxidant activity; the greater the concentration of Mo (V) solution, the greater the absorbency of the solution. With the increase in absorbance of the solutions, there was an increase in the concentrations of the sample, causing the antioxidant activity to increase significantly. The microcapsules were analysed and the results indicate that the choice of emulsifier significantly affects the size and effectiveness of incorporating the microcapsules. Figure 24a presents the FTIR spectra observed in CS, CEOs, and four groups of microcapsules prepared with different emulsifiers. In the CEOs curve, the peak at 886 cm−1 corresponds to the absorption of limonene. The strong methylene/methyl band occurs at 1435 cm−1, and at 1646 cm−1, the C=O stretching vibration appears. Peaks corresponding to the asymmetric and symmetrical modes of the CH2 elongation vibration appear for Span 80 and Tween 60, and the new connections can be seen at 2922 cm−1. Through these results, it was possible to observe that the CEOs were incorporated in the microcapsules, showing benefits for inhibiting them from oxidation and volatilisation. A second step in the characterisation of the microcapsules was the analysis of the crystallographic structure. Through X-ray diffraction (XRD) analysis (Figure 24b), it was possible to observe that CS exhibits a diffraction pattern with a broad band centred at 2θ 20°, thus indicating the existence of an amorphous structure. Comparing the CS with the microcapsule groups, the latter exhibit a significant reduction in this broad band. This reduction in intensity is due to the destruction of the CS structure, which can be attributed to a change in the arrangement of the molecules in the crystalline chain. To develop a new use of functional EOs, Karimi Sani et al. studied the influence of process parameters on the characteristics of microencapsulated essential oil Melissa officinalis using whey protein isolate (WPI) and sodium caseinate (NaCS). The impacts of these variables were examined using the response surface methodology. Smaller particle sizes were obtained for higher amounts of WPI with the lowest level of applied sonication power. The results of the desirability function indicate that the maximum amount of WPI with an ultrasound power of 50 W led to the smallest particle size and the lowest zeta potential and turbidity. In this study, the ultrasonic technique showed potential in the use of milk proteins to produce microparticles with OEs. The obtained results showed that the microcapsules loaded with Melissa officinalis can preserve the bioactive compounds and induce flavour stability, enabling their use in food formulations and pharmaceutical products. Mehran et al. carried out a study of the microencapsulation of spearmint essential oil (SEO), using a mixture of inulin and arabic gum as wall material in order to be used in the food and pharmaceutical industry [209]. The technique used for the formation of the microcapsules was spray drying. The microcapsules were characterised for oil retention, encapsulation efficiency, hygroscopicity, and carbon content, having as ideal conditions 35% solid wall, 4% essential oil concentration, and 110 °C inlet temperature, with maximum retention of 91% of oil. To confirm that the SEO was encapsulated, this group of researchers used differential scanning calorimetry (DSC) and FTIR characterisation techniques. The infrared spectra of pure SEO, pure matrix (containing inulin and arabic gum), and microcapsules are shown in Figure 25. In the SEO spectrum, the characteristic peaks at 801 cm−1 and 894 cm−1 are ascribed to =CH vibrations. The C-O-C elongation corresponds to the peak at 1109 cm−1 and the C=O elongation corresponds to the peak 1675 cm−1. In the matrix spectrum, a wide band at 3392 cm−1 is related to the hydroxylated group. In relation to the peak at 1030 cm−1, it can be associated with the strong absorption bands of the C-O-C elongation. In the microcapsule spectrum, it can be observed that it is quite similar to the matrix, and that the peaks related to the SEO disappear or are absent, which may be related to the overlap of the peaks of the matrix and SEO due to the low weight fraction of SEO in the total weight of the microcapsules. Through this spectrum, it was possible to verify the successful encapsulation of the SEO (peaks at 1673 cm−1 and 900 cm−1). The double barrier release system is a method used for essential oils that have antifungal activity, even against drug-resistant fungi. However, there are some limitations due to the sensitivity to pH, temperature, and light. Adepu et al. encapsulated three essential oils (thymol, eugenol, and carvacrol) in a polylactic acid shell with high encapsulation efficiency to achieve their synergistic antifungal activity using the coacervation phase separation method. These were incorporated into bacterial cellulose (a nanofibre fibrous hydrogel) [210]. An antifungal test was performed on the Candida albicans fungus model (a cause of common oral and vaginal infections). Another test was carried out—a transvaginal drug release study in vitro—to compare the release of microcapsules like colloids and composites, where the latter exhibited a controlled release. Through several studies, such as the SEM technique, it was found that the average size and size distribution of the microcapsules depends on the concentration of the used polymer (poly(lactic acid)) and surfactant (poloxamer). SEM images of BC loaded with microcapsules demonstrate a regular distribution and spherical shape, appearing to be well separated and stable in the stages of the preparation process (Figure 26). From the highest magnification image, it was observed that the microcapsules were anchored to the nanofibre matrix. Repellent essential oils are becoming increasingly widespread due to their low toxicity and customer approval. Its application in textile materials has been widely developed. To optimise their application efficiency, it is important to develop long-lasting repellent textiles using OEs. Specos et al. obtained citronella-loaded gelatin microcapsules through the complex coacervation method, which were applied to cotton fabrics in order to study the repellent effectiveness of the obtained fabrics [70]. The release of citronella by the treated tissues was monitored and the repellent activity evaluated by exposing a human hand and arm covered with the treated tissues to Aedes aegypti mosquitoes. It was found that the tissues treated with citronella microcapsules present greater and more lasting protection against insects in comparison to fabrics sprayed with an ethanol solution of essential oil. Repellent textiles were obtained by filling cotton fabrics with microcapsule sludge, using a conventional drying method. This methodology does not require additional investments for the textile finishing industries, which is a desirable factor in developing countries. Figure 27A shows the morphology of blackberry-type microcapsules in a fresh paste with diameters ranging from 25 to 100 µm, while Figure 27B shows SEM micrographs of spray-dried microcapsules revealing two types of structures, with small spherical units of less than 10 µm and clusters ranging from 25 to 100 µm. In 2016, Ribeiro et al. investigated the functionalisation of photocatalytic titanium dioxide nanoparticles on the surface of polymeric microcapsules as a way to control the release of citronella by solar radiation, thus obtaining a release of a repellent without mechanical intervention [211]. These authors used a modified hydrothermal sol-gel method to synthesise TiO2 nanoparticles. Through several characterisation techniques, these authors were able to observe the surface of the microcapsules and the release efficiency. Using in vitro biological assays with live mosquitoes, the controlled release repellence effect of these photocatalytic microcapsules was reinforced by the inhibition of these vectors. According to the results, it was shown that functionalising the microcapsules with photocatalytic nanoparticles on the surface, and then exposing them to ultraviolet radiation, effectively increased the emission of citronella into the air, repelling mosquitoes. Table 6 shows an overview of illustrative examples of EO microencapsulation oils, wall materials, and microencapsulation methods with industrial importance. 5. Conclusions This review summarises different types of EO structures and describes their extraction and application methodology. In addition, different techniques for microencapsulating essential oils are described and some reports are presented to provide a basis for research and industrial development. As described in this paper, EOs are used in several applications in the pharmaceutical, cosmetic, agricultural, and food industries, as they are natural metabolites produced by plants with interesting properties. Furthermore, EOs are being explored as an alternative to synthetic products due to their ecological factors and the fact that their characteristics are different from the corresponding synthetic product. For example, synthesised oil may have the same odour as natural oils, but may not have the same therapeutic characteristics. Currently, there is growing interest in the application of EO microencapsulation, making it an effective and important tool in the preparation of high-quality products, improving their chemical, oxidative, and thermal stability. Besides these advantages, the shelf life, biological activity, functional activity, controlled release, physicochemical properties, and general quality of oils can also be improved with microencapsulation technology. Based on the scientific studies available and presented throughout this paper, it can be concluded that the microencapsulation of EOs is an emerging trend for industrial applications. However, this development has limitations, such as the low diversity of wall materials and their incompatibility with microencapsulation methods. Many of the encapsulating agents available present a high cost for production on an industrial scale. In future research, microencapsulation must also be directed to encapsulate a different mixture of oils by different techniques, in order to disguise the flavour of the oils and to improve safety, quality, and nutritional value. Funding Vânia I. Sousa and Joana F. Parente are grateful to the Project ReleaseME-POCI-01-0247-FEDER-033268, for their research grants from the Agência Nacional de Inovação, co-funded by the European Regional Development Fund (ERDF), through the Operational Programme for Competitiveness and Internationalisation (COMPETE 2020), under the PORTUGAL 2020 Partnership Agreement. Juliana F. Marques and Marta A. Forte are grateful to the Fundação para a Ciência e Tecnologia (FCT) of Portugal for their Ph.D. grants, SFRH/BD/112868/2015 and PD/BD/128491/2017, respectively. The authors also acknowledge the funding from FCT/PIDDAC through the Strategic Funds project reference UIDB/04650/2020-2023. This research was funded by the project Repel+: New solutions for mosquito repellence as an application for malaria control (project number 47036) from the Agência Nacional de Inovação, co-funded by the European Regional Development Fund (ERDF), through the Operational Programme for Competitiveness and Internationalisation (COMPETE 2020), under the PORTUGAL 2020 Partnership Agreement. Data Availability Statement The raw/processed data required to reproduce these findings cannot be shared at this time due to technical or time limitations. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Biological activities of essential oils. Figure 2 Structures of terpenes and terpenoids: acyclic monoterpenes (2a), cyclic monoterpenes (2b), diterpenes (2c), triterpenes (2d), and terpenoids (2e). Figure 3 Schematic representation of hydrodistillation. Figure 4 Experimental setup used in steam distillation. Figure 5 Schematic representation of organic solvent extraction using the Soxhlet method. Figure 6 Schematic representation of a microcapsule and a microsphere. Figure 7 Objectives of microencapsulation. Figure 8 Main microencapsulation methods. Figure 9 Illustration of emulsion systems. Figure 10 Schematic illustration of the coacervation method. Figure 11 Schematic illustration in situ polymerisation method (adapted from [168]). Figure 12 Schematic representation of spray drying. Figure 13 Schematic diagram of a freeze dryer (adapted from [163]). Figure 14 Flow chart of supercritical fluid technology. Figure 15 Schematic representation of microencapsulation process by coaxial electrospraying (adapted from [144]). Figure 16 Schematic representation of bottom spray fluidised bed coating process (adapted from [166]). Figure 17 Layer-by-layer (LbL) self-assembly microcapsules (adapted from [197]). Figure 18 SEM micrographs (a–c) of phenol formaldehyde microcapsules containing neem oil [201]. Figure 19 Controlled release of core material over time [201]. Figure 20 Scanning electron micrographs of the particles containing rosemary essential oil [203]. (A): arabic gum; (B): arabic gum/maltodextrin; (C): arabic gum/inulin; (D): starch; (E): modified starch/maltodextrin; (F): modified starch/inulin. Figure 21 SEM micrographs of UC, CC, and CB gelatin-based microcapsules: (a) whole UC; (b) whole CC; (c) whole CB; (d) external surface of UC; (e) external surface of CC; (f) external surface of CB; (g) inner edge of UC; (h) inner edge of CC; (i) inner edge of CB [204]. Figure 22 TGA curves of (a) plain β-cyclodextrin (β-CD), (b) microencapsulated cinnamon (CIN/β-CD), and (c) oregano (OREG/β-CD) [205]. Figure 23 Absorbance of antioxidants in different concentrations of 2,6-ditert-butylphenol (BHT, square), propyl gallate (PG, circle) and Laurus nobilis essential oil (LEO, triangle) [207]. Figure 24 (a) FTIR spectroscopy, (b) X-ray diffraction of pure chitosan (CS), control group (CK), Tween 60, Tween 20/Span 80, and Span 80 [208]. Figure 25 FTIR spectra of (A) pure SEO, (B) pure matrix, and (C) inulin and arabic gum-based microcapsules [209]. Figure 26 Low- and high-magnification SEM micrographs of (a,a’) BC-PLA1.5-Pol5.0, (b,b’) BC-PLA1.5-Pol2.5, (c,c’) BC-PLA3.0-Pol5.0, and (d,d’) BCPLA3.0-Pol2.5 [210]. Figure 27 (A) Optical micrographs of gelatin microcapsules containing citronella essential oil (100× magnification) and (B) SEM microphotographs of spray-dried microcapsules containing citronella essential oil (500× magnification) [70]. polymers-14-01730-t002_Table 2 Table 2 Advantages and disadvantages of each essential oil extraction method [46]. Type of Method Method Advantages Disadvantages Conventional Hydrodistillation - Versatile and simple; - Easy implementation; - Selectivity. - Complete extraction is not possible; - High energy consumption; - Long extraction time. Steam distillation - Extraction time and loss of polar molecules are reduced (compared with hydrodistillation). - Longer extractions; - Present non-appreciable and higher cost compounds due to the long process time. Organic solvent extraction - Simple, cheap, and reasonably efficient; - Appropriate for small scale. - Time consuming; - High solvent consumption; - Does not allow agitation to speed up the process; - Organic solvents can cause chemical changes or toxic effects in final product. Cold pressing - Simple and inexpensive; - Suitable for the production of citrus oils. - Oil extraction is not complete; - Not feasible for low-oil samples. Innovative Supercritical fluid extraction - Reduced time; - Low toxicity solvents; - Solvent-free extract. - High cost of equipment, installation, and maintenance operations. Microwave-assisted extraction - High reproducibility; - Simple manipulation; - Low solvent consumption; - Lower energy input; - Improved extraction yield. - Filtration or centrifuging required at the end. Ultrasound-assisted extraction - Simple and inexpensive (compared to SCFE and MAE); - Reduced extraction time; - Low solvent consumption; - Mass transfer intensification; - Improvement of solvent penetration. - Difficult to scale up; - High power consumption. polymers-14-01730-t003_Table 3 Table 3 Essential oil plants, species, and main components, and their molecular structure and biological properties. Plant Species EO Major Components Chemical Structures of EOs Components Some Biological Properties Bay Laurel Laurus nobilis Cineol, pinene, linalool, terpineol acetate, methyleugenol Antioxidant, antibacterial, antifungal, insect repellent Citronella Cymbopogon nardus geraniol, citronellal, geranyl acetate, limonene, camphene Antimicrobial, antifungal, antioxidant, antibacterial, insect and stimulate repellent Hops Humulus lupulus Humulene, myrcene, caryophyllene, farnesene Antibacterial, antifungal, anti-cancer, repellent Lemon Balm Melissa officinalis Geraniol, citral, citronellol, eugenol, linalyl acetate Antibacterial, antifungal, antimicrobial, sedative, antipyretic, antispasmodic, anti-hypertensive, anti-Alzheimer, antiseptic Neem Azadirachta indica Azadiractin Insecticide, antiviral, antibacterial, antimicrobial Pennyroyal Mentha Pulegium Pulegone, menthol, iso-mentone, octanol, piperitenone, trans-iso-pulegone Insect repellent, antimicrobial, antioxidant, antibacterial, anti-tumour Star Anise Illicium verum Trans-anethole Antimicrobial, antioxidant, diuretic, anti-cancer potential, insect repellent Valerian Valeriana officinalis Borneol, camphene, α and β-pinene, valeranone, valerenol Antibacterial, antimicrobial, antifungal, antioxidative, sedative polymers-14-01730-t004_Table 4 Table 4 Some pharmacological actions of essential oils [114]. Condition Essential Oil Anxiety, agitation, stress, challenging behaviours Angelica, labdanum, bergamot, sweet orange, palmarosa, lavender, basil, geranium, and valerian End-of-life agitation Lavender, sandalwood, and frankincense Fatigue Labdanum, grapefruit, coriander, citronella, black peppermint, gully gum, spearmint, geranium, Scots pine, clary, and ginger Insomnia Angelica, lemon, mandarin, sweet orange, lavender, lemon balm, myrtle, basil, sweet marjoram, and valerian Mental exhaustion, burnout Peppermint, basil, and everlasting Memory loss May Chang, peppermint, and rosemary Pain management Gully gum, lavender, German chamomile, sweet marjoram, rosemary, and ginger polymers-14-01730-t005_Table 5 Table 5 Pests/pesticides and their corresponding essential oil [145]. Pests Essential Oil Ants Peppermint, mint Aphids Cedar, hyssop, peppermint, mint Beetles Peppermint, thyme Caterpillars Peppermint, mint Mites Lavender, lemongrass, sage, thyme Fleas Peppermint, lemon grass, mint, lavender Flies Lavender, mint, rosemary, sage Mosquitoes Patchouli, mint Lice Cedar, peppermint, mint Moths Cedar, hyssop, lavender, peppermint, mint Slugs Cedar, hyssop, pine Snails Cedar, pine, patchouli Spiders Peppermint, mint Ticks Lavender, lemongrass, sage, thyme polymers-14-01730-t006_Table 6 Table 6 Overview of essential oil microencapsulation, methods, wall materials, and industrial applications. Microencapsulation Method Wall Material(s) Essential Oil/Source Applications Reference Emulsification Hydroxypropyl methyl cellulose/ chitosan/silica Peppermint oil Medical [212] Polydopamine Turpentine Agrochemical [213] β-cyclodextrin Thyme Food [214] β-cyclodextrin/sugar beet pectin Garlic Food [215] Ionic Gelation Cassava starch/poly(butylene adipate-co-terephthalate) Oregano Food [216] Simple Coacervation Gelatin Basil Agrochemical [217] Poly(vinyl alcohol) Lemongrass Food and pharmaceutical [206] Complex Coacervation Gelatin/gum arabic Citronella Anti-mosquito textile [149] Gelatin/sodium alginate Citronella Anti-mosquito textile [218] Gelatin/gum arabic Lavender Cosmetics [219] Chitosan/gum arabic/maltodextrin Peppermint Cosmetics [220] Chitosan/k-carrageenan Pimenta dioica Food [221] Gelatin/chia mucilage Oregano Food [222] Mung bean protein isolate/apricot peel pectin Rose Food [223] Sichuan pepper seed soluble dietary fibre/soybean protein isolate Sichuan pepper Food [224] Whey protein isolate/arabic gum Orange Food [225] Interfacial Polymerization Polyurea Lemongrass, lavender, sage and thyme Agrochemical [200] In situ polymerisation Silicon dioxide/poly(melamine formaldehyde) Cinnamon Agrochemical [226] Melamine/formaldehyde Thyme Food [169] Melamine/formaldehyde Lavandin and tea tree Paints [227] Extrusion Sodium alginate Nutmeg Pharmaceutical [228] Sodium alginate Rosemary Agrochemical [229] Spray Drying Acacia gum Citronella Cosmetics [230] Palm trunk/ chitosan Ginger Food [231] Maltodextrin Citrus Food [232] Gum arabic/maltodextrin/sodium alginate/whey protein concentrate Juniper berry Food [233] Whey protein isolate/maltodextrin/sodium alginate Cinnamon Food [234] Hydroxypropyl methyl cellulose/maltodextrin Oregano Food [235] Gelatin/arabic gum Citronella Textile [70] Gum arabic/starch/maltodextrin/inulin Rosemary Food [203] Gum arabic/modified starch Black pepper Food [236] Gum arabic/maltodextrin/whey protein isolate Basil Food [237] Freeze-drying Urushiol Not mentioned Medical and pharmaceutical [238] β-cyclodextrin Litsea cubeba Cosmetics and pharmaceutical [239] Maltodextrin/gelation Lemongrass Cosmetics, pharmaceutical and food [240] Gum arabic/collagen hydrolysate Origanum onites L. Environmental [241] Whey protein isolate/carboxymethylcellulose Orange Food [242] Supercritical Fluid Technology Starch Oregano Food [243] n-octenyl succinic/modified starches Lavandin Agrochemical [244] Modified starch Limonene Food [245] Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Benchaar C. Greathead H. Essential oils and opportunities to mitigate enteric methane emissions from ruminants Anim. Feed Sci. Technol. 2011 166–167 338 355 10.1016/j.anifeedsci.2011.04.024 2. Benchaar C. Calsamiglia S. Chaves A.V. Fraser G.R. Colombatto D. McAllister T.A. Beauchemin K.A. A review of plant-derived essential oils in ruminant nutrition and production Anim. Feed Sci. Technol. 2008 145 209 228 10.1016/j.anifeedsci.2007.04.014 3. Raut J.S. Karuppayil S.M. A status review on the medicinal properties of essential oils Ind. Crops Prod. 2014 62 250 264 10.1016/j.indcrop.2014.05.055 4. Lubbe A. Verpoorte R. Cultivation of medicinal and aromatic plants for specialty industrial materials Ind. Crop. Prod. 2011 34 785 801 10.1016/j.indcrop.2011.01.019 5. daCunha A.P. Nogueira M.T. Rodrigues Roque O. Barroso J.M.G. Plantas Aromáticas e Óleos Essenciais: Composição e Aplicações Fundação Calouste Gulbenkian Lisbon, Portugal 2012 978-972-31-1450-8 6. Teixeira B. Marques A. Ramos C. Neng N.R. Nogueira J.M.F. Saraiva J.A. Nunes M.L. Chemical composition and antibacterial and antioxidant properties of commercial essential oils Ind. Crops Prod. 2013 43 587 595 10.1016/j.indcrop.2012.07.069 7. Sacchetti G. Maietti S. Muzzoli M. Scaglianti M. Manfredini S. Radice M. Bruni R. Comparative evaluation of 11 essential oils of different origin as functional antioxidants, antiradicals and antimicrobials in foods Food Chem. 2005 91 621 632 10.1016/j.foodchem.2004.06.031 8. Bakkali F. Averbeck S. Averbeck D. Idaomar M. Biological effects of essential oils—A review Food Chem. Toxicol. 2008 46 446 475 10.1016/j.fct.2007.09.106 17996351 9. Warnke P.H. Becker S.T. Podschun R. Sivananthan S. Springer I.N. Russo P.A.J. Wiltfang J. Fickenscher H. Sherry E. The battle against multi-resistant strains: Renaissance of antimicrobial essential oils as a promising force to fight hospital-acquired infections J. Cranio-Maxillofac. Surg. 2009 37 392 397 10.1016/j.jcms.2009.03.017 10. De Billerbeck V.G. Huiles essentielles et bactéries résistantes aux antibiotiques Phytotherapie 2007 5 249 253 10.1007/s10298-007-0265-z 11. Burt S. Essential oils: Their antibacterial properties and potential applications in foods—A review Int. J. Food Microbiol. 2004 94 223 253 10.1016/j.ijfoodmicro.2004.03.022 15246235 12. Nakatsu T. Lupo A.T. Chinn J.W. Kang R.K.L. Biological activity of essential oils and their constituents Stud. Nat. Prod. Chem. 2000 21 571 631 10.1016/S1572-5995(00)80014-9 13. Dima C. Dima S. Essential oils in foods: Extraction, stabilization, and toxicity Curr. Opin. Food Sci. 2015 5 29 35 10.1016/j.cofs.2015.07.003 14. Hogenbom J. Jones A. Wang H.V. Pickett L.J. Faraone N. Synthesis and characterization of β-cyclodextrin-essential oil inclusion complexes for tick repellent development Polymers 2021 13 1892 10.3390/polym13111892 34200230 15. Ponce Cevallos P.A. Buera M.P. Elizalde B.E. Encapsulation of cinnamon and thyme essential oils components (cinnamaldehyde and thymol) in β-cyclodextrin: Effect of interactions with water on complex stability J. Food Eng. 2010 99 70 75 10.1016/j.jfoodeng.2010.01.039 16. Saberi-Riseh R. Moradi-Pour M. Mohammadinejad R. Thakur V.K. Biopolymers for biological control of plant pathogens: Advances in microencapsulation of beneficial microorganisms Polymers 2021 13 1938 10.3390/polym13121938 34200966 17. Lengyel M. Kállai-Szabó N. Antal V. Laki A.J. Antal I. Microparticles, microspheres, and microcapsules for advanced drug delivery Sci. Pharm. 2019 87 20 10.3390/scipharm87030020 18. Council of Europe European Directorate for the Quality of Medicines & HealthCare (EDQM) European Pharmacopoeia 10th ed. EDQM Strasbourg, France 2021 19. Svoboda K.P. Greenaway R.I. Investigation of volatile oil glands of Satureja hortensis L. (summer savory) and phytochemical comparison of different varieties Int. J. Aromather. 2003 13 196 202 10.1016/S0962-4562(03)00038-9 20. Dajic Stevanovic Z. Sieniawska E. Glowniak K. Obradovic N. Pajic-Lijakovic I. Natural Macromolecules as Carriers for Essential Oils: From Extraction to Biomedical Application Front. Bioeng. Biotechnol. 2020 8 1 24 10.3389/fbioe.2020.00563 32039188 21. Figueiredo A.C. Pedro L.G. Barroso J.G. Plantas aromáticas e medicinais: Factores que afetam a produção Potencialidades e Aplicações das Plantas Aromáticas e Medicinais Plant Biotechnology Center—Faculty of Sciences of the University of Lisbon Lisbon, Portugal 2007 978-972-9348-16-7 22. Peixoto L.R. Rosalen P.L. Ferreira G.L.S. Freires I.A. de Carvalho F.G. Castellano L.R. de Castro R.D. Antifungal activity, mode of action and anti-biofilm effects of Laurus nobilis Linnaeus essential oil against Candida spp. Arch. Oral Biol. 2017 73 179 185 10.1016/j.archoralbio.2016.10.013 27771586 23. Eslahi H. Fahimi N. Sardarian A.R. Chemical Composition of Essential Oils: Chemistry, Safety and Applications Essential Oils in Food Processing: Chemistry, Safety and Applications Hashemi S.M.B. Khaneghah A.M. de Souza Sant’Ana A. Wiley Hoboken, NJ, USA 2017 119 171 9781119149392 24. Dhifi W. Bellili S. Jazi S. Bahloul N. Mnif W. Essential Oils’ Chemical Characterization and Investigation of Some Biological Activities: A Critical Review Medicines 2016 3 25 10.3390/medicines3040025 25. Chang T.H. Hsieh F.L. Ko T.P. Teng K.H. Liang P.H. Wang A.H.J. Structure of a heterotetrameric geranyl pyrophosphate synthase from mint (Mentha piperita) reveals intersubunit regulation Plant Cell 2010 22 454 467 10.1105/tpc.109.071738 20139160 26. Noriega P. Terpenes in Essential Oils: Bioactivity and Applications Terpenes and Terpenoids—Recent Advances Perveen S. Al-Taweel A.M. Licensee IntechOpen London, UK 2020 27. Neelam Khatkar A. Sharma K.K. Phenylpropanoids and its derivatives: Biological activities and its role in food, pharmaceutical and cosmetic industries Crit. Rev. Food Sci. Nutr. 2020 60 2655 2675 10.1080/10408398.2019.1653822 31456411 28. Liu H.-W.B. Begley T.P. Comprehensive Natural Products III 3rd ed. Elsevier Amsterdam, The Netherlands 2020 978-0-08-102691-5 29. Grassmann J. Elstner E.F. Essential Oils—Properties and Uses Encyclopedia of Food Sciences and Nutrition Caballero B. Academic Press Cambridge, MA, USA 2003 978-0-12-227055-0 30. Stratakos A. Koidis A. Essential Oils in Food Preservation, Flavor and Safety Methods for Extracting Essential Oils Preedy V.R. Academic Press Cambridge, MA, USA 2016 31 38 31. Silva M.G.F. Atividade antioxidante e antimicrobiana in vitro de óleos essenciais e extratos hidroalcóolicos de manjerona (Origanum majorana L.) e manjericão (Ocimum basilicum L.) Bachelor’s Thesis Graduation Conclusion Work Federal Technological University of Paraná Londrina, Brazil December 2011 32. Sartor R.B. Modelagem, Simulação e Otimização de uma Unidade Industrial de Extração de Óleos Essenciais por Arraste a Vapor Universidade Federal do Rio Grande do Sul Porto Alegre, Brazil 2009 33. Masango P. Cleaner production of essential oils by steam distillation J. Clean. Prod. 2005 13 833 839 10.1016/j.jclepro.2004.02.039 34. Luque de Castro M.D. García-Ayuso L.E. Soxhlet extraction of solid materials: An outdated technique with a promising innovative future Anal. Chim. Acta 1998 369 1 10 10.1016/S0003-2670(98)00233-5 35. Elyemni M. Louaste B. Nechad I. Elkamli T. Bouia A. Taleb M. Chaouch M. Eloutassi N. Extraction of Essential Oils of Rosmarinus officinalis L. by Two Different Methods: Hydrodistillation and Microwave Assisted Hydrodistillation Sci. World J. 2019 2019 3659432 10.1155/2019/3659432 36. Rezaei K. Hashemi N.B. Sahraee S. Use of Hydrodistillation as a Green Technology to Obtain Essential Oils From Several Medicinal Plants Belonging to Lamiaceae (Mint) Family Phytopharmaceuticals Wiley Hoboken, NJ, USA 2021 59 75 10.1002/9781119682059.ch3 37. Atti-Santos A.C. Rossato M. Serafini L.A. Cassel E. Moyna P. Extraction of essential oils from lime (Citrus latifolia tanaka) by hydrodistillation and supercritical carbon dioxide Brazilian Arch. Biol. Technol. 2005 48 155 160 10.1590/S1516-89132005000100020 38. Baldin V.P. Bertin de Lima Scodro R. Mariano Fernandez C.M. Ieque A.L. Caleffi-Ferracioli K.R. Dias Siqueira V.L. de Almeida A.L. Gonçalves J.E. Garcia Cortez D.A. Cardoso R.F. Ginger essential oil and fractions against Mycobacterium spp. J. Ethnopharmacol. 2019 244 112095 10.1016/j.jep.2019.112095 31325601 39. Fitriady M.A. Sulaswatty A. Agustian E. Salahuddin Aditama D.P.F. Steam distillation extraction of ginger essential oil: Study of the effect of steam flow rate and time process AIP Conf. Proc. 2017 1803 020032-1 020032-10 10.1063/1.4973159 40. Jeliazkova E. Zheljazkov V.D. Kačániova M. Astatkie T. Tekwani B.L. Sequential elution of essential oil constituents during steam distillation of hops (Humulus lupulus L.) and influence on oil yield and antimicrobial activity J. Oleo Sci. 2018 67 871 883 10.5650/jos.ess17216 29877222 41. Gavahian M. Chu Y.H. Ohmic accelerated steam distillation of essential oil from lavender in comparison with conventional steam distillation Innov. Food Sci. Emerg. Technol. 2018 50 34 41 10.1016/j.ifset.2018.10.006 42. Donelian A. Carlson L.H.C. Lopes T.J. Machado R.A.F. Comparison of extraction of patchouli (Pogostemon cablin) essential oil with supercritical CO2 and by steam distillation J. Supercrit. Fluids 2009 48 15 20 10.1016/j.supflu.2008.09.020 43. Guzmán E. Lucia A. Essential Oils and Their Individual Components in Cosmetic Products Cosmetics 2021 8 114 10.3390/cosmetics8040114 44. Sharmeen J.B. Mahomoodally F.M. Zengin G. Maggi F. Essential oils as natural sources of fragrance compounds for cosmetics and cosmeceuticals Molecules 2021 26 666 10.3390/molecules26030666 33514008 45. Carvalho I.T. Estevinho B.N. Santos L. Application of microencapsulated essential oils in cosmetic and personal healthcare products—A review Int. J. Cosmet. Sci. 2016 38 109 119 10.1111/ics.12232 25923295 46. Asbahani A.E. Miladi K. Badri W. Sala M. Addi E.H.A. Casabianca H. Mousadik A.E. Hartmann D. Jilale A. Renaud F.N.R. Essential oils: From extraction to encapsulation Int. J. Pharm. 2015 483 220 243 10.1016/j.ijpharm.2014.12.069 25683145 47. Li G. Xiang S. Pan Y. Long X. Cheng Y. Han L. Zhao X. Effects of Cold-Pressing and Hydrodistillation on the Active Non-volatile Components in Lemon Essential Oil and the Effects of the Resulting Oils on Aging-Related Oxidative Stress in Mice Front. Nutr. 2021 8 1 15 10.3389/fnut.2021.689094 48. Menichini F. Tundis R. Bonesi M. De Cindio B. Loizzo M.R. Conforti F. Statti G.A. Menabeni R. Bettini R. Menichini F. Chemical composition and bioactivity of Citrus medica L. cv. Diamante essential oil obtained by hydrodistillation, cold-pressing and supercritical carbon dioxide extraction Nat. Prod. Res. 2011 25 789 799 10.1080/14786410902900085 21337254 49. Ferhat M.A. Meklati B.Y. Chemat F. Comparison of different isolation methods of essential oil from Citrus fruits: Cold pressing, hydrodistillation and microwave ‘dry’ distillation Flavour Fragr. J. 2007 22 494 504 10.1002/ffj.1829 50. Yousefi M. Rahimi-Nasrabadi M. Pourmortazavi S.M. Wysokowski M. Jesionowski T. Ehrlich H. Mirsadeghi S. Supercritical fluid extraction of essential oils TrAC—Trends Anal. Chem. 2019 118 182 193 10.1016/j.trac.2019.05.038 51. De Corato U. Maccioni O. Trupo M. Di Sanzo G. Use of essential oil of Laurus nobilis obtained by means of a supercritical carbon dioxide technique against post harvest spoilage fungi Crop Prot. 2010 29 142 147 10.1016/j.cropro.2009.10.012 52. Irmak S. Solakyildirim K. Hesenov A. Erbatur O. Study on the stability of supercritical fluid extracted rosemary (Rosmarinus Offcinalis L.) essential oil J. Anal. Chem. 2010 65 899 906 10.1134/S1061934810090030 53. Glisic S. Ivanovic J. Ristic M. Skala D. Extraction of sage (Salvia officinalis L.) by supercritical CO2: Kinetic data, chemical composition and selectivity of diterpenes J. Supercrit. Fluids 2010 52 62 70 10.1016/j.supflu.2009.11.009 54. Uquiche E. Cirano N. Millao S. Supercritical fluid extraction of essential oil from Leptocarpha rivularis using CO2 Ind. Crops Prod. 2015 77 307 314 10.1016/j.indcrop.2015.09.001 55. Zhao S. Zhang D. Supercritical fluid extraction and characterisation of Moringa oleifera leaves oil Sep. Purif. Technol. 2013 118 497 502 10.1016/j.seppur.2013.07.046 56. Martínez-Abad A. Ramos M. Hamzaoui M. Kohnen S. Jiménez A. Garrigós M.C. Optimisation of sequential microwave-assisted extraction of essential oil and pigment from lemon peels waste Foods 2020 9 1493 10.3390/foods9101493 57. Ferhat M.A. Meklati B.Y. Smadja J. Chemat F. An improved microwave Clevenger apparatus for distillation of essential oils from orange peel J. Chromatogr. A 2006 1112 121 126 10.1016/j.chroma.2005.12.030 16384566 58. Bayramoglu B. Sahin S. Sumnu G. Extraction of essential oil from laurel leaves by using microwaves Sep. Sci. Technol. 2009 44 722 733 10.1080/01496390802437271 59. Lucchesi M.E. Smadja J. Bradshaw S. Louw W. Chemat F. Solvent free microwave extraction of Elletaria cardamomum L.: A multivariate study of a new technique for the extraction of essential oil J. Food Eng. 2007 79 1079 1086 10.1016/j.jfoodeng.2006.03.029 60. Filly A. Fernandez X. Minuti M. Visinoni F. Cravotto G. Chemat F. Solvent-free microwave extraction of essential oil from aromatic herbs: From laboratory to pilot and industrial scale Food Chem. 2014 150 193 198 10.1016/j.foodchem.2013.10.139 24360439 61. Okoh O.O. Sadimenko A.P. Afolayan A.J. Comparative evaluation of the antibacterial activities of the essential oils of Rosmarinus officinalis L. obtained by hydrodistillation and solvent free microwave extraction methods Food Chem. 2010 120 308 312 10.1016/j.foodchem.2009.09.084 62. Lucchesi M.E. Chemat F. Smadja J. Solvent-free microwave extraction of essential oil from aromatic herbs: Comparison with conventional hydro-distillation J. Chromatogr. A 2004 1043 323 327 10.1016/j.chroma.2004.05.083 15330107 63. Sandhu H.K. Sinha P. Emanuel N. Kumar N. Sami R. Khojah E. Al-Mushhin A.A.M. Effect of ultrasound-assisted pretreatment on extraction efficiency of essential oil and bioactive compounds from citrus waste by-products Separations 2021 8 244 10.3390/separations8120244 64. Zheng S. Zhang G. Wang H.J. Long Z. Wei T. Li Q. Progress in ultrasound-assisted extraction of the value-added products from microorganisms World J. Microbiol. Biotechnol. 2021 37 1 14 10.1007/s11274-021-03037-y 33763773 65. Ramos C. Teixeira B. Batista I. Matos O. Serrano C. Neng N.R. Nogueira J.M.F. Nunes M.L. Marques A. Antioxidant and antibacterial activity of essential oil and extracts of bay laurel Laurus nobilis Linnaeus (Lauraceae) from Portugal Nat. Prod. Res. 2012 26 518 529 10.1080/14786419.2010.531478 21756182 66. Dadalioǧlu I. Evrendilek G.A. Chemical compositions and antibacterial effects of essential oils of Turkish oregano (Origanum minutiflorum), bay laurel (Laurus nobilis), Spanish lavender (Lavandula stoechas L.), and fennel (Foeniculum vulgare) on common foodborne pathogens J. Agric. Food Chem. 2004 52 8255 8260 10.1021/jf049033e 15612826 67. Shaaya E. Ravid U. Paster N. Juven B. Zisman U. Pissarev V. Fumigant toxicity of essential oils against four major stored-product insects J. Chem. Ecol. 1991 17 499 504 10.1007/BF00982120 24258800 68. Cosimi S. Rossi E. Cioni P.L. Canale A. Bioactivity and qualitative analysis of some essential oils from Mediterranean plants against stored-product pests: Evaluation of repellency against Sitophilus zeamais Motschulsky, Cryptolestes ferrugineus (Stephens) and Tenebrio molitor (L.) J. Stored Prod. Res. 2009 45 125 132 10.1016/j.jspr.2008.10.002 69. Lawless J. The Encyclopedia of Essential Oils: The Complete Guide to the Use of Aromatic Oils In Aromatherapy, Herbalism, Health, and Well Being Conari Press Newburyport, MA, USA 2013 9780007145188 70. Specos M.M.M. García J.J. Tornesello J. Marino P. Vecchia D.M. Tesoriero D.M.V. Hermida L.G. Microencapsulated citronella oil for mosquito repellent finishing of cotton textiles Trans. R. Soc. Trop. Med. Hyg. 2010 104 653 658 10.1016/j.trstmh.2010.06.004 20673937 71. Sakulku U. Nuchuchua O. Uawongyart N. Puttipipatkhachorn S. Soottitantawat A. Ruktanonchai U. Characterization and mosquito repellent activity of citronella oil nanoemulsion Int. J. Pharm. 2009 372 105 111 10.1016/j.ijpharm.2008.12.029 19162149 72. Tawatsin A. Wratten S.D. Scott R.R. Thavara U. Techadamrongsin Y. Repellency of volatile oils from plants against three mosquito vectors J. Vectory Ecol. 2001 26 76 82 73. Nakahara K. Alzoreky N.S. Yoshihashi T. Nguyen H.T.T. Trakoontivakorn G. Chemical Composition and Antifungal Activity of Essential Oil from Cymbopogon nardus (Citronella Grass) Jpn. Agric. Res. Q. 2003 37 249 252 10.6090/jarq.37.249 74. Andrade M.A. Das Graças Cardoso M. Batista L.R. Mallet A.C.T. Machado S.M.F. Essential oils of Cinnamomum zeylanicum, Cymbopogon nardus and Zingiber officinale: Composition, antioxidant and antibacterial activities Rev. Ciência Agronômica 2012 43 399 408 10.1590/S1806-66902012000200025 75. Simpson W.J. Smith A.R.W. Factors affecting antibacterial activity of hop compounds and their derivatives J. Appl. Bacteriol. 1992 72 327 334 10.1111/j.1365-2672.1992.tb01843.x 1517174 76. Zanoli P. Zavatti M. Pharmacognostic and pharmacological profile of Humulus lupulus L. J. Ethnopharmacol. 2008 116 383 396 10.1016/j.jep.2008.01.011 18308492 77. Miranda C.L. Stevens J.F. Helmrich A. Henderson M.C. Rodriguez R.J. Yang Y.H. Deinzer M.L. Barnes D.W. Buhler D.R. Antiproliferative and cytotoxic effects of prenylated flavonoids from hops (Humulus lupulus) in human cancer cell lines Food Chem. Toxicol. 1999 37 271 285 10.1016/S0278-6915(99)00019-8 10418944 78. Reher T. Van Kerckvoorde V. Verheyden L. Wenseleers T. Beliën T. Bylemans D. Martens J.A. Evaluation of hop (Humulus lupulus) as a repellent for the management of Drosophila suzukii Crop Prot. 2019 124 104839 10.1016/j.cropro.2019.05.033 79. Bedini S. Flamini G. Girardi J. Cosci F. Conti B. Not just for beer: Evaluation of spent hops (Humulus lupulus L.) as a source of eco-friendly repellents for insect pests of stored foods J. Pest Sci. 2015 88 583 592 10.1007/s10340-015-0647-1 80. Abdel-Naime W.A. Fahim J.R. Fouad M.A. Kamel M.S. Antibacterial, antifungal, and GC–MS studies of Melissa officinalis S. Afr. J. Bot. 2019 124 228 234 10.1016/j.sajb.2019.05.011 81. Shakeri A. Sahebkar A. Javadi B. Melissa officinalis L.—A review of its traditional uses, phytochemistry and pharmacology J. Ethnopharmacol. 2016 188 204 228 10.1016/j.jep.2016.05.010 27167460 82. Moradkhani H. Sargsyan E. Bibak H. Naseri B. Sadat-Hosseini M. Fayazi-Barjin A. Meftahizade H. Melissa officinalis L., a valuable medicine plant: A review J. Med. Plants Res. 2010 4 2753 2759 83. Vitalini S. Tomè F. Fico G. Traditional uses of medicinal plants in Valvestino (Italy) J. Ethnopharmacol. 2009 121 106 116 10.1016/j.jep.2008.10.005 18996175 84. Biswas K. Banerjee R.K. Bandyopadhyay U. Chattopadhyay I. Biological activities and medicinal properties of Neem (Azadirachta indica) Curr. Sci. 2001 82 1336 1445 85. Schmutterer H. Properties and Potencial of Natural Pesticides from Neem Tree, Azadirachta Indica Anu. Rev. Entomol. 1990 35 271 297 10.1146/annurev.en.35.010190.001415 2405771 86. Rao A. Sukumar S. Paramasivam T. Kamalakshi S. Parashuraman M. Shantha A.R. Study of antiviral activity of tender leaves of Margosa tree (Melia azadericta) on vaccinia and variola virus: A preliminary report Indian J Med Res. 1969 57 495 502 4309947 87. Chopra I. Gupta K. Nazir B. Preliminary study of anti-bacterial substances from Melia azidirachta Indian J. Med. Res. 1952 40 511 515 13061059 88. Almas K. The antimicrobial effects of extracts of Azadirachta indica (Neem) and Salvadora persica (Arak) chewing sticks Indian J. Dent. Res. 1999 10 23 26 10865390 89. Salem N. Sriti J. Bachrouch O. Msaada K. Khammassi S. Hammami M. Selmi S. Boushih E. Ouertani M. Hachani N. Phenological stage effect on phenolic composition and repellent potential of Mentha pulegium against Tribolium castaneum and Lasioderma serricorne Asian Pac. J. Trop. Biomed. 2018 8 207 216 10.4103/2221-1691.231283 90. Salem N. Bachrouch O. Sriti J. Msaada K. Khammassi S. Hammami M. Selmi S. Boushih E. Koorani S. Abderraba M. Fumigant and repellent potentials of Ricinus communis and Mentha pulegium essential oils against Tribolium castaneum and Lasioderma serricorne Int. J. Food Prop. 2018 20 2899 2913 10.1080/10942912.2017.1382508 91. Soares S.F. Borges L.M.F. de Sousa Braga R. Ferreira L.L. Louly C.C.B. Tresvenzol L.M.F. de Paula J.R. Ferri P.H. Repellent activity of plant-derived compounds against Amblyomma cajennense (Acari: Ixodidae) nymphs Vet. Parasitol. 2010 167 67 73 10.1016/j.vetpar.2009.09.047 19897309 92. Brahmi F. Abdenour A. Bruno M. Silvia P. Alessandra P. Danilo F. Drifa Y.G. Fahmi E.M. Khodir M. Mohamed C. Chemical composition and in vitro antimicrobial, insecticidal and antioxidant activities of the essential oils of Mentha pulegium L. and Mentha rotundifolia (L.) Huds growing in Algeria Ind. Crops Prod. 2016 88 96 105 10.1016/j.indcrop.2016.03.002 93. Teixeira B. Marques A. Ramos C. Batista I. Serrano C. Matos O. Neng N.R. Nogueira J.M.F. Saraiva J.A. Nunes M.L. European pennyroyal (Mentha pulegium) from Portugal: Chemical composition of essential oil and antioxidant and antimicrobial properties of extracts and essential oil Ind. Crops Prod. 2012 36 81 87 10.1016/j.indcrop.2011.08.011 94. Ait-Ouazzou A. Lorán S. Arakrak A. Laglaoui A. Rota C. Herrera A. Pagán R. Conchello P. Evaluation of the chemical composition and antimicrobial activity of Mentha pulegium, Juniperus phoenicea, and Cyperus longus essential oils from Morocco Food Res. Int. 2012 45 313 319 10.1016/j.foodres.2011.09.004 95. Khaled-Khodja N. Boulekbache-Makhlouf L. Madani K. Phytochemical screening of antioxidant and antibacterial activities of methanolic extracts of some Lamiaceae Ind. Crops Prod. 2014 61 41 48 10.1016/j.indcrop.2014.06.037 96. Pehlivan Karakaş F. Yildirim A. Türker A. Biological screening of various medicinal plant extracts for antibacterial and antitumor activities Turk. J. Biol. 2012 36 641 652 10.3906/biy-1203-16 97. Lee S.W. Li G. Lee K.S. Jung J.S. Xu M.L. Seo C.S. Chang H.W. Kim S.K. Song D.K. Son J.K. Preventive Agents against Sepsis and New Phenylpropanoid Glucosides from the Fruits of Illicium verum Planta Med. 2003 69 861 864 10.1055/s-2003-43210 14598217 98. Ohira H. Torii N. Aida T.M. Watanabe M. Smith R.L. Rapid separation of shikimic acid from Chinese star anise (Illicium verum Hook. f.) with hot water extraction Sep. Purif. Technol. 2009 69 102 108 10.1016/j.seppur.2009.07.005 99. Yang J.F. Yang C.H. Chang H.W. Yang C.S. Wang S.M. Hsieh M.C. Chuang L.Y. Chemical composition and antibacterial activities of Illicium verum against antibiotic-resistant pathogens J. Med. Food 2010 13 1254 1262 10.1089/jmf.2010.1086 20828316 100. Padmashree A. Roopa N. Semwal A.D. Sharma G.K. Agathian G. Bawa A.S. Star-anise (Illicium verum) and black caraway (Carum nigrum) as natural antioxidants Food Chem. 2007 104 59 66 10.1016/j.foodchem.2006.10.074 101. Xin S. Xiongmin L. Chunlin F. Qiuxia L. Extraction, characterization and antitumor effect of the polysaccharides from star anise (Illicium verum Hook. f.) J. Med. Plants Res. 2010 4 2666 2673 10.5897/JMPR09.438 102. Lee J.S. Park M.A. Yoon C.S. Na J.H. Han J. Characterization and Preservation Performance of Multilayer Film with Insect Repellent and Antimicrobial Activities for Sliced Wheat Bread Packaging J. Food Sci. 2019 84 3194 3203 10.1111/1750-3841.14823 31589346 103. Park M.A. Chang Y. Choi I. Bai J. Ja-hyun N. Han J. Development of A Comprehensive Biological Hazard-Proof Packaging Film with Insect-Repellent, Antibacterial, and Antifungal Activities J. Food Sci. 2018 83 3035 3043 10.1111/1750-3841.14397 30457667 104. Brown J. Valeriana officinalis Aust. J. Herb. Med. 2012 24 136 139 10.1007/978-3-540-71095-0_10944 105. Joanne Barnes Herbal Medicines: A Guide for Health Care Professionals 2nd ed. Pharmaceutical Press London, UK 1996 0853692890 106. Grant K.L. Patient education and herbal dietary supplements Am. J. Health Pharm. 2000 57 1997 2003 10.1093/ajhp/57.21.1997 11094655 107. Wang J. Zhao J. Liu H. Zhou L. Liu Z. Wang J. Han J. Yu Z. Yang F. Chemical analysis and biological activity of the essential oils of two valerianaceous species from China: Nardostachys chinensis and valeriana officinalis Molecules 2010 15 6411 6422 10.3390/molecules15096411 20877232 108. Khademian R. Karimzadeh F. Moradi P. Asghari B. Anti-microbial Properties of Valeriana officinalis, Satureja bachtiarica and Thymus daenensis Methanolic Extracts against Helicobacter pylori J. Pharm. Res. Int. 2019 26 1 7 10.9734/jpri/2019/v26i630153 109. Holubec H. Baines A.T. Bayse J.L. Waite S. Bhattacharyya A.K. Clark L.C. Payne C.M. Earnest D.L. Nelson M.A. Chemoprevention of colon carcinogenesis by dietary selenomethionine Gastroenterology 1998 114 A612 110. Trombetta D. Castelli F. Sarpietro M.G. Venuti V. Cristani M. Daniele C. Saija A. Mazzanti G. Bisignano G. Mechanisms of antibacterial action of three monoterpenes Antimicrob. Agents Chemother. 2005 49 2474 2478 10.1128/AAC.49.6.2474-2478.2005 15917549 111. Khan M.T.H. Ather A. Thompson K.D. Gambari R. Extracts and molecules from medicinal plants against herpes simplex viruses Antivir. Res. 2005 67 107 119 10.1016/j.antiviral.2005.05.002 16040137 112. Peana A.T. D’Aquila P.S. Panin F. Serra G. Pippia P. Moretti M.D.L. Anti-inflammatory activity of linalool and linalyl acetate constituents of essential oils Phytomedicine 2002 9 721 726 10.1078/094471102321621322 12587692 113. Ziegler H.L. Jensen T.H. Christensen J. Stærk D. Hägerstrand H. Sittie A.A. Olsen C.E. Staalsø T. Ekpe P. Jaroszewski J.W. Possible artefacts in the in vitro determination of antimalarial activity of natural products that incorporate into lipid bilayer: Apparent antiplasmodial activity of dehydroabietinol, a constituent of Hyptis suaveolens Planta Med. 2002 68 547 549 10.1055/s-2002-32548 12094303 114. Sienkiewicz M. Głowacka A. Poznańska-Kurowska K. Kaszuba A. Urbaniak A. Kowalczyk E. The effect of clary sage oil on staphylococci responsible for wound infections Postep. Dermatol. I Alergol. 2015 32 21 26 10.5114/pdia.2014.40957 25821423 115. Fujita K.I. Chavasiri W. Kubo I. Anti-Salmonella Activity of Volatile Compounds of Vietnam Coriander Phyther. Res. 2015 29 1081 1087 10.1002/ptr.5351 25870012 116. Wan J. Wilcock A. Coventry M.J. The effect of essential oils of basil on the growth of Aeromonas hydrophila and Pseudomonas fluorescens J. Appl. Microbiol. 1998 84 152 158 10.1046/j.1365-2672.1998.00338.x 9633630 117. Takarada K. Kimizuka R. Takahashi N. Honma K. Okuda K. Kato T. A comparison of the antibacterial efficacies of essential oils against oral pathogens Oral Microbiol. Immunol. 2004 19 61 64 10.1046/j.0902-0055.2003.00111.x 14678476 118. Hammer K.A. Dry L. Johnson M. Michalak E.M. Carson C.F. Riley T.V. Susceptibility of oral bacteria to Melaleuca alternifolia (tea tree) oil in vitro Oral Microbiol. Immunol. 2003 18 389 392 10.1046/j.0902-0055.2003.00105.x 14622345 119. Hammer K.A. Carson C.F. Riley T.V. Antifungal activity of the components of Melaleuca alternifolia (tea tree) oil J. Appl. Microbiol. 2003 95 853 860 10.1046/j.1365-2672.2003.02059.x 12969301 120. Deans S.G. Ritchie G. Antibacterial properties of plant essential oils Int. J. Food Microbiol. 1987 5 165 180 10.1016/0168-1605(87)90034-1 121. Samber N. Khan A. Varma A. Manzoor N. Synergistic anti-candidal activity and mode of action of Mentha piperita essential oil and its major components Pharm. Biol. 2015 53 1496 1504 10.3109/13880209.2014.989623 25853964 122. Mejía-Garibay B. Palou E. López-Malo A. Composition, Diffusion, and Antifungal Activity of Black Mustard (Brassica nigra) Essential Oil When Applied by Direct Addition or Vapor Phase Contact J. Food Prot. 2015 78 843 848 10.4315/0362-028X.JFP-14-485 25836415 123. Prakash B. Singh P. Goni R. Raina A.K.P. Dubey N.K. Efficacy of Angelica archangelica essential oil, phenyl ethyl alcohol and α- terpineol against isolated molds from walnut and their antiaflatoxigenic and antioxidant activity J. Food Sci. Technol. 2015 52 2220 2228 10.1007/s13197-014-1278-x 25829603 124. Trindade L.A. de Araújo Oliveira J. de Castro R.D. de Oliveira Lima E. Inhibition of adherence of C. albicans to dental implants and cover screws by Cymbopogon nardus essential oil and citronellal Clin. Oral Investig. 2015 19 2223 2231 10.1007/s00784-015-1450-3 25804885 125. Barkatullah Ibrar M. Muhammad N. De Feo V. Chemical composition and biological activities of the essential oil of Skimmia laureola leaves Molecules 2015 20 4735 4745 10.3390/molecules20034735 25786161 126. Mahboubi M. Kazempour N. The antifungal activity of Artemisia sieberi essential oil from different localities of Iran against dermatophyte fungi J. Mycol. Med. 2015 25 65 71 10.1016/j.mycmed.2015.02.042 25770648 127. Baratta M.T. Damien H.J. Deans S.G. Biondi D.M. Ruberto G. Chemical composition, antimicrobial and antioxidative activity of laurel, sage, rosemary, oregano and coriander essential oils J. Essent. Oil Res. 1998 10 618 627 10.1080/10412905.1998.9700989 128. Gkinis G. Tzakou O. Iliopoulou D. Roussis V. Chemical Composition and Biological Activity of Nepeta parnassica Oils and Isolated Nepetalactones Z. Für Nat. C. 2003 58 681 686 10.1515/znc-2003-9-1015 129. Grunfeld E. Gresty M.A. Relationship between motion sickness, migraine and menstruation in crew members of a “round the world” yacht race Brain Res. Bull. 1998 47 433 436 10.1016/S0361-9230(98)00099-9 10052571 130. Howes M.-J.R. Houghton P.J. Barlow D.J. Pocock V.J. Milligan S.R. Assessment of estrogenic activity in some common essential oil constituents J. Pharm. Pharmacol. 2010 54 1521 1528 10.1211/002235702216 131. Ali B. Al-Wabel N.A. Shams S. Ahamad A. Khan S.A. Anwar F. Essential oils used in aromatherapy: A systemic review Asian Pac. J. Trop. Biomed. 2015 5 601 611 10.1016/j.apjtb.2015.05.007 132. Bhalla Y. Gupta V.K. Jaitak V. Anticancer activity of essential oils: A review J. Sci. Food Agric. 2013 93 3643 3653 10.1002/jsfa.6267 23765679 133. Edris A.E. Pharmaceutical and Therapeutic Potentials of Essential Oils and Their Individual Volatile Constituents: A Review Phyther. Res. 2007 21 308 323 10.1002/ptr.2072 134. Jaganathan S.K. Supriyanto E. Antiproliferative and Molecular Mechanism of Eugenol-Induced Apoptosis in Cancer Cells Molecules 2012 17 6290 6304 10.3390/molecules17066290 22634840 135. Gupta S. Khan S. Muzafar M. Kushwaha M. Yadav A.K. Gupta A.P. Encapsulation: Entrapping essential oil/flavors/aromas in food Encapsulations Grumezescu A.M. Elsevier Amsterdam, The Netherlands 2016 229 268 9780128043073 136. Reis D.R. Ambrosi A. Luccio M. Di Encapsulated essential oils: A perspective in food preservation Future Foods 2022 10.1016/j.fufo.2022.100126 137. Ribeiro-Santos R. Andrade M. Sanches-Silva A. Application of encapsulated essential oils as antimicrobial agents in food packaging Curr. Opin. Food Sci. 2017 14 78 84 10.1016/j.cofs.2017.01.012 138. Zinoviadou K.G. Koutsoumanis K.P. Biliaderis C.G. Physico-chemical properties of whey protein isolate films containing oregano oil and their antimicrobial action against spoilage flora of fresh beef Meat Sci. 2009 82 338 345 10.1016/j.meatsci.2009.02.004 20416718 139. Suppakul P. Sonneveld K. Bigger S.W. Miltz J. Efficacy of polyethylene-based antimicrobial films containing principal constituents of basil LWT—Food Sci. Technol. 2008 41 779 788 10.1016/j.lwt.2007.06.006 140. Vu K.D. Hollingsworth R.G. Leroux E. Salmieri S. Lacroix M. Development of edible bioactive coating based on modified chitosan for increasing the shelf life of strawberries Food Res. Int. 2011 44 198 203 10.1016/j.foodres.2010.10.037 141. De Melo A.A.M. Geraldine R.M. Silveira M.F.A. Lopes M.C. Silva C. Fernandes T.H. Oliveira A.N. De Microbiological Quality and Other Characteristics of Refrigerated Chicken Meat in Contact with Cellulose Acetate-based Film Incorporated with Rosemary Essential Oil Braz. J. Microbiol. 2012 16 1419 1427 10.1590/S1517-83822012000400025 142. Souza A.C. Goto G.E.O. Mainardi J.A. Coelho A.C.V. Tadini C.C. Cassava starch composite films incorporated with cinnamon essential oil: Antimicrobial activity, microstructure, mechanical and barrier properties LWT—Food Sci. Technol. 2013 54 346 352 10.1016/j.lwt.2013.06.017 143. Aburjai T. Natsheh F.M. Plants Used in Cosmetics Phyther. Res. 2003 17 987 1000 10.1002/ptr.1363 144. Minozzo M. Steffens J. Backes G.T. Paroul N. Cansian R.L. Biological potential and microencapsulation of Cinnamomum cassia essential oil as an alternative for pest control in stored maize Res. Soc. Dev. 2021 10 e530101422334 10.33448/rsd-v10i14.22334 145. Malešević V.K. Vaštag Ž. Radulović-Popović L. Senka M.P. Periĉin-Starĉević I. Microencapsulation technology and essential oil pesticides for food plant production Essential Oils in Food Preservation, Flavor and Safety Preedy V.R. Academic Press Cambridge, MA, USA 2016 123 129 9780124166448 146. Govindarajan M. Sivakumar R. Laboratory evaluation of Indian medicinal plants as repellents against malaria, dengue, and filariasis vector mosquitoes Parasitol. Res. 2015 114 601 612 10.1007/s00436-014-4222-0 25399815 147. Khanna S. Chakraborty J.N. Mosquito repellent activity of cotton functionalized with inclusion complexes of β-cyclodextrin citrate and essential oils Fash. Text. 2018 5 1 18 10.1186/s40691-017-0125-x 148. Soroh A. Owen L. Rahim N. Masania J. Abioye A. Qutachi O. Goodyer L. Shen J. Laird K. Microemulsification of essential oils for the development of antimicrobial and mosquito repellent functional coatings for textiles J. Appl. Microbiol. 2021 131 2808 2820 10.1111/jam.15157 34022108 149. Bezerra F.M. Carmona O.G. Carmona C.G. Lis M.J. de Moraes F.F. Controlled release of microencapsulated citronella essential oil on cotton and polyester matrices Cellulose 2016 23 1459 1470 10.1007/s10570-016-0882-5 150. Sariişik M. Kartal G.E. Erkan G. Etkeser S. Alternative methods for transferring mosquito repellent capsules containing bio-based citronella oil to upholstery fabrics: Coating and printing J. Coat. Technol. Res. 2022 19 323 336 10.1007/s11998-021-00529-2 151. Singh M.N. Hemant K.S.Y. Ram M. Shivakumar H.G. Microencapsulation: A promising technique for controlled drug delivery Res. Pharm. Sci. 2010 5 65 77 21589795 152. Gonçalves A. Estevinho B.N. Rocha F. Microencapsulation of vitamin A: A review Trends Food Sci. Technol. 2016 51 76 87 10.1016/j.tifs.2016.03.001 153. Dubey R. Shami T.C. Rao K.U.B. Microencapsulation Technology and Applications Def. Sci. J. 2009 59 82 95 10.14429/dsj.59.1489 154. Paulo F. Santos L. Design of experiments for microencapsulation applications: A review Mater. Sci. Eng. C 2017 77 1327 1340 10.1016/j.msec.2017.03.219 155. Mehta N. Kumar P. Verma A.K. Umaraw P. Kumar Y. Malav O.P. Sazili A.Q. Lorenzo J.M. Microencapsulation as a Noble Technique for the Application of Bioactive Compounds in the Food Industry: A Comprehensive Review Appl. Sci. 2022 12 1424 10.3390/app12031424 156. Nesterenko A. Alric I. Silvestre F. Durrieu V. Vegetable proteins in microencapsulation: A review of recent interventions and their effectiveness Ind. Crops Prod. 2013 42 469 479 10.1016/j.indcrop.2012.06.035 157. Petrusic S. Koncar V. Controlled release of active agents from microcapsules embedded in textile structures Smart Textiles and Their Applications Koncar V. Elsevier Amsterdam, The Netherlands 2016 89 114 9780081005835 158. Zhang Y. Rochefort D. Characterisation and applications of microcapsules obtained by interfacial polycondensation J. Microencapsul. 2012 29 636 649 10.3109/02652048.2012.676092 22494033 159. Hoyos-Leyva J.D. Bello-Pérez L.A. Alvarez-Ramirez J. Garcia H.S. Microencapsulation using starch as wall material: A review Food Rev. Int. 2018 34 148 161 10.1080/87559129.2016.1261298 160. Jamekhorshid A. Sadrameli S.M. Farid M. A review of microencapsulation methods of phase change materials (PCMs) as a thermal energy storage (TES) medium Renew. Sustain. Energy Rev. 2014 31 531 542 10.1016/j.rser.2013.12.033 161. Fang Z. Bhandari B. Encapsulation of polyphenols—A review Trends Food Sci. Technol. 2010 21 510 523 10.1016/j.tifs.2010.08.003 162. Mcclements D.J. Decker E.A. Park Y. Weiss J. Structural design principles for delivery of bioactive components in nutraceuticals and functional foods Crit. Rev. Food Sci. Nutr. 2009 49 577 606 10.1080/10408390902841529 19484636 163. Augustin M.A. Sanguansri L. Bode O. Maillard reaction products as encapsulants for fish oil powders J. Food Sci. 2006 71 E25 E32 10.1111/j.1365-2621.2006.tb08893.x 164. Li M. Rouaud O. Poncelet D. Microencapsulation by solvent evaporation: State of the art for process engineering approaches Int. J. Pharm. 2008 363 26 39 10.1016/j.ijpharm.2008.07.018 18706988 165. O’Donnell P.B. McGinity J.W. Preparation of microspheres by the solvent evaporation technique Adv. Drug Deliv. Rev. 1997 28 25 42 10.1016/S0169-409X(97)00049-5 10837563 166. Bakry A.M. Abbas S. Ali B. Majeed H. Abouelwafa M.Y. Mousa A. Liang L. Microencapsulation of Oils: A Comprehensive Review of Benefits, Techniques, and Applications Compr. Rev. Food Sci. Food Saf. 2016 15 143 182 10.1111/1541-4337.12179 33371581 167. Timilsena Y.P. Akanbi T.O. Khalid N. Adhikari B. Barrow C.J. Complex coacervation: Principles, mechanisms and applications in microencapsulation Int. J. Biol. Macromol. 2019 121 1276 1286 10.1016/j.ijbiomac.2018.10.144 30352231 168. Hwang J. Kim J. Wee Y. Yun J. Jang H. Kim S. Ryu H. Preparation and characterization of melamine-formaldehyde resin microcapsules containing fragrant oil Biotechnol. Bioprocess Eng. 2006 11 332 336 10.1007/BF03026249 169. Chung S.K. Seo J.Y. Lim J.H. Park H.H. Yea M.J. Park H.J. Microencapsulation of essential oil for insect repellent in food packaging system J. Food Sci. 2013 78 E709 E714 10.1111/1750-3841.12111 23551154 170. Sánchez-Navarro M.M. Cuesta-Garrote N. Arán-Áis F. Orgilés-Barceló C. Microencapsulation of melaleuca alternifolia (tea tree) oil as biocide for footwear applications J. Dispers. Sci. Technol. 2011 32 1722 1727 10.1080/01932691.2011.616126 171. Long Y. York D. Zhang Z. Preece J.A. Microcapsules with low content of formaldehyde: Preparation and characterization J. Mater. Chem. 2009 19 6882 6887 10.1039/b902832c 172. Bansode S.S. Banarjee S.K. Gaikwad D.D. Jadhav S.L. Thorat R.M. Microencapsulation: A review Int. J. Pharm. Sci. Rev. Res. 2010 1 38 43 173. Lam P.L. Gambari R. Advanced progress of microencapsulation technologies: In vivo and in vitro models for studying oral and transdermal drug deliveries J. Control. Release 2014 178 25 45 10.1016/j.jconrel.2013.12.028 24417967 174. Huang H.J. Chen X.D. Yuan W.K. Microencapsulation based on emulsification for producing pharmaceutical products: A literature review Dev. Chem. Eng. Miner. Process. 2006 14 515 544 10.1002/apj.5500140318 175. Estevinho B.N. Rocha F. Santos L. Alves A. Microencapsulation with chitosan by spray drying for industry applications—A review Trends Food Sci. Technol. 2013 31 138 155 10.1016/j.tifs.2013.04.001 176. Gharsallaoui A. Roudaut G. Chambin O. Voilley A. Saurel R. Applications of spray-drying in microencapsulation of food ingredients: An overview Food Res. Int. 2007 40 1107 1121 10.1016/j.foodres.2007.07.004 177. Mohammed N.K. Tan C.P. Manap Y.A. Muhialdin B.J. Hussin A.S.M. Spray Drying for the Encapsulation of Oils—A Review Molecules 2020 25 3873 10.3390/molecules25173873 32858785 178. Hazarika U. Gosztola B. Lyophilization and its Effects on the Essential Oil Content and Composition of Herbs and Spices—A Review Acta Sci. Pol. Technol. Aliment. 2020 19 467 473 10.17306/J.AFS.2020.0853 33179486 179. Gardeli C. Evageliou V. Poulos C. Yanniotis S. Komaitis M. Drying of fennel plants: Oven, freeze drying, effect of freeze-drying time, and use of biopolymers Dry. Technol. 2010 28 542 549 10.1080/07373931003622321 180. Vujanović M.D. Đurović S.D. Radojković M.M. Chemical Composition of Essential Oils of Elderberry (Sambucus Nigra L.) Flowers and Fruits Acta Period. Technol. 2021 52 229 237 10.2298/APT2152229V 181. de Araújo J.S.F. de Souza E.L. Oliveira J.R. Gomes A.C.A. Kotzebue L.R.V. da Silva Agostini D.L. de Oliveira D.L.V. Mazzetto S.E. da Silva A.L. Cavalcanti M.T. Microencapsulation of sweet orange essential oil (Citrus aurantium var. dulcis) by liophylization using maltodextrin and maltodextrin/gelatin mixtures: Preparation, characterization, antimicrobial and antioxidant activities Int. J. Biol. Macromol. 2020 143 991 999 10.1016/j.ijbiomac.2019.09.160 31669659 182. Sebaaly C. Greige-Gerges H. Stainmesse S. Fessi H. Charcosset C. Effect of composition, hydrogenation of phospholipids and lyophilization on the characteristics of eugenol-loaded liposomes prepared by ethanol injection method Food Biosci. 2016 15 1 10 10.1016/j.fbio.2016.04.005 183. Cocero M.J. Martín Á. Mattea F. Varona S. Encapsulation and co-precipitation processes with supercritical fluids: Fundamentals and applications J. Supercrit. Fluids 2009 47 546 555 10.1016/j.supflu.2008.08.015 184. Jyothi N.V.N. Prasanna P.M. Sakarkar S.N. Prabha K.S. Ramaiah P.S. Srawan G.Y. Microencapsulation techniques, factors influencing encapsulation efficiency J. Microencapsul. 2010 27 187 197 10.3109/02652040903131301 20406093 185. Zhang S. Kawakami K. Yamamoto M. Masaoka Y. Kataoka M. Yamashita S. Sakuma S. Coaxial electrospray formulations for improving oral absorption of a poorly water-soluble drug Mol. Pharm. 2011 8 807 813 10.1021/mp100401d 21395264 186. Koo S.Y. Cha K.H. Song D.G. Chung D. Pan C.H. Microencapsulation of peppermint oil in an alginate-pectin matrix using a coaxial electrospray system Int. J. Food Sci. Technol. 2014 49 733 739 10.1111/ijfs.12358 187. Zhang L. Huang J. Si T. Xu R.X. Coaxial electrospray of microparticles and nanoparticles for biomedical applications Expert Rev. Med. Devices 2012 9 595 612 10.1586/erd.12.58 23249155 188. Lastow O. Balachandran W. Numerical simulation of electrohydrodynamic (EHD) atomization J. Electrostat. 2006 64 850 859 10.1016/j.elstat.2006.02.006 189. Anwar S.H. Kunz B. The influence of drying methods on the stabilization of fish oil microcapsules: Comparison of spray granulation, spray drying, and freeze drying J. Food Eng. 2011 105 367 378 10.1016/j.jfoodeng.2011.02.047 190. Teunou E. Poncelet D. Fluid-Bed Coating Encapsulated and Powdered Foods Onwulata C. Taylor & Francis Abingdon, UK 2005 18 9780429120152 191. Desai K.G.H. Jin Park H. Recent developments in microencapsulation of food ingredients Dry. Technol. 2005 23 1361 1394 10.1081/DRT-200063478 192. Sadovoy A.V. Lomova M.V. Antipina M.N. Braun N.A. Sukhorukov G.B. Kiryukhin M.V. Layer-by-layer assembled multilayer shells for encapsulation and release of fragrance ACS Appl. Mater. Interfaces 2013 5 8948 8954 10.1021/am401871u 24000789 193. Xie Y.L. Wang M.J. Yao S.J. Preparation and characterization of biocompatible microcapsules of sodium cellulose sulfate/chitosan by means of layer-by-layer self-assembly Langmuir 2009 25 8999 9005 10.1021/la9014338 19583223 194. Ge L. Tan X. Sheng R. Xiao J. Layer-by-layer self-assembly of giant polyelectrolyte microcapsules templated by microbubbles as potential hydrophilic or hydrophobic drug delivery system Colloids Interface Sci. Commun. 2022 47 10.1016/j.colcom.2022.100603 195. Zhang Z. Zhang S. Su R. Xiong D. Feng W. Chen J. Controlled Release Mechanism and Antibacterial Effect of Layer-By-Layer Self-Assembly Thyme Oil Microcapsule J. Food Sci. 2019 84 1427 1438 10.1111/1750-3841.14610 31070787 196. Pan H.M. Subramanian A. Ochs C.J. Dewavrin J.Y. Beyer S. Trau D.W. Edible polyelectrolyte microcapsules with water-soluble cargo assembled in organic phase RSC Adv. 2014 4 35163 35166 10.1039/C4RA04750H 197. Verma G. Hassan P.A. Self assembled materials: Design strategies and drug delivery perspectives Phys. Chem. Chem. Phys. 2013 15 17016 17028 10.1039/c3cp51207j 23907560 198. Aguiar M.C.S. das Graças Fernandes da Silva M.F. Fernandes J.B. Forim M.R. Evaluation of the microencapsulation of orange essential oil in biopolymers by using a spray-drying process Sci. Rep. 2020 10 1 11 10.1038/s41598-020-68823-4 31913322 199. Martins I.M. Barreiro M.F. Coelho M. Rodrigues A.E. Microencapsulation of essential oils with biodegradable polymeric carriers for cosmetic applications Chem. Eng. J. 2014 245 191 200 10.1016/j.cej.2014.02.024 200. Scarfato P. Avallone E. Iannelli P. De Feo V. Acierno D. Synthesis and characterization of polyurea microcapsules containing essential oils with antigerminative activity J. Appl. Polym. Sci. 2007 105 3568 3577 10.1002/app.26420 201. Bagle A.V. Jadhav R.S. Gite V.V. Hundiwale D.G. Mahulikar P.P. Controlled release study of phenol formaldehyde microcapsules containing neem oil as an insecticide Int. J. Polym. Mater. Polym. Biomater. 2013 62 421 425 10.1080/00914037.2012.719142 202. Ahsaei S.M. Rodríguez-Rojo S. Salgado M. Cocero M.J. Talebi-Jahromi K. Amoabediny G. Insecticidal activity of spray dried microencapsulated essential oils of Rosmarinus officinalis and Zataria multiflora against Tribolium confusum Crop Prot. 2020 128 104996 10.1016/j.cropro.2019.104996 203. Fernandes R.V.D.B. Borges S.V. Botrel D.A. Gum arabic/starch/maltodextrin/inulin as wall materials on the microencapsulation of rosemary essential oil Carbohydr. Polym. 2014 101 524 532 10.1016/j.carbpol.2013.09.083 24299808 204. Chitprasert P. Sutaphanit P. Holy basil (ocimum sanctum linn.) Essential oil delivery to swine gastrointestinal tract using gelatin microcapsules coated with aluminum carboxymethyl cellulose and beeswax J. Agric. Food Chem. 2014 62 12641 12648 10.1021/jf5019438 25382222 205. Munhuweyi K. Caleb O.J. van Reenen A.J. Opara U.L. Physical and antifungal properties of β-cyclodextrin microcapsules and nanofibre films containing cinnamon and oregano essential oils LWT—Food Sci. Technol. 2018 87 413 422 10.1016/j.lwt.2017.09.012 206. Leimann F.V. Gonçalves O.H. Machado R.A.F. Bolzan A. Antimicrobial activity of microencapsulated lemongrass essential oil and the effect of experimental parameters on microcapsules size and morphology Mater. Sci. Eng. C 2009 29 430 436 10.1016/j.msec.2008.08.025 207. Li R. Jiang S. Jiang Z.T. β-Cyclodextrin supramolecular microcapsules and antioxidant activity of Laurus nobilis essential oil J. Essent. Oil-Bear. Plants 2017 20 1511 1524 10.1080/0972060X.2017.1354724 208. Li Y. Wu C. Wu T. Wang L. Chen S. Ding T. Hu Y. Preparation and characterization of citrus essential oils loaded in chitosan microcapsules by using different emulsifiers J. Food Eng. 2018 217 108 114 10.1016/j.jfoodeng.2017.08.026 209. Mehran M. Masoum S. Memarzadeh M. Microencapsulation of Mentha spicata essential oil by spray drying: Optimization, characterization, release kinetics of essential oil from microcapsules in food models Ind. Crops Prod. 2020 154 112694 10.1016/j.indcrop.2020.112694 210. Adepu S. Khandelwal M. Bacterial cellulose with microencapsulated antifungal essential oils: A novel double barrier release system Materialia 2020 9 100585 10.1016/j.mtla.2020.100585 211. Ribeiro A.D. Marques J. Forte M. Correia F.C. Parpot P. Oliveira C. Pereira A.I. Andrade L. Azenha C. Mendes A. Microencapsulation of citronella oil for solar-activated controlled release as an insect repellent Appl. Mater. Today 2016 5 90 97 10.1016/j.apmt.2016.09.003 212. Lai H. Liu Y. Huang G. Chen Y. Song Y. Ma Y.Q. Yue P. Fabrication and antibacterial evaluation of peppermint oil-loaded composite microcapsules by chitosan-decorated silica nanoparticles stabilized Pickering emulsion templating Int. J. Biol. Macromol. 2021 183 2314 2325 10.1016/j.ijbiomac.2021.05.198 34087300 213. Tang C. Li Y. Pun J. Mohamed Osman A.S. Tam K.C. Polydopamine microcapsules from cellulose nanocrystal stabilized Pickering emulsions for essential oil and pesticide encapsulation Colloids Surf. A Physicochem. Eng. Asp. 2019 570 403 413 10.1016/j.colsurfa.2019.03.049 214. Cai C. Ma R. Duan M. Lu D. Preparation and antimicrobial activity of thyme essential oil microcapsules prepared with gum arabic RSC Adv. 2019 9 19740 19747 10.1039/C9RA03323H 35519357 215. Emadzadeh B. Ghorani B. Naji-Tabasi S. Charpashlo E. Molaveisi M. Fate of β-cyclodextrin-sugar beet pectin microcapsules containing garlic essential oil in an acidic food beverage Food Biosci. 2021 42 101029 10.1016/j.fbio.2021.101029 216. de Medeiros J.A.S. Blick A.P. Galindo M.V. Alvim I.D. Yamashita F. Ueno C.T. Shirai M.A. Grosso C.R.F. Corradini E. Sakanaka L.S. Incorporation of Oregano Essential Oil Microcapsules in Starch-Poly (Butylene Adipate Co-Terephthalate) (PBAT) Films Macromol. Symp. 2019 383 1800052 10.1002/masy.201800052 217. Ngamekaue N. Chitprasert P. Effects of beeswax-carboxymethyl cellulose composite coating on shelf-life stability and intestinal delivery of holy basil essential oil-loaded gelatin microcapsules Int. J. Biol. Macromol. 2019 135 1088 1097 10.1016/j.ijbiomac.2019.06.002 31173831 218. De Matos E.F. Scopel B.S. Dettmer A. Citronella essential oil microencapsulation by complex coacervation with leather waste gelatin and sodium alginate J. Environ. Chem. Eng. 2018 6 1989 1994 10.1016/j.jece.2018.03.002 219. Xiao Z. Liu W. Zhu G. Zhou R. Niu Y. Production and characterization of multinuclear microcapsules encapsulating lavender oil by complex coacervation Flavour Fragr. J. 2014 29 166 172 10.1002/ffj.3192 220. Baiocco D. Preece J.A. Zhang Z. Microcapsules with a fungal chitosan-gum Arabic-maltodextrin shell to encapsulate health-beneficial peppermint oil Food Hydrocoll. Health 2021 1 100016 10.1016/j.fhfh.2021.100016 221. Dima C. Cotârlet M. Alexe P. Dima S. Microencapsulation of essential oil of pimento [Pimenta dioica (L) Merr.] by chitosan/k-carrageenan complex coacervation method Innov. Food Sci. Emerg. Technol. 2014 22 203 211 10.1016/j.ifset.2013.12.020 222. Hernández-Nava R. López-Malo A. Palou E. Ramírez-Corona N. Jiménez-Munguía M.T. Encapsulation of oregano essential oil (Origanum vulgare) by complex coacervation between gelatin and chia mucilage and its properties after spray drying Food Hydrocoll. 2020 109 106077 10.1016/j.foodhyd.2020.106077 223. Qiu L. Zhang M. Adhikari B. Chang L. Microencapsulation of rose essential oil in mung bean protein isolate-apricot peel pectin complex coacervates and characterization of microcapsules Food Hydrocoll. 2022 124 107366 10.1016/j.foodhyd.2021.107366 224. Chen K. Zhang M. Adhikari B. Wang M. Microencapsulation of Sichuan pepper essential oil in soybean protein isolate-Sichuan pepper seed soluble dietary fiber complex coacervates Food Hydrocoll. 2022 125 107421 10.1016/j.foodhyd.2021.107421 225. Rojas-Moreno S. Cárdenas-Bailón F. Osorio-Revilla G. Gallardo-Velázquez T. Proal-Nájera J. Effects of complex coacervation-spray drying and conventional spray drying on the quality of microencapsulated orange essential oil J. Food Meas. Charact. 2018 12 650 660 10.1007/s11694-017-9678-z 226. Li Y. Liu J. He X. Kong D. Zhou C. Wu H. Yang Z. Yang Z. Hu Y. Preparation of Cinnamon Oil-Loaded Antibacterial Composite Microcapsules by In Situ Polymerization of Pickering Emulsion Templates Macromol. Mater. Eng. 2020 305 1900851 10.1002/mame.201900851 227. Revuelta M.V. Bogdan S. Gámez-Espinosa E. Deyá M.C. Romagnoli R. Green antifungal waterborne coating based on essential oil microcapsules Prog. Org. Coat. 2021 151 106101 10.1016/j.porgcoat.2020.106101 228. Matulyte I. Kasparaviciene G. Bernatoniene J. Development of new formula microcapsules from nutmeg essential oil using sucrose esters and magnesium aluminometasilicate Pharmaceutics 2020 12 628 10.3390/pharmaceutics12070628 32635497 229. Dolçà C. Ferrándiz M. Capablanca L. Franco E. Mira E. López F. García D. Microencapsulation of Rosemary Essential Oil by Co-Extrusion/Gelling Using Alginate as a Wall Material J. Encapsulation Adsorpt. Sci. 2015 5 121 130 10.4236/jeas.2015.53010 230. Yingngam B. Kacha W. Rungseevijitprapa W. Sudta P. Prasitpuriprecha C. Brantner A. Response surface optimization of spray-dried citronella oil microcapsules with reduced volatility and irritation for cosmetic textile uses Powder Technol. 2019 355 372 385 10.1016/j.powtec.2019.07.065 231. Maulidna Wirjosentono B. Tamrin Marpaung L. Microencapsulation of ginger-based essential oil (Zingiber cassumunar roxb) with chitosan and oil palm trunk waste fiber prepared by spray-drying method Case Stud. Therm. Eng. 2020 18 100606 10.1016/j.csite.2020.100606 232. Ren W. Tian G. Zhao S. Yang Y. Gao W. Zhao C. Zhang H. Lian Y. Wang F. Du H. Effects of spray-drying temperature on the physicochemical properties and polymethoxyflavone loading efficiency of citrus oil microcapsules LWT—Food Sci. Technol. 2020 133 109954 10.1016/j.lwt.2020.109954 233. Bajac J. Nikolovski B. Lončarević I. Petrović J. Bajac B. Đurović S. Petrović L. Microencapsulation of juniper berry essential oil (Juniperus communis L.) by spray drying: Microcapsule characterization and release kinetics of the oil Food Hydrocoll. 2022 125 107430 10.1016/j.foodhyd.2021.107430 234. Hu Q. Li X. Chen F. Wan R. Yu C.W. Li J. McClements D.J. Deng Z. Microencapsulation of an essential oil (cinnamon oil) by spray drying: Effects of wall materials and storage conditions on microcapsule properties J. Food Process. Preserv. 2020 44 e14805 10.1111/jfpp.14805 235. Asensio C.M. Paredes A.J. Martin M.P. Allemandi D.A. Nepote V. Grosso N.R. Antioxidant Stability Study of Oregano Essential Oil Microcapsules Prepared by Spray-Drying J. Food Sci. 2017 82 2864 2872 10.1111/1750-3841.13951 29095492 236. Shaikh J. Bhosale R. Singhal R. Microencapsulation of black pepper oleoresin Food Chem. 2006 94 105 110 10.1016/j.foodchem.2004.10.056 237. Ozdemir N. Bayrak A. Tat T. Yanık Z.N. Altay F. Halkman A.K. Fabrication and characterization of basil essential oil microcapsule-enriched mayonnaise and its antimicrobial properties against Escherichia coli and Salmonella Typhimurium Food Chem. 2021 359 129940 10.1016/j.foodchem.2021.129940 33957334 238. Qi Z. Zhou H. Xue X. Zhang C. Chen H. Yuan H. Wang C. Fabrication of epoxy sustained-release coatings loaded with urushiol microcapsules containing essential oil for inhibition on drug-resistant Helicobacter pylori and Staphylococcus aureus Prog. Org. Coat. 2021 161 106459 10.1016/j.porgcoat.2021.106459 239. Yang Y.H. Li X.Z. Zhang S. Preparation methods and release kinetics of: Litsea cubeba essential oil microcapsules RSC Adv. 2018 8 29980 29987 10.1039/C8RA05769A 240. Martins W.S. de Araújo J.S.F. Feitosa B.F. Oliveira J.R. Kotzebue L.R.V. Agostinic D.L.S. Oliveira D.L.V. Mazzetto S.E. Cavalcanti M.T. daSilva A.L. Lemongrass (Cymbopogon citratus DC. Stapf) essential oil microparticles: Development, characterization, and antioxidant potential Food Chem. 2021 355 129644 10.1016/j.foodchem.2021.129644 33799254 241. Ocak B. Gum arabic and collagen hydrolysate extracted from hide fleshing wastes as novel wall materials for microencapsulation of Origanum onites L. essential oil through complex coacervation Environ. Sci. Pollut. Res. 2020 27 42727 42737 10.1007/s11356-020-10201-8 32720020 242. Rojas-Moreno S. Osorio-Revilla G. Gallardo-Velázquez T. Cárdenas-Bailón F. Meza-Márquez G. Effect of the cross-linking agent and drying method on encapsulation efficiency of orange essential oil by complex coacervation using whey protein isolate with different polysaccharides J. Microencapsul. 2018 35 165 180 10.1080/02652048.2018.1449910 29513057 243. Almeida A.P. Rodríguez-Rojo S. Serra A.T. Vila-Real H. Simplicio A.L. Delgadilho I. Beirão Da Costa S. Beirão Da Costa L. Nogueira I.D. Duarte C.M.M. Microencapsulation of oregano essential oil in starch-based materials using supercritical fluid technology Innov. Food Sci. Emerg. Technol. 2013 20 140 145 10.1016/j.ifset.2013.07.009 244. Varona S. Kareth S. Martín Á. Cocero M.J. Formulation of lavandin essential oil with biopolymers by PGSS for application as biocide in ecological agriculture J. Supercrit. Fluids 2010 54 369 377 10.1016/j.supflu.2010.05.019 245. Machado L.C. Pelegati V.B. Oliveira A.L. Study of simple microparticles formation of limonene in modified starch using PGSS—Particles from gas-saturated suspensions J. Supercrit. Fluids 2016 107 260 269 10.1016/j.supflu.2015.09.023

PMC009xxxxxx/PMC9099682.txt

==== Front Molecules Molecules molecules Molecules 1420-3049 MDPI 10.3390/molecules27092997 molecules-27-02997 Article Exploring the Parallel G-Quadruplex Nucleic Acid World: A Spectroscopic and Computational Investigation on the Binding of the c-myc Oncogene NHE III1 Region by the Phytochemical Polydatin https://orcid.org/0000-0003-4217-7990 Greco Francesca 1† https://orcid.org/0000-0001-7624-1933 Musumeci Domenica 23† https://orcid.org/0000-0003-0216-9814 Borbone Nicola 14 Falanga Andrea Patrizia 1 https://orcid.org/0000-0002-8100-3726 D’Errico Stefano 1 https://orcid.org/0000-0001-6367-2419 Terracciano Monica 1 Piccialli Ilaria 5 https://orcid.org/0000-0002-6978-542X Roviello Giovanni Nicola 2* https://orcid.org/0000-0003-0765-2127 Oliviero Giorgia 6 Eritja Ramon Academic Editor Montesarchio Daniela Academic Editor Terrazas Montserrat Academic Editor 1 Department of Pharmacy, University of Naples Federico II, Via Domenico Montesano 49, 80131 Naples, Italy; francesca.greco@unina.it (F.G.); nicola.borbone@unina.it (N.B.); andreapatrizia.falanga@unina.it (A.P.F.); stefano.derrico@unina.it (S.D.); monica.terracciano@unina.it (M.T.) 2 Institute of Biostructures and Bioimaging, Italian National Council for Research (IBB-CNR), Area di Ricerca Site and Headquarters-Via Pietro Castellino 111, 80131 Naples, Italy; domenica.musumeci@unina.it 3 Department of Chemistry, University of Naples Federico II, Via Vicinale Cupa Cintia 21, 80126 Naples, Italy 4 ISBE-IT, University of Naples Federico II, 80138 Naples, Italy 5 Division of Pharmacology, Department of Neuroscience, Reproductive and Odontostomatological Sciences, University of Naples Federico II, Via Sergio Pansini 5, 80131 Naples, Italy; ilaria.piccialli@unina.it 6 Department of Molecular Medicine and Medical Biotechnologies, University of Naples Federico II, Via Sergio Pansini 5, 80131 Naples, Italy; golivier@unina.it * Correspondence: giroviel@unina.it † These authors contributed equally to this work. 07 5 2022 5 2022 27 9 299711 3 2022 05 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Trans-polydatin (tPD), the 3-β-D-glucoside of the well-known nutraceutical trans-resveratrol, is a natural polyphenol with documented anti-cancer, anti-inflammatory, cardioprotective, and immunoregulatory effects. Considering the anticancer activity of tPD, in this work, we aimed to explore the binding properties of this natural compound with the G-quadruplex (G4) structure formed by the Pu22 [d(TGAGGGTGGGTAGGGTGGGTAA)] DNA sequence by exploiting CD spectroscopy and molecular docking simulations. Pu22 is a mutated and shorter analog of the G4-forming sequence known as Pu27 located in the promoter of the c-myc oncogene, whose overexpression triggers the metabolic changes responsible for cancer cells transformation. The binding of tPD with the parallel Pu22 G4 was confirmed by CD spectroscopy, which showed significant changes in the CD spectrum of the DNA and a slight thermal stabilization of the G4 structure. To gain a deeper insight into the structural features of the tPD-Pu22 complex, we performed an in silico molecular docking study, which indicated that the interaction of tPD with Pu22 G4 may involve partial end-stacking to the terminal G-quartet and H-bonding interactions between the sugar moiety of the ligand and deoxynucleotides not included in the G-tetrads. Finally, we compared the experimental CD profiles of Pu22 G4 with the corresponding theoretical output obtained using DichroCalc, a web-based server normally used for the prediction of proteins’ CD spectra starting from their “.pdb” file. The results indicated a good agreement between the predicted and the experimental CD spectra in terms of the spectral bands’ profile even if with a slight bathochromic shift in the positive band, suggesting the utility of this predictive tool for G4 DNA CD investigations. Pu22 G-quadruplex c-myc phytochemicals circular dichroism in silico simulations molecular docking CD prediction Department of Pharmacy-University of Naples Federico II grant “Sostegno allo Sviluppo della Ricerca Dipartimentale” (N.B., M.T.)This research was supported by the Department of Pharmacy-University of Naples Federico II grant “Sostegno allo Sviluppo della Ricerca Dipartimentale” (N.B., M.T.). ==== Body pmc1. Introduction Among the noncanonical secondary structures of DNA, G-quadruplex (G4) is an appealing therapeutic target being found in specific regions of the genome such as telomeres and the regulatory regions of many oncogenes including c-kit, c-myc, and bcl-2 [1,2,3,4,5,6,7,8,9,10,11,12,13,14]. Remarkably, the promoter region of c-myc—an oncogene over-expressed in the majority of solid tumors and closely associated with cancer cell apoptosis, proliferation, invasion, cell-cycle arrest, and metastasis—can form a parallel G4 structure via Hoogsteen hydrogen bonds under specific conditions, and has been proposed as an effective target for antitumor drugs [15,16,17,18,19,20,21,22,23]. Particularly, it was found that molecules capable of binding and stabilizing this type of G4 downregulate the expression of c-myc, finally resulting in the apoptosis of cancer cells with great benefit in anticancer therapy [24,25,26]. Trans-polydatin (tPD, Figure 1a), the 3-β-D-glucoside of the well-known nutraceutical trans-resveratrol [27], is a natural polyphenol with documented anti-cancer, anti-inflammatory, cardioprotective, and immunoregulatory effects [28,29]. In a recent work, the G4-binding of tPD was explored toward three cancer-related G-rich DNA sequences, including c-myc, in comparison with a model duplex [30]. Interestingly, tPD displayed a clear binding ability with all the G4s and a higher ability, with respect to its aglycone derivative trans-resveratrol, to discriminate G4 over duplex DNA. Moreover, in vitro assays on melanoma cells proved that tPD significantly reduced telomerase activity, and inhibited cancer cell proliferation [30]. However, the adopted experimental conditions did not allow the detection of any significant conformational changes of the analyzed G4 DNA upon binding with tPD. Moreover, it was not possible to estimate the thermal stability of both c-myc and its complex with tPD, as needed for evaluating any stabilizing or destabilizing effects of the polyphenol on the G4-folded c-myc promoter [30]. On the other hand, other studies clearly indicated that the anticancer effects (including inhibition of cell proliferation and metastasis) of tPD took place through suppressing the c-myc expression, as proven in a model of human cervical cancer [31]. Therefore, conscious of the role of G4-structure binding and stabilization by ligands in c-myc deregulation [32], we decided to examine in more detail the molecular recognition of c-myc G4 by tPD through an approach differing from that previously reported in the literature from both experimental and in silico perspectives. To this aim, the interaction of tPD with c-myc DNA was studied in the present work focusing on the G4 structure formed by the Pu22 region having the sequence 5′-TGAGGGTGGGTAGGGTGGGTAA-3′, a mutated and shorter analog of the sequence known as Pu27 located in the promoter of the c-myc oncogene and associated with the regulation of promoter activity and gene transcription. Circular dichroism (CD) spectra of Pu22, either unliganded or in complex with the tPD, were recorded at variable temperatures in a buffer containing a lower concentration of potassium ions than previously reported [30]. CD spectroscopy is a technique typically employed to verify the formation of several secondary structures of nucleic acids and their analogs [33,34,35,36], including the G4 structure in G-rich DNA sequences [37,38,39,40,41], which also allows one to determine whether the denaturing temperature of a DNA secondary structure is affected by potential ligands [42]. Being aware of the utility of molecular docking in identifying DNA ligands [43,44] through verification of the favored binding sites in a complex, and in the estimation of the binding affinity, we decided to further characterize the molecular recognition of the G4 by tPD, by docking experiments of the tested polyphenol against the c-myc G4. The CD spectrum of the parallel G4 structure of Pu22 was further predicted by DichroCalc software [45], with the aim to verify whether the experimental profile could be reproduced by simulation as described below. 2. Results The effective binding of polydatin to Pu22 had been unequivocally shown by some of us by using various techniques including fluorescence [30]; however, with our work, we aimed to explore some biophysical characteristics of the complex, such as its thermal stability, and give more insights into the molecular aspects of the interaction by using in silico approaches. Our combined experimental and computational work started with the examination of the CD spectral features of Pu22 DNA and its complex with tPD. Moreover, a thermal denaturation study was conducted with both unliganded Pu22 and tPD-Pu22. The observations from CD spectroscopy were then interpreted in the light of the docking experiments performed by us on tPD-Pu22, but also on (tPD-Pu22)-Pu22, (Pu22)2, and tPD-(Pu22)2 molecular systems. 2.1. CD Spectroscopic Analysis of the Binding of Pu22 by tPD With the aim to shed light on the possible mechanisms underlying the previously reported anticancer activity of tPD [31], we evaluated the potential of this polyphenol in binding Pu22. In our CD study, we observed a spectrum for Pu22 corresponding to a G4 with parallel topology, as identified by the characteristic positive band at ~265 nm and the negative one at 240 nm (Figure 1b, black line) [46]. In the presence of the polyphenol, we observed an increase in the positive CD signal at 263 nm accompanied with a 1 nm red-shift in the band maximum, and a concomitant reduction in the CD minimum at 240 nm (Figure 1b, red line). In addition, some differences in the CD spectra were evidenced in the 280–300 nm region. Overall, in the studied conditions, tPD induced a greater degree of structuration in the Pu22 G-quadruplex, as evidenced by the “difference” CD spectrum obtained by subtracting the CD spectrum of the Pu22 G4 to that of the tPD-Pu22 complex (inset of Figure 1b). Then, we studied the effect of tPD on the stability of this G4 DNA by recording, for Pu22 and its mixture with the polyphenol, the CD values at 265 nm as a function of temperature (Figure 1c). We found a slight thermal stabilization in the presence of tPD, detectable by the increased value of the G4 melting temperature (Tm = 64 °C) with respect to the unliganded Pu22 G4 reference (Tm = 62 °C), leading to a ΔTm of +2 °C (Figure 1d and Table 1). Furthermore, the overall variation in the CD signal at the λmax upon heating, i.e., between 40 (folded state) and 90 °C (unfolded), for Pu22 alone or in complex with tPD, was 1.99 and 2.23, respectively, again confirming a higher structuration degree of the quadruplex when bound by the ligand. Specifically, the highest difference in the ΔCD for the two systems was evidenced between 40 and 50 °C (Figure 2, black-line dashed squares, and Table 1). Some differences between the two systems were also detected in the CD spectra recorded at the various temperatures in the 280–300 nm spectral region (Figure 2a,c). 2.2. Molecular Docking Studies The importance of phytochemicals in drug discovery [47] prompted the scientific community to investigate the potential of a plethora of natural products in anticancer strategies by using in silico approaches for a rapid screening or to corroborate and describe at a molecular level in vitro observations. In this context, we used herein in silico methods, and more specifically molecular docking, in analogy to other recent literature examples using polyphenols as anticancer drug candidates [48,49], to deeper analyze the interaction between tPD and the target Pu22 G4, whose sequence is located in a regulatory region of the c-myc oncogene. More in detail, we exploited the Hdock software [50,51] for the computational studies involving DNA. Hdock is used for both macromolecule–macromolecule [50] and small molecule–macromolecule [52] dockings, including those involving DNA and RNA G4s [53,54]. It is worth noting that the docking software provides dimensionless scores (Hdock scores) that are correlated to binding affinities [55]. This allows the comparison to made of the binding affinity of ligands for a given target by simply comparing their docking scores, with the most negative values being associated with the highest affinity ligands [55]. We found by Hdock docking that tPD bound the G4 target in proximity of the G4, G8, G13, and G17 nucleotides (Figure 3, Table 2). Comparing the Hdock scores for the top-1 poses, we can predict that the ligand bound Pu22 with a lower affinity than its aglycone resveratrol (tRES, Table 2), as experimentally shown in the literature [30]. The interactions that emerged by analyzing the top-1–3 poses are held by H-bondings with aromatic rings involving the tPD H1 (Figure 1a) and the guanine residues 4 (3.15 Å, ligand H1–G-ring; π donor H-bond) and 8 (3.05 Å, ligand H1–G ring; π donor H-bond), respectively, in poses 1 and 3 (Figure 3b,d). In pose 2, a H-bond between ligand H2 (Figure 1a) and the O6 (2.14 Å) of guanine residue 10 was also detected (Figure 3c). Interestingly, unlike pose 2 (Figure 3c,e), poses 1 and 3 show the tPD aromatic moieties laying almost parallel to the terminal quartets of the quadruplex (Figure 3), thus suggesting a partial end-stacking interaction of the polyphenol to the G4. The dimerization of Pu22 G4 was described in the literature under some conditions; for example, a quadruplex dimer was clearly evidenced in the solid state [56], whereas in solution, this G4 is present mainly as a monomer [57]. Nonetheless, Jana and Weisz [58] using nondenaturing polyacrylamide gel electrophoresis showed that in solution, MYC-Δ1,6 and, albeit to a much lesser extent, Pu22 (indicated by them as “MYC-Δ1,6[1.2.1]”, carrying two G replacements by T with respect to MYC-Δ1,6) presented dimeric forms corresponding to slower migrating bands, more evident in the former case and somewhat faint, but still detectable, in the case of Pu22 [58]. Similarly, the electrophoretic assays of Moriyama et al. showed for Pu22 (indicated by them as c-myc) a main band and two slower migrating bands [59]. The presence of dimeric Pu22 in solution was suggested also by size exclusion chromatography (SEC), revealing two main SEC peaks for the Pu22 solution that led to the hypothesis of the coexistence of monomeric and dimeric forms in solution [60]. Therefore, we hypothesize that Pu22 in solution is found mainly as a monomer, which justifies its usage in biomolecular studies as a model of G4 DNA not prone to undesirable multimerization, but also, albeit at a much lesser degree, as a dimer. G4 DNA dimer binding by ligands could, in principle, alter the monomer-dimer equilibrium, and importantly, some ligands can induce dimerization of truncated parallel c-myc G-quadruplexes [61]. With all the above considerations in mind, we decided to explore by molecular docking also the propensity of tPD to bind the (Pu22)2 dimer model. We found for the top-ranked pose, as well as poses 1–3 of this docking, less negative Hdock scores (−103.2 and −102.7 ± 0.5, respectively; Table 2) with respect to those found in the case of the docking of the same ligand with the monomeric G4 (−112.1 and −111.7 ± 0.3), suggesting a slightly higher affinity of tPD for the most abundant monomeric form of the Pu22 G4 structure. We also performed DNA–DNA dockings to explore the dimerization of Pu22 G-quadruplex to obtain (Pu22)2 and the effects of tPD on this process. To this scope, in the first case, we docked Pu22 G4 to a second Pu22 G4 unit, set as the target (Figure 4a), while in the second docking, we used the pre-docked tPD-Pu22 G4 for docking to a second Pu22 G4 unit (Figure 4b). We found that tPD-Pu22 G4 binds Pu22 G4 with an affinity 1.3 times lower than that showed by unliganded Pu22 G4 with the same target (Hdock scores (mean of top-1–3 values): −460.1 ± 14.1 vs. −601.3 ± 7.3, respectively). In other terms, it seems that tPD hinders Pu22 G4 dimerization that, in its absence, is more favored (Figure 4a,b), and binds the Pu22 G4 monomer with slightly higher affinity than the dimeric (Pu22)2 G4 (Figure 4b,d). Remarkably, the dimeric form of Pu22 G4 with tPD (Figure 4b and Figure 5a) was predicted to show considerable structural differences with respect to the unliganded (Pu22)2 G4 dimer (Figure 4a). In this regard, it is worth noting how 18 hydrophobic/π–π stacking Pu22-Pu22 intermolecular interactions (pink, Figure 5b) along with six intermolecular H-bonds (green) are predicted to sustain the trimeric complex structure. 2.3. CD Predictions and Comparison with Experimental Spectroscopic Data Furthermore, the solution NMR structure of the monomeric model of Pu22 G4 formed in human c-myc promoter [57] was used to simulate its CD spectrum by DichroCalc [45]. This software is routinely used for obtaining simulated CD spectra of proteins starting from their PDB structure files. In our approach, we applied the method to the prediction of the spectroscopic profile of the G4 structure object of our study. In particular, a positive band at 268 nm and a negative one at 244 were predicted by DichroCalc (Figure 6a), which were, to some extent, in analogy to what we experimentally found by CD (Figure 6b) and was previously described in the literature for the parallel G4 structure of Pu22, though with a bathochromic shift in the bands by about 5 nm. 3. Discussion With this investigation, we aimed to give more insights into the interaction of the natural polyphenol tPD with the G4-forming DNA model of the c-myc promoter Pu22, as the anticancer effects of this phytochemical compound were previously associated to c-myc deregulation [31]. Specifically, a possible mechanism of anticancer activity could be the stabilization of a G4 structure within a regulatory region of this oncogene. Previous attempts [30] in this regard failed to show any stabilization effects of tPD due to the experimental conditions used and notably because of the particularly K+-rich buffer [30]. In this work, we decided to substitute the previously used buffer with PBS, which corresponds to an overall 4.5 mM K+ concentration. The binding of tPD with the parallel Pu22 G4 was confirmed by CD spectroscopy, which showed changes in the CD spectrum of this DNA secondary structure under our experimental conditions, especially in the characteristic positive band centered at 263 nm (Figure 1b). The overall variation in the CD spectrum of Pu22 when bound by tPD was significant and evidenced by the “difference” CD spectrum (inset of Figure 1b). The thermal denaturation profiles in PBS revealed for both Pu22 and tPD-Pu22 sigmoidal shapes with transition midpoint temperatures (Tms) of 62 and 64 °C, respectively (Figure 1c,d, Table 1), denoting a stabilization effect of tPD on the G4. Furthermore, by examining the variations in the CD curves upon heating, we observed that tPD in the complex slowed down the unfolding process of the G4 structure, especially in the 40–50 °C range. Then, to give a tentative interpretation of the experimental findings, we conducted a molecular docking study on different systems including Pu22 monomeric and dimeric G4s and tPD. The most interesting docking results revealed that tPD may bind the monomeric G4 c-myc model (PDB ID: 6AU4 [56]) used in our CD experiments by partial stacking to the terminal G-quartet of the 22-mer sequence Pu22 (Figure 3). The binding involves a region similar to that described previously [30] for the 24-mer G4 structure (PDB ID: 2A5P), in the proximity of nucleotides common to both computational studies, such as G13 [30] (Table 2). The tPD-Pu22 complex is held also by H-bonding interactions with the aromatic rings [62] between the tPD hydrogen H1 and the guanines in positions 4 and 8. There is also an H-bond between ligand H2 and the O6 of guanine 10 (pose 2), but we cannot exclude that other intermolecular forces (for instance, hydrophobic interactions) contribute to the complex formation. Interestingly, the tPD aromatic moieties in two poses out of three lay almost parallel to the quartets of G4 (Figure 3), thus suggesting a partial π–π stacking of the polyphenol to the terminal G4 quartet (end-stacking). In our hypothesis, the partial end-stacking of tPD to the G-quartet could have a role in the experimental CD thermal behavior observed, as this interaction could reinforce the G4 stabilizing it. Our docking suggests that polydatin could bind the G4 structure by end-stacking as reported in the literature for other stilbene derivatives [63]. Interestingly, Esaki et al. [64] found that naphthalene derivatives are able to stack with the quadruplex G-quartet and afforded thermal stabilizations similar to those observed by us with polydatin, which are also comparable to those recorded for polydatin and resveratrol with the G4s tel26 and hTERT [30]. When tPD is bound to Pu22, it leads to the formation of a complex in which 18 hydrophobic/π–π stacking intermolecular interactions along with six intermolecular H-bonds sustain a trimeric structure of polydatin-(Pu22)2, although it prevents Pu22 G4 dimerization with a full eight-floor coplanar system (Figure 4a). In this regard, other ligands of G-quadruplex DNAs were able to induce dimerization in monomeric G4-forming sequences, such as truncated c-myc promoter DNAs [61], leading also to G4 thermal stabilization [65]. 4. Materials and Methods 4.1. Materials All the reagents and solvents were of the highest commercially available quality and were used as received from Sigma-Aldrich (Merck S.r.l., Milan, Italy). Pu22 DNA sequence d[TGAGGGTGGGTAGGGTGGGTAA], purchased by Eurofins (Turin, Italy) in lyophilized and desalted form, was dissolved in nuclease-free bidistilled water and quantified by UV measurements of the absorbance at 260 nm at 95 °C using as extinction molar coefficient of 228,700 M−1 cm−1 (ssDNA, nn model, https://atdbio.com/tools/oligo-calculator, accessed on 3 March 2022). The DNA stock solution had a 200 µM concentration. Stock solutions of tPD ligand (kind gift of Prof. G Ravagnan) were prepared at 8 mM concentration in DMSO. 4.2. CD Studies Circular dichroism (CD) spectra were registered with procedures similar to previous literature reports [66] on a Jasco J-810 (Jasco Europe S.R.L., Cremella, Italy) spectropolarimeter, equipped with a Peltier ETC-505 T temperature controller, in a Hellma (Milan, Italy) quartz cell with a light path of 0.1 cm. The spectra were recorded within the 240–320 nm wavelength range and corrected by subtracting the contribution of the solvents. All experiments were performed in PBS buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, and 1.8 mM KH2PO4, pH 7.4; Sigma Aldrich, Milan, Italy), using 2.5 µM DNA (Pu22), diluted from the stock solution in water, and 125 µM tPD (50 equiv. respect to the DNA), diluted from the stock solution in DMSO. 4.3. CD Denaturation Studies All G4-containing solutions were annealed by heating them at 95 °C for 5 min and then letting them slowly cool down to room temperature (over 16 h). The presented melting curves (obtained by recording CD265nm vs. T in the 40–90 °C temperature range) are representative of three independent experiments. Melting temperature (Tm) values were determined as the temperatures relative to the minima of the 1st derivative plots of the denaturation curves. All experiments were repeated at least three times and all spectra were recorded in triplicate. 4.4. Molecular Docking We conducted our blind molecular docking with the program HDOCK [50,51] using default parameters for all dockings and the PDB entry 6AU4 that is suitable for studies involving dimerization (selecting one of the two G4 monomers) of the Pu22 G4 structure [56]. The HDOCK server uses the iterative knowledge-based scoring function ITScore-PP to rank the top-ten poses provided after each docking run. The HDOCK score furnished by the program is an energy score whose values are listed as dimensionless, and larger negative numbers of the HDOCK score indicate stronger binding interactions between the interacting ligand/macromolecules, which was reported to correlate well to experimental binding affinities. We used the 3D structure of the Pu22 DNA with the PDB (Protein Data Bank) ID: 6AU4 [56]. The 3D structure, including H-atoms, for the natural compound trans-polydatin, was retrieved by us from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/, accessed on 8 November 2021). More details on the HDOCK docking server and on the procedures for docking experiments can be found at http://hdock.phys.hust.edu.cn/ (accessed on 9 November 2021). We analyzed the top-ranked pose (Top-1) and the top-three ranked poses for the complexes predicted by HDOCK according to the energy scores provided by the program as explained in the Results section. Ligand/DNA complexes were visualized by Discovery Studio 2021 software (Accelrys, San Diego, CA, USA) [67] that was used also for analyzing H-bonding between tPD and G4 DNA. 4.5. CD Predictions The prediction of the CD spectrum of the monomeric Pu22 G4 structure was performed using the DichroCalc [45] web server starting from the PDB file of the NMR structure deposited with PDB ID 1XAV. At first, the 1XAV.pdb file was manually edited by replacing the unrecognized “DA, DC, DG, DT” text for deoxyribonucleotides with “A, C, G, T”. Then, the edited PDB file was uploaded as the input file in DichroCalc obtaining the predicted CD spectrum file, which was edited with SpectraGryph 1.2 [68]. The predicted CD spectrum from the “ds” format was finally visualized in Jasco Spectra Manager (JASCO Corporation, Sendai, Japan). 5. Conclusions Here, we described a combined approach including in silico (molecular docking) and experimental (CD binding assay/CD thermal denaturation) analyses, through which we verified that tPD can interact with Pu22, a G4-forming sequence related to the promoter region of the c-myc oncogene, stabilizing this DNA structure. The tPD anticancer activity previously observed in vitro correlates with its stabilizing effects on this cancer-related target. The interaction of tPD with the parallel quadruplex has been proven by CD, showing changes in the CD spectrum of this DNA secondary structure under our experimental conditions, especially in the characteristic positive band centered at 263 nm. Moreover, slight thermal stabilization effects on the G4 by tPD have been revealed by CD melting studies. The binding with the DNA structure has been described in more detail in silico by molecular docking, which suggests that the interaction of tPD with Pu22 G4 may take place through partial end-stacking to the terminal quartet involving deoxynucleotides placed in the external regions of the G4 and the sugar moiety of the ligand. Finally, the exploitation of the DichroCalc web-based server, normally used for the prediction of CD spectra of proteins, for the computation of CD spectra of Pu22 revealed the feasibility of the method for the predictions of CD spectra of G4 DNA. Acknowledgments We thank Antonietta Gargiulo for her technical assistance and help in the literature search. We also thank Glures s.r.l., a spin-off company of the Italian National Research Council (CNR), and Giampietro Ravagnan for kindly providing tPD. Author Contributions All authors contributed to the conceptualization, experimental design, methodology, data analysis, writing, and editing and reviewing of the article. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Sample Availability Not applicable. Figure 1 (a) Chemical structure of tPD; some atoms are numbered as in the docking program. (b) CD spectra of Pu22 2.5 μM (black) and its complex with tPD (red) at 40 °C. Inset shows the “difference” CD spectrum (tPD-Pu22 (mdeg)–Pu22 (mdeg)). (c) CD thermal denaturation curves (CD265 (mdeg) vs. T (°C)) and (d) their first derivatives vs. T plots for Pu22 (2.5 μM, black) and its complex with tPD (red). All experiments were run in PBS, pH 7.4 (optical path length = 0.1 cm). Figure 2 CD spectra of Pu22 (2.5 μM) (a) and its complex with tPD (c) recorded in the 40–90 °C temperature range. Plots of the CD signal at λmax (in mdeg) vs. temperature (in °C) for Pu22 (b) and its complex with tPD (d) derived from panels a and c. All experiments were run in PBS, pH 7.4 (optical path length = 0.1 cm). Figure 3 The docked structures of the tPD-Pu22, with the Pu22 PDB ID: 6AU4, corresponding to the top-1–3 ranked poses: (a,b) pose 1; (c) pose 2; (d) pose 3. Note how in poses 1 and 3, tPD seems to interact by end-stacking and H-bondings with the nucleotides represented in yellow in panels (b,d). Panel (e) reports a different depiction of pose 2 in which the backbone of Pu22 is represented as a white arrow and the base pairs as ladders for clarity. Figure 4 Docking of Pu22 (a) or tPD-Pu22 (b) to another Pu22 unit. Docking of tPD to Pu22 monomer (c) or dimer (d). Hdock scores (mean of top-1–3 values) were also indicated. Figure 5 (a) Detailed pose view of the trimeric complex (tPD-Pu22)-Pu22 of Figure 4b; tPD structure is indicated. (b) Enlargement of the area delimited by the blue rectangle. Figure 6 Theoretical CD spectrum of Pu22 G4 (a) as simulated by DichroCalc [45] using the PDB ID 1XAV, compared with the experimental counterpart (b) obtained for Pu22 at 2.5 μM in PBS. molecules-27-02997-t001_Table 1 Table 1 Summary of the CD and CD melting data for Pu22 and the complex tPD-Pu22. ΔTm is the variation in the melting temperature of the complex with respect to the Pu22 reference; ΔCDmax 40–90 is the difference in the CD value at the λmax between 40 (folded state) and 90 °C (unfolded), whereas ΔCDmax 40–50 is the one between 40 and 50 °C. Entry ΔTm * (°C) ΔCDmax 40–90 °C (mdeg) ΔCDmax 40–50 °C (mdeg) Pu22 0 1.99 0.54 tPD-Pu22 +2 2.23 0.15 * Tm Pu22 = 62 °C. molecules-27-02997-t002_Table 2 Table 2 HDOCK docking scores (for the top-ranked pose and mean value from the top-1–3 poses). The interface nucleotide residues within 5.0 Å from the ligand in the top-1–3 complexes are reported in the last column. Complex HDOCK Score * Top-1 Ranked Pose HDOCK Score Mean Value (Top-1–3 Poses) ± SD Interface Residues tPD/Pu22 –112.1 –111.7 ± 0.3 G4, G6, G8, G10, G13, G15, T16, G17, G19, T20, A21 tRES/Pu22 –120.6 –112.9 ± 7.3 G6, T7, G10, G15, T16, G19, T20, A21 tPD/(Pu22)2 –103.2 –102.7 ± 0.5 G6, G10, T11, G15, T16, G19, T20, A21, G’14, G’15, T’16, G’17, G’18, G’19, T’20, A’21 * The docking energy scores. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Patel D.J. Phan A.T. Kuryavyi V. Human telomere, oncogenic promoter and 5′-UTR G-quadruplexes: Diverse higher order DNA and RNA targets for cancer therapeutics Nucleic Acids Res. 2007 35 7429 7455 10.1093/nar/gkm711 17913750 2. Chen Z.-F. Qin Q.-P. Qin J.-L. Liu Y.-C. Huang K.-B. Li Y.-L. Meng T. Zhang G.-H. Peng Y. Luo X.-J. Stabilization of G-quadruplex DNA, inhibition of telomerase activity, and tumor cell apoptosis by organoplatinum (II) complexes with oxoisoaporphine J. Med. Chem. 2015 58 2159 2179 10.1021/jm5012484 25650792 3. Qin Q.-P. Qin J.-L. Meng T. Lin W.-H. Zhang C.-H. Wei Z.-Z. Chen J.-N. Liu Y.-C. Liang H. Chen Z.-F. High in vivo antitumor activity of cobalt oxoisoaporphine complexes by targeting G-quadruplex DNA, telomerase and disrupting mitochondrial functions Eur. J. Med. Chem. 2016 124 380 392 10.1016/j.ejmech.2016.08.063 27597414 4. Amato J. Pagano B. Borbone N. Oliviero G. Gabelica V. Pauw E.D. D’Errico S. Piccialli V. Varra M. Giancola C. Targeting G-quadruplex structure in the human c-Kit promoter with short PNA sequences Bioconjugate Chem. 2011 22 654 663 10.1021/bc100444v 5. Esposito V. Galeone A. Mayol L. Oliviero G. Virgilio A. Randazzo L. A topological classification of G-quadruplex structures Nucleosides Nucleotides Nucleic Acids 2007 26 1155 1159 10.1080/15257770701527059 18058556 6. Neidle S. The structures of quadruplex nucleic acids and their drug complexes Curr. Opin. Struct. Biol. 2009 19 239 250 10.1016/j.sbi.2009.04.001 19487118 7. Onel B. Lin C. Yang D. DNA G-quadruplex and its potential as anticancer drug target Sci. China Chem. 2014 57 1605 1614 10.1007/s11426-014-5235-3 27182219 8. Siddiqui-Jain A. Grand C.L. Bearss D.J. Hurley L.H. Direct evidence for a G-quadruplex in a promoter region and its targeting with a small molecule to repress c- MYC transcription Proc. Natl. Acad. Sci. USA 2002 99 11593 11598 10.1073/pnas.182256799 12195017 9. Seenisamy J. Rezler E.M. Powell T.J. Tye D. Gokhale V. Joshi C.S. Siddiqui-Jain A. Hurley L.H. The Dynamic Character of the G-Quadruplex Element in the c-MYC Promoter and Modification by TMPyP4 J. Am. Chem. Soc. 2004 126 8702 8709 10.1021/ja040022b 15250722 10. Rankin S. Reszka A.P. Huppert J. Zloh M. Parkinson G.N. Todd A.K. Ladame S. Balasubramanian S. Neidle S. Putative DNA Quadruplex Formation within the Human c-kit Oncogene J. Am. Chem. Soc. 2005 127 10584 10589 10.1021/ja050823u 16045346 11. Fernando H. Reszka A.P. Huppert J. Ladame S. Rankin S. Venkitaraman A.R. Neidle S. Balasubramanian S. A Conserved Quadruplex Motif Located in a Transcription Activation Site of the Human c-kit Oncogene Biochemistry 2006 45 7854 7860 10.1021/bi0601510 16784237 12. Huppert J.L. Balasubramanian S. G-quadruplexes in promoters throughout the human genome Nucleic Acids Res. 2007 35 406 413 10.1093/nar/gkl1057 17169996 13. Phan A.T. Kuryavyi V. Burge S. Neidle S. Patel D.J. Structure of an Unprecedented G-Quadruplex Scaffold in the Human c-kit Promoter J. Am. Chem. Soc. 2007 129 4386 4392 10.1021/ja068739h 17362008 14. Chen Y. Agrawal P. Brown R.V. Hatzakis E. Hurley L. Yang D. The Major G-Quadruplex Formed in the Human Platelet-Derived Growth Factor Receptor β Promoter Adopts a Novel Broken-Strand Structure in K+ Solution J. Am. Chem. Soc. 2012 134 13220 13223 10.1021/ja305764d 22866911 15. Francisco A.P. Paulo A. Oncogene expression modulation in cancer cell lines by DNA G-quadruplex-interactive small molecules Curr. Med. Chem. 2017 24 4873 4904 10.2174/0929867323666160829145055 27573064 16. Ciribilli Y. Singh P. Spanel R. Inga A. Borlak J. Decoding c-Myc networks of cell cycle and apoptosis regulated genes in a transgenic mouse model of papillary lung adenocarcinomas Oncotarget 2015 6 31569 10.18632/oncotarget.5035 26427040 17. Zhang Z. Wu Q. Wu X.-H. Sun F.-Y. Chen L.-M. Chen J.-C. Yang S.-L. Mei W.-J. Ruthenium (II) complexes as apoptosis inducers by stabilizing c-myc G-quadruplex DNA Eur. J. Med. Chem. 2014 80 316 324 10.1016/j.ejmech.2014.04.070 24793882 18. Platella C. Mazzini S. Napolitano E. Mattio L.M. Beretta G.L. Zaffaroni N. Pinto A. Montesarchio D. Dallavalle S. Plant-Derived Stilbenoids as DNA-Binding Agents: From Monomers to Dimers Chem.–A Eur. J. 2021 27 8832 8845 10.1002/chem.202101229 19. Chaudhuri R. Bhattacharya S. Dash J. Bhattacharya S. Recent Update on Targeting c-MYC G-Quadruplexes by Small Molecules for Anticancer Therapeutics J. Med. Chem. 2020 64 42 70 10.1021/acs.jmedchem.0c01145 33355454 20. Sheikh-Zeineddini N. Safaroghli-azar A. Salari S. Bashash D. C-Myc inhibition sensitizes pre-B ALL cells to the anti-tumor effect of vincristine by altering apoptosis and autophagy: Proposing a probable mechanism of action for 10058-F4 Eur. J. Pharmacol. 2020 870 172821 10.1016/j.ejphar.2019.172821 31770526 21. Mazzini S. Gargallo R. Musso L. De Santis F. Aviñó A. Scaglioni L. Eritja R. Di Nicola M. Zunino F. Amatulli A. Stabilization of c-KIT G-Quadruplex DNA Structures by the RNA Polymerase I Inhibitors BMH-21 and BA-41 Int. J. Mol. Sci. 2019 20 4927 10.3390/ijms20194927 31590335 22. Chanvorachote P. Sriratanasak N. Nonpanya N. C-myc Contributes to Malignancy of Lung Cancer: A Potential Anticancer Drug Target Anticancer Res. 2020 40 609 618 10.21873/anticanres.13990 32014901 23. Dallavalle S. Mattio L.M. Artali R. Musso L. Aviñó A. Fàbrega C. Eritja R. Gargallo R. Mazzini S. Exploring the Interaction of Curaxin CBL0137 with G-Quadruplex DNA Oligomers Int. J. Mol. Sci. 2021 22 6476 10.3390/ijms22126476 34204214 24. Georgiades S.N. Abd Karim N.H. Suntharalingam K. Vilar R. Interaction of metal complexes with G-quadruplex DNA Angew. Chem. Int. Ed. 2010 49 4020 4034 10.1002/anie.200906363 25. Brooks T.A. Hurley L.H. Targeting MYC expression through G-quadruplexes Genes Cancer 2010 1 641 649 10.1177/1947601910377493 21113409 26. Yan S. Zheng Z. Cui Y. Mi Y. Liu H. Zhao X. Luo D. C-myc g-quadruplex stabilization and cytotoxicity of an oxadiazole-bearing ruthennium (II) complex Rev. Roum. De Chim. 2021 66 423 433 27. Platella C. Guida S. Bonmassar L. Aquino A. Bonmassar E. Ravagnan G. Montesarchio D. Roviello G.N. Musumeci D. Fuggetta M.P. Antitumour activity of resveratrol on human melanoma cells: A possible mechanism related to its interaction with malignant cell telomerase Biochim. Et Biophys. Acta (BBA)-Gen. Subj. 2017 1861 2843 2851 10.1016/j.bbagen.2017.08.001 28. Quagliariello V. Berretta M. Buccolo S. Iovine M. Paccone A. Cavalcanti E. Taibi R. Montopoli M. Botti G. Maurea N. Polydatin reduces cardiotoxicity and enhances the anticancer effects of sunitinib by decreasing pro-oxidative stress, pro-inflammatory cytokines, and nlrp3 inflammasome expression Front. Oncol. 2021 11 680758 10.3389/fonc.2021.680758 34178667 29. Lanzilli G. Cottarelli A. Nicotera G. Guida S. Ravagnan G. Fuggetta M.P. Anti-inflammatory effect of resveratrol and polydatin by in vitro IL-17 modulation Inflammation 2012 35 240 248 10.1007/s10753-011-9310-z 21369944 30. Platella C. Raucci U. Rega N. D’Atri S. Levati L. Roviello G.N. Fuggetta M.P. Musumeci D. Montesarchio D. Shedding light on the interaction of polydatin and resveratrol with G-quadruplex and duplex DNA: A biophysical, computational and biological approach Int. J. Biol. Macromol. 2020 151 1163 1172 10.1016/j.ijbiomac.2019.10.160 31747572 31. Bai L. Ma Y. Wang X. Feng Q. Zhang Z. Wang S. Zhang H. Lu X. Xu Y. Zhao E. Polydatin Inhibits Cell Viability, Migration, and Invasion Through Suppressing the c-Myc Expression in Human Cervical Cancer Front. Cell Dev. Biol. 2021 9 587218 10.3389/fcell.2021.587218 33912552 32. Mathad R.I. Hatzakis E. Dai J. Yang D. c-MYC promoter G-quadruplex formed at the 5′-end of NHE III 1 element: Insights into biological relevance and parallel-stranded G-quadruplex stability Nucleic Acids Res. 2011 39 9023 9033 10.1093/nar/gkr612 21795379 33. Roviello V. Musumeci D. Mokhir A. Roviello G.N. Evidence of protein binding by a nucleopeptide based on a thyminedecorated L-diaminopropanoic acid through CD and in silico studies Curr. Med. Chem. 2021 28 5004 5015 10.2174/0929867328666210201152326 33593247 34. Musumeci D. Mokhir A. Roviello G.N. Synthesis and nucleic acid binding evaluation of a thyminyl l-diaminobutanoic acid-based nucleopeptide Bioorganic Chem. 2020 100 103862 10.1016/j.bioorg.2020.103862 35. Musumeci D. Ullah S. Ikram A. Roviello G.N. Novel insights on nucleopeptide binding: A spectroscopic and In Silico investigation on the interaction of a thymine-bearing tetrapeptide with a homoadenine DNA J. Mol. Liq. 2022 347 117975 10.1016/j.molliq.2021.117975 36. Amato J. Stellato M.I. Pizzo E. Petraccone L. Oliviero G. Borbone N. Piccialli G. Orecchia A. Bellei B. Castiglia D. PNA as a potential modulator of COL7A1 gene expression in dominant dystrophic epidermolysis bullosa: A physico-chemical study Mol. BioSyst. 2013 9 3166 3174 10.1039/c3mb70283a 24121392 37. Pirota V. Platella C. Musumeci D. Benassi A. Amato J. Pagano B. Colombo G. Freccero M. Doria F. Montesarchio D. On the binding of naphthalene diimides to a human telomeric G-quadruplex multimer model Int. J. Biol. Macromol. 2021 166 1320 1334 10.1016/j.ijbiomac.2020.11.013 33166559 38. Platella C. Capasso D. Riccardi C. Musumeci D. DellaGreca M. Montesarchio D. Natural compounds from Juncus plants interacting with telomeric and oncogene G-quadruplex structures as potential anticancer agents Org. Biomol. Chem. 2021 19 9953 9965 10.1039/D1OB01995C 34747958 39. Oliviero G. Borbone N. Amato J. D’Errico S. Galeone A. Piccialli G. Varra M. Mayol L. Synthesis of quadruplex-forming tetra-end-linked oligonucleotides: Effects of the linker size on quadruplex topology and stability Biopolym. Orig. Res. Biomol. 2009 91 466 477 10.1002/bip.21153 40. Oliviero G. Amato J. Borbone N. Galeone A. Varra M. Piccialli G. Mayol L. Synthesis and characterization of DNA quadruplexes containing T-tetrads formed by bunch-oligonucleotides Biopolym. Orig. Res. Biomol. 2006 81 194 201 10.1002/bip.20399 41. Fuggetta M.P. De Mico A. Cottarelli A. Morelli F. Zonfrillo M. Ulgheri F. Peluso P. Mannu A. Deligia F. Marchetti M. Synthesis and enantiomeric separation of a novel spiroketal derivative: A potent human telomerase inhibitor with high in vitro anticancer activity J. Med. Chem. 2016 59 9140 9149 10.1021/acs.jmedchem.6b01046 27592503 42. Arounaguiri S. Easwaramoorthy D. Ashokkumar A. Dattagupta A. Maiya B.G. Cobalt (III), nickel (II) and ruthenium (II) complexes of 1, 10-phenanthroline family of ligands: DNA binding and photocleavage studies J. Chem. Sci. 2000 112 1 17 10.1007/BF02704295 43. Ricci C.G. Netz P.A. Docking studies on DNA-ligand interactions: Building and application of a protocol to identify the binding mode J. Chem. Inf. Model. 2009 49 1925 1935 10.1021/ci9001537 19655805 44. Mulliri S. Laaksonen A. Spanu P. Farris R. Farci M. Mingoia F. Roviello G.N. Mocci F. Spectroscopic and in silico studies on the interaction of substituted pyrazolo [1, 2-a] benzo [1, 2, 3, 4] tetrazine-3-one derivatives with c-myc G4-DNA Int. J. Mol. Sci. 2021 22 6028 10.3390/ijms22116028 34199659 45. Bulheller B.M. Hirst J.D. DichroCalc—Circular and linear dichroism online Bioinformatics 2009 25 539 540 10.1093/bioinformatics/btp016 19129206 46. Kong D.-M. Wu J. Wang N. Yang W. Shen H.-X. Peroxidase activity–structure relationship of the intermolecular four-stranded G-quadruplex–hemin complexes and their application in Hg2+ ion detection Talanta 2009 80 459 465 10.1016/j.talanta.2009.07.010 19836504 47. Egbuna C. Kumar S. Ifemeje J.C. Ezzat S.M. Kaliyaperumal S. Phytochemicals as Lead Compounds for New Drug Discovery Elsevier Amsterdam, The Netherlands 2019 48. Mechchate H. Costa de Oliveira R. Es-Safi I. Vasconcelos Mourão E.M. Bouhrim M. Kyrylchuk A. Soares Pontes G. Bousta D. Grafov A. Antileukemic Activity and Molecular Docking Study of a Polyphenolic Extract from Coriander Seeds Pharmaceuticals 2021 14 770 10.3390/ph14080770 34451867 49. Artali R. Beretta G. Morazzoni P. Bombardelli E. Meneghetti F. Green tea catechins in chemoprevention of cancer: A molecular docking investigation into their interaction with glutathione S-transferase (GST P1-1) J. Enzym. Inhib. Med. Chem. 2009 24 287 295 10.1080/14756360802177282 50. Yan Y. Zhang D. Zhou P. Li B. Huang S.-Y. HDOCK: A web server for protein–protein and protein–DNA/RNA docking based on a hybrid strategy Nucleic Acids Res. 2017 45 W365 W373 10.1093/nar/gkx407 28521030 51. Yan Y. Tao H. He J. Huang S.-Y. The HDOCK server for integrated protein–protein docking Nat. Protoc. 2020 15 1829 1852 10.1038/s41596-020-0312-x 32269383 52. Majumder A. Mondal S.K. Mukhoty S. Bag S. Mondal A. Begum Y. Sharma K. Banik A. Virtual Screening and Docking Analysis of Novel Ligands for Selective Enhancement of Tea (Camellia sinensis) Flavonoids Food Chem. X 2022 13 100212 10.1016/j.fochx.2022.100212 35498963 53. Roy A. Chatterjee O. Banerjee N. Roychowdhury T. Dhar G. Mukherjee G. Chatterjee S. Curcumin arrests G-quadruplex in the nuclear hyper-sensitive III1 element of c-MYC oncogene leading to apoptosis in metastatic breast cancer cells J. Biomol. Struct. Dyn. 2021 1 17 10.1080/07391102.2021.1940284 54. Ji D. Juhas M. Tsang C.M. Kwok C.K. Li Y. Zhang Y. Discovery of G-quadruplex-forming sequences in SARS-CoV-2 Brief. Bioinform. 2021 22 1150 1160 10.1093/bib/bbaa114 32484220 55. Stoddard S.V. Wallace F.E. Stoddard S.D. Cheng Q. Acosta D. Barzani S. Bobay M. Briant J. Cisneros C. Feinstein S. In silico design of peptide-based SARS-CoV-2 fusion inhibitors that target wt and mutant versions of SARS-CoV-2 HR1 Domains Biophysica 2021 1 311 327 10.3390/biophysica1030023 56. Stump S. Mou T.-C. Sprang S.R. Natale N.R. Beall H.D. Crystal structure of the major quadruplex formed in the promoter region of the human c-MYC oncogene PLoS ONE 2018 13 e0205584 10.1371/journal.pone.0205584 30312328 57. Ambrus A. Chen D. Dai J. Jones R.A. Yang D. Solution structure of the biologically relevant G-quadruplex element in the human c-MYC promoter. Implications for G-quadruplex stabilization Biochemistry 2005 44 2048 2058 10.1021/bi048242p 15697230 58. Jana J. Weisz K. A Thermodynamic Perspective on Potential G-Quadruplex Structures as Silencer Elements in the MYC Promoter Chem.–A Eur. J. 2020 26 17242 17251 10.1002/chem.202002985 59. Moriyama K. Yoshizawa-Sugata N. Masai H. Oligomer formation and G-quadruplex binding by purified murine Rif1 protein, a key organizer of higher-order chromatin architecture J. Biol. Chem. 2018 293 3607 3624 10.1074/jbc.RA117.000446 29348174 60. Marzano M. Falanga A.P. Marasco D. Borbone N. D’Errico S. Piccialli G. Roviello G.N. Oliviero G. Evaluation of an Analogue of the Marine ε-PLL Peptide as a Ligand of G-quadruplex DNA Structures Mar. Drugs 2020 18 49 10.3390/md18010049 61. Funke A. Karg B. Dickerhoff J. Balke D. Müller S. Weisz K. Ligand-Induced Dimerization of a Truncated Parallel MYC G-Quadruplex ChemBioChem 2018 19 505 512 10.1002/cbic.201700593 29228465 62. Brinkley R.L. Gupta R.B. Hydrogen bonding with aromatic rings AIChE J. 2001 47 948 953 10.1002/aic.690470417 63. O’Hagan M.P. Peñalver P. Gibson R.S.L. Morales J.C. Galan M.C. Stiff-Stilbene Ligands Target G-Quadruplex DNA and Exhibit Selective Anticancer and Antiparasitic Activity Chem.–A Eur. J. 2020 26 6224 6233 10.1002/chem.201905753 32030823 64. Esaki Y. Islam M.M. Fujii S. Sato S. Takenaka S. Design of tetraplex specific ligands: Cyclic naphthalene diimide Chem. Commun. 2014 50 5967 5969 10.1039/C4CC01005A 24752309 65. Bianchi F. Comez L. Biehl R. D’Amico F. Gessini A. Longo M. Masciovecchio C. Petrillo C. Radulescu A. Rossi B. Structure of human telomere G-quadruplex in the presence of a model drug along the thermal unfolding pathway Nucleic Acids Res. 2018 46 11927 11938 10.1093/nar/gky1092 30407585 66. Pradeep T.P. Tripathi S. Barthwal R. Molecular recognition of parallel quadruplex [d-(TTGGGGT)] 4 by mitoxantrone: Binding with 1: 4 stoichiometry leads to telomerase inhibition RSC Adv. 2016 6 71652 71661 10.1039/C6RA05266E 67. Pawar S.S. Rohane S.H. Review on discovery studio: An important tool for molecular docking Asian J. Res. Chem. 2021 14 86 88 10.5958/0974-4150.2021.00014.6 68. Menges F. Spectragryph-Optical Spectroscopy Software, Version 1.2. 8 2018 Available online: http://www.effemm2.de/spectragryph (accessed on 1 August 2018)

PMC009xxxxxx/PMC9099683.txt

==== Front Foods Foods foods Foods 2304-8158 MDPI 10.3390/foods11091196 foods-11-01196 Article Surface Modification via Dielectric Barrier Discharge Atmospheric Cold Plasma (DBD–ACP): Improved Functional Properties of Soy Protein Film https://orcid.org/0000-0002-0426-6653 Li Zhibing 1 Deng Shanggui 1 Chen Jing 12* Cobos Angel Academic Editor 1 Zhejiang Provincial Key Laboratory of Health Risk Factors for Seafood, Collaborative Innovation Center of Seafood Deep Processing, College of Food and Pharmacy, Zhejiang Ocean University, Zhoushan 316022, China; omar18355099192@163.com (Z.L.); dengshanggui@163.com (S.D.) 2 Key Laboratory of Health Risk Factors for Seafood of Zhejiang Province, Zhoushan 316022, China * Correspondence: chenjing1979@126.com 20 4 2022 5 2022 11 9 119624 2 2022 18 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Atmospheric cold plasma (ACP), a novel technology, has been widely adopted as an efficient approach in surface modification of the film. The effect of ACP treatment on the physicochemical and structural properties of soy protein film were investigated. As a result, the optimal conditions for the preparation of the film were determined for soy protein (10%), glycerol (2.8%), ACP treatment at 30 kV for 3 min, on the basis of elongation at the break, and water vapor permeability. Under the optimal conditions, the ACP–treated films exhibited enhanced polarity according to the increased values of solubility, swelling index, and moisture content, compared with the untreated counterpart. An increase in the hydrophilicity is also confirmed by the water contact angle analysis, which decreased from 87.9° to 77.2° after ACP pretreatment. Thermostability was also improved by ACP exposure in terms of DSC analysis. SEM images confirmed the tiny pores and cracks on the surface of film could be lessened by ACP pretreatment. Variations in the Fourier transform infrared spectroscopy indicated that some hydrophilic groups were formed by ACP pretreatment. Atomic force microscopy data revealed that the roughness of soy protein film which was pretreated by ACP was lower than that of the control group, with an Rmax value of 88.4 nm and 162.7 nm for the ACP- treated and untreated samples, respectively. The soy protein film was characterized structurally by FT–IR and DSC, and morphological characterization was done by SEM and AFM. The soy protein film modified by ACP was more stable than the control group. Hence, the great potential in improving the properties of the film enables ACP treatment to be a feasible and promising alternative to other modification methods. atmospheric cold plasma soy protein film modification properties ==== Body pmc1. Introduction With the increasing harm of plastics to the environment, the security of the living environment has gained considerable attention [1]. Nowadays, searching for alternatives to plastics has become a hot issue. Bio-based materials evoked intensive interests due to ease of processability, abundant resources, and eco-friendliness. Among them, soy protein, a widely distributed vegetable protein, has been undoubtedly considered as the most important material, ascribed to its extremely low price, easy availability, high protein content, and excellent film-forming properties [2]. Hence, films made from soy protein can be applied in edible packaging and food preservation. It has gradually attracted attention as an alternative to plastics [3]. Nevertheless, films prepared from soy protein are difficult to be employed in food packaging unless its drawbacks, such as high rigidity, easiness to grow bacteria, and poor stability, are improved [4,5]. In this regard, a series of strategies has been developed to modify the soy protein film, thus to achieve desirable functions. Among them, addition of exogenous compounds has been extensively utilized to improve the film properties. Glycerol, 1- and 2- propylene glycol, and sorbitol were reported to be added into the soy protein solution to enhance the mechanical strength of the film [6]. The supplement of citric acid, nalidixic acid [7], and mandelic acid [8] to soy protein could enhance the antibacterial properties of the film. The barrier properties of the soy protein film were improved through adding CaCl2 [9], ZnSe [10], and ZnS [11] nanoparticles [12]. However, chemical additives may have adverse health effects, such as toxicity, food allergies, and impaired nutrition. Heating, as the most common physical modification of protein, would lead to the dissociation of quaternary structures and protein subunit denaturation [13]. Other solvent-free techniques, such as pressure treatment, ultrasound, pulsed electric field, microwaves, and gramma irradiation, have also been reported to modify protein functionally by destroying the internal chemical bonds of protein and making it re-cross-linking [14]. These non-thermal processes demonstrated great potential in protein modification, but some of them still cannot meet the requirement of desirable functions. Therefore, it is necessary to develop a novel approach for current food and material industries. Atmospheric cold plasma (ACP) is a novel non-thermal and pollution-free technology [15]. It can produce a large number of reactive particles, including charged particles, free radicals, photons, and various radiations, through high-voltage discharge under normal temperature conditions [16]. ACP has been widely applied in the food industry owing to its high microbial decontamination effectiveness. As reported, ACP exhibited excellent eliminating effects on Escherichia coli [17], Listeria [18], Salmonella, Shigella [19], Staphylococcus aureus [20], and Thermobacteria [21]. Besides, it is worth noting that cold plasma is capable of selective modification of starch and protein [16]. As a promising modification alternative, ACP displayed several advantages, including no generation of thermal damage, free of additives, and inexpensive operation costs [22]. Moreover, ACP treatment was also proved to be an efficient modification technology without destroying bulk attributes, since the treatment only affected a few nanometers below the surface of the materials [23]. ACP treatment for wheat flour at 80 kV for 5–30 min resulted in an improvement of hydration properties [24]. It was demonstrated that emulsion stability and water holding capacity were prominently enhanced for peanut isolated protein subjected to ACP treatment [25]. Pankaj et al. [26] observed that the roughness of polylactic acid film was increased after ACP treatment. It was also observed that strength and surface hydrophobicity of zein film could be reinforced by ACP treatment [27]. Similar results were also obtained by Oh et al. [28], who described that ACP treatment resulted in improved tensile strength and a moisture barrier of soybean meal-based edible film. Although several studies have reported that cold plasma was able to interact with protein film, the effect of cold plasma technology on the soy protein film has been scarcely investigated [29]. In order to provide a better understanding of ACP interactions with this edible film, in this study, dielectric barrier discharge (DBD) employing atmospheric air as a working gas was used to generate CP. Soy protein was subjected to ACP treatment prior to preparing the film. The physical properties, thermal properties, and microscopic morphology of the soy protein film were evaluated. 2. Materials and Methods 2.1. Materials Soy protein was purchased from Linyi Shansong Biological Products Co., Ltd. (Linyi, China). All chemicals were of analytical grade purity and commercially available. 2.2. ACP Treatment of Soy Protein Solution The protein solution was produced in four steps. First, 100 mL of deionized water was mixed with 10 g of soy protein powder. The pH of the soy protein solution was adjusted to 9.5–10, which was placed in a 40 °C water bath for 15 min and then stirred with a magnetic stirrer for 15 min. The protein solution, after cooling for 30 min, was filtered with three layers of gauze to remove air bubbles. The protein solution was subjected to DBD–ACP (Phenix Technologies, Accident, MD, USA) treatment which was reported in our previous study [30]. The schematic diagram of the device is shown in Figure 1. Parameters including ACP treatment voltage and exposure time were investigated. Soy protein was treated at a fixed exposure time (3 min) with varying voltages of 10, 20, 30, 40, and 50 kV. The exposure time (1, 2, 3, 4, 5 min) was also assessed at a fixed voltage (30 kV). 2.3. Preparation of Soy Protein Film Soy protein film was prepared by hot-air drying based on the method provided by Kumar et al. [8], with slight modification. Glycerol was added to the soy protein solution after ACP treatment. The sample was stirred with a magnetic stirrer for 15 min and then it was allowed to stand for 1 h. Three layers of gauze were used to remove scum in the soy protein solution. About 40 mL of the treated sample solution was collected on a constant-temperature platform and poured into a 20 cm × 20 cm silicone mold. The samples were heated at 35 °C on a constant temperature platform for 4 h and then transferred to a desiccator. Hot air was blown until the solvent was completely volatilized (25 °C). The soy protein film was stored in a desiccator at 25 ± 0.1 °C (relative humidity 50%) for 24 h prior to the test. Film prepared from soy protein without ACP treatment was considered as the control group. The amount of soy protein (8%, 9%, 10%, 11%, and 12%) was investigated with a fixed concentration of glycerol (2.4%). Meanwhile, the amount of glycerol (1.6%, 2.0%, 2.4%, 2.8%, and 3.2%) was also evaluated with a fixed amount of soy protein (10%). 2.4. Mechanical Properties of Soy Protein Film Mechanical properties of the soy protein film were tested by an electric tensile testing machine (Dongguan Zhitake Precision Instrument Co., Ltd., Dongguan, China), according to Dong et al. [27], with a slight modification. A special knife was used to cut the film sample with a width of 24 mm and length of 50 mm prior to measurement. The test speed was 10 mm/min. Elongation at the break (E, %) was determined using Equation (1). (1) E=L1−L0L0×100% where L0 (mm) is the initial length of the film, and L1 (mm) is the length at break. 2.5. Water Vapor Permeability (WVP) The moisture permeability of the control and ACP -treated soy protein film was evaluated by water vapor permeation instrument (PERMATRAN -W 1/50, Mocon, Minneapolis, MN, USA) according to the procedure of Wiles et al. [31], with minor modifications. The film (100 mm2 film area) was placed uniformly in a closed container with a relative humidity of 50% ± 1% at 25 °C and for 48 h. Water vapor permeability (WVP) was expressed as Equation (2). (2) WVP=WVTR×tPw where WVP is the water vapor transmission coefficient (g·mm/m2·d·kPa), WVTR is the water vapor transmission rate (g/m2·d), t is the average thickness of the film (mm), and PW is the water pressure difference of vapor passing from one side of the film to the other side (kPa). 2.6. Moisture Content (MC) Soy protein films were weighed before and after drying in an electric oven at 110 °C until a constant weight was obtained [32]. MC was expressed as Equation (3). (3) MC(%)=M−M1M×100% where MC is the moisture content (%), M is the initial weight of soy protein film (g), and M1 is the constant weight of soy protein film. 2.7. Solubility (So) The solubility in water of soy protein film was determined according to the method described by Peng et al. [32]. Soy protein film (2 cm × 2 cm) was soaked in 30 mL of pure water. Then, the sample was placed in an oscillation box at 25 °C for 24 h and dried to a constant weight. The solubility was calculated as Equation (4):(4) So(%)=M−M1M×100% where So is the solubility (%), M is the initial weight of soy protein film (g), and M1 is the constant weight of soy protein film (g). 2.8. Swelling Index (Si) The swelling index (Si) reflects the water absorption capacity of soy protein film [33]. Film pieces (2 cm × 2 cm) were dried at 70 °C for 24 h in a vacuum oven to obtain the initial dry weight. The sample was transferred into a 100 mL beaker with 30 mL pure water and soaked in the water at 25 °C for 24 h. Subsequently, the moisture on the film surface was sucked with filter paper and weighed up. The swelling index is expressed as Equation (5). (5) Si(%)=M1−MM×100% where Si is the swelling index (%), M1 is the quality of soy protein film after soaking (g), and M is initial dry weight of soy protein film (g). 2.9. Differential Scanning Calorimetry (DSC) The thermodynamic properties of soy protein film were determined by NETZSCH 200 F3 DSC (NETZSCH, Selb, Germany), adopting the protocol of Dong et al. [27]. Approximately 8–12 mg of the sample was weighed and placed in an aluminum pan with an empty aluminum pan as reference. Samples were subjected to two heating–cooling cycles from 20 °C to 200 °C at a rate of 10 °C/min. Denaturation temperature (Td), glass transition temperature (Tg), heat, and Denaturation Enthalpy were calculated with 89TA-60WS software. 2.10. Scanning Electron Microscopy (SEM) The surface morphology of the control and ACP-treated soy protein film were recorded by a cold-field emission scanning electron microscope (Su8010, Hitachi, Japan) [3]. Prior to the measurement, all samples were fixed on the stage and coated with a layer of gold for 5 min. Afterwards, the sample was placed on the scanning electron microscope under an acceleration voltage of 5.0 kV to capture images. 2.11. Fourier Transform Infrared Spectroscopy (FTIR) The infrared spectrum of the soy protein film was analyzed by a Fourier spectrometer (FT/IR-650, Thermo Nicolet Inc., Waltham, MA, USA) within a range of 4000 to 500 cm−1, at a resolution of 4 cm−1 [5]. The KBr tablet method was adopted, and a mixture composed of soy protein film (1 mg) and KBr (50–100 mg) was grounded into fine powder. The powder was compressed into tablets and placed in the sample chamber. 2.12. Atomic Force Microscopy (AFM) Surface morphology of the soy protein film was assessed following the procedure adapted from Pankaj et al. [26], with some modifications by Atomic Force Microscope (Dimension Icon, Bruker, Germany) operating in intermittent contact (tap) mode. Images were collected at a fixed scan rate of 0.5 Hz. The AFM images provided topographic images as well as quantitative data. The scanning area was 25 μm2. 2.13. Water Contact Angle (WCA) Water contact angle was used to evaluate the hydrophilicity and hydrophobicity of the surface of the soy protein film [8]. The measurement was conducted by Optical Contact Angle Measuring Device (OCA20, Germany Dataphysics, Filderstadt, Germany). About 5.5 μL of distilled water was dropped on the film surface with a micro-syringe. 2.14. Statistical Analysis All experiments were carried out in triplicate. Analysis of variance (ANOVA) was used to evaluate differences among means, and statistical software was adopted to compare the means of different groups by Turkey’s test at the 5% significance level (p < 0.05). Data for each treatment were collected in triplicate, and their results were expressed as average ± standard deviation and analyzed using the descriptive statistics function in Origin 8.5 software (Origin Lab, Northampton, MA, USA). 3. Results 3.1. Factors Affecting Elongation at Break and Water Vapor Permeability of the Soy Protein Film The effect of several parameters, including soybean protein concentration, glycerol concentration, ACP treatment voltage, and exposure time, were investigated in terms of elongation at break and water vapor permeability, which are regarded as important indexes to display the packaging performance of soy protein film. As is known, elongation at the break was used to evaluate the mechanical properties of the film, while water vapor permeability (WVP) can be adopted to indicate the barrier property. 3.1.1. Effect of Soybean Protein Concentration Films with different amounts of soybean protein were prepared and the optimal concentration was investigated when glycerol concentration was fixed as 2.4%, and ACP treatment voltage and time were set as 30 kV and 3 min, respectively. As depicted in Figure 2, elongation at break and water vapor transmittance of the soy protein film were obtained under different soybean protein concentrations. As the soy protein concentration increased, the elongation at the break enhanced at first and then declined for the ACP-treated film, while a steady decrease was observed for the counterpart. The maximum value of 227.37% was achieved for the ACP-treated film when the content of soybean protein was 10%. As reported, soy protein concentration was highly related to the properties of the protein film [30]. The poor toughness was a serious problem, which limited the widespread utilization of soy protein film [34]. For all the tested concentrations, ranging from 8% to 12%, elongation at the break of the treated film was significantly higher (p < 0.05) than that of the control group, demonstrating that cold plasma treatment was efficient in modifying the target protein. Elongation at the break of the treated group was 143.73% higher than that of the control group. Meanwhile, according to Figure 2b, water vapor permeability of the film exerted an extremely slight decrease with the increase in soy protein concentration both for the ACP-treated and untreated group. However, an obvious enhancement of WVP was detected between the ACP-treated and untreated group for all the tested soy protein concentrations, confirming the modification efficiency by ACP. Hence, the optimum soy protein content was selected as 10%. 3.1.2. Effect of Glycerol Concentration As a plasticizer, glycerol is commonly used in edible films due to its small molecular size and abundant functional hydroxyl groups. Hence, the addition of glycerol is of great importance in the formation of soy protein film, and the optimal concentration was evaluated under a fixed soy protein content of 10%, and ACP treatment conditions of 30 kV for 3 min. As shown in Figure 3a, with the increase of glycerol concentration, the elongation at the break increased first and then decreased for the ACP–treated film, while a gradual increase was detected for the untreated film. The elongation at the break reached the peak (223.32%), with the glycerol addition of 2.8%. It was found that when the addition of glycerol was 1.6%, the film was difficult to uncover and easily broken. Furthermore, the rigidity and brittleness of the soy protein film turned out to be relatively high. The packaging performance of the film became extremely poor. As is reported, glycerol was embedded in the soy protein to enhance the plasticity of the film, which made the film soft and elastic [35]. Moreover, as the glycerol concentration increased, WVP of soy protein film displayed a gradual downward trend both for the ACP–treated and untreated samples (Figure 3b). It was observed that WVP dropped from 13.48 g·mm/m2·d·kPa to 6.02 g·mm/m2·d·kPa when the glycerol concentration raised from 1.6% to 3.6%, which was predominantly due to the higher capacity of higher concentration of glycerol in promoting protein cross-linking. Cracks and numerous pores would appear after the protein was prepared in the film, especially for the films with low concentrations or without the addition of glycerol. The compactness of the film would be improved owing to the supplement of glycerol, which was reported to be embedded in the gaps of the soy protein [5]. No significant difference (p > 0.05) of WVP was recorded between the treated and control samples when the glycerol content varied from 2.4% to 3.2%. Thus, taken with the effect on elongation at the break, the optimal glycerol content of soy protein film was finally determined to be 2.8%. Similarly, ACP treatment resulted in a remarkable increase in elongation at the break and a notable decrease of WVP, compared with the counterpart, further implying that cold plasma is capable of modifying the protein film. 3.2. Effects of Different ACP Treatment Voltages on Soy Protein Film The effect of ACP voltage level on the mechanical properties of the soy protein film was investigated at a fixed treatment time of 3 min with the optimal soybean and glycerol concentration of 10% and 2.8%, respectively. As illustrated in Figure 4a, elongation at the break increased at a rapid rate and subsequently decreased at a fast rate, with a maximum value of 211.53% at 30 kV. An increase in the treatment voltage exhibited a profitable effect, which could improve the elongation at the break of soy protein film and help the formation of gels. In this study, ACP could promote the interaction between proteins, which was one of rationales for the capacity of modifying soy protein film [25]. Higher treatment voltage, however, would decrease the elongation at the break of the soy protein film. The possible reason is that ACP treatment would destroy the active sites on the surface of soy protein at higher voltage levels, resulting in the denaturation of some soy protein and disability of participation in the cross-linking reaction. Meanwhile, WVP of the treated soy protein film decreased slightly at first but increased significantly (p < 0.05) after the treatment voltage exceeded 30 kV (Figure 4b). WVP was found to be 7.07 g·mm/m2·d·kPa when the soy protein was treated at 30 kV, while this value boosted to 40 kV, and WVP was increased to 9.29 g·mm/m2·d·kPa. The deduction made was that active particle (O2−, O2+, H3O+) interacted with the surface of soy protein, leading to an increase in WVP [29]. This phenomenon was consistent with that of Pankaj et al. [26], who also found an increase in the water vapor transmission rate of the film made from polylactic acid under a higher voltage level. Based on the comprehensive appeal, it is concluded that 30 kV is the best ACP processing voltage. 3.3. Effects of Different ACP Treatment Times on Soy Protein Film ACP treatment time was also optimized by varying from 1 min to 5 min at the above obtained optimal conditions. Variations in the elongation at the break and water vapor transmittance of soy protein films treated with different voltages were displayed in Figure 5. The elongation at the break increased significantly (p < 0.05) as the treatment time was prolonged from 0 to 3 min. After 3 min of exposure to ACP at 30 kV, the elongation at the break reached the peak value (199.64%). This finding was in agreement with the results of Jahromi et al. [36], who also observed an increase of casein film in elongation at the break after ACP treatment. The elongation at the break no longer increased when the treatment time was further extended to 5 min. In contrast, it reduced significantly (p < 0.05) at the extension of exposure time. The structure of protein was destroyed and the molecules were broken down, which in turn caused the proteins to be unable to cross-link well, with extended treatment time. In the case of WVP, it slightly declined and then significantly enhanced (p < 0.05) with the increased treatment time (Figure 5b). Extended exposure to ACP caused modification of the soy protein to a large extent, which was harmful to the interaction between proteins. A similar phenomenon was also reported by Dong et al. [27], who found that the original smooth surface of zein gradually disappeared, and a considerable number of irregular distortions and ruptures occurred after ACP treatment. To sum up, the exposure time of 3 min was selected for further analysis. Based on the above observations, it was concluded that the optimum conditions for preparation of the film were as follows: soy protein concentration of 10%, glycerol concentration of 2.8%, treatment voltage of 30 kV, and treatment time of 3 min. The characteristics of both the control and ACP–treated film were investigated by DSC, FTIR, SEM, AFM, and WCA. 3.4. Solubility, Swelling Index, and Moisture Content of Soy Protein Film Moisture content, solubility, and swelling index are three important characteristics of protein film, which can positively reflect the hydrophilicity of soy protein film. The results of solubility, swelling index, and water content of soy protein film were listed in Table 1. The control film displayed a lower solubility and swelling degree than those of the ACP-treated film. As reported, the solubility mostly depended on the hydrophilic properties of films, while the swelling index was relevant to water diffusion and ionization of amino or carboxyl groups [33]. The data revealed that the solubility of soy protein increased from 34.15% to 37.35% by ACP treatment. Air, as the working gas during ACP treatment, is ionized and –OH and –NH can be generated. ACP treatment exposed the active sites on the surface of the soy protein and promoted the hydration reaction. Ji et al. [25] also found that ACP treatment could enhance solubility of peanut protein isolate. The swelling index of soy protein was increased after ACP treatment. Though some unstable compounds in soy protein film have been dissolved in the water after ACP treatment, swelling ability did not weaken as expected. The increase in swelling ability might be predominantly due to the improvement of hydrophilicity of soy protein film by the reactive species induced by ACP treatment, thus allowing more water to be absorbed. As for water content, it was observed that the ACP–treated soy protein film was 6.01% higher than that of the control group. Hydrogen bonds were more likely to be formed after ACP treatment, and thus, more water molecules could be locked in the film-forming process by the force of hydrogen bonds [5]. 3.5. Thermal Properties of Soy Protein Film Figure 6 and Table 2 presented the thermal profiles of the control and treated film by DSC. The denaturation temperature (Td) of the control sample exhibited was 113.4 °C. In contrast, the corresponding peak of the ACP–treated film shifted to 123.8 °C. There was a slight increase in Tg value, which increased from 98.2 °C to 100.3 °C after ACP treatment. It required a higher temperature and more heat for denaturation of ACP treatment film, indicating that the soy protein film was more stable after ACP treatment. The high–energy plasma discharge could break part of protein structure and generate some new bond [26]. ACP treatment would result in the increase in the hydrogen bonds of soy protein. During the discharge, the newly increased bonds of soy protein could enhance inter molecular force, leading to tighter cross-links between the proteins [35]. Jahromi et al. [36] reported that ACP can increase the thermal stability of casein membranes after treatment at 50 kV for 5 min. 3.6. Microstructure of Soy Protein Film As shown in Figure 7, the surface of the control group was uneven and had obvious wrinkles under the scanning electron microscope. Besides, numerous cracks and pores could be detected for the control group. The uneven size of the protein clusters of the untreated sample, as well as the lack of good cross-linking between the molecules during film formation, resulted in the wrinkles, cracks, and pores. However, compared with the control counterpart, it could be observed the ACP-treated soy protein film has a smoother surface without visible cracks and pores. High–energy electrons excited by ACP decomposed the soy protein during ACP treatment, which changed its tertiary and quaternary structures. Changes in structure of soy protein had a momentous influence on the morphology of film. In general, ACP treatment would increase the roughness of the film surface observed by the scanning electron microscope. It has been stated by Chen et al. [37] that DBD plasma increased the roughness of zein film because of the intensification of the etching effect. The opposite phenomenon might be attributed to direct treatment on protein solution prior to film formation rather than on film itself. Soy protein would distribute more evenly in the solution, and the dispersed soy protein cross-linked into network by self-rearranging, which reduce the etching effect on the film. Dong et al. [38] evaluated zein in aqueous ethanol under ACP treatment, and they found plasma could decrease the aggregation degree of zein micelles. The observation of uniform surface with negligible pores and cracks in this finding could well explain for higher elongation at the break for the ACP–treated group. Furthermore, it was detected that elongation at the break remarkably increased with the ACP treatment in the above experiment. This might be related to the results of the scanning electron microscope, which indicated that ACP treatment on a protein solution largely lessened the cracks and pores. Moreover, the surface of the treated film became smoother. Hence, the observation of uniform surface with negligible pores and cracks in this finding could well explain for higher elongation at the break for the ACP–treated group. 3.7. Fourier Transform Infrared Spectroscopy Analysis of Soy Protein Film The FT-IR results of the control and ACP–treated film are shown in Figure 8. An increase in the intensity of bands at 3200 cm−1 and 3400 cm−1 was observed after ACP treatment. The broad band caused by O–H stretching and N–H stretching was assigned to amide A of the soy protein film. The intensity of the C–H stretching band was observed at 2930 cm−1, and also increased. The carbonyl band appeared at 1668 cm−1, which was assigned to amide I. The amide II and amide III bands of soy protein were observed at 1546 cm−1 and 1256 cm−1, respectively [10]. The finding might be ascribed to the fact that considerable charged particles were generated by ACP, which bombard the surface of soy protein and bound to the active sites of soy protein [30]. Though no new peak was generated in both of the control and ACP treatment samples, the secondary structure components of two groups were slightly different. Hence, ACP treatment did not denature the soy protein or did not generate new functional groups [39]. Based on the results, it could be concluded that ACP treatment altered the film secondary structure involving the amide A region. The changes in the soy protein structure were caused by hydrogen bonds produced by high-energy particles, whereby hydrogen bonds helped to maintain the structure of the film. A similar phenomenon was also observed by Chen et al. [40], who found that the DBD-ACP pre-treatment led to a change in the protein conformation and promoted hydrogen bonding interactions between zein and polylactic acid. 3.8. Surface Roughness of Soy Protein Film Surface morphology of the film was observed by an atomic force microscope (AFM). Maximum roughness (Rmax), root mean square roughness (Rms), and roughness average (Ra) were used to quantify the surface roughness changes [40]. As presented in Figure 9, the Rmax of the control and ACP-treated film were 162.7 nm and 88.4 nm, respectively. As summarized in Table 3, the Ra and Rms value of the control group were 17.1 nm and 22 nm, respectively. However, these two values appeared to be 10.9 nm and 13.4 nm, respectively. It was evident that the roughness of the film remarkably decreased after ACP treatment, which was inconsistent with others research. Pankaj et al. [26] demonstrated a dramatic increase in the surface roughness of the chitosan film subjected to DBD–ACP treatment. The main reason might be that film was directly exposed to ACP treatment in other research, while in this study, soy protein solution was disposed by ACP prior to preparing into the film. As is known, high–energy particles, including radicals, electrons, ions, neutrals, excited atoms, and UV radiations, exerted an etching effect on the surface of the film, which led to the increase in roughness. Nevertheless, direct treatment on the soy protein solution, rather than the film, avoided this etching impact, which in turn effectively reduced the damage to the film and thus resulted in the reduction of the roughness in our present investigation. Dong et al. [38] demonstrated that the discharge space, the shockwave, as well as high-energy particles induced by ACP treatment, might cause a breakage in intermolecular forces between zein micelles, and thus disperse the originally clumped zein protein in aqueous ethanol. Similarly, the soy protein solution treated by ACP was more uniformly dispersed in the solution after ACP treatment. The direct treatment on the protein was more favorable for its rearrangement to form a regular and uniform film structure. These results were well in accordance with those obtained from scanning electron microscope analysis of the film. 3.9. Surface Hydrophilicity of Soy Protein Film The water contact angle (WCA), an important variable, was used to evaluate the hydrophilicity and hydrophobicity of the film surface. A higher WCA stands for stronger hydrophobicity of the film [4]. As is known, soy protein film exhibited strong hydrophobicity [5]. Low WCA value of the film was observed in the case of ACP treatment, while relatively higher value of WCA was obtained for the control sample (Figure 10). Based on Table 4, the WCA of the control and ACP–treated film was 87.9°and 77.2°, respectively. The deduction was presumably owing to the interaction of the active particles with the soy protein, leading to the increase in polarity [29]. Based on FT–IR analysis, a large number of –OH groups and hydrogen bonds were formed on the film after ACP modification. The hydrophilic groups on the protein surface would result in hydration by the water molecules [41]. A number of studies indicated enhancement of surface roughness occurred with the improved surface hydrophilicity when the film was directly subjected to ACP treatment [42]. As afore mentioned, direct treatment on film itself might lead to a certain damage to the barrier properties. In terms of AFM observation in our study, surface roughness of the film decreased due to the pretreatment on the soy protein solution rather than the film. Furthermore, increased values of solubility, swelling index, and moisture content illustrated the stronger hydrophobicity, which was inconsistent with the lower WCA, which also indicated the better affinity for water molecules. A low water contact angle and high hydrophilic character of film were obtained for ACP treatment [43]. Therefore, the increase of hydrophilic groups could account for the decrease in WCA. The proposed method of using ACP to modify the soy protein could effectively improve the surface hydrophilicity of the film. 4. Conclusions Though ACP has been successfully applied to improve the packaging properties of protein film, few data can be achieved concerning direct modification of a protein solution by plasma. In this investigation, a soy protein solution was subjected to DBD-ACP treatment prior to preparation in the edible film. The optimal conditions were found for soy protein (10%), glycerol (2.8%), and ACP treatment at 30 kV for 3 min, under which, changes in physicochemical and structure characteristics were observed. An enhanced elongation at thebreak and reduced WVP occurred, accompanied by the reinforcement of thermal properties, demonstrating the cross-linking with the soy protein matrix. ACP pretreatment also resulted in the increase of hydrophilicity, which was proved by the improvement of So, Si, MC, and the decline of the water contact angle. Moreover, compared with the control group, ACP treatment could obviously reduce the voids and cracks of the film by SEM images. Fourier transform infrared spectroscopy suggested that ACP pretreatment would form more hydrophilic groups, such as hydroxyl groups and hydrogen bonds. Interestingly, roughness decreased for the ACP treated sample, owing to the direction treatment on the protein solution rather than the film. Therefore, it was expected that ACP treatment on the protein solution could serve as a feasible and effective surface modification approach that would enlarge edible film application in the package industry with desirable functions. Acknowledgments We thank the Key Laboratory of Seafood Health Risk Factors of Zhejiang Ocean University for providing the scientific research platform. Author Contributions Z.L.: investigation, methodology, formal analysis, data curation, and writing—original draft. J.C.: validation and writing—review and editing. S.D.: validation and writing—review and editing. All authors have read and agreed to the published version of the manuscript. Funding This research was funded by National Natural Science Foundation of China: (Project No. 31901764). Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement The data that support the findings of this study are available from the corresponding author at reasonable request. Conflicts of Interest The authors declare no conflict of interest. Abbreviations DBD–ACP Dielectric barrier discharge Atmospheric cold plasma DBD Dielectric barrier discharge ACP Atmospheric cold plasma CP Cold plasma WVP Water vapor permeability MC Moisture content So Solubility Si Swelling index DSC Differential scanning calorimetry Td Denaturation temperature Tg Glass transition temperature SEM Scanning electron microscopy FTIR Fourier transform infrared spectroscopy Rmax Maximum roughness Rms Root mean square roughness Figure 1 Schematic diagram of experimental equipment for DBD plasma system. Figure 2 Effect of soy protein concentration on mechanical properties of the film. (a) Elongation at break of ACP–treated and control group. (b) WVP of ACP–treated and control group. Different lowercase letters denote significant differences (p < 0.05). Figure 3 Effect of glycerol concentration on elongation at break and WVP of the soy protein film. (a) Elongation at break of ACP–treated and control group. (b) WVP of ACP–treated and control group. Different lowercase letters denote significant differences (p < 0.05). Figure 4 Effects of ACP treatment voltages on elongation at break and WVP of the soy protein film. (a) Elongation at break; (b) WVP. Different lowercase letters denote significant differences (p < 0.05). Figure 5 Effects of different ACP treatment times on elongation at break and WVP of the soy protein film. (a) Elongation at break; (b) WVP. Different lowercase letters denote significant differences (p < 0.05). Figure 6 DSC curve of control and ACP–treated film at 30 kV for 3 min. Untreated soy protein film was considered as the control. The amount of soy protein was 10% and glycerol was 2.8%. Figure 7 Scanning electron microscope images of the ACP–treated film and control. (a) Untreated soy protein film. (b) Soy protein film treated by ACP at 30 kV for 3 min. Untreated soy protein film was considered as the control. The amount of soy protein was 10% and glycerol was 2.8%. Figure 8 Fourier near-infrared spectra of control and ACP–treated film at 30 kV for 3 min. Untreated soy protein film was considered as the control. The amount of soy protein was 10% and glycerol was 2.8%. Figure 9 Atomic force microscope images of control and ACP–treated film at 30 kV for 3 min. (a) Untreated soy protein film. (b) Soy protein film treated by ACP at 30 kV for 3 min. Untreated soy protein film was considered as the control. The amount of soy protein was 10% and glycerol was 2.8%. Figure 10 Photos of water contact angle of the control and ACP–treated samples. (a) Untreated soy protein film. (b) Soy protein film treated at 30 kV for 3 min. Untreated soy protein film was considered as the control. The amount of soy protein was 10% and glycerol was 2.8%. foods-11-01196-t001_Table 1 Table 1 Solubility, swelling index, and moisture content of soy protein film. ACP treatment is soy protein film treated for 3 min, at 30 kV. The control is untreated soy protein film. Stage So (%) Si (%) MC (%) Control 34.15 ± 0.07 134.89 ± 0.41 25.13 ± 0.16 ACP treatment 37.35 ± 0.15 153.52 ± 1.32 31.44 ± 0.21 So: Solubility. Si: swelling index. MC: moisture content. The values are expressed as mean ± SD. (n = 3). foods-11-01196-t002_Table 2 Table 2 Thermodynamic parameters of control and ACP–treated film, with the amount of soy protein being 10% and glycerol being 2.8%. Thermodynamic Parameters Tg (°C) Td (°C) Peak Area (μVs/mg) Peak (μV/mg) Enthalpy (J/g) Control 98.2 113.4 512.6 3.7829 10.338 ACP treatment 100.3 123.8 674.9 4.9177 14.025 foods-11-01196-t003_Table 3 Table 3 Roughness parameters for the control and ACP–treated samples with the amount of soy protein being 10% and glycerol being 2.8%. Roughness Index Rmax (nm) Ra (nm) Rms (nm) Control 162.7 ± 2.14 17.1 ± 0.45 22.0 ± 0.54 ACP treatment 88.4 ± 1.33 10.9 ± 27 13.4 ± 0.19 The values are expressed as mean ± SD. (n = 3). foods-11-01196-t004_Table 4 Table 4 Water contact angle of the control and ACP–treated samples. The ACP treatment is soy protein film treated at 3 min, at 30 kV. The control is untreated soy protein film. Water Contact Angle Right Left Control 86.9° ± 0.53 87.9° ± 0.47 ACP treatment 78.7° ± 0.32 77.2° ± 0.39 The values are expressed as mean ± SD. (n = 3). Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Kontominas M.G. Use of alginates as food packaging materials Foods 2020 9 1440 10.3390/foods9101440 33053627 2. Xu L. Li T. Cao W. Wu Y. Chi Y. Zhang H. Liu Y. Properties of soy protein isolate antimicrobial films and its application in preservation of meat Emir. J. Food Agric. 2017 29 589 600 10.9755/ejfa.2017-05-107 3. Lee E.J. Kim H. Lee J.Y. Ramachandraiah K. Hong G.P. β-Cyclodextrin-mediated beany flavor masking and textural modification of an isolated soy protein-based yuba film Foods 2020 9 818 10.3390/foods9060818 32580273 4. Xie D.Y. Song F. Zhang M. Wang X.L. Wang Y.Z. Soy protein isolate films with improved property via a facile surface coating Ind. Crops Prod. 2014 54 102 108 10.1016/j.indcrop.2014.01.017 5. Kumar R. Zhang L. Soy protein films with the hydrophobic surface created through non-covalent interactions Ind. Crops Prod. 2009 29 485 494 10.1016/j.indcrop.2008.09.010 6. Cho S.Y. Rhee C. Sorption characteristics of soy protein films and their relation to mechanical properties LWT-Food Sci. Technol. 2002 35 151 157 10.1006/fstl.2001.0829 7. Rani S. Singh A.K. Paswan R.R. Kumar K.D. Kumar R. Preparation, characterization and antibacterial evaluation of soy protein isolate biopolymeric films loaded with nalidixic acid J. Polym. Environ. 2020 28 1841 1850 10.1007/s10924-020-01729-4 8. Kumar R. Rani P. Dinesh Kumar K. Soy protein isolate film by incorporating mandelic acid as Well as through fermentation mediated by Bacillus Subtilis J. Renew. Mater. 2019 7 103 115 10.32604/jrm.2019.00027 9. Bosquez-Molina E. Tomás S.A. Rodríguez-Huezo M.E. Influence of CaCl2 on the water vapor permeability and the surface morphology of mesquite gum based edible films LWT-Food Sci. Technol. 2010 43 1419 1425 10.1016/j.lwt.2010.04.023 10. Kumar R. Praveen R. Rani S. Sharma K. Tiwary P.K. Dinesh Kumar K. ZnSe nanoparticles reinforced biopolymeric soy protein isolate film J. Renew. Mater. 2019 7 749 761 10.32604/jrm.2019.06286 11. Kumar R. Anjum K.N. Rani S. Sharma K. Tiwary K.P. Kumar K.D. Material properties of ZnS nanoparticles incorporated soy protein isolate biopolymeric film Plast. Rubber Compos. 2019 48 448 456 10.1080/14658011.2019.1651125 12. Vieira T.M. Moldão-Martins M. Alves V.D. Design of chitosan and alginate emulsion-based formulations for the production of monolayer crosslinked edible films and coatings Foods 2021 10 1654 10.3390/foods10071654 34359524 13. Barać M. Stanojević S. Jovanović S. Pešić M. Soy protein modification: A review Acta Period. Technol. 2004 35 3 16 10.2298/APT0435003B 14. Wang J.M. Yang X.Q. Yin S.W. Zhang Y. Tang C.H. Li B.S. Yuan D.B. Guo J. Structural rearrangementof ethanol-denatured soy proteins by high hydrostatic pressure treatment J. Agric. Food Chem. 2011 59 7324 7332 10.1021/jf201957r 21609024 15. Los A. Ziuzina D. Van Cleynenbreugel R. Boehm D. Bourke P. Assessing the biological safety of atmospheric cold plasma treated wheat using cell and insect models Foods 2020 9 898 10.3390/foods9070898 32650404 16. Segat A. Misra N.N. Cullen P.J. Innocente N. Atmospheric pressure cold plasma (ACP) treatment of whey protein isolate model solution Innov. Food Sci. Emerg. 2015 29 247 269 10.1016/j.ifset.2015.03.014 17. Liao X. Li J. Muhammad A.I. Suo Y. Chen S. Ye X. Liu D. Ding T. Application of a dielectric barrier discharge atmospheric cold plasma (Dbd-Acp) for eshcerichia coli inactivation in apple juice J. Food Sci. 2018 83 401 408 10.1111/1750-3841.14045 29355961 18. Trevisani M. Berardinelli A. Cevoli C. Cecchini M. Ragni L. Pasquali F. Effects of sanitizing treatments with atmospheric cold plasma, SDS and lactic acid on verotoxin-producing Escherichia coli and Listeria monocytogenes in red chicory (radicchio) Food Control 2017 78 138 143 10.1016/j.foodcont.2017.02.056 19. Timmons C. Pai K. Jacob J. Zhang G. Ma L.M. Inactivation of Salmonella enterica, Shiga toxin-producing Escherichia coli, and Listeria monocytogenes by a novel surface discharge cold plasma design Food Control 2018 84 455 462 10.1016/j.foodcont.2017.09.007 20. Chingsungnoen A. Maneerat S. Chunpeng P. Poolcharuansin P. Nam-Matra R. Antimicrobial treatment of Escherichia coli and Staphylococcus aureus in herbal tea using low-temperature plasma J. Food Protect. 2018 81 1503 1507 10.4315/0362-028X.JFP-18-062 21. Patange A. Boehm D. Bueno-Ferrer C. Cullen P.J. Bourke P. Controlling Brochothrix thermosphacta as a spoilage risk using in-package atmospheric cold plasma Food Microbiol. 2017 66 48 74 10.1016/j.fm.2017.04.002 28576372 22. Pérez-Andrés J.M. Álvarez C. Cullen P.J. Tiwari B.K. Effect of cold plasma on the techno-functional properties of animal protein food ingredients. Innov Food Sci. Emerg. 2019 58 102205 10.1016/j.ifset.2019.102205 23. Bußler S. Steins V. Ehlbeck J. Schlüter O. Impact of thermal treatment versus cold atmospheric plasma processing on the techno-functional protein properties from Pisum sativum ‘Salamanca’ J. Food Eng. 2015 167 166 174 10.1016/j.jfoodeng.2015.05.036 24. Chaple S. Sarangapani C. Jones J. Carey E. Causeret L. Genson A. Duffy B. Bourke P. Effect of atmospheric cold plasma on the functional properties of whole wheat (Triticum aestivum L.) grain and wheat flour Innov. Food Sci. Emerg. 2020 66 102529 10.1016/j.ifset.2020.102529 25. Ji H. Han F. Peng S. Yu J. Li L. Liu Y. Chen Y. Li S. Chen Y. Behavioral solubilization of peanut protein isolate by atmospheric pressure cold plasma (ACP) treatment Food Bioprocess Technol. 2019 12 2018 2027 10.1007/s11947-019-02357-0 26. Pankaj S.K. Bueno-Ferrer C. Misra N.N. O’Neill L. Jiménez A. Bourke P. Cullen P.J. Characterization of polylactic acid films for food packaging as affected by dielectric barrier discharge atmospheric plasma Innov. Food Sci. Emerg. 2014 21 107 113 10.1016/j.ifset.2013.10.007 27. Dong S. Gao A. Zhao Y. Li Y.T. Chen Y. Characterization of physicochemical and structural properties of atmospheric cold plasma (ACP) modified zein Food Bioprod Process. 2017 106 65 74 10.1016/j.fbp.2017.05.011 28. Oh Y.A. Roh S.H. Min S.C. Cold plasma treatments for improvement of the applicability of defatted soybean meal-based edible film in food packaging Food Hydrocoll. 2016 58 150 159 10.1016/j.foodhyd.2016.02.022 29. Chizoba Ekezie F.G. Sun D.W. Cheng J.H. A review on recent advances in cold plasma technology for the food industry: Current applications and future trends Trends Food Sci. Technol. 2017 69 46 58 10.1016/j.tifs.2017.08.007 30. Chen J. Wang S.Z. Chen J.Y. Chen D.Z. Deng S.G. Xu B. Effect of cold plasma on maintaining the quality of chub mackerel (Scomber japonicus): Biochemical and sensory attributes J. Sci. Food Agric. 2019 99 39 46 10.1002/jsfa.9138 29786860 31. Wiles J.L. Vergano P.J. Barron F.H. Bunn J.M. Testin R.F. Water vapor transmission rates and sorption behavior of chitosan films J. Food Sci. 2000 65 1175 1179 10.1111/j.1365-2621.2000.tb10261.x 32. Peng Y. Wu Y. Li Y. Development of tea extracts and chitosan composite films for active packaging materials Int. J. Biol. Macromol. 2013 59 282 289 10.1016/j.ijbiomac.2013.04.019 23603075 33. Zhang H. Wang L. Li H. Chi Y. Zhang H. Xia N. Ma Y. Jiang L. Zhang X. Changes in properties of soy protein isolate edible films stored at different temperatures: Studies on water and glycerol migration Foods 2021 10 1797 10.3390/foods10081797 34441574 34. dos Santos Paglione I. Galindo M.V. de Souza K.C. Yamashita F. Grosso C.R.F. Sakanaka L.S. Shirai M.A. Optimization of the conditions for producing soy protein isolate films Emir. J. Food Agric. 2019 31 297 303 10.9755/ejfa.2019.v31.i4.1933 35. Cao N. Fu Y. He J. Preparation and physical properties of soy protein isolate and gelatin composite films Food Hydrocoll. 2007 21 1153 1162 10.1016/j.foodhyd.2006.09.001 36. Jahromi M. Niakousari M. Golmakani M.T. Ajalloueian F. Khalesi M. Effect of dielectric barrier discharge atmospheric cold plasma treatment on structural, thermal and techno-functional characteristics of sodium caseinate Innov. Food Sci. Emerg. 2020 66 102542 10.1016/j.ifset.2020.102542 37. Chen G. Dong S. Zhao S. Li S. Chen Y. Improving functional properties of zein film via compositing with chitosan and cold plasma treatment Ind. Crops Prod. 2019 129 318 326 10.1016/j.indcrop.2018.11.072 38. Dong S. Guo P. Chen Y. Chen G.Y. Ji H. Ran Y. Li S.H. Chen Y. Surface modification via atmospheric cold plasma (ACP): Improved functional properties and characterization of zein film Ind. Crops Prod. 2018 115 124 133 10.1016/j.indcrop.2018.01.080 39. Sharma S. Singh R.k. Cold plasma treatment of dairy proteins in relation to functionality enhancement Trends Food Sci. Technol. 2020 102 30 36 10.1016/j.tifs.2020.05.013 40. Chen G. Chen Y. Jin N. Li J. Dong S. Li S. Zhang Z. Chen Y. Zein films with porous polylactic acid coatings via cold plasma pre-treatment Ind. Crops Prod. 2020 150 112382 10.1016/j.indcrop.2020.112382 41. Dong S. Wang J.M. Cheng L.M. Lu Y.L. Li S.H. Chen Y. Behavior of zein in aqueous ethanol under atmospheric pressure cold plasma treatment J. Agric. Food Chem. 2017 65 7352 7360 10.1021/acs.jafc.7b02205 28759211 42. Pankaj S.K. Bueno-Ferrer C. Misra N.N. O’Neill L. Bourke P. Cullen P.J. Effects of cold plasma on surface, thermal and antimicrobial release properties of chitosan film J. Renew. Mater. 2017 5 14 20 10.7569/JRM.2016.634105 43. Karam L. Casetta M. Chihib N.E. Bentiss F. Maschke U. Jama C. Optimization of cold nitrogen plasma surface modification process for setting up antimicrobial low density polyethylene films J. Taiwan Inst. Chem. E 2016 64 299 305 10.1016/j.jtice.2016.04.018

PMC009xxxxxx/PMC9099684.txt

==== Front Sensors (Basel) Sensors (Basel) sensors Sensors (Basel, Switzerland) 1424-8220 MDPI 10.3390/s22093156 sensors-22-03156 Article Airspeed-Aided State Estimation Algorithm of Small Fixed-Wing UAVs in GNSS-Denied Environments https://orcid.org/0000-0002-9226-5676 Ye Xiaoyu 1 Zeng Yifan 2 Zeng Qinghua 1* Zou Yijun 1 Rizos Chris Academic Editor 1 School of Aeronautics and Astronautics, Sun Yat-sen University, Shenzhen 518107, China; yexy39@mail2.sysu.edu.cn (X.Y.); zouyj5@mail2.sysu.edu.cn (Y.Z.) 2 School of Systems Science and Engineering, Sun Yat-sen University, Guangzhou 510275, China; zengyf29@mail2.sysu.edu.cn * Correspondence: zqinghua@mail.sysu.edu.cn 20 4 2022 5 2022 22 9 315604 3 2022 18 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Aimed at improving the navigation accuracy of the fixed-wing UAVs in GNSS-denied environments, this paper proposes an algorithm of nongravitational acceleration estimation based on airspeed and IMU sensors, which use a differential tracker (TD) model to further supplement the effect of linear acceleration for UAVs under dynamic flight. We further establish the mapping relationship between vehicle nongravitational acceleration and the vehicle attitude misalignment angle and transform it into the attitude angle rate deviation through the nonlinear complementary filtering model for real-time compensation. It can improve attitude estimation precision significantly for vehicles in dynamic conditions. Furthermore, a lightweight complementary filter is used to improve the accuracy of vehicle velocity estimation based on airspeed, and a barometer is fused on the height channel to achieve the accurate tracking of height and the lift rate. The algorithm is actually deployed on low-cost fixed-wing UAVs and is compared with ACF, EKF, and NCF by using real flight data. The position error within 30 s (about 600 m flying) in the horizontal channel flight is less than 30 m, the error within 90 s (about 1800 m flying) is less than 50 m, and the average error of the height channel is 0.5 m. The simulation and experimental tests show that this algorithm can provide UAVs with good attitude, speed, and position calculation accuracy under UAV maneuvering environments. GNSS-denied navigation nonlinear complementary filter sensor fusion the fix-wing UAV nongravitational acceleration estimation airspeed ==== Body pmc1. Introduction Small fixed-wing unmanned aerial vehicles (UAVs) play an essential role in future air battles and civilian fields due to their low cost, small size, full autonomy, and dense formation. The normal autonomous flight of UAVs is inseparable from the accurate ground velocity and absolute position provided by the Global Navigation Satellite System (GNSS), and especially the real-time kinematic (RTK) measurement can provide centimeter-level position using the technique of dispersion assignment. The traditional navigation mode of the Inertial Navigation System (INS) assisted by GNSS is greatly challenged, since the GNSS is easily interfered with by complex external environments or human factors [1,2]. State estimation of UAVs can be defined as the process of tracking the current attitude, velocity, and position of the vehicle [3]. The design of the motion estimation algorithm is necessary for flight control and is a crucial step in the development of autonomous flying machines. The accuracy of the attitude and heading reference systems (AHRS) have a crucial role. With the continuous improvement in the chip manufacturing process, the current low-cost UAV airborne sensors mainly use microelectromechanical systems (MEMS), which are limited by the performance of low-cost sensors and inevitably introduce interference noise and random noise. The rapid development of modern wireless communication systems provides technical support for various UAV positioning systems [4], and provides good and fast data communication links for various positioning methods such as the Ultra-Wide Band (UWB) and GNSS, which rely on antenna to receive data [5]. The traditional GNSS/INS-integrated navigation of UAVs use GNSS positioning measurement information to estimate the bias of the gyroscope, accelerometer, and attitude error, then feedback [6]. According to the different fusion levels of the GNSS/INS-integrated navigation system, it can be divided into loose integration, tight integration, and deep integration [7]. This method can improve the accuracy of attitude estimation significantly. However, in GNSS-denied environments, the UAV cannot rely on GNSS positioning information for guidance and control, which has a great impact on the autonomous flight of the UAV. In recent years, many researchers have been exploring autonomous navigation techniques in GNSS-denied environments to improve positioning and attitude accuracy. Visual-based position does not depend on external equipment support and has high autonomy. Visual and inertial navigation integration technologies have made great strides in recent years. Mourikis et al. [8] presented an EKF-based algorithm for real-time vision-aided inertial navigation. Tong Qin et al. [9] proposed a tightly coupled, nonlinear optimization-based method used to obtain highly accurate visual–inertial odometry by fusing preintegrated IMU measurements and feature observations, which successfully solved the navigation problem of rotorcraft in GNSS-denied environments. At present, fixed-wing UAVs have the following difficulties in navigating GNSS-denied environments: (1) High flying speed [10]: Higher speeds bring higher flight dynamics and vibrations caused by airflow, which puts higher requirements on the UAV damping design. The G-sensitivity of the gyroscope is more difficult to process directly at the algorithm level, and direct inertial integration will cause rapid dispersion of attitude position. (2) The high speed and high altitude of fixed-wing UAVs results in a vast field of view for the airborne camera. Consequently, capturing high-quality feature points becomes more complex, making it hard to apply vision-based inertial navigation algorithms to fixed-wing UAVs directly. Most commonly used attitude estimation algorithms can be concluded to three kinds: the extended Kalman filter, the gradient descent algorithm, and the complementary filter algorithms. The EKF is more precise in the process of the state error transfer and the bias error, and the process noise of the gyroscope and accelerometer are modeled, thence the error parameters are estimated and compensated by other sensors, but it obviously adds computational complexity [11]. Leutenegger, S et al. [12] used an extended Kalman filter estimation framework to replace GPS updates with airspeed measurement under GPS-denied environments. The experiment demonstrates that the position error enlarges linearly with time. Despite its widespread use in UAV navigation, the EKF is subject to limitations. The local linearization of the process dynamic models and measurement models for feature points can degrade with the increasing nonlinearity in the system dynamics [13]. The gradient descent algorithm uses a quaternion representation, allowing the accelerometer and magnetometer data to be used in an analytically-derived and optimized gradient descent algorithm to compute the direction of the gyroscope measurement error as a quaternion derivative [14]. Based on the principle of dual vector gravity and magnetic field fixation in the complementary filters algorithm, the accelerometer data is considered as an approximate observation of the local gravity vector, and the accumulated error of the heading angle is corrected by magnetic field measurements [15,16]. Mahony, R. et al. [15] first proposed and established the complementary filter on the special orthogonal group (SO3) and proved the Lyapunov stability to ensure the global stability of the observer error. However, this algorithm is sensitive to nongravitational acceleration, which may lead to the wrong attitude correction in maneuvering environments. In [17,18], the GPS velocity measurement was used to establish the model of the nongravitational acceleration of the vehicle, and the covariance of the measured noise was increased in the absence of the GPS signal, which did not solve the problem of the accurate estimation of nongravitational acceleration under GPS-denied conditions. Euston, M. et al. [19] used airspeed measurement in vehicle nongravitational acceleration observation for the first time. By establishing the centripetal force model, the result of the gravity vector observation can be ameliorated, and the accuracy of attitude estimation enhanced under GPS-denied conditions can be maintained for a short time. Unfortunately, a large attitude error will be caused in maneuvering environments since the influence of linear acceleration was not considered in the gravitational acceleration estimation. Moreover, it is only used as attitude estimation without calculating the reliability of the velocity and position estimation. Li, X. et al. [20] proposed a method to estimate the external acceleration with the purpose of improving navigation performance under dynamic conditions. Marantos et al. [21] fully combined the visual algorithm and multisensor speed/position estimation with an adaptive complementary filter, which gave the algorithm a low computational complexity. Compared with the convenience of rotorcraft to deploy intelligent algorithms related to vision, and lidar for simultaneous localization and mapping (SLAM) due to its low speed and more stable flight performance, there has been less work on the navigation and position of low-cost fixed-wing UAVs in the GNSS-denied environments because of the reasons mentioned above. Most of the previous works simply provide stable attitude output for UAVs in denied environments. The main contribution of this paper is to explore the provision of UAV stable state estimation in denied environments. The main work and innovations are as follows: (1) Based on [19], the filtering model further improves the accuracy of dynamic modeling, and an estimation algorithm of UAV nongravitational acceleration using airspeed and inertial sensors is proposed. We then further establish the mapping relationship between vehicle nongravitational acceleration and the vehicle attitude misalignment angle by combining the magnetometer. (2) Subsequently, the data of the barometer are fused to realize the stable tracking of the UAV in the altitude and lifting rate channels. (3) Aiming at the defect that the horizontal velocity and position errors of UAVs are easy to accumulate, a complementary filter for inertial navigation speed correction using airspeed assistance is designed, which greatly elevates the accuracy of the velocity position estimation of the vehicle. The framework of the algorithm is shown as Figure 1. As a fully autonomous navigation solution, the algorithm proposed in this paper has been verified by real flight, which can be used as a key switch to airspeed compensation when GNSS is denied, and thus provides a more stable navigation result. 2. Airspeed-Aided Navigation Filter 2.1. Estimation of Nongravitational Acceleration Euston, M. et al. [19] proposed a model that use airspeed and gyroscope measurements to estimate the centripetal acceleration of a vehicle, which the following equation can express. (1) a^n=ωibb×V^TASb, where ωibb is the 3-axis angular rate vector measured by the gyroscopes, V^TASb is the projection of the airspeed vector in the body frame, and a^n is the vector of centripetal acceleration. Only considering centripetal acceleration in flight is not sufficient to describe the maneuvering process of the vehicle accurately. When the vehicle speeds up or slows down, the effect of linear acceleration also needs to be taken into consideration. (2) a^L=dV^TASbdt, where a^L is the linear acceleration vector of the vehicle. The airspeed measured directly by the pitot tube is a scalar quantity defined in the velocity coordinate frame a. The longitudinal plane difference between frame a and the body frame b is the angle of attack α, and the horizontal difference is the sideslip angle β. Accurate angle-of-attack calculation requires unique sensors. The angle of attack can be estimated by flight dynamics approximation on low-cost UAVs. We just consider the vertical channel of the UAV, which can be described as follows:(3) α=φ−θθ =arcsin(vhv), where θ is the flight path angle, which can be calculated from the triaxial velocity, and φ is the pitch Angle. The airspeed vector can be described in the airflow frame as vaira=[0vair0]T, the transfer to body frame as vairb=Cabvaira, where Cab denotes the attitude rotation matrix from the airflow frame to the body frame. Therefore, vairb=[−vcosαcosβvcosαcosβ−vsinα]T Cab=[cosβ−cosαcosβ−sinαsinβsinβcosαcosβsinαcosβ0−sinαcosα] For small fixed-wing UAVs, the angle of attack and the sideslip angle are difficult to measure directly by sensors because accurate measurements require atmospheric parameter sensors, but they are not suitable for small vehicles. We notice that if the sideslip angle β in flight is approximately no more than 10 degrees (in reference [22], as for small fixed-wing UAVs, the sideslip angle estimation is no more than 5 degrees), and the cosine of 10 degrees is equal to 0.9848, it makes only 1.52% velocity errors if we assume the effect of the sideslip angle is ignored. In contrast, the angle of attack α is the angle between the incoming direction of the flow vector and chord line of an airfoil. As the angle of attack increases, the relative lift of the airfoil increases. When the UAV makes a turn, additional centripetal acceleration is provided by increasing the angle of attack. To ensure centripetal acceleration at the turn, the vehicle enters a Bank-to-Turn (BTT) inclined turn mode where the increased lift from the wing is decomposes into a vertical component and a horizontal component. In order for the vehicle to maintain altitude, the vertical component of lift must counteract gravity, which requires increasing α to gain additional lift. So, the angle of attack α cannot be ignored, especially when the vehicle in turning. Obviously, the field winds are dynamic and inevitable. Due to the fact that small fixed-wing UAVs are lightweight, they are not suitable to fly in high field wind. Moreover, the wind speed is as hard to estimate as the angle of attack or sideslip, so we try to ignore the effect of the wind. Figure 2 shows the comparison between the true airspeed and the ground velocity measured by RTK in real flight. The linear acceleration of the vehicle can be calculated from the differentiation of the linear velocity. Random noise inevitably exists in the pitot airspeed measurement, which leads to an additional error in acceleration estimation. The function of the differential tracker (TD) of the Active Disturbance Rejection Control (ADRC) is to extract differential signals properly from those polluted by noise, so the second-order differential tracker [23] in the ADRC is used to achieve data filtering and differential signal extraction. The second-order differential tracker is described as follows:(4) {x1(k+1)=x1(k)+Tx2(k)x2(k+1)=x2(k)+T⋅fst, where x1(k) tracks the original signal, x2(k) calculates the differential value of the original signal, and the fst is calculated as follows:(5) {δ=rhδ0=δhy=x0−u+hx2a0=δ2+8r|y|a={x2+yh,|y|≤δ0x2+0.5(a0−δ)sgn(y),|y|>δ0fst={−raδ,|a|≤δ−rsgn(a),|a|>δ, here, T is the period of the input signal and h is the filter factor; when h=T, the algorithm is close to the first-order difference. The higher the value of h, the better the filter effect will be. Still, it will bring the corresponding time delay. Factor r is the rate factor, which can be used to adjust the tracking speed. The speed will raise as the factor r increases, but the signal noise will be amplified. Based on the above equation, the vehicle acceleration model is:(6) a^=ωibb×V^TASb+dvTASadt 2.2. Attitude Calculation Model Based on External Acceleration Correction The inertial navigation-specific force equation under local horizontal frame is written as:(7) v˙n=Cbnfb+gn−(2Ωien+Ωenn)vn, here, the corner marks b and n, respectively, denote the East-North-Up (ENU) and the body frame. Vn represents the speed of the vehicle, fb represents the specific force vector under the body frame, gn denotes the gravity field vector in the ENU frame, and Cbn is the coordinate transformation matrix from the body frame to the ENU frame. Moreover, Ωien denotes the Earth-rotation skew-symmetric matrix in the local navigation frame, and Ωenn denotes the transport rate skew-symmetric matrix from the rotation of the local-frame to the center frame, which can be neglected due to the short flight span. (8) an=Cbnfb+gn, here, an is the nongravitational acceleration of the vehicle in the local horizontal frame. The Cbn is the theoretical value of the attitude rotation matrix. Due to the measurement error of the gyroscope, the relationship between Cbn and the real calculated value C˜bn is shown as follows [24]:(9) an=(I+ϕn×)C˜bnfb+gn, here, ϕn is the projection of the attitude misalignment angle vector in the ENU frame. Because the gyroscope bias accumulates large attitude errors over time, it is necessary to estimate and compensate the gyroscope bias in real-time. The bias of the accelerometer is usually small, and we only want to calculate the nongravitational acceleration of the vehicle, which is not cumulative, so we assume that the accelerometer measurement error is negligible as fb≈f˜b. Then, the equation written as:(10) f˜n×ϕn=gn+f˜n−an The left f˜n can be approximated as f˜n≈an−gn, where gn=[00−g]T. Construct the above equation in component form as:(11) [aEaNaU−g]×[ϕEϕNϕU]=[fE−aEfN−aNfU−g−aU] Due to rank((a−g)×)=2<3, Equation (11) can only solve two attitude misalignment angles. Because horizontal acceleration only provides information about horizontal misalignment angle, the heading misalignment angle cannot be observed. Accordingly, make ϕU=0. Then, the misalignment angle under the navigation system can be described as Equation (12). The e3=[001]T presents the z-axis unit vector. Convert to the body frame:(12) ϕb=fb−abaU−g×(Cnbe3)=fb−abaU−g×C3T, where C3 is the third-row vector of the Cbn. The specific force fb, the acceleration of air velocity measurement ab, and the acceleration of gravity g are all vectors whose errors do not diverge with time. Since the attitude rotation matrix Cnb is obtained by integrating the angular rate, it will generate cumulative errors over time. The cumulative errors can be converted into the projection of the horizontal attitude misalignment angle ϕφγb under the body frame by multiplying these two vectors. For the yaw channel, the magnetometer complementary filter is adopted. (13) ϕψb=mb|mb|×(Cnbm^n), (14) ϕb=ϕφγb[110]T+ϕψb[001]T, similarly, ϕψb is the projection of the heading error angle under the system, and mb is the triaxial magnetometer measurement vector. Take the misalignment angle into the complementary filter, and the measurement error generated by the gyroscope can be corrected in the next step. The negative feedback model of the complementary filters uses a PI controller. It can be defined as:(15) ωbias=kpϕb+kI∫ϕb. the gains kp and kI are proportional and integral gains, respectively. Adjusting the appropriate kp gain can make the system track the angular motion quickly and compensate the attitude misalignment angle continuously. (16) qk+1=qk⊗Δqk, here, Δqk is the delta quaternion, which can be defined as:(17) Δqk=cosΔθ2+Δθ‖Δθ‖sin(Δθ2), (18) Δθ=(ωibb+ωbiasb)⋅ΔT At high sampling times, the delta angle Δθ from k moment to k+1 moment is usually tiny and can be approximated as follows:(19) qk≈ [1Δθx2Δθy2Δθz2]T The three-axis attitude can be solved from the quaternion. 2.3. Adaptive Complementary Fusion in Horizontal Channel UAV velocity estimation can be obtained recursively through a specific force equation:(20) vn=∫(Cbnfb+gn)dt Due to the errors of inertial measurement and attitude calculation accumulation with time, the velocity position estimation will become meaningless over a long time. A complementary filter is used to smooth and correct the horizontal velocity by airspeed measurement. (21) v^E=vTASsin(ψ)+v˜windE v^N=vTAScos(ψ)+v˜windN Using the above Equation (21), the airspeed can be converted to the local frame. (22) vIk,N/E=vIk−,N/E+KTASv(v^TASk−,N/E−vIk−,N/E), here, KTASv is the gain of the complementary filter fused with airspeed. The state estimator parameters have adaptive functions to obtain the best performance based on sensor characteristics. We use the following function to adjust the gain KTASv and use an activation function to smooth the gain KTASv switching process of the observer. The adaptive strategy is given in the following equation. (23) KTASv={0, if t≤t0G1+e−(t−t0−t1),if t>t0, where t0 is the time switching threshold, and the velocity solved by the inertial navigation algorithm can maintain a low error when t≤t0. This error is lower than that in the direct estimation of the ground velocity from the airspeed, so the gain should be set to a low value, indicating complete trust in the vehicle velocity calculated by the inertial integration. When t>t0, the inertial velocity integration error gradually accumulates and is greater than the estimated value using the airspeed, and at this time should improve the gain correction effect. G is the gain value and t1 factors the control the curve smoothness. 2.4. High Channel Kalman Filtering Model For the flight control level, the fixed-wing UAVs require a high accuracy of altitude position and lift rate, which affects the climbing, landing, and cruising performance of the UAV. The GNSS-denied environments are limited by the inertial navigation accuracy and the inability to measure the local gravitational acceleration precisely. Using low-precision inertial guidance alone for altitude solving, the altitude error will diverge significantly over time. The barometer is susceptible to high-frequency noise from the atmospheric environment, so fusing the barometer to the inertial navigation for correction is necessary. The inertial sensor, ADIS16488B, has a built-in barometer to measure static atmospheric pressure. The simplified conversion equation of atmospheric pressure to altitude is shown below [25]. The performance parameters of the barometer are shown as Table 1. (24) Hb=44,300×[1−(PsP0)15.255], here, Hb is the altitude we require. P0=1.01325 bar  is the value of standard atmosphere and Ps is the measured value of the barometer. Barometer measurement error is mainly affected by airflow intensity and temperature, and the changes of temperature make the barometric output drift. Using the temperature control system in the flight control can achieve the heat balance before the data fusion solution. In contrast, the ADIS16488B sensor embedded in the flight control component has been indirectly isolated from airflow, so the error correlation is significantly reduced. The height measurement error and the rate of change error correlation use the first-order Markov (Markov) process. (25) δh˙baro=−1τbaroδhbaro+ωbaroδv˙hbaro=−1τhbaroδvhbaro+ωhbaro, where δhbaro and δvhbaro, respectively, denote the error of barometer height and lift rate. τbaro and τhbaro denote the correlation time coefficient. ωbaro and ωhbaro represent white noise. (26) δv˙n=fsfn×f+vn×(2δωien+δωenn)−(2ωien+ωenn)×δvn+δfsfn+δgn≈fsfn×f+δfsfn+δgn The above Equation (26) is the error transfer model of the inertial navigation in the altitude channel. After ignoring the small error caused by rotation, the error equation of inertial navigation in the altitude channel is established as Equation (27). (27) δv˙U=−fNϕE+fEϕN+δg+ΔU, here, ϕN and ϕE  are the horizontal misalignment angle and δg denotes the gravity acceleration error. ΔU represents the altitude channel bias of the accelerometer. For the horizontal misalignment angle, we consider that it has been compensated in Equation (15). In addition, the error of the gravitational acceleration term is also ignored, so we consider that the velocity error in the altitude channel comes from the bias of the accelerometer. We describe the z axis velocity error by using the first-order Markov process. (28) δv˙U=−1τΔδvU+ωUδh˙INS=δvU+ωh, where δvU and δhINS denote the inertial navigation z-axis velocity error and the altitude error, respectively. τΔ is the correlation time coefficient. ωU and ωh are the white noise. According to Equations (25) and (28), the state equation of the system is established as follows:(29) X˙(t)=F(t)X(t)+W(t) The state is X(t)=[δvUδvbaroδHINSδHbaro]T The system translation matrix is:F(t)=[−1τΔz0000−1τvbaro001000000−1τhbaro]4×4 The system observation equation is:Z(t)=H(t)X(t)+V(t), here, H(t)=[1−100001−1]. The system equation is discretized and solved by the Kalman filter equation. The algorithm flow chart of the filter is shown as Figure 3:(30) v^U=vU−K1δvUH^INS=HINS−K2δHINS, where K=[K1K2]T is the error feedback coefficient. Adjusting the appropriate gain can make the signal smoother. 3. Experiment The small fixed-wing drone was used for experiments to verify the correct functionality in a practical scenario, and the vehicle is shown in Figure 4 and the UAV parameters are shown in Table 2. As an algorithm-verified vehicle, the UAV flight is fully autonomous on route. The flight control system was synthesized on an OMAP-L138 C6000 using the MATLAB/Simulink code generation design tool to build the embedded code. A serial-to-parallel interface (SPI) was developed to connect directly with the ADIS16488B sensor, which consisted of the three-axis gyro, three-axis accelerometer, three-axis magnetometer, and the barometer sensor. This IMU error indicator is shown in Table 3. We can see in Figure 2 that the maximum wind speed under this experiment is 3 m/s. Both the flight control algorithm and the multisensor fusion algorithm are arranged into the flight control hardware platform. The flow chart of the experimental tasks of the whole system is shown in Figure 5. The main parameter values of the algorithm are shown in Table 4. Since no higher accuracy inertial navigation is applied as the flight attitude reference, the combined GNSS/INS mode is still used for comparison and analysis in the UAV navigation control loop. The algorithm result data of the fused TAS/INS/BARO are saved to the flight log, and the vehicle completes the maneuvers of turning, circling, pulling up, and descending in the air independently to verify the navigation accuracy. Finally, the flight log is read after the vehicle lands for data comparison and analysis. The algorithm proposed in this paper is compared with the two-vector EKF model (denoted as EKF/TAS) proposed by [16], the centripetal force compensation model fused with airspeed presented in [19] (denoted as NCF/TAS), and the ACF model with adaptive adjustment weights estimated from external acceleration (denoted as ACF). Since there is little difference between the offline solution and online real-time calculation, we use offline processing to compare the accuracy of different algorithms. The results of the GPS/INS combination are used as true values for error calculation analysis. The results are shown below. Figure 6 and Figure 7, respectively, show the Euler angles calculation and attitude error comparison of the four algorithms. It can be seen from Table 5 and Figure 7 that the attitude error of the ACF is relatively stable, but it is challenging to correct the attitude error directly with the accelerometer because the external acceleration estimation of the air velocity is not accurate. It is obvious that attitude errors accumulate over time. EKF/TAS, NCF/TAS, and the algorithm proposed in this paper add airspeed measurement into the filter. Due to the influence of wind, higher fluctuations in attitude error can be seen. As EKF/TAS and NCF/TAS algorithms do not consider the effect of linear acceleration, the difference in attitude error among the three is not significant when the UAV flight speed is close to constant. At about 200 s, the UAV began to descend. It can be seen that the pitch error increased rapidly, and the instantaneous maximum reached nearly 11°. In 190 s, it can be seen in Table 5 that the MAE and RMSE of the roll error are 1.1388° and 1.4195°, the MAE and RMSE of pitch error are 1.1114° and 1.4672°, and the MAE and RMSE of yaw error are 4.7935° and 5.5443°, respectively. The yaw angle preformed worse due to the accuracy of the yaw estimation and was affected by the precision of the magnetometer. (It is difficult to calibrate the magnetic field around the UAV accurately, and the magnetometer is highly susceptible to disturbances from the electromagnetic environment, which makes the heading angle accuracy worse.) Generally, the error of the proposed algorithm is stable. The results suggest that this algorithm can adapt to the attitude estimation of the UAV under flight dynamics. Figure 8 shows the comparison among the velocity solutions of the different algorithms and RTK truth values. The horizontal velocity of the model in this paper can well track RTK velocity measurement with an error within 2 m/s. The defect of horizontal velocity error accumulated over time can be changed by using complementary filter and integrated airspeed. The vertical channel is integrated with a barometer to gain the maximum error of 0.5 m/s, which has obvious advantages over the other three algorithms. Figure 9 shows a comparison of the positions settled by RTK and the other four algorithms. Figure 10 compares the position errors of the different algorithms, while Figure 11 depicts the comparison between the track solution and the real track. It can be seen that the new algorithm can better track the position of the RTK. The error of 30 s (about 600 m flying) in the horizontal channel flight is within 30 m, the error of 90 s (about 1800 m flying) is within 50 m, and the average error of the height channel is 0.5 m, with higher accuracy than the other three. We also notice that during 140–190 s, the north position error of the proposed algorithm is a little more than the ACF and NCF/TAS. The low-precision INS position error transfer equation is established as Equation (31). (31) ΔR˙n=∫(fUϕE−fEϕU−2ωUΔvE)dt, where the ϕE and ϕU is the east and up direction attitude misalignment angle, respectively, which are the main source of north position errors. From Figure 7 and Figure 11, after the first turn, the roll angle error caused by random wind disturbance is converted to the north position cumulative error. After 160 s, the adaptive complementary filter in Equation (22) makes sense, and the north error stops growing. The algorithm eliminates the accumulated error caused by attitude misalignment error to position as much as possible. We also compared the offline computing efficiency of the four algorithms. The host computer with the 3.2 GHz AMD Ryzen 7 5800H CPU was used to run the four models with MATLAB 2021A. The total running time was set to 200 s with each step length of 0.005 s. Table 6 shows the comparison of the running time and its actual ratio among the different algorithms. As can be seen from the Table 6, the adaptive complementary filter algorithm (ACF) has the highest computational efficiency, accounting for only 2.12% of the actual operating time of the algorithm. It is usually a nice choice for a lightweight sensor fusion algorithm. The value of the fusion airspeed nonlinear complementary filter (NCF/TAS) is 3.21%. Due to differential tracking of airspeed data and other algorithm modules, the ratio of the running time in the proposed algorithm is 4.09%, slightly higher than ACF and NCF/TAS. Although the computation time is slightly longer, this processing improves the navigation accuracy, and this algorithm can be deployed in our flight control equipment for real autonomous flight verification. As for the EKF algorithm, it requires several high-dimensional matrix operations and is not superior in operational efficiency, accounting for 6.31%. The proposed algorithm consumes fewer computing resources than the EKF/TAS and can provide the higher precision attitude, position, and speed solutions than the EKF/TAS algorithm. The performance of the fusion algorithm is very satisfactory. 4. Conclusions This paper proposes a robust and universal sensor fusion algorithm, including an IMU, barometer, magnetometer, and airspeed sensor. The contributions of this paper include the following: (1) We use airspeed to improve the estimation accuracy of the nongravitational acceleration of vehicle, subsequently, to optimize the nonlinear complementary filter model of the vehicle’s attitude misalignment angle based on observability derivation, which can adapt to the state estimation accuracy of the UAV under different maneuvers. In the flight test, the MAE and RMSE of roll error are 1.1388° and 1.4195°, the MAE and RMSE of pitch error are 1.1114° and 1.4672°, and the MAE and RMSE of yaw error are 4.7935° and 5.5443°, respectively, and, when compared with the more commonly used EKF algorithm, is improved. (2) At the level of the horizontal velocity fusion, a complementary filtering model using airspeed correction is established to suppress the accumulated errors caused by the calculation speed of INS. (3) At the height level, the Kalman filter model is designed using barometer data so that the vehicle can obtain the accurate solution of the lift rate and altitude without GNSS. The average error of the height channel is 0.5 m, and the maximum error of the lift rate is 0.5 m/s. This design idea uses a cascade fusion strategy that combines the benefits of an individual systems model using a cascade fusion strategy, combining the advantages of a single system. Compared against the other three conventional methods, the proposed method shows superior performance, providing good attitude velocity and position estimation, even in GNSS-denied environments. In addition, the algorithm proposed in this paper consumes lower computing resources and is suitable for common embedded systems. Taken as a whole, the new approach provides a feasible solution for the navigation and positioning of small UAVs, as much as possible in GNSS-denied environments; however, the result is still not very precise. In further works, we will continue to explore multiple fusion navigation technologies, and explore the use of low-cost camera sensors to enhance the robustness and fault tolerance of navigation systems. Author Contributions Conceptualization, X.Y. and Q.Z.; methodology, X.Y.; software, X.Y.; validation, Q.Z.; formal analysis, X.Y.; investigation, X.Y.; resources, Q.Z.; data curation, X.Y.; writing—original draft preparation, X.Y.; writing—review and editing, X.Y., Y.Z. (Yijun Zou) and Y.Z. (Yifan Zeng); visualization, Y.Z. (Yifan Zeng) and Y.Z. (Yijun Zou); supervision, Q.Z.; project administration, Q.Z.; funding acquisition, Q.Z. All authors have read and agreed to the published version of the manuscript. Funding This research was funded by the National Natural Science Foundation of China, grant number 61174120. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement The data presented in this study are available on request from the corresponding author. The data are not publicly available due to confidentiality agreement for the project. Conflicts of Interest The authors declare no conflict of interest. Figure 1 The framework of the multisensor fusing algorithm in this paper is divided into four parts. The green area shows the sensors we used in this study; the original sensor’s data will first be preprocessed, including using the Butterworth low-pass filter to reduce the high-frequency noise of the gyroscope and accelerometer, and using a differential tracker (DT) for the airspeed and barometer. The blue part is the attitude fusion frame, mainly divided into three parts: the main filter, the gravitational acceleration estimation described in Section 2.1, and the attitude misalignment angle calculation and the error feedback compensation are described in Section 2.2. Meanwhile, the yellow area presents the velocity fusion frame, and in Section 2.3, the filter which combines the barometer and airspeed is mainly described. The gray area represents the altitude fusion frame by the Kalman filter, which is established in Section 2.4. Figure 2 The comparisons between the true airspeed (TAS) measured by pilot tube and the ground velocity measured by RTK. The orange line shows the original signal of TAS. The red line and the green line represent the TAS data of the filtered and differential tracker (TD) modules, respectively. Figure 3 The flow chart of the Kalman filter. At each calculated step size, the Kalman algorithm runs in two steps: time updating and measurement updating. The time-update step predicts the navigation states vector and its covariance matrix by propagation through a model of the system dynamics. The measurement-update uses the data from the sensors and is incorporated to correct the prediction and output an optimal estimation by calculating the optimal Kalman gain. The system error state can be estimated by the Kalman filter, then the system error can be corrected by closed-loop feedback. Figure 4 The experimental platform of the small fixed-wing UAV. Figure 5 The system experimental task flow chart. The flight control system needs to wait for 5 s for bias correction and self-alignment after power-on. When receiving RTK signal, the system enters GNSS/INS cooperative mode. The algorithm switches to TAS/INS/BARO fusion mode during the flight after receiving the navigation switching command from the ground station. Figure 6 Three-axis attitude solution of five algorithms in flight is shown for pitch, roll, and yaw angle (from (a–c), respectively). Figure 7 GPS/INS integrated navigation result is used as true values. Three-axis attitude error of other four algorithms in flight is shown for pitch, roll, and yaw error angle (from (a–c), respectively). Figure 8 Three-axis velocity solution of five algorithms in flight is shown for east, north, and up velocity (from (a–c), respectively). In order to find difference of horizontal position, NCF/TAS and what we proposed use the same height model, so the curve of NCF/TAS is not drawn in the bottom figure. Figure 9 Three-axis position solution of five algorithms in flight is shown for east position, north position, and altitude (from (a–c), respectively). In order to find difference of horizontal position, NCF/TAS and what we proposed use the same height model, so the curve of NCF/TAS is not drawn in the figure (c). Figure 10 GPS navigation result is used as true values. Three-axis position error of the other four algorithms in flight is shown for east position, north position, and altitude error angle (from (a–c), respectively). Figure 11 Comparison of 190 s long trajectories obtained from different algorithms, and what we proposed is the most accurate tracking. sensors-22-03156-t001_Table 1 Table 1 The performance parameters of the barometer we use. The errors of the barometer mainly consist of the measurement error and measurement noise. The absolute error of measurement reaches 40 m. Measurement Error Parameters The Parameter Values Measurement range 300−1100 mbar −700∼9165 m Measurement error 4.5 mbar About 40 m (About 500 m above sea level) Measurement noise 0.025 mbar About 0.2 m (About 500 m above sea level) sensors-22-03156-t002_Table 2 Table 2 Parameters of fixed-wing UAV. Parameters Value Total Weight 6.9 kg Span 2100 mm Body length 1620 mm Power Electric drive sensors-22-03156-t003_Table 3 Table 3 Specifications of ADIS16488B. Parameters Typical Value In-Run Bias Stability of Gyroscope 6.25°/hr Angular Random Walk 0.3°/hr In-Run Bias Stability of Accelerometer 0.1 mg Velocity Random Walk 0.029 m/s/hr sensors-22-03156-t004_Table 4 Table 4 The main parameter values of the algorithm we used in the following experiment. Categories Variable Definition Value Differential tracker filter T Period of the input signal 0.005 h Filter factor 0.15 r Rate factor 900 Attitude calculation model kp The ϕ γ compensation gain 0.05 kI The ϕ γ compensation integral gain 0.01 kp The ψ compensation gain 0.2 kI The ψ compensation integral gain 0.01 Horizontal channel adaptive complementary fusion KTASv The gain factor of the complementary filter 0.9 G The gain value 1 t0 The time switching threshold 30 t1 The curve smoothness control factor 10 High Channel Kalman Filtering Q The error covariance matrix diag ([0.1, 0.5, 0.1, 2])2 R The measurement noise covariance matrix diag ([1, 10])2 P0 The initial covariance matrix diag ([0.1, 1, 0.1, 10])2 K1 The error feedback coefficient of lift rate 0.2 K1 The error feedback coefficient of altitude 0.8 sensors-22-03156-t005_Table 5 Table 5 In the experiment, the attitude error evaluation indexes of MAE (Mean Absolute Error) and RMSE (Root Mean Square Error) calculated by the proposed fusion algorithm are compared with those calculated by the adaptive complementary filter (ACF), the two-vector extended Kalman filter fused with airspeed (EKF/TAS), and the nonlinear complementary filter fused with airspeed (NCF/TAS). Methods Attitude (deg) Roll (γ) Pitch (ϕ) Yaw (ψ) ACF MAE 0.9871 1.8859 9.4879 RMSE 1.1324 2.0623 9.8084 EKF/TAS MAE 1.2389 1.8450 7.0168 RMSE 1.5633 3.1493 8.4507 NCF/TAS MAE 1.2564 1.8770 6.2642 RMSE 1.5557 3.2114 7.0386 PROP MAE 1.1388 1.1114 4.7935 RMSE 1.4195 1.4672 5.5443 sensors-22-03156-t006_Table 6 Table 6 Comparison of the computational efficiency among different algorithms. Algorithm Total Time (s) The Ratio of the Actual Running Time of the Algorithm to the Total Simulation Time ACF 4.2402 2.12% EKF/TAS 12.6251 6.31% NCF/TAS 6.4150 3.21% PROP 8.1774 4.09% Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Bhowmick J. Singh A. Gupta H. Nallanthighal R. A Novel Approach to Computationally Lighter GNSS-Denied UAV Navigation Using Monocular Camera Proceedings of the 2021 International Conference on Automation, Robotics and Applications, ICARA 2021 Prague, Czech Republic 4–6 February 2021 114 121 2. Viswanathan A. Pires B.R. Huber D. Vision-based robot localization across seasons and in remote locations Proceedings of the IEEE International Conference on Robotics and Automation Stockholm, Sweden 16–21 May 2016 4815 4821 3. Qu Y.H. Yang M.H. Zhang J.Q. Xie W. Qiang B.H. Chen J.L. An Outline of Multi-Sensor Fusion Methods for Mobile Agents Indoor Navigation Sensors 2021 21 1605 10.3390/s21051605 33668886 4. Alibakhshikenari M. Virdee B.S. See C.H. Abd-Alhameed R.A. Falcone F. Limiti E. Super-Wide Impedance Bandwidth Planar Antenna for Microwave and Millimeter-Wave Applications Sensors 2019 19 2306 10.3390/s19102306 31109122 5. Shirkolaei M.M. Ghalibafan J. Scannable Leaky-Wave Antenna Based on Ferrite-Blade Waveguide Operated Below the Cutoff Frequency IEEE Trans. Magn. 2021 57 10.1109/tmag.2021.3060683 6. Konrad T. Gehrt J.J. Lin J.Y. Zweigel R. Abel D. Advanced state estimation for navigation of automated vehicles Annu. Rev. Control 2018 46 181 195 10.1016/j.arcontrol.2018.09.002 7. Niu X. Ban Y. Zhang T. Liu J. Research progress and prospects of GNSS/INS deep integration Acta Aeronaut. Astronaut. Sin. 2016 37 2895 2908 8. Mourikis A.I. Roumeliotis S.I. A multi-state constraint Kalman filter for vision-aided inertial navigation Proceedings of the IEEE International Conference on Robotics and Automation Rome, Italy 10–14 April 2007 3565 9. Qin T. Li P. Shen S. VINS-Mono: A Robust and Versatile Monocular Visual-Inertial State Estimator IEEE Trans. Robot. 2018 34 1004 1020 10.1109/TRO.2018.2853729 10. Mohta K. Sun K. Liu S. Watterson M. Pfrommer B. Svacha J. Mulgaonkar Y. Taylor C.J. Kumar V. Experiments in Fast, Autonomous, GPS-Denied Quadrotor Flight Proceedings of the IEEE International Conference on Robotics and Automation Brisbane, Australia 21–25 May 2018 7832 7839 11. Young A.D. Comparison of Orientation Filter Algorithms for Realtime Wireless Inertial Posture Tracking Proceedings of the 6th International Workshop on Wearable and Implantable Body Sensor Networks Berkeley, CA, USA 3–5 June 2009 59 64 12. Leutenegger S. Siegwart R.Y. A low-cost and fail-safe inertial navigation system for airplanes Proceedings of the IEEE International Conference on Robotics and Automation Saint Paul, MN, USA 14–18 May 2012 612 618 13. Nazarahari M. Rouhani H. 40 years of sensor fusion for orientation tracking via magnetic and inertial measurement units: Methods, lessons learned, and future challenges Inf. Fusion 2021 68 67 84 10.1016/j.inffus.2020.10.018 14. Madgwick S.O.H. Harrison A.J.L. Vaidyanathan A. Estimation of IMU and MARG orientation using a gradient descent algorithm Proceedings of the 2011 IEEE International Conference on Rehabilitation Robotics Zurich, Switzerland 29 June–1 July 2011 Volume 2011 5975346 10.1109/icorr.2011.5975346 15. Mahony R. Hamel T. Pflimlin J.M. Nonlinear complementary filters on the special orthogonal group IEEE Trans. Autom. Control 2008 53 1203 1218 10.1109/TAC.2008.923738 16. Sabatelli S. Galgani M. Fanucci L. Rocchi A. A Double-Stage Kalman Filter for Orientation Tracking With an Integrated Processor in 9-D IMU IEEE Trans. Instrum. Meas. 2013 62 590 598 10.1109/TIM.2012.2218692 17. Liu X. Liu X. Zhang W. Yang Y. UAV attitude calculation algorithm based on acceleration correction model J. Northwestern Polytech. Univ. 2021 39 175 181 10.1051/jnwpu/20213910175 18. Hemerly E.M. Maciel B.C.O. Milhan A.D. Schad V.R. Attitude and heading reference system with acceleration compensation Aircr. Eng. Aerosp. Technol. 2012 84 87 93 10.1108/00022661211207893 19. Euston M. Coote P. Mahony R. Jonghyuk K. Hamel T. A complementary filter for attitude estimation of a fixed-wing UAV Proceedings of the 2008 IEEE/RSJ International Conference on Intelligent Robots and Systems Nice, France 22–26 September 2008 340 345 20. Li X. Li Q. External Acceleration Elimination for Complementary Attitude Filter Proceedings of the IEEE International Conference on Information and Automation (ICIA) Macau, China 18–20 July 2017 208 212 21. Marantos P. Koveos Y. Kyriakopoulos K.J. UAV State Estimation Using Adaptive Complementary Filters IEEE Trans. Control Syst. Technol. 2016 24 1214 1226 10.1109/TCST.2015.2480012 22. Leutenegger S. Melzer A. Alexis K. Siegwart R. Robust state estimation for small unmanned airplanes Proceedings of the 2014 IEEE Conference on Control Applications, CCA 2014 Juan Les Antibes, France 8–10 October 2014 1003 1010 23. Jing-Qing H. Wei W. Nonlinear Tracking-Differentiator Inst. Systerm Sci. 1994 14 177 183 24. Yan G.M. Qin Y.Y. Initial Alignment Accuracy Analysis and Simulation of Strapdown Inertial Navigation System on a Stationary Base Comput. Simul. 2006 23 36 25. Alberi M. Baldoncini M. Bottardi C. Chiarelli E. Fiorentini G. Raptis K.G.C. Realini E. Reguzzoni M. Rossi L. Sampietro D. Accuracy of Flight Altitude Measured with Low-Cost GNSS, Radar and Barometer Sensors: Implications for Airborne Radiometric Surveys Sensors 2017 17 1889 10.3390/s17081889 28813023

PMC009xxxxxx/PMC9099685.txt

==== Front Animals (Basel) Animals (Basel) animals Animals : an Open Access Journal from MDPI 2076-2615 MDPI 10.3390/ani12091081 animals-12-01081 Article Long-Distance Counter Calling in Maned Wolves: Friends or Foes? Ferreira Luane S. 1 Sábato Victor 1 Pinheiro Thiago A. 1 Neto Edvaldo 1 Rocha Luciana H. 1 Baumgarten Júlio 2* https://orcid.org/0000-0002-4797-0085 Rodrigues Flávio H. 3 https://orcid.org/0000-0002-2638-1695 Sousa-Lima Renata S. 1 Hart Lynette A. Academic Editor 1 Laboratory of Bioacoustics, Department of Physiology & Behavior, Universidade Federal do Rio Grande do Norte, Natal 59078-970, RN, Brazil; fsluane@gmail.com (L.S.F.); victorsabato@gmail.com (V.S.); thiago4jazz@gmail.com (T.A.P.); nettoBlack@gmail.com (E.N.); lua.lupin@gmail.com (L.H.R.); sousalima.renata@gmail.com (R.S.S.-L.) 2 Applied Ecology and Conservation Lab, Universidade Estadual de Santa Cruz, Ilhéus 45662-900, BA, Brazil 3 Departamento de Biologia Geral, Universidade Federal de Minas Gerais, Belo Horizonte 31270-901, MG, Brazil; rodriguesfhg@gmail.com * Correspondence: baumgarten.julio@gmail.com 21 4 2022 5 2022 12 9 108117 12 2021 11 2 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Simple Summary Maned wolves generally maintain long distances between individuals and are difficult to see in the wild. To understand how they interact, we recorded sequences of alternating long-distance calls (roar-barks) of maned wolves in captivity and in the wild. In natural habitat recordings, we detected more interactions during the mating and initial parental care periods, suggesting communication among mated pairs and, later, among parental caregivers. In captivity, almost all vocal interactions involved both sexes and males presented longer roar-barks compared to females. We measured the same parameter in wild maned wolves and found that the participants in such vocal interactions differ with respect to the duration of their calls, suggesting that maned wolves engaging in long-distance counter-calling are mates. Such acoustic-based inferences of wildlife behavior are cost-effective and can be a useful tool in the conservation efforts to protect this vulnerable canid, and potentially other endangered species that are difficult to observe in the wild. Abstract Maned wolves (Chrysocyon brachyurus) are monogamous and display biparental care for their young, although adults rarely spend time in close proximity. To better understand vocal interactions of maned wolves over long-distances, we passively recorded >10 months of audio data in the species’ natural habitat and analyzed manual recordings of captive animals, covering the reproductive and non-reproductive seasons. In the natural habitat recordings, we found that maned wolves engage in vocal exchanges (termed interactive sequences) more often during the mating season, suggesting the existence of a partner attraction/reunion/guarding function, and also during the initial parental care period, suggesting communication among caregivers. We analyzed 21 interactive sequences, which were the only instances in which we could distinguish individuals, and found that the individuals interacting differed significantly in their roar-bark parameters, including duration, which also differed between males and females in captivity (male vocalizations were, on average, 0.124 s longer). We also found that interactive sequences in captive animals, involving two or more participants, almost always involved both sexes. These results suggest that acoustic interacting maned wolves are most likely male–female dyads. Chrysocyon brachyurus canid vocal interaction acoustic parameters roar-bark captivity free ranging ==== Body pmc1. Introduction The maned wolf (Chrysocyon brachyurus, Illiger 1815) is the only large canid from South America (70–90 cm shoulder height, 20–30 kg weight [1]). It is an omnivorous and mainly solitary animal that inhabits savanna-like environments [2]. Due to its low densities (1–8/100 km2 [3]), nocturnal–crepuscular habits, and shy nature, the species is hard to study in the wild [4]. Exploring the long-range acoustic communication of cryptic species is an interesting monitoring alternative [5] that has already proven efficient for studying maned wolves [6,7]. Maned wolves produce an explosive roar-bark [8], similar to a longer bark from a domestic dog, emitted in sequences of 5–15 units separated by relatively long intervals of 2–6 s [7]. Roar-bark sequences can often travel more than 1 km, potentially over 3 km [9], making this species more easily detected acoustically than visually [10] (and LSF personal observation). Nonetheless, maned wolves are not highly vocal. In captivity, individuals emit 0.68 roar-bark sequences by night in the reproductive season and 0.28 sequences by night outside the reproductive season [11]. It is hypothesized that this vocalization is used for territorial announcement and defense, mediating same sex spacing [8,11,12]. Maned wolves are monogamous, and pairs use and defend the same home-range, patrolling it separately [13,14]. The species shows biparental care and yearlings may also participate as helpers [4,15]. Therefore, long-range calling could potentially function in intra-familiar group communication [10,16,17]. Contrary to other indirect signs of a species’ presence (e.g., footprints and urine/scat), vocal signs allow the study of real-time interactions between individuals [18], a potential that has been little explored in the maned wolf. Although most roar-bark sequences are emitted by a single individual [8,10,19], 12–33% of sequences involve roar-bark alternations between two or, more rarely, three or more maned wolves (referred to here as interactive, answered, or group sequences [8,10,19]). During intervals between roar-barks, wolves seem to wait and aurally attend to answers [10], suggesting that the long internote intervals may function to facilitate those long-distance vocal exchanges [20]. Pair-mates exchange roar-barks both in captivity [17,19] and in the wild [10,15,17]. Maned wolves may also visually search for their partner after roar-barking [16] and, during estrous, they emit roar-barks whenever the partner is out of visual range [15]. Often partners reunite after the emission of roar-barks or the receiver moves towards the caller [10,19]. Therefore, some authors propose that group vocalizations are primarily pair-mate communication, at least during the breeding season [7,10,19]. Contrary to that, others [12] have found that roar-bark sequences are more often answered by same sex individuals. Free ranging wolves with adjacent home-ranges have been heard exchanging roar-barks and calling toward perceived threats from conspecifics or humans [19]. Maned wolves vocally responded to male and female roar-bark playbacks, although elicited sequences did not alternate with played back roar-barks [9]. These last three pieces of evidence are in accordance with the suggestion that group sequences of roar-barks mediate resource disputes. The majority of the studies mentioned were conducted in captivity, or short opportunistic observations and thus, long term acoustic monitoring should help elucidate maned wolves’ long-distance acoustic interactions. A previous study exploring our long-term acoustic dataset [7] shows that interactive roar-bark sequences follow the seasonal (but not the lunar or nightly) general pattern of solo/unanswered sequences. However, at that time, recordings during the non-reproductive season were not available. Here we were interested in determining who are the individuals interacting vocally and we predict that if interactive sequences are reproductive male and female communication, then they would be more common during the reproductive season, or some of its periods (mating: March–April, gestation: May; parturition: June; initial parental care: July–September), than during the non-reproductive season. Different rates of interactive sequences depending on the period of the year would also be observed if interactive sequences are disputes for seasonal reproduction-related resources, e.g., territorial disputes may be more intense in critical periods, such as mating and lactation. In this case, we would expect same sex vocal interactions (disputing for mates, similar to jackals [21]) or no sex bias in individuals participating in roar-bark exchanges (interfamilial resource disputes). If interacting maned wolves are male and female dyads, then the animals’ difference in sexually dimorphic acoustic parameters will be large (significantly different from zero). Finally, if interactive sequences are mostly emitted in one of the aforementioned contexts, then we should find the same trend in captivity data, where we have the identity of participants. 2. Materials and Methods 2.1. Natural Habitat Area Recordings were done at the Serra da Canastra National Park, in Minas Gerais state, Brazil (Figure 1). The Park is mainly composed of highland Cerrado open savannas with a cold, dry season (April–September) and a hot, rainy season (October–March) [22]. Previous capture-recapture studies indicate a density of 8 maned wolves per 100 km2 (considered high [23]), with home ranges averaging 80 km2 [13]. 2.2. Natural Habitat Data All natural habitat recordings were made passively, with autonomous recorders SongMeter SM2+ (Wildlife Acoustics, Inc., Concord, MA, USA) coupled with a single SMX-II omnidirectional weatherproof microphone each (Wildlife Acoustics, Inc., Concord, MA, USA). Recorders were distributed broadly in high places (1373.0 ± 56.6 m altitude) and attached to 1.4 m wooden stakes. The equipment was set to record with +36 dB gain, 8 kHz sample rate, and 16-bit wave format coding. In 2014 we recorded between 5 April and 8 August, during the maned wolf reproductive season. We deployed 12 autonomous recorders (Figure 1, “2014” labels) set to record from 18:00 p.m. to 6:00 a.m. every 24 h. In 2016 we recorded from 9 March to 1 July, also during the maned wolf reproductive season. We deployed 13 autonomous recorders (Figure 1, “2016” labels) set to record from 17:00 p.m. to 5:00 a.m. every 24 h. Between 1 December 2016 and 31 January 2017, during the maned wolf non-reproductive season, we deployed 8 autonomous recorders (Figure 1, “2016–2017” labels) programmed to record during the first 3 h of the night (17:45–20:45 p.m.: the period of highest maned wolf roar-bark activity [6]). The non-reproductive recording scheme differed from the previous ones because it was used for a multi-objective project. Equipment malfunction, due to low battery and equipment wear, restricted the use of all the data. Nights during which at least half of the deployed recorders were active were considered for analyses. During 2014, 118 nights were considered with 11.9 ± 0.3 (mean ± SD) average active recorders, in 2016 we considered data from 105 nights with 12.3 ± 1.8 average active recorders and, in 2016–2017, we considered 37 nights with 6.3 ± 1.2 average active recorders. Since the number of recorders and the times recorded differed among years, all sequence counts were divided by the number of active recorders, and only the first 3 h of the night were selected for subsequent analyses. That is, between 17:45 p.m. and 20:45 p.m. in 2016 and 2016–2017 and from 18:00 p.m. to 21:00 p.m. in 2014. Some areas recorded could be more intensely used by maned wolves than others, which would influence the probability of detecting their roar-barks. As the recorders’ sites varied between years (Figure 1), and not all of them were active during the same nights, the analysis with the full dataset could be spatially biased/unbalanced. In an effort to control for that, we made a subsampled data including only the 5 recording sites that were used in all years (sites with no label in Figure 1), and only nights in which all of those 5 recorders were active. We analyzed both the data with all recorders and the subsampled with only 5. 2.3. Roar-Bark Sequence Detection, Counting and Classification Roar-barks were detected in the audio files using an automatic detector built in XBAT-R7 (Extensible Bioacoustic Tool [24]) extension for Matlab R2011a (MathWorks, Inc., Natick, MA, USA). Detections were manually validated according to a published protocol [25]. The end of a sequence was arbitrarily defined as the point when there was more than 10 s between any roar-barks, independent of the emitter [6]. This definition is important for moments of high vocal activity. For each sequence found we noted the date, begin time, and the number of animals involved. Multiple participating maned wolves can be detected by differences in cadence, relative intensity, time of arrival at different sensors, and spectral characteristics of roar-barks, as well as occasional overlaps which indicate calls came from different individuals (Figure 2; Supplementary Material Audio S1–S3). As the recordings were made passively, that is, with autonomous recorders and no visual information, interactive roar-bark sequences are the only instances we can distinguish between free ranging individuals. Although determining which individual emitted which roar-bark is usually easy when there are only two animals, sequences involving more than two jeopardizes individual discrimination of roar-barks, as the calls signal-to-noise ratio is commonly too low to allow any comparison. We classified sequences as “solo” when there was no indication of more than one animal participating, or “interactive” when there was more than one maned wolf alternating roar-barks. Sequences in which the presence of another animal was uncertain (e.g., a solo sequence with one faint mark in between that may or may not be a roar-bark) were not included in either category. When those categories are not mentioned the data refers to all vocal activity, combining solo, interactive, and uncertain sequences. We also use the term “dyad sequence” to refer to any interactive sequence involving only 2 animals. For the natural recordings the term is almost a synonym of interactive sequences, as more than 2 animals in the same sequence is very rare in our wild dataset. However, the term is very useful for the captivity dataset. We divided our sample in parts within the reproductive cycle of the species that we refer to as ‘periods’. We considered the mating period from 1 March to 20 April and the gestational period from 21 April to 31 May (shortened because the exact time of conception and parturition varies). The parturition period was considered to be June and the initial parental care was in July (our latest records during the reproductive season). These periods are based on the reported mating period for the species [15], a gestation of 65 days [26], and the reported peak in births for the Serra da Canastra National Park in June [4,19,27]. We have confirmation that at least one female in the area was lactating in July 2014 and 2016 (R. C. de Paula, personal communication). The 2016–2017 dataset was considered as the non-reproductive period. Besides calculating nightly sequences by recorder and nightly percentage of interactive sequences, we made 4 comparisons between periods: the proportion of solo/interactive sequences (number of roar-bark sequences), the proportion of nights with interactive sequences versus with solo sequences only (nights without vocal activity are not included), the proportion of nights with versus without sequences (nights with solo sequences only and without vocal activity are both included in the last category), and the proportion of all four categories of nights (without vocal activity, with solo sequences only, with interactive sequences only, and with any kind of vocal activity). 2.4. Relative Distance Estimation Sometimes the same sequence was registered by more than one autonomous recorder (Figure 3). As the recorders were not time synchronized, time cannot be used to guarantee it was the same sequence. Certainty about it stems from several idiosyncrasies in maned wolf roar-barks sequences, e.g., unique inter-roar-bark intervals. This way, if we can temporally align roar-barks recorded in two or more different recorders, it becomes easy to evaluate if it is the same sequence. To avoid counting sequences twice in the seasonal analysis we only considered the recording with the highest signal-to-noise ratio. In some rare cases an interactive sequence was registered by more than one recorder. In most of those cases the alignment of roar-barks between recorders was only possible with the roar-barks of a single animal at a time (Figure 2). This happens because the acoustically interacting maned wolves are not at the same position [7]. Therefore roar-barks of each animal will travel different distances to reach each recorder and will arrive at different times in relation to each other. We used the time of arrival (TOA) difference between the roar-barks of the different animals in each recorder to make estimates of the relative distance between vocalizing wolves. If animals are vocalizing together, the TOA difference will be zero and as animals are further away from each other it will increase. For instance, in Figure 2 the time difference could be calculated by the time between a3 and b3 in recorder I minus the time between the same roar-barks in recorder II. Considering a sound speed of 343 m/s and that the time difference is created in both recorders, we used the formula “(time difference/2) × 343” to calculate a distance in meters. We recalculated this estimation for all interactive sequences registered by multiple recorders that had enough quality for the time difference measure. We used any interactive sequence that met these criteria, independent of hour or time of the year. 2.5. Captivity Data We used data collected in 2011 [11] at two facilities in the state of Minas Gerais (Brazil): the Criadouro Científico de Fauna Silvestre para Fins de Conservação da Companhia Brasileira de Metalurgia e Mineração (CC-CBMM) and the Zoológico da Associação Esportiva e Recreativa dos Funcionários das Usinas Siderúrgicas de Minas Gerais (ZOO-USMG). Animals were recorded with a unidirectional microphone, Sennheiser K6-module ME-66 (40–20,000 Hz ± 2.5 dB flat response frequency), connected to a Marantz PMD-661 solid state recorder using 96 kHz sampling rate and 24-bit wav encoding format. At CC-CBMM, acoustic monitoring was done for 40 nights during the breeding season (between April and June) sampling calls from four adult captive maned wolves females (SA, FI, JU, RO) and 2 males (SH, NE). Two pairs were housed together (SH + SA, NE + FI) and the remaining females were housed separately in enclosures with no other wolves. RO is the mother of NE and JU which have different fathers, and JU is the mother of SA. The males that sired the captive animals were no longer in the facility. At ZOO-USMG the acoustic monitoring was done for 20 nights outside the breeding season (November). Four adult captive maned wolves were recorded. They were housed in two pairs (male + female: GA + LU and GI + BA). The males are litter siblings. An example of each captive maned wolf roar-bark is shown in Figure 4. Data from captive animals was used to determine sexually dimorphic acoustic parameters. We additionally counted the number of solo and interactive roar-bark sequences recorded in each season, as well as the sex and identity of captive participants. 2.6. Acoustic Parameters We looked for acoustic parameters that were robust against the effects of propagation through distance, or that were little affected relative to the variation due to individual differences (in mammals [28,29]; in maned wolves [20,30]). Parameters measured in the roar-barks were taken automatically after each call was manually selected in Raven Pro 1.6 (Bioacoustics Research Program, 2014. Ithaca, NY, USA: The Cornell Lab of Ornithology; http://www.birds.cornell.edu/raven, accessed on 13 January 2022). The acoustic parameters selected were (Figure 4): the roar-bark Total Duration, in seconds; the InterQuartile Duration, in seconds (the duration between the moment of 25% energy accumulation in time and the moment of 75% energy accumulation in time); the 2nd Band Peak Frequency, in Hertz (the frequency of highest intensity of the second frequency band, which is usually at 600–1000 Hz); the 2nd Band 1st Frequency Quartile, in Hertz (the frequency that accumulates 25% energy of the second band); and the 2nd Band 3rd Frequency Quartile, in Hertz (the frequency that accumulates 75% energy of the second band). The parameters descriptions were taken from Raven Pro User’s Manual [31]. The acoustic parameters of the captivity dataset were measured in spectrograms built in Raven Pro 1.6, using Hann window, 4096 window size, 50% overlap, 50% brightness, and 75% contrast. For the natural habitat recordings, we used spectrograms with 512 window size, 45–55% brightness, and 60% contrast. One selection box was made comprising the first two frequency bands of roar-barks, fixed from 200 Hz to 2000 Hz (Figure 3 blue box), for the extraction of the Total Duration and the InterQuartile Duration. A second selection box was made comprising only the roar-bark second frequency band (Figure 3 green box), around 620 Hz to 1000 Hz depending on the roar-bark, for the extraction of the remaining three frequency parameters. Twenty good quality roar-barks (high signal-to-noise ratio) of each captive individual were used (total: 200 roar-barks). We tried to include roar-barks of as many different sequences as possible, but some animals emitted very few of them. The number of sequences from which we selected roar-barks varied from 3 to 19 per individual, totaling 56 distinct sequences, of which 23 were solo and 33 interactive sequences (in some of which, we selected calls from more than one wolf). To control for those factors, we noted for each captivity roar-bark the emitting individual, the sequence it belonged to and how many participants the sequence had. In the recordings from the natural habitat, we used any good quality interactive sequence, independent of hour or time of the year. We selected interactive sequences in which both animals had at least 3 good-quality and confidently identified (as belonging to the same animal) roar-barks each. Our goal was 5 good quality roar-barks, but that was not always possible. We averaged the acoustic parameters’ values from roar-barks of the same animal in a sequence in order to avoid pseudoreplication, and also to stabilize parameters possibly affected by varying noise characteristics (wind, other zoophony, etc.). Pairwise comparisons were conducted, that is, each animal was only compared to the one it was vocally interacting with in the interactive sequence considered. We tried to select homogeneously distributed interactive sequences across all periods, registered at different sites (recorders) and avoided sequences emitted less than a night apart from each other. This was done to avoid biasing the data with many sequences from the same individuals. In total, 21 interactive sequences from the natural habitat recordings were selected for analyses. To visually show the distribution (in box plots) of the roar-bark parameters in the natural recordings, we classified them based on which part of the spectrum they occupy: for each interaction the wolf with the smaller value of “duration” between the 2 was assigned to the “shortest” set, and the other to the “longest”; the wolf with the smaller value of “frequency” between the 2 was assigned to the “lowest” set, and the other to the “highest”. 2.7. Statistical Analysis Data was generally not normally distributed (tested with Shapiro–Wilk normality test), so we used Kruskal–Wallis tests, followed by pairwise Wilcoxon tests adjusted by the Benjamini–Hochberg method [32], for the seasonal analysis. To compare the seasonal variance in proportions of sequences and of night composition we used chi-square tests, followed by Fisher’s exact test if any category had less than 5 counts, and a comparison of Pearson’s residuals to find the groups and categories that deviated from the expected proportion. To compare captive male and female acoustic parameters we transformed the non-normal parameters using the Yeo-Johnson method [33] and then fitted linear mixed models to each parameter controlling for the random factor sequence nested in individual. We also included in the fixed factors the number of participants (one/two/three or more) to evaluate if the sequence type influences the acoustic parameter. The final formula was acoustic parameter ~sex + participants + (1|individual/sequence). We used ANOVAs followed by Tukey contrasts to test the effect of fixed factors and their levels. To test the difference in the acoustic parameters of participants of interactive sequences in the natural habitat recordings we used paired Wilcoxon tests. All statistics were computed in R (R version 4.1.1 (10 August 2021)—“Kick Things” Copyright © 2021 The R Foundation for Statistical Computing, Vienna, Austria). 3. Results 3.1. Seasonal Distribution of Roar-Bark Sequences: When They Interact Considering only the first 3 h of the night, we found a total of 523 maned wolf roar-bark sequences in the combined datasets; of those, 58 were interactive sequences (11%). Only two interactive sequences showed engagement of three or more animals instead of two. The distribution pattern during the reproductive season (Figure 5a) shows a high vocal activity during the mating period, a decrease during the gestational period, and a smaller increase circa parturition (with an occasional peak in the initial parental care period). The number of interactive sequences in general followed the vocal activity pattern, except for the non-reproductive season that presented only four interactive sequences in two nights (Table 1). There were many nights with high percentages of interactive sequences in the reproductive season, especially in March and from June on (Figure 5b). Unexpectedly, the vocal activity during the non-reproductive season was as high as in the mating period (Figure 5a). However, visually the occurrence of interactive sequences was rare during this period (Figure 5b). Considering the entire periods, for the complete dataset, the initial parental care presented a tendency for a higher proportion of interactive sequences compared to the expected solo/interactive sequence proportion (Chi-squared test = 8.823, df = 4, p = 0.0657). The mating period presented a higher proportion of nights with interactive sequences compared to the expected proportion of nights with and without any interactive sequence (Figure 6a; X2 = 10.939, df = 4, p = 0.0273, Fisher’s exact test: p = 0.0329) and a tendency of a higher proportion of nights with both solo and interactive sequences compared to the expected proportion of night composition (no sequence, solo only, interactive only, both; Figure 6c; X2 = 20.684, df = 12, p = 0.0552, Fisher’s exact test: simulated p = 0.0705). For the dataset including only the five recorders in common, the initial parental care period presented a higher proportion of interactive sequences compared to the expected solo/interactive sequence proportion (X2 = 20.003, df = 4, p = 0.0005, Fisher’s exact test: p = 0.0019), a higher proportion of nights with interactive sequences compared to the expected proportion of nights with interactive sequences and with only solo sequences (X2 = 10.003, df = 4, p = 0.0404, Fisher’s exact test: p = 0.0352), a higher proportion of nights with interactive sequences compared to the expected proportion of nights with and without any interactive sequence (Figure 6b; X2 = 12.269, df = 4, p = 0.0155, Fisher’s exact test: p = 0.0083), and a higher proportion of nights with interactive sequences only compared to the expected proportion of night composition (Figure 6d; X2 = 22.114, df = 12, p = 0.0363, Fisher’s exact test: simulated p = 0.0485). In addition, the gestational period presented a smaller proportion of nights with interactive sequences compared to the expected proportion of nights with and without any interactive sequence (Figure 6b; X2 = 12.269, df = 4, p = 0.0155, Fisher’s exact test: p = 0.0083). Figure 7a shows the sequences by night pooled together by period. The Kruskal–Wallis test (Kruskal–Wallis chi-squared = 12.778, df = 4, p = 0.0124), followed by pairwise Wilcoxon comparisons, indicated that during the mating period there were significantly more sequences by recorder by night than in the gestational (BH adjusted p = 0.0097 Mate × Gest) and the initial parental care period (p = 0.0215 Mate x Pups). The mating period was not higher in vocal activity than the non-reproductive and the parturition period (p > 0.05). Despite the visual impressions of Figure 7b, the nights with sequences (any) did not differ significantly in their proportion of interactive sequences in the night across periods (K-W X2 = 8.1282, df = 4, p = 0.08699). The subsampled data with only the five common recorders presented no significant difference in the number of sequences by night between periods (K-W X2 = 7.4627, df = 4, p = 0.1134), but the proportion of interactive sequences by night was significantly greater in the initial parental care period than in the gestational period (K-W X2 = 11.318, df = 4, p = 0.0232; p = 0.0170 Gest × Pups). In the comparative Table 1 it is possible to see that, on average, the percentage of interactive sequences by night and the percentage of nights with interactive sequences was low in the non-reproductive and gestational periods. The average interactive sequences by night (by recorder in the full sample) is high in the mating and parental care periods, and particularly low in the gestational period. 3.2. Distance Estimation: How Close to Each Other They Interact? We were able to estimate the distance between participating maned wolves in 19 interactive roar-bark sequences. In four of those cases the sequence was registered with enough quality in three recorders (Figure 8: non-black bars). For those we could calculate three distances (A–B, A–C, B–C) instead of just one. Those four cases were consecutive interactive sequences from 01:19 a.m. to 02:05 a.m. of 19 July 2014 (one of them is the example in Figure 3). The estimated distance between animals decreases along these four interactions (1523 m, 885 m, 576 m, 326 m: larger of the three distances), suggesting animals walked to meet each other (approximately 0.43 m/s if only one moved). Including all relative distance values calculated (N = 27), the estimated distance between individuals was 449.32 ± 374.36 m (mean ± SD). The two smallest values were 11 m and 16 m and the largest 1523 m. The shortest distance interactions could be considered as animals together since the estimation is not precise and those distances allow visual contact among individuals. Note that distances calculated for the same interactive sequence varied in the values of estimated distance (e.g., 885 m for recorders A–B and 48 m for recorders B–C). This happens because the estimated distance is relative to the axis formed by the aligned recorders used for the calculation. Therefore, any of those distances may not be the true maximum linear distance between animals, although they should be close to the minimum. 3.3. Captivity Dataset: Who Is Interacting in Captivity? There was a total of 89 recorded roar-bark sequences at the captivity facility during the breeding season (CC-CBMM: four females, two males). Of those, 60.7% were solo sequences, 41 emitted by males, 10 by females, and 3 by unidentified maned wolves from areas outside the facility. A proportion of 39.3% of those were interactive sequences, 15.7% dyad sequences and the remaining 23.6% involving three or more animals. Thirteen out of fourteen dyad sequences involved a male and a female. The remaining dyad sequence was an interaction between mother and daughter (JU/SA, adults, not housed together). The random chance of any dyad sequence being composed of a male and a female was 53.3%. Nine dyad sequences were made by the same male/female (SH/JU, that were not housed together) and three by this male and another female (2 with FI and 1 with RO, neither of these females housed with SH). The female housed with SH (SA) was the least vocal animal, normally only roar-barking when many maned wolves were engaged in long distance calling. The only dyad sequence the other male (NE) made was with female FI, his enclosure mate. SH and JU were the most vocally active captive maned wolves, participating in 71 (39 solos/12 dyad sequences/20 3+ individuals) and 30 (3/10/17) sequences respectively. FI was the third most vocal animal, participating in 26 sequences (7/4/15). Considering sequences involving more than two animals (N = 21), 18 had the participation of both males (SH and NE) and two of just SH. Female JU was engaged in 17 out of those 21 sequences, female FI in 16, and in 12, both females were participating. The participant composition of the remaining sequences with more than two animals varied. Except for the JU/SA dyad sequence, all interactive sequences had the participation of both sexes. At the facility recorded outside the breeding season (ZOO-USMG: two females, two males) only 10 roar-bark sequences were recorded, three were solo sequences (two emitted by male GA and one by female LU). All dyad sequences (N = 4) involved a male and a female, two being made by GA/LU (enclosure mates), one by GA/BA, and one by GI/BA (enclosure mates). The chance that a random dyad sequence would involve a male and a female was 2/3 in this case. The remaining three sequences were made by GA/GI/BA. 3.4. Acoustic Parameters: How Do Participants of Interactions Differ? Total Duration was the only normally distributed parameter (Shapiro–Wilk W = 0.9923, p = 0.3748), and therefore not transformed, and the only whose model is in Table 2. The linear mixed model of the captivity roar-barks Total Duration revealed a significant effect of factor sex when controlling for individual and sequence (ANOVA: numDF = 1, denDF = 8, F = 5.8582, p = 0.0418). Males were longer than females (Table 2; Tukey contrasts: m-f Z = 2.483, single-step adjusted p = 0.015). In no other model was sex a significant factor, although there was a tendency in the model for InterQuartile Duration (ANOVA: numDF = 1, denDF = 8, F = 4.3681, p = 0.0700) in the same direction (males on average 0.030 s longer: untransformed values). The type of sequence was not a significant factor in any parameter model, except for the InterQuartile Duration (ANOVA: numDF = 2, denDF = 77, F = 5.1698, p = 0.0078). For this model, roar-barks of sequences with three or more participants were slightly longer (Tukey ‘3more’–’one’: Z = 2.808, p = 0.0136; ‘3more’–’two’: Z = 2.377, p = 0.0454; untransformed average difference of 0.020 s from ‘3more’ to ’one’). Individuals were responsible for a considerable variation in the time parameters (see Table 2 for Total Duration), but almost none in frequency parameters nor the factor sequence in any parameter (all estimates of those random factors are at least four orders of magnitude smaller than the intercept estimate; see Table 2 for Total Duration). In the natural habitat we could measure the acoustic parameters of 21 interactive roar-bark sequences in the recordings (from 160). Unfortunately, none of the four interactive sequences in the non-reproductive period had enough quality (signal-to-noise ratio) to be used. All other months have at least one sample. In the natural habitat recordings, all five acoustic parameters were significantly different between participants of interactive sequences in the paired Wilcoxon tests (Total Duration: V = 0; InterQuartile Duration: V = 0; Peak Frequency of the 2nd Band: V = 210; 1st Frequency Quartile of the 2nd Band: V = 231; 3rd Frequency Quartile of the 2nd Band: V = 231; all p < 0.0001). The comparison of roar-bark parameters of free-ranging animals with those parameters from captive females and males can be seen in Figure 9. 4. Discussion In this work we aimed to determine when and who engages in acoustic interactions in free-ranging maned wolves. Using passive acoustic monitoring we discovered that there were more interactive roar-bark sequences (two animals engaging in vocal exchange) during the mating and initial parental care periods. This finding is in accordance with our prediction that interactive roar-bark sequences mediate intrafamilial group long-distance communication. We also found that participants of these interactive sequences differed in the acoustic parameters of their roar-barks, including their duration, which in captivity differed between males (longer) and females. Additionally, we found that in captivity, almost all roar-bark dyad sequences (two participants) were between a male and a female and interactive sequences with three or more participants always included both sexes. Taken together, our evidence indicates that maned wolves in our free-ranging recordings that are acoustically interacting at long distances are most likely male-female dyads. Maned wolves usually remain several hundred meters from each other [14], and female estrus lasts only five days [26]. During estrus, long distance vocal exchanges could help receptive individuals to find each other to mate. Our findings suggest that when maned wolves interact vocally, they are far away from each other (mean of 450 m), which supports the function of roar-barks as means of individual location. The evolution of such a strategy can positively influence the maintenance of low-density populations subjected to the Allee effect (positive density dependence in mating probability [34,35]). Mate limitation reduces reproduction when males and females have difficulty finding each other, thus acoustic communication may be crucial for the conservation of the species in the wild. In addition to location, roar-barking whenever the partner vocalizes can serve as a mate guarding strategy. That is, if the first caller is announcing its receptivity, then the second animal could be announcing to third parties that the mate is taken and will be defended. Besides the mating period, we found many interactive sequences circa parturition and parental care periods. Interestingly, the only interactive sequence registered in a study in the ecological station of Itirapina (SP/Brazil) was on 19 June [36], in accordance with our findings. During these periods, the function of interactive roar-bark sequences could be communication among caregivers. The most reported direct paternal care involves bringing food to the lactating female or offspring [14,16,19,37]. The female’s response to the male calling may help him locate the den, as she may move the pups frequently from den to den [16,19]. When the pups are around 4–5 weeks old they start following the female [16], and therefore the male would still benefit from real-time clues of their positions. Otherwise, the female may be vocally soliciting care. Vocal exchanges between the mated pair may be a form of parental care negotiation or manipulation [38,39]. Future studies investigating which sex initiates these interactive sequences would help clarify this issue. Lastly, the simultaneous vocalizations, at any period, could be a verification that the other pair member is still alive and present in the territory. If positive, it serves as an announcement to neighbors and potential intruders, similar to one of the proposed functions of avian duet [40,41]. Maned wolves have been observed extending their range [14,37] or invading territories after the death/disappearance of one of the owners of the breeding pair [10,19]. Thus, the vocal presence of both individuals in a mated pair may reinforce their resource-holding potential to individuals looking to establish new territories or expand their boundaries. During the reproductive season (particularly the mating and parental care periods in our case) territorial defense may become more important, and therefore pairs would make more interactive sequences at this period, when (reproductive) stakes are highest. Territorial defense during this last period could also be considered an indirect form of parental care [42]. A small number of the interactive sequences in our natural habitat dataset may involve same sex individuals. Such events have been reported [10,19], especially in captivity [8,12], and suggest that roar-barks are multifunctional signals [6,7]. There are also reports of juveniles from the same home range participating in interactive vocal sequences [10]. Thus, those potential helpers would eventually create same sex vocal exchanges. Furthermore, juveniles probably have an underdeveloped vocal apparatus, which could bias our acoustic parameter analyses. Finally, we could only analyze the acoustic parameters of 21 interactive sequences (of 160 recorded), and none in the non-reproductive season, limiting our ability to conclude anything about sex discrimination in these vocal exchanges. Nevertheless, this is a general problem of studies in natural habitats, and we consider ours very extensive, in area and time recorded. Our captivity data supports the idea that female–male communication is the main function of long-distance vocal interactions. Except for a mother–daughter dyad sequence, all other dyad sequences (N = 17) were female–male dyads. We would expect many more female–female dyad sequences, as the random chance for that was around 37%. The chance for a male–male dyad sequence was lower (7.5%), but as males are more vocally active [11] we would still expect at least one interaction of this kind. It should be noted, though, that most dyad sequences (13/17) were not made by enclosure mates, which apparently does not support the intrapair communication hypothesis. However, one of the alleged benefits of such exchanges stems from relocating mates separated by long distances, and thus, enclosed in a limited area, the need for such an encounter facilitation mechanism is absent. Additionally, captive pairs are mated artificially, not involving natural sexual selection by the animals. Therefore, the lack of same-enclosure dyad sequences could also be the result of weak pair bonds. Another aspect to consider in captivity is that, contrary to our natural habitat findings, interactive sequences with more than two animals (N = 24) were more common than dyad sequences. As calling by one individual stimulates calling in other maned wolves [8,11,12], group sequences may be more common in artificially closer proximity situations. Curiously, sequences with more than two animals always involved both sexes. Hence, our data does not support the idea that maned wolves respond more to same sex individuals (as found by [8,12]). Instead, captive maned wolves could respond more to any opposite sex individual and the composition of interactive sequences may only reflect the individuals’ disposition to vocalize. This last characteristic could be a product of dominance status and/or individual differences (i.e., personalities). Larger vocal folds and vocal apparatus can produce sounds of longer wavelength [43] and more voluminous lungs may allow to sustain a loud vocalization longer [44]. The last is in agreement with our captivity findings that males produce longer roar-barks and are slightly larger (on average 2 kg and 2.5 cm bigger [14]). Contrary to what we had expected, though, the difference in frequency was not significant and more likely due to individual variance. Maybe the small sexual dimorphism in the species is not enough to produce detectable frequency discrepancies or that differences may only become significant with a larger sample size. This can also mean the difference in duration may be less related to anatomy and instead be related to motivation [11,45,46]. In the context of opposite sex communication, the motivation would be to advertise sexual quality and/or territorial holding potential (for joint territorial announcement). In any of those cases, female and male motivation would be the same and vocalizations would tend to be more similar in duration between sexes. As this is not the case, maybe males are more motivated to attract extra-pair mating or be subject to more intense competition/selection. More studies are needed to elucidate this matter, but it is important to reflect that if the differences are motivational, then it is possible that captivity and natural habitat roar-bark acoustic parameters trends should differ, as motivation is context dependent. This would be a caveat for our assumptions about sex discrimination in free-ranging animals. 5. Conclusions Here we started from passive acoustic recordings with no visual information and ended up with new information on real-time interactions between maned wolves, based on temporal distributions and differences in acoustic parameters in their long-distance calls. Despite being usually considered a solitary canid, a growing body of evidence, including the present work, is revealing how complex the long-range acoustic communication of maned wolves is and how we can use sequences of acoustic events to gain extensive information about cryptic species. Acknowledgments We thank Jean Pierre Santos and Marcello Montagno do Valle for equipment deployment and maintenance. We are also thankful for the Serra da Canastra National Park staff for their availability and support, and the UFRN Laboratory of Bioacoustics team for the expertise and intellectual contributions. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ani12091081/s1, Acoustic_parameters_dataset, Audio S1: interactive roar-bark sequence Casc_20140619_015530, Audio S2: interactive roar-bark sequence Bif_20160409_185414, Audio S3: interactive roar-bark sequence Rol_20160314_200004. Click here for additional data file. Author Contributions Conceptualization, L.S.F. and R.S.S.-L.; methodology, L.S.F., V.S., L.H.R. and R.S.S.-L.; formal analysis, L.S.F., V.S., T.A.P., E.N. and L.H.R.; investigation, L.S.F., V.S., T.A.P., E.N. and L.H.R.; resources, L.H.R., J.B., F.H.R. and R.S.S.-L.; writing—original draft preparation, L.S.F.; writing—review and editing, L.S.F., V.S., R.S.S.-L.; visualization, L.S.F.; supervision, J.B., F.H.R. and R.S.S.-L.; project administration, R.S.S.-L.; funding acquisition, L.H.R., J.B., F.H.R. and R.S.S.-L. All authors have read and agreed to the published version of the manuscript. Funding This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior—Brasil (CAPES)—Finance Code 001, by PhD CAPES scholarship 23001011 and PNPD/CAPES postdoc fellowship 88887.472468/2019-00 to L.S.F. This study was also supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) Edital Universal grant number 474000/2010-9 and CNPq Bolsa Produtividade em Pesquisa grant number 301665/2011-7 to F.H.R. and 312763/2019-0 to R.S.S.-L., by the Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG) through PPM Programme to F.H.R., by the Idea Wild, and the Graduate Programme in Psychobiology—UFRN. Institutional Review Board Statement Captive animals were not manipulated in any way and recordings in the natural habitat were made with autonomous equipment that precludes the need for researchers in the field, therefore our study is completely non-invasive. Wild maned wolf acoustic data collection at the Serra da Canastra National Park was authorized by Instituto Chico Mendes de Conservação da Biodiversidade (ICMBio; SISBIO license number 41329-2). Informed Consent Statement Not applicable. Data Availability Statement Raw tabulated data for the acoustic parameters is available in the Supplementary Materials. The audio files are available on request from the corresponding author. The data are not publicly available because the complete audio dataset is too large and not yet available in a public archive. Conflicts of Interest The authors declare no conflict of interest. Figure 1 The location of acoustic recorders at Serra da Canastra National Park in Minas Gerais state, Brazil. Microphone symbols indicate the position of the autonomous recorders (Wildlife Acoustics SongMeters SM2+). Recorders with no noted years were active in 2014 (April–July), 2016 (March–June) and 2016–2017 (December–January) periods. The blue line indicates the park boundaries. Image from Google Earth © 2019 TerraMetrics. Figure 2 Three examples of maned wolf interactive roar-bark sequences recorded passively at the Serra da Canastra National Park, MG/BR. Each was registered in a different recorder site, on 19 July 2014 at 1:55 a.m., 9 April 2016 at 19:21 p.m., and 14 March 2016 at 20:00 p.m., respectively from top to bottom. The audio files of those examples are in Supplementary Material Audio S1–S3. Figure 3 Interactive roar-bark sequence with two participating maned wolves (a and b). This sequence was registered in two recorders (I and II). Roar-barks (numbers) only align in recorder II one animal at a time (top and bottom) because individuals are at different positions. The spectrogram shown in recorder I (middle) is a portion of the first roar-bark sequence example in Figure 2. Figure 4 Roar-barks of captive maned wolves. The first box (blue) exemplified in SA shows the entire selection (200–2000 Hz), in which the Total Duration (“Dur” in the image) and InterQuartile Duration (“QQt”) are measured. The second box (green) exemplified in BA shows the selection of the second frequency band, in which the 1st Frequency Quartile (“1qF”), the Peak Frequency (“pkF”), and the 3rd Frequency Quartile (“3qF”) are measured. Spectrogram made in Raven Pro. 1.6. Figure 5 (a) Seasonal distribution of maned wolf roar-bark sequences recorded at the Serra da Canastra National Park, Brazil. Each point is a night (the first 3 h). Periods are separated by the grayscale in the dates (non-reproductive, mating, gestation, parturition, and initial parental care, consecutively). The number of sequences found was divided by the number of active recorders at that moment (4–13 Wildlife Acoustic SongMeters 2). (b) Percentage of sequences in which 2 maned wolves were participating. Figure 6 Mosaic plots showing the proportion of nights in each roar-bark activity category through time. Each mosaic tile area is proportional to the observed number of nights in each category (height) and the number of nights sampled in each time period (width). Time periods are as follows: NRep = non-reproductive; Mate = mating; Gest = gestation; Part = parturition; and Pups = initial parental care. Pearson residuals were used in a Chi-Square Test of Independence to verify if there is a difference among observed and expected values. Blue tiles have significantly greater frequency values than expected, dark grey tiles have greater frequency values than expected, but marginally significant, and red tiles have significantly smaller frequency values than expected. All other tiles show no significant difference among observed and expected values. (a) Proportions of nights with and without interactive roar-bark sequences using the entire dataset; (b) Proportions of nights with and without interactive roar-bark sequences using the subsampled dataset considering only the 5 recorders common to all periods; (c) Proportion of each category of vocal activity using the entire dataset; (d) Proportions of each vocal activity category using the subsampled dataset considering only the 5 recorders common to all periods. Categories of vocal activity are as follows: None = no vocal activity; Solo = only solo roar-bark sequences; Interact. = only interactive roar-bark sequences; and Both = solo and interactive sequences. Plots made in R with function mosaic {vcd}. Figure 7 (a) Number of maned wolf roar-bark sequences by recorder by night; and (b) Proportion of interactive sequences by night during each time period. Time periods are as follows: NRep = non-reproductive; Mate = mating; Gest = gestation; Part = parturition; and Pups = initial parental care. Boxes’ widths are proportional to the number of nights sampled. Figure 8 Estimated distance (in meters) between maned wolves exchanging roar-barks in sequences recorded simultaneously in 2 recorders. Non-black bars represent cases (4) where the same interactive sequence was recorded with enough quality in 3 recorders, and thus 3 different distances were calculated for each of them (using the time difference between recorders A × B, B × C, and A × C.). Figure 9 Acoustic parameters of maned wolf roar-barks analyzed separately for captive (recorded manually in 2 facilities at Minas Gerais state, Brazil) and for free ranging wolves (recorded passively in the Serra da Canastra National Park, MG/BR). In captivity the differences between females (red) and males (blue) were tested using a linear mixed model controlling for individual and sequence (males are longer, no other parameters significantly differ). The parameters of free ranging wolves were measured and paired within each dyad interactive roar-bark sequence (2 animals) using paired Wilcoxon tests. Individuals were only compared to the one they were interacting at the moment (participants differ significantly in all parameters). For visualization only, for each comparison the shortest in duration and highest in frequency of the dyad was set in the first box (purple) of the free ranging data and the other in the next box (lavender). (a) Total duration; (b) InterQuartile duration; (c) 1st frequency quartile of the 2nd band; (d) Peak frequency of the 2nd band; and (e) 3rd frequency quartile of the 2nd band. animals-12-01081-t001_Table 1 Table 1 Maned wolves’ roar-bark sequences recorded passively at the Serra da Canastra National Park. At the top is the complete dataset (all recorders) and at the bottom the subsample of the 5 common recording sites used in all periods (only nights in which all 5 were active). All Recorders Periods: Non-Reproductive Mating Gestation Parturition Initial Parental Care Nights recorded 37 53 82 57 31 Sequences 67 202 93 104 49 Interactive sequences 4 24 10 9 11 Average recorders 6.3 ± 1.2 12.4 ± 1.0 12.1 ± 1.2 11.8 ± 1.6 12 ± 0 Sequences by night by recorder 0.286 ± 0.522 0.303 ± 0.402 0.093 ± 0.120 0.154 ± 0.215 0.132 ± 0.362 Interactive sequences by night by recorder 0.014 ± 0.057 0.036 ± 0.087 0.011 ± 0.036 0.014 ± 0.044 0.030 ± 0.063 % of interactive sequences 2.4 ± 6.9% 14.8 ± 29.0% 10.5 ± 26.8% 14.1 ± 32.5% 33.1 ± 44.4% Nights with sequences 16 38 44 32 16 % of nights with sequences 43.2% 71.7% 53.7% 56.1% 51.6% Nights with interactive sequences 2 13 8 6 7 % of nights with interactive sequences 5.4% 24.5% 9.8% 10.5% 22.6% Five Recorders Periods: Non-Reproductive Mating Gestation Parturition Initial Parental Care Nights recorded 21 50 69 59 21 Sequences 46 76 36 64 11 Interactive sequences 4 12 1 10 6 Sequences by night 2.190 ± 3.669 1.520 ± 2.801 0.522 ± 1.001 1.085 ± 3.218 0.524 ± 0.814 Interactive sequences by night 0.190 ± 0.602 0.240 ± 0.916 0.014 ± 0.120 0.169 ± 0.530 0.286 ± 0.644 % of interactive sequences 4.2 ± 10.4% 14.5 ± 31.5% 1.2 ± 5.5% 21.0 ± 38.8% 45.8 ± 50.2% Nights with sequences 11 22 21 24 8 % of nights with sequences 52.4% 44.0% 30.4% 40.7% 38.1% Nights with interactive sequences 2 6 1 4 4 % of nights with interactive sequences 9.5% 12.0% 1.4% 6.8% 19.0% animals-12-01081-t002_Table 2 Table 2 Intervals of the linear mixed model of Total Duration of captive maned wolves roar-barks (N = 200). Values are shown in seconds. ‘Participants’ is the number of vocally interacting animals in the sequence from which the roar-bark was selected. There are 10 individuals, 6 females and 4 males. Total number of different sequences combined with individuals is 89. Total Duration ~Sex + Participants + (1|Individual/Sequence) Approximate 95% Confidence Intervals Lower Estimate Upper Intercept (Female, one) 0.463 0.530 0.597 Fixed factors Sex (males) 0.006 0.124 0.241 Participants (two) −0.023 0.008 0.038 Participants (3more) −0.011 0.017 0.044 Random factors Individual 0.047 0.078 0.129 Sequence in individual 1.69 × 10−201 1.78 × 10−07 1.87 × 10+187 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Silveira L. Ecologia e Conservação da Comunidade de Carnívoros do Parque Nacional das Emas, GO Master’s Thesis Universidade Federal de Goiás Goiânia, Brazil 1999 2. De Paula R.C. Adequabilidade Ambiental dos Biomas Brasileiros à Ocorrência do Lobo-Guará (Chrysocyon brachyurus) e Efeitos da Composição da Paisagem em sua Ecologia Espacial, Atividade e Movimentação Ph.D. Thesis Universidade de São Paulo São Paulo, Brazil 2016 10.11606/t.11.2016.tde-05072016-114911 3. Paula R.D. Medici P. Morato R.G. Plano de Ação para a Conservação do Lobo-Guará—Análise de Viabilidade Populacional e de Habitat CENAP/ICMBio Brasília, Brazil 2008 1 158 4. de Melo L.F.B. Sábato M.A.L. Magni E.M.V. Young R.J. Coelho C.M. Secret lives of maned wolves (Chrysocyon brachyurus Illiger 1815): As revealed by GPS tracking collars J. Zoöl. 2006 271 27 36 10.1111/j.1469-7998.2006.00176.x 5. Blumstein D.T. Mennill D.J. Clemins P. Girod L. Yao K. Patricelli G. Deppe J.L. Krakauer A.H. Clark C. Cortopassi K.A. Acoustic monitoring in terrestrial environments using microphone arrays: Applications, technological considerations and prospectus J. Appl. Ecol. 2011 48 758 767 10.1111/j.1365-2664.2011.01993.x 6. Rocha L.H.S. Ferreira L. Venticinque E.M. Rodrigues F.H.G. Sousa-Lima R.S. Temporal and environmental influences on long-distance calling by free-ranging maned wolves (Chrysocyon brachyurus) J. Mammal. 2016 98 302 311 10.1093/jmammal/gyw171 7. Ferreira L.S. Rocha L.H. Duarte D. Neto E. Baumgarten J.E. Rodrigues F.H. Sousa-Lima R.S. Temporal and spatial patterns of the long-range calls of maned wolves (Chrysocyon brachyurus) Mastozoología Neotropical 2020 27 081 095 10.31687/saremMN.20.27.1.0.24 8. Kleiman D.G. Social Behavior of the Maned Wolf (Chrysocyon brachyurus) and Bush Dog (Speothos venaticus): A Study in Contrast J. Mammal. 1972 53 791 806 10.2307/1379214 9. Ferreira L.S. Damo J.S. Sábato V. Baumgarten J.E. Rodrigues F.H.G. Sousa-Lima R. Using playbacks to monitor and investigate the behaviour of wild maned wolves Bioacoustics 2019 30 74 92 10.1080/09524622.2019.1691655 10. Emmons L.H. Maned Wolves of Noel Kempff Mercado National Park Smithson. Contrib. Zoöl. 2012 639 1 135 10.5479/si.00810282.639 11. Sábato V. Aspectos do Comportamento Acústico do Lobo-guará Chrysocyon Brachyurus (Illiger 1815) Master’s Thesis Universidade Federal de Minas Gerais Belo Horizonte, Brazil 2011 12. Brady C.A. The vocal repertoires of the bush dog (Speothos venaticus), crab-eating fox (Cerdocyon thous), and maned wolf (Chrysocyon brachyurus) Anim. Behav. 1981 29 649 669 10.1016/S0003-3472(81)80001-2 13. Azevedo F.C. Home Range and Spatial Organization of Maned Wolf (Chrysocyon brachyurus) in the Serra da Canastra National Park Region, Minas Gerais, Brazil Master’s Thesis Universidade Federal de Minas Gerais Belo Horizonte, Brazil 2008 14. Jácomo A.T.D.A. Kashivakura C.K. Ferro C. Furtado M.M. Astete S.P. Tôrres N.M. Sollmann R. Silveira L. Home Range and Spatial Organization of Maned Wolves in the Brazilian Grasslands J. Mammal. 2009 90 150 157 10.1644/07-MAMM-A-380.1 15. Rodden M. Rodrigues F. Bestelmeyer S. Maned wolf (Chrysocyon brachyurus) Canids: Foxes, Wolves, Jackals and Dogs, Status Survey and Conservation Action Plan 1st ed. IUCN Canid Specialist Group Cambridge, UK 2004 38 43 10.2307/j.ctv39x6vm.3 16. Bestelmeyer S.V. Solitary, Reproductive, and Parental Behavior of Maned Wolves (Chrysocyon brachyurus) Ph.D. Thesis Col-orado State University Fort Collins, CO, USA 2000 17. Balieiro F. Monticelli P.F. Evidence of individual discrimination in the maned wolf long-distance extended-bark Behav. Process. 2019 158 219 227 10.1016/j.beproc.2018.12.004 18. Bradbury J.W. Vehrencamp S.L. Principles of Animal Communication 1st ed. Sinauer Sunderland, MA, USA 1998 19. Dietz J.M. Ecology and social organization of the maned wolf (Chrysocyon brachyurus) Smithson. Contrib. Zoöl. 1984 392 1 51 10.5479/si.00810282.392 20. Ferreira L.M.S. The Call of the (Neotropical) Wild: Maned Wolf Long-Range Acoustic Ecology Ph.D. Thesis Universidade Federal do Rio Grande do Norte Natal, Brazil 2019 Available online: https://repositorio.ufrn.br/handle/123456789/27378 (accessed on 12 January 2022) 21. Moehlman P.D. Social Organization in Jackals: The complex social system of jackals allows the successful rearing of very dependent young Am. Sci. 1987 75 366 375 22. Queirolo D. Motta-Junior J.C. Prey availability and diet of maned wolf in Serra da Canastra National Park, southeastern Brazil Acta Thériol. 2007 52 391 402 10.1007/BF03194237 23. Paula R.C. Rodrigues F.H.G. Queirolo D. Jorge R.P.S. Lemos F.G. De Almeida Rodrigues L. Avaliação do risco de extinção do lobo guará Chrysocyon brachyurus (Illiger, 1815) no Brasil Biodiv. Br. 2013 1 146 159 24. Figueroa H. Extensible BioAcoustic Tool. XBAT R7 Cornell Bioacoustics Research Program Ithaca, NY, USA 2007 Available online: http://www.birds.cornell.edu/brp/software/xbatinstallation (accessed on 30 November 2016) 25. Rocha L.H. Ferreira L. Paula B.C. Rodrigues F.H.G. Sousa-Lima R.S. An evaluation of manual and automated methods for detecting sounds of maned wolves (Chrysocyon brachyurus Illiger 1815) Bioacoustics 2015 24 185 198 10.1080/09524622.2015.1019361 26. Carvalho C.T.D. Vasconcellos L.E. Disease, food and reproduction of the maned wolf: Chrysocyon brachyurus (Illiger) (Carnivora, Canidae) in southeast Brazil R. Bras. Zool. 1995 12 627 640 10.1590/S0101-81751995000300018 27. De Melo L.F.B. Sábato M.A.L. Magni E.M.V. Young R.J. Coelho C.M. First observations of nest attendance behavior by wild maned wolves, Chrysocyon brachyurus Zoo Biol. 2009 28 69 74 10.1002/zoo.20213 19358320 28. Charlton B.D. Reby D. Ellis W.A.H. Brumm J. Fitch W.T. Estimating the Active Space of Male Koala Bellows: Propagation of Cues to Size and Identity in a Eucalyptus Forest PLoS ONE 2012 7 e45420 10.1371/journal.pone.0045420 23028996 29. Matrosova V.A. Volodin I.A. Volodina E.V. Vasilieva N.A. Stability of acoustic individuality in the alarm calls of wild yellow ground squirrels Spermophilus fulvus and contrasting calls from trapped and free-ranging callers Naturwissenschaften 2010 97 707 715 10.1007/s00114-010-0686-7 20559613 30. Sábato V.R. Ferreira L.S. Baumgarten J. Rodrigues F.H. Sousa-Lima R.S. Assessing Identity and Sex Information on Calls from Rare and Solitary Species Laboratory of Bioacoustics, Universidade Federal do Rio Grande do Norte Natal, Brazil 2022 to be submitted 31. Charif R.A. Waack A.M. Strickman L.M. Raven Pro 1.4 User’s Manual Cornell Lab of Ornithology Ithaca, NY, USA 2010 Available online: http://www.birds.cornell.edu/raven (accessed on 13 January 2022) 32. Benjamini Y. Hochberg Y. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing J. R. Stat. Soc. Ser. B 1995 57 289 300 10.1111/j.2517-6161.1995.tb02031.x 33. Yeo I.-K. Johnson R. A new family of power transformations to improve normality or symmetry Biometrika 2000 87 954 959 10.1093/biomet/87.4.954 34. Allee W.C. Animal Aggregations Q. Rev. Biol. 1927 2 367 398 10.1086/394281 35. Allee W.C. Animal Aggregations: A Study in General Sociology University of Chicago Press Chicago, IL, USA 1931 431 36. Paula B. Monticelli P. Maned wolf duet? Record of two free-ranging maned wolves vocalizing simultaneous Acad. Lett. 2021 1720 10.20935/al1720 37. Rodrigues F.H.G. Biologia e Conservação do Lobo-guará na Estação Ecológica de Águas Emendadas, DF Ph.D. Thesis Universidade Estadual de Campinas Campinas, Brazil 2002 38. Wachtmeister C.-A. Display in monogamous pairs: A review of empirical data and evolutionary explanations Anim. Behav. 2001 61 861 868 10.1006/anbe.2001.1684 39. Baldassarre D.T. Greig E.I. Webster M.S. The couple that sings together stays together: Duetting, aggression and extra-pair paternity in a promiscuous bird species Biol. Lett. 2016 12 20151025 10.1098/rsbl.2015.1025 26911342 40. Koloff J. Mennill D.J. The Responses of Duetting Antbirds to Stereo Duet Playback Provide Support for the Joint Territory Defence Hypothesis Ethology 2013 119 462 471 10.1111/eth.12084 41. Diniz P. Macedo R.H. Webster M. Duetting correlates with territory quality and reproductive success in a suboscine bird with low extra-pair paternity Auk 2019 136 004 10.1093/auk/uky004 42. Kleiman D.G. Malcolm J.R. The Evolution of Male Parental Investment in Mammals Parental Care in Mammals Gubernick D.J. Klopfer P.H. Springer Boston, MA, USA 1981 347 387 43. Morton E.S. On the Occurrence and Significance of Motivation-Structural Rules in Some Bird and Mammal Sounds Am. Nat. 1977 111 855 869 10.1086/283219 44. Fitch W.T. Primate vocal production and its implications for auditory research Primate Audition: Ethology and Neurobiology Ghazanfar A. CRC Press Boca Raton, FL, USA 2002 87 108 45. Vehrencamp S.L. Handicap, index, and conventional signal elements of bird song Animal Signals: Signalling and Signal Design in Animal Communication Espmark Y. Amundsen T. Rosenqvist G. Tapir Academic Press Trondheim, Norway 2000 277 300 46. Linhart P. Slabbekoorn H. Fuchs R. The communicative significance of song frequency and song length in territorial chiffchaffs Behav. Ecol. 2012 23 1338 1347 10.1093/beheco/ars127

PMC009xxxxxx/PMC9099686.txt

==== Front Nutrients Nutrients nutrients Nutrients 2072-6643 MDPI 10.3390/nu14091971 nutrients-14-01971 Article Changes in Serum Levels of Ketone Bodies and Human Chorionic Gonadotropin during Pregnancy in Relation to the Neonatal Body Shape: A Retrospective Analysis Noshiro Kiwamu 1 https://orcid.org/0000-0001-9236-9945 Umazume Takeshi 1* Hattori Rifumi 2 Kataoka Soromon 3 Yamada Takashi 4 Watari Hidemichi 1 Tako Elad Academic Editor 1 Department of Obstetrics and Gynecology, Graduate School of Medicine, Hokkaido University, Sapporo 060-8638, Japan; ichigoichie_finalfantasy@hotmail.com (K.N.); watarih@med.hokudai.ac.jp (H.W.) 2 Department of Obstetrics and Gynecology, Obihiro-Kosei General Hospital, Obihiro 080-0024, Japan; rrrrrr4756@yahoo.co.jp 3 Department of Obstetrics and Gynecology, Hakodate Central General Hospital, Hakodate 040-8585, Japan; sorokata@hakochu-hp.gr.jp 4 Department of Obstetrics and Gynecology, Japan Community Health Care Organization Hokkaido Hospital, Sapporo 062-8618, Japan; yamatakashi@me.com * Correspondence: takeuma@med.hokudai.ac.jp; Tel.: +81-11-706-5941 09 5 2022 5 2022 14 9 197126 3 2022 05 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Among the physiological changes occurring during pregnancy, the benefits of morning sickness, which is likely mediated by human chorionic gonadotropin (HCG) and induces serum ketone production, are unclear. We investigated the relationship between serum levels of ketone bodies and HCG in the first, second, and third trimesters and neonatal body shape (i.e., birth weight, length, head circumference, and chest circumference) in 245 pregnant women. Serum levels of 3-hydroxybutyric acid peaked in late-stage compared with early stage pregnancy (27.8 [5.0–821] vs. 42.2 [5.0–1420] μmol/L, median [range], p < 0.001). However, serum levels of ketone bodies and HCG did not correlate with neonatal body shape. When weight loss during pregnancy was used as an index of morning sickness, a higher pre-pregnancy body mass index was associated with greater weight loss. This study is the first to show that serum ketone body levels are maximal in the third trimester of pregnancy. As the elevation of serum ketone bodies in the third trimester is a physiological change, high serum levels of ketone bodies may be beneficial for mothers and children. One of the possible biological benefits of morning sickness is the prevention of diseases that have an increased incidence due to weight gain during pregnancy. ketone body pregnancy morning sickness hyperemesis gravidarum human chorionic gonadotropin (HCG) Scientific Research from the Ministry of Education, Science, Sports, and Culture of JapanJP19K1865709 Japan Science and Technology AgencyJPMJPF2108 This study was supported by a Grant-in Aid for Scientific Research from the Ministry of Education, Science, Sports, and Culture of Japan (No. JP19K1865709), as well as Japan Science and Technology Agency (Grant Number JPMJPF2108). This work was not associated with any commercial entity. ==== Body pmc1. Introduction Approximately 80% of pregnant women develop morning sickness, with 0.5–0.8% becoming severe enough to necessitate hospitalization [1,2,3]. The incidence of severe morning sickness in East Asian women is approximately 3.6% [4], which is slightly higher than the incidence range (0.3–2.0%) in other races [4]. Various endocrine substances, including progesterone and human chorionic gonadotropin (HCG), have been investigated as causative agents of morning sickness; however, no definitive causative agent has been identified yet [5]. Physiological changes during pregnancy have some advantages for the mother and child: for example, anemia during pregnancy is the result of increased circulating blood volume and facilitates the prevention of blood clots. The higher the circulating plasma volume, the less likely gestational hypertension or intrauterine growth restriction will occur [6]. However, the physiological benefits of morning sickness are unknown, and severe cases can even be fatal for the mother. In patients with severe morning sickness, ketone bodies are detected on urinalysis. Ketone bodies can pass through the placenta and become a source of nutrition for the fetus [7]. In adults, ketones have been shown to be beneficial for the central nervous system and can affect many metabolic processes associated with aging and apoptosis. One of the ketone bodies, 3-hydroxybutyric acid, blocks NLR family pyrin domain containing 3 (NLRP3) inflammasome and attenuates caspase-1 and interleukin-1β secretion in mouse models and is believed to reduce hypoglycemia-related neuronal apoptosis, increase the number of motor neurons, increase neuronal activity and angiogenesis, and protect neuronal cell cultures from the development of amyloid pathology [8,9,10]. Ketone bodies pass through the placenta and are consumed by the brain, and thereby contribute to the growth of the fetus’ central nervous system [11]. We focused on the hypothesis that morning sickness begins at the same time as when the maternal serum ketone levels increase and neural tube formation occurs in the fetus. It has been reported that cell movement of the surrounding tissue (non-neural ectoderm), which does not become a neural tube, is also essential for neural tube formation [12], and it was thought that the effect of ketone bodies on neural tube formation also affects tissues other than the neural tube and may affect the fetal body shape. The HCG level increases in early stage pregnancy, and this increase is considered one of the causes of morning sickness [13]—a theory that is supported by the fact that the highest incidence of morning sickness is observed at peak HCG levels, as well as in twin pregnancies and hydatidiform moles, which induce high HCG levels [14]. The incidence of hyperemesis gravidarum in singleton pregnancies is reported as 1.4%, while that in twins is 2.7%, which is about twice as high [14]. Therefore, the current data in the literature suggests a relationship between morning sickness and high HCG levels; however, the role HCG plays in the pathogenesis of morning sickness remains unclear [5]. We hypothesized that an increase in the levels of maternal serum ketone bodies and HCG in early pregnancy is possibly beneficial for the fetus. Therefore, this study was conducted to investigate the biological benefits of morning sickness by comparing the levels of serum ketone bodies and HCG during early pregnancy with factors related to neonatal growth. 2. Materials and Methods This study was approved by the Institutional Review Board of the Hokkaido University Hospital (019-0390). All participants provided written informed consent prior to their participation in the study. The study enrolled 379 pregnant women who were scheduled to give birth in Obihiro-Kosei General Hospital, Hakodate Central General Hospital, or the Japan Community Health Care Organization Hokkaido Hospital between October 2018 and April 2019. Physicians at the study sites selected participants based on the following inclusion criteria: (1) those aged 18 years or above, (2) those scheduled to give birth at participating institutions, and (3) those who agreed to participate in the study. The exclusion criteria were as follows: (1) those with complications from earlier pregnancy and (2) those with a gestational age of 14 weeks or above. Blood samples were collected at the time of regular blood tests in the first (8–12 gestational weeks), second (24–27 gestational weeks), and third trimesters (35–37 gestational weeks), and the serum levels of ketone bodies (total ketones, acetoacetic acid, and 3-hydroxybutyric acid) and HCG were quantified. In addition, the participant’s age, height, pre-pregnancy weight, pre-pregnancy body mass index (BMI), weight at blood sampling, weight at delivery, BMI at delivery, pregnancy and delivery history, weeks of delivery, delivery pattern, placental weight, and neonatal findings (e.g., sex, birth weight, birth length, birth head circumference, birth chest circumference) were recorded from the medical records. The relationship of morning sickness with serum levels of ketone bodies and HCG was retrospectively examined. In the analysis phase, pregnant women who delivered after 36 weeks of gestation were included, whereas patients with twin pregnancies, thyroid disease, preeclampsia, kidney disease, and defect data were excluded. 2.1. Biochemical Procedures Serum was stored at −80 °C until assays were conducted for the following four blood variables: total ketone bodies, acetoacetic acid, 3-hydroxybutyric acid, and HCG levels were measured using enzyme linked immune sorbent assay kits TKB-L (KAINOS, Tokyo, Japan), 3HB-L (KAINOS, Tokyo, Japan), and II HCG (TOSOH, Tokyo, Japan), respectively. 2.2. Statistical Methods Statistical analyses were performed using the JMP Pro14© statistical software package (SAS, Cary, NC, USA). Changes in variables within a group were compared using the t-test and the Tukey–Kramer method. Single regression analysis was used to investigate the relationship between ketone bodies and perinatal prognosis. In all analyses, statistical significance was set at p < 0.05. 3. Results Among the 379 pregnant women who participated in the study, 197, 94, and 88 were screened from Obihiro-Kosei General Hospital, Hakodate Central General Hospital, and the Japan Community Health Care Organization Hokkaido Hospital, respectively. However, after screening, 105 eligible women who missed blood sampling during the first, second, or third trimester (including preterm birth at <36 weeks of gestation), 6 women with twin pregnancies, 15 with thyroid disease, 4 with preeclampsia, 2 with kidneys disease, and 2 with missing data in the medical records were excluded; thus, a total of 245 women, including 11 with gestational diabetes, were included in the final analysis dataset (Figure 1). 3.1. Demographic Characteristics Among the 245 participants, 85 were primiparas. The mean (standard deviation) age and gestational period at delivery were 31.9 ± 5.0 years and 39.1 ± 1.2 gestational weeks, respectively. The timing of blood sampling was 10.2 ± 1.5, 25.8 ± 1.3, and 36.2 ± 0.9 gestational weeks for the first, second, and third trimesters, respectively. There were 13 deliveries at 36 gestational weeks, 191 transvaginal deliveries, and 54 cesarean deliveries (Table 1). 3.2. Hematological Parameters during Pregnancy The median 3-hydroxybutyric acid levels were 27.8, 21.2, and 42.2 μmol /L in the first, second, and third trimesters, respectively. The median concentrations of acetoacetic acid, which reflects the ketone body index for a longer period than the 3-hydroxybutyric acid, were 18.6, 19.9, and 24.1 μmol/L in the first, second, and third trimesters, respectively. The median HCG levels were 132,000, 16,400, and 22,300 IU/L for the first, second, and third trimesters, respectively. The 3-hydroxybutyric acid concentration decreased in the second trimester, but the acetoacetic acid levels showed a tendency to gradually increase from the first to the third trimester. As a 50 g glucose challenge test (GCT) is conducted at the time of regular blood sampling in the second trimester, it was considered that the blood sampling at 60 min after consuming 50 g glucose affected the levels of total ketone bodies and decreased the 3-hydroxybutyric acid level. However, as acetoacetic acid has a long half-life, it is considered to have been less affected by the glucose challenge test (Table 2). 3.3. Effect of Glucose Challenge Test on Ketone Body Concentrations The 50 g GCT in the second trimester was not conducted in the Obihiro-Kosei General Hospital, and the random blood glucose level was used to screen for gestational diabetes. We compared the blood parameters of participants in the second trimester at the Obihiro-Kosei General Hospital (without GCT) and the other participants (with GCT). The median 3-hydroxybutyric acid level was significantly higher among the participants in the without GCT group than among those in the with GCT group (26.0 vs. 16.3 µmol/L, p < 0.001) (Table 3). Therefore, when the change in the median 3-hydroxybutyric acid level during pregnancy was confirmed only in the without GCT group, that is, where there was no glucose intake prior to blood sampling, the level gradually increased in the order of the first, second, and third trimester (25.4, 27.1, and 41.7 µmol/L, respectively; Figure 2). The 3-hydroxybutyric acid level during pregnancy in the without glucose challenge test group gradually increased in the order of the first, second, and third trimester. 3.4. Changes in HCG during Pregnancy The serum HCG level was high in the first trimester, peaked at 8 weeks of gestation, and decreased gradually, but did not change considerably from the second to the third trimester (Figure 3 and Table S1). The serum HCG level was high in the first trimester, peaked at 8 weeks of gestation, and decreased gradually 3.5. Relationship between Ketone Body and HCG in Early Pregnancy and the Neonatal Body Shape We examined the correlations between the neonatal body shape and the serum levels of 3-hydroxybutyric acid, acetoacetic acid, and HCG in the first trimester, but found no significant correlation with the birth weight, birth length, birth head circumference, or birth chest circumference (Figure 4). There were no correlations between the neonatal body shape and the serum levels of 3-hydroxybutyric acid, acetoacetic acid, and HCG in the first trimester. 3.6. Relationship between Placental Weight and Other Factors No correlation was found between placental weight and the serum levels of 3-hydroxybutyric acid or HCG. There was a positive correlation between the placental weight and birth weight (Figure 5). 3.7. Gestational Characteristics of the Weight-Loss and Non-Weight-Loss Groups For an evaluation of the severity of morning sickness, we compared whether there was any weight loss during pregnancy: 88 women lost weight, whereas 157 did not. In the weight-loss group, the levels of total ketone bodies, 3-hydroxybutyric acid, and acetoacetic acid in the first trimester were significantly higher than the levels in the non-weight-loss group; however, there was no significant intergroup difference in the perinatal prognosis. Although the weight loss group had a significantly high pre-pregnancy BMI, there was no significant intergroup difference in the BMI at delivery (Table 4). 4. Discussion This study showed that: (1) serum levels of ketone bodies gradually increased during pregnancy; (2) the serum levels of ketone bodies and HCG in the first trimester did not affect the neonatal body shape; and (3) women who lose weight in early pregnancy have a higher pre-pregnancy BMI, which may lead to physiological benefits from morning sickness that could prevent complications. To our knowledge, this is the first report to show that the serum levels of ketone bodies gradually increased during pregnancy. The abovementioned trend was particularly evident in institutions that did not perform the GCT during the second trimester. There was a significant difference in the serum levels of ketone bodies between institutions that performed the GCT in the second trimester and those that did not, suggesting that sugar intake affected the results of the blood tests (Table 3). Ketone bodies pass through the placenta and are consumed by the brain, thereby contributing to the growth of the fetal central nervous system [11]. Furthermore, ketone bodies act protectively on brain cells in Alzheimer’s and Parkinson’s diseases [15]. Moreover, a ketogenic diet may reduce the frequency of epileptic seizures [16,17]. In contrast, a study showed that an elevated 3-hydroxybutyric acid level in the third trimester adversely influenced the intelligence of the child [18]. In our study, an increase in the serum levels of ketone bodies in the third trimester of pregnancy was observed as a physiological change, which suggested that a high level of ketone bodies is unlikely to adversely affect the fetus. In this study, we examined the hypothesis that ketone bodies, which are by-products of morning sickness, and, HCG, which is a possible cause of morning sickness, may confer benefits on fetal development. However, there was no correlation between the serum levels of ketone bodies or HCG in the first trimester of pregnancy, when morning sickness is most severe, and the body shape, as ascertained by the birth weight, length, head circumference, and chest circumference of the newborn (Figure 4). We inferred that the higher the level of ketone bodies in early pregnancy, the larger is the fetal brain. By focusing on the neuroprotective effects of ketone bodies [14,15,16], we predicted that a higher ketone body concentration in early pregnancy, the greater would be the head circumference of the neonate; however, we found no such relationship. Nonetheless, birth head circumference only may be insufficient as an indicator of brain growth, but it is difficult to assess brain function at this timepoint. HCG is produced by the villous synctiotrophoblast cell layer [19] and, therefore, we expected a correlation between the HCG concentration and the placental weight, but there was no significant relationship. There was a positive correlation between placental weight and birth weight, which was previously reported [20,21] (Figure 5). HCG levels peaked at 8 weeks of gestation, which is consistent with reports from international studies [22] and shows that the trend of HCG levels is the same in Japanese women as in Western women. We focused on weight loss during pregnancy as a method for evaluating the severity of morning sickness. On comparing the group that had weight loss and the group that did not, we found that the group with weight loss had significantly higher serum 3-hydroxybutyric acid concentrations in the first trimester. The pre-pregnancy BMI was slightly higher in the group with weight loss during pregnancy than in the non-weight-loss group (Table 4). This result was consistent with previous studies, which showed that a higher pre-pregnancy BMI increases the risk of severe morning sickness [23]. However, this finding was contradicted by a study that only included pregnant women in East Asia and found that women with a lower pre-pregnancy BMI were more likely to experience hyperemesis gravidarum [4]. Considering the biological benefits, it may be logical for pregnant women with a high pre-pregnancy BMI to lose weight during pregnancy. The higher the pre-pregnancy BMI, the greater the risk of gestational diabetes and preeclampsia [24,25]. Furthermore, it is well known that greater weight gain during pregnancy is associated with a higher likelihood of developing gestational diabetes and preeclampsia [24,26,27]. One of the benefits of morning sickness may be to reduce weight gain during pregnancy. In future studies, it may be necessary to distinguish whether there are physiological reasons for the weight loss mediated by morning sickness. There are several limitations of this study. First, weight loss during pregnancy was used as an evaluation method for morning sickness. However, weight loss information is not an adequate indicator of morning sickness. Second, we also excluded cases of preterm delivery, preeclampsia, and several diseases; therefore, the relationship between initial ketone body levels and these diseases remains unclear. Third, a research design to explore the function aspect would be desirable, but it would require monitoring of the growth of newborns and changes in neurological development, which was difficult considering the short study period. Since formation of the neural tube requires cell movement in the surrounding tissues that do not form the neural tube [12], we thought that ketone bodies affect the formation of not only the neural tube but also the whole fetus, and thus we performed this study. However, assessment of neonatal morphology is insufficient as a substitute for assessment of function, and further studies with sufficient time are needed. Of note, it is possible to infer the effects of ketone bodies during pregnancy on the development of the child’s brain, and further research is needed to clarify the relationship. 5. Conclusions During pregnancy, serum levels of ketone bodies increased from early to late pregnancy and peaked in late pregnancy. As the elevation of serum levels of ketone bodies in the third trimester is a physiological change, high serum levels of ketone bodies may be beneficial for both the mother and the child. However, serum levels of ketone bodies and HCG in early pregnancy had no effect on the neonatal body shape. One of the biological benefits of morning sickness in early pregnancy was the prevention of diseases that possibly increase in frequency due to excessive weight gain during pregnancy. This study did not examine the effects of maternal ketone bodies on fetal neurodevelopment, and further studies are warranted. Acknowledgments We thank the doctors, nurses, and clerks of Obihiro-Kosei General Hospital, Hakodate Central General Hospital, and Japan Community Health Care Organization Hokkaido Hospital for supporting this study. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/nu14091971/s1, Table S1. Human chorionic gonadotropin data of 245 pregnant women. Click here for additional data file. Author Contributions All authors, K.N., T.U., R.H., S.K., T.Y. and H.W., participated in the design of the study and collected data on each pregnant woman who participated in this study. K.N. and T.U. participated in the design of the study, performed statistical analyses, and drafted the manuscript. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Hokkaido University Hospital (019-0390). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Study design. Figure 2 Median 3-hydroxybutyric acid levels in the group without the glucose challenge test during pregnancy. Figure 3 Serum human chorionic gonadotropin levels during pregnancy. Figure 4 Relationship between serum levels of ketone bodies or human chorionic gonadotropin in the first trimester with the neonatal body shape. Figure 5 Correlation of placental weight with serum levels of human chorionic gonadotropin and 3-hydroxybutyric acid, and birth weight. nutrients-14-01971-t001_Table 1 Table 1 Baseline characteristics of 245 pregnant women. Nulliparous women 85, 35% Age, years 31.9 (5.0) Height, m 1.58 (0.05) Pre-pregnancy weight, kg 53.1 (10.3) Pre-pregnancy body mass index, kg/m2 21.3 (3.5) Weight gain in pregnancy, kg 11.2 (3.5) Gestational week at delivery, week 39.1 (1.2) Preterm delivery at 36 gestational weeks 13 (5.3%) Vaginal delivery 191 (78%) Cesarean delivery 54 (22%) Infant sex   Male 125, 51% Birth weight, kg 3.1 (0.4) Birth length, cm 49.4 (1.8) Birth head circumference, cm 33.4 (1.3) Birth chest circumference, cm 32.2 (1.6) Timing of the tests   First trimester, week 10.2 (1.5)   Second trimester, week 25.8 (1.3)   Third trimester, week 36.2 (0.9) Data are presented as the means (standard deviation). nutrients-14-01971-t002_Table 2 Table 2 Hematological data of 245 pregnant women. 1st Trimester 2nd Trimester 3rd Trimester p-Value Clinical data   Maternal body weight, kg 53.8 (10.4) 48.6 (9.8) 62.6 (10.1) <0.001   Weight gain during pregnancy, kg 0.7 (2.4) 5.6 (3.2) 9.5 (3.9) <0.001   Weight gain, % 1.4 (4.5) 11.1 (6.6) 18.8 (8.4) <0.001   Body mass index, kg/m2 21.6 (3.6) 23.5 (3.3) 25.0 (3.4) <0.001   Gestational age, week 10.2 (1.5) 25.8 (1.3) 36.2 (0.9) <0.001 Hematological data   Total ketones, µmol/L 39.3 (5.0–934) 32.8 (5.0–812) 58.8 (5.0–1460) <0.001   3-hydroxybutyric acid, µmol/L 27.8 (5.0–821) 21.2 (5.0–690) 42.2 (5.0–1420) <0.001   Acetoacetic acid, µmol/L 18.6 (5.0–140) 19.9 (5.0–146) 24.1 (5.0–161) 0.083   Human chorionic gonadotropin, IU/L 132,000 (2620–341,000) 16,400 (166–216,000) 22,300 (31.9–113,000) <0.001 Clinical data are presented as mean (standard deviation). Hematological data are presented as median (range). Data below the measured value were set to 5. nutrients-14-01971-t003_Table 3 Table 3 Comparison of hematological parameters in the second trimester in the groups with and without a glucose challenge test. without GCT with GCT p-Value Clinical Data   Number of women 105 140   Age, years 31.8 (4.7) 32.2 (5.2) 0.196   Height, m 157.2 (0.5) 158.3 (0.5) 0.115   Pre-pregnancy weight, kg 52.3 (9.8) 53.7 (10.8) 0.277   Body weight in the second trimester, kg 58.5 (9.5) 58.9 (10.2) 0.732   Body mass index, kg/m2 23.6 (3.1) 23.5 (3.5) 0.758   Weight gain in the second trimester, kg 6.1 (3.2) 5.1 (3.1) 0.014   Weight gain, % 12.3 (6.9) 10.2 (6.3) 0.012   Gestational age, weeks 26.8 (0.6) 25.0 (1.1) <0.001 Hematological data   Total ketones, µmol/L 46.7 (12.3–812) 18.9 (5–296) <0.001   3-hydroxybutyric acid, µmol/L 26.0 (5–690) 16.3 (5–271) <0.001   Acetoacetic acid, µmol/L 19.3 (5–146) 5.0 (5–25.3) <0.001   Human chorionic gonadotropin, IU/L 16,300 (1890–92,900) 16,550 (166–216,000) 0.191 Clinical data are presented as mean (standard deviation). Hematological data are presented as median (range). Data below the measured value were set to 5. GCT, glucose challenge test. nutrients-14-01971-t004_Table 4 Table 4 Characteristics of the weight-loss and non-weight-loss groups during pregnancy. Weight Loss during Pregnancy Non-Weight Loss p-Value Clinical data   Number of women 88 157   Height, m 1.58 (0.05) 1.58 (0.06) 0.345   Pre-pregnancy body weight, kg 54.8 (11.8) 52.2 (9.3) 0.058   Pre-pregnancy body mass index, kg/m2 22.0 (4.2) 20.9 (3.1) 0.025   Minimum body weight, kg 53.0 (11.4) 52.2 (9.3) 0.537   Weight loss, kg 1.78 (2.0) not available.   Weight gain during pregnancy, % 9.8 (2.9) 12.1 (3.6) <0.001   Body weight at delivery, kg 62.8 (11.1) 64.3 (9.8) 0.274   Body mass index at delivery, kg/m2 25.2 (3.8) 25.8 (3.2) 0.193   Gestational age at delivery, week 39.2 (1.1) 39.0 (1.2) 0.362   Birth length, cm 49.7 (1.6) 49.3 (1.9) 0.091   Birth weight, kg 3.08 (0.35) 3.08 (0.41) 0.981   Birth head circumstance, cm 33.6 (1.3) 33.3 (1.3) 0.114   Birth chest circumstance, cm 32.3 (1.5) 32.2 (1.7) 0.397 Hematological data at first trimester   Total ketones, µmol/L 46.0 (5–934) 34.0 (5–541) <0.001   3-hydroxybutyric acid, µmol/L 33.0 (5–821) 23.4 (5–539) <0.001   Acetoacetic acid, µmol/L 5.0 (5–140) 5 (5–113) 0.046   Human chorionic gonadotropin, IU/L 137,000 (4640–341,000) 129,000 (2620–284,000) 0.072 Clinical data are presented as mean (standard deviation). Hematological data are presented as median (range). Data below the measured value were set to 5. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Quinla J.D. Hill D.A. Nausea and vomiting of pregnancy Am. Fam. Phys. 2003 68 121 128 2. Bailit J.L. Hyperemesis gravidarium: Epidemiologic findings from a large cohort Am. J. Obstet. Gynecol. 2005 193 811 814 10.1016/j.ajog.2005.02.132 16150279 3. Fell D.B. Dodds L. Joseph K.S. Allen V.M. Butler B. Risk factors for hyperemesis gravidarum requiring hospital admission during pregnancy Obstet. Gynecol. 2006 107 277 284 10.1097/01.AOG.0000195059.82029.74 16449112 4. Matsuo K. Ushioda N. Nagamatsu M. Kimura T. Hyperemesis gravidarum in eastern Asian population Gynecol. Obstet. Investig. 2007 64 213 216 10.1159/000106493 17664884 5. Verberg M.F.G. Gillott D.J. Al-Fardan N. Grudzinskas J.G. Hyperemesis gravidarum, a literature review Hum. Reprod. Update 2005 11 527 539 10.1093/humupd/dmi021 16006438 6. De Haas S. Ghossein-Doha C. Van Kuijk S.M.J. Van Drongelen J. Spaanderman M.E.A. Physiological adaption of maternal plasma volume during pregnancy: A systematic review and meta-analysis Ultrasound Obstet. Gynecol. 2017 49 177 187 10.1002/uog.17360 28169502 7. Robertson A.F. Karp W.B. Placental transport of nutrients South. Med. J. 1976 69 1358 1362 10.1097/00007611-197610000-00031 982117 8. Sussman D. Germann J. Henkelman M. Gestational ketogenic diet programs brain structure and susceptibility to depression & anxiety in the adult mouse offspring Brain Behav. 2015 5 e00300 10.1002/brb3.300 25642385 9. Van der Auwera I. Wera S. Van Leuven F. Henderson S.T. A ketogenic diet reduces amyloid beta 40 and 42 in a mouse model of alzheimer’s disease Nutr. Metab. 2005 2 28 10.1186/1743-7075-2-28 10. Youm Y.H. Nguyen K.Y. Grant R.W. Goldberg E.L. Bodogai M. Kim D. D’agostino D. Planavsky N. Lupfer C. Kanneganti T.D. Ketone body β-hydroxybutyrate blocks the NLRP3 inflammasome-mediated inflammatory disease Nat. Med. 2015 21 263 269 10.1038/nm.3804 25686106 11. Bronisz A. Ozorowski M. Hagner-Derengowska M. Pregnancy ketonemia and development of the fetal central nervous system Int. J. Endocrinol. 2018 2018 1232901 10.1155/2018/1242901 29971100 12. Morita H. Kajiura-Kobayashi H. Takagi C. Yamamoto T.S. Nonaka S. Ueno N. Cell movements of the deep layer of non-neural ectoderm underlie complete neural tube closure in xenopus Development 2012 139 1417 1426 10.1242/dev.073239 22378637 13. Tamay A.G. Kuşçu N.K. Hyperemesis gravidarum: Current aspect J. Obstet. Gynaecol. 2011 31 708 712 10.3109/01443615.2011.611918 22085059 14. Dypvik J. Pereira A.L. Tanbo T.G. Eskild A. Maternal human chorionic gonadotrophin concentrations in very early pregnancy and risk of hyperemesis gravidarum: A retrospective cohort study of 4372 pregnancies after in vitro fertilization Eur. J. Obstet. Gynecol. Reprod. Biol. 2018 221 12 16 10.1016/j.ejogrb.2017.12.015 29227846 15. Kim D.Y. Davis L.M. Sullivan P.G. Maalouf M. Simeone T.A. van Brederode J. Rho J.M. Ketone bodies are protective against oxidative stress in neocortical neurons J. Neurochem. 2007 101 1316 1326 10.1111/j.1471-4159.2007.04483.x 17403035 16. van der Louw E.J. Williams T.J. Henry-Barron B.J. Olieman J.F. Duvekot J.J. Vermeulen M.J. Bannink N. Williams M. Neuteboom R.F. Kossoff E.H. Ketogenic diet therapy for epilepsy furing pregnancy: A case series Seizure 2017 45 198 201 10.1016/j.seizure.2016.12.019 28110175 17. Simeone T.A. Simeone K.A. Stafstrom C.E. Rho J.M. Do ketone bodies mediate the anti-seizure effects of the ketogenic diet? Neuropharmacology 2018 133 233 241 10.1016/j.neuropharm.2018.01.011 29325899 18. Rizzo T. Metzger B.E. Burns W.J. Burns K. Correlation between antepartum maternal metabolism and intelligence of offspring N. Engl. J. Med. 1991 325 911 916 10.1056/NEJM199109263251303 1881416 19. Cole L.A. Biological functions of hCG and hCG-related molecules Reprod. Biol. Endocrinol. 2010 8 102 10.1186/1477-7827-8-102 20735820 20. Nascente L.M.D.P. Grandi C. Aragon D.C. Cardoso V.C. Plcental measurements and their association with birth weight in a brazilian cohort Rev. Bras. Epidemiol. 2020 23 e200004 10.1590/1980-549720200004 21. Salafia C.M. Zhang J. Charles A.K. Bresnahan M. Shrout P. Sun W. Maas E.M. Placental characteristics and birthweight Paediatr. Perinat. Epidemiol. 2008 22 229 239 10.1111/j.1365-3016.2008.00935.x 18426518 22. Korevaar T.I.M. Steegers E.A. de Rijke Y.B. Schalekamp-Timmermans S. Visser W.E. Hofman A. Jaddoe V.W. Tiemeier H. Visser T.J. Medici M. Reference ranges and determinants of total hCG levels furing pregnancy: The generation R study Eur. J. Epidemiol. 2015 30 1057 1066 10.1007/s10654-015-0039-0 25963653 23. Nurmi M. Rautava P. Gissler M. Vahlberg T. Polo-Kantola P. Incidence and risk factors of hyperemesis gravidarum: A national register-based study in Finland, 2005–2017 Acta Obstet. Gynecol. Scand. 2020 99 1003 1013 10.1111/aogs.13820 32030718 24. Bouvier D. Forest J.C. Dion-Buteau E. Bernard N. Bujold E. Pereira B. Giguère Y. Association of maternal weight and gestational weight gain with maternal and neonate outcomes: A prospective cohort study J. Clin. Med. 2019 8 2074 10.3390/jcm8122074 25. Dude A.M. Kominiarek M.A. Haas D.M. Iams J. Mercer B.M. Parry S. Reddy U.M. Saade G. Silver R.M. Simhan H. Weight gain in early, mid, and late pregnancy and hypertensive disorder of pregnancy Pregnancy Hypertens. 2020 20 50 55 10.1016/j.preghy.2020.03.001 32172170 26. Goldstein R.F. Abell S.K. Ranasinha S. Misso M. Boyle J.A. Black M.H. Li N. Hu G. Corrado F. Rode L. Association of gestational weight gain with maternal and infant outcomes: A systematic review and meta-analysis JAMA 2017 317 2207 2225 10.1001/jama.2017.3635 28586887 27. Hung T.-H. Hsieh T.T. Pregestational body mass index, gestational weight gain, and risks for adverse pregnancy outcomes among taiwanese women: A retrospective cohort study Taiwan J. Obstet. Gynecol. 2016 55 575 581 10.1016/j.tjog.2016.06.016 27590385

PMC009xxxxxx/PMC9099687.txt

==== Front Int J Mol Sci Int J Mol Sci ijms International Journal of Molecular Sciences 1422-0067 MDPI 10.3390/ijms23094846 ijms-23-04846 Article Molecular Cloning and Characterization of MbMYB108, a Malus baccata MYB Transcription Factor Gene, with Functions in Tolerance to Cold and Drought Stress in Transgenic Arabidopsis thaliana Yao Chunya 1 Li Wenhui 1 Liang Xiaoqi 1 Ren Chuankun 1 Liu Wanda 2 Yang Guohui 1 Zhao Mengfei 1 Yang Tianyu 1 Li Xingguo 1* https://orcid.org/0000-0002-0132-3018 Han Deguo 1* Moon Yong-Hwan Academic Editor Ruiz Lozano Juan Manuel Academic Editor 1 Key Laboratory of Biology and Genetic Improvement of Horticultural Crops (Northeast Region), Ministry of Agriculture and Rural Affairs/National-Local Joint Engineering Research Center for Development and Utilization of Small Fruits in Cold Regions/College of Horticulture & Landscape Architecture, Northeast Agricultural University, Harbin 150030, China; 13753449701@163.com (C.Y.); lwh_neau@126.com (W.L.); a13989332297@163.com (X.L.); ren473475964@163.com (C.R.); yangguohui_neau@126.com (G.Y.); z1991689991@163.com (M.Z.); yty12980@163.com (T.Y.) 2 Horticulture Branch of Heilongjiang Academy of Agricultural Sciences, Harbin 150040, China; haaslwd@126.com * Correspondence: xingguoli@neau.edu.cn (X.L.); deguohan@neau.edu.cn (D.H.); Tel.: +86-451-55190781 (D.H.) 27 4 2022 5 2022 23 9 484629 3 2022 25 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). The MYB transcription factor (TF) family is one of the largest transcription families in plants, which is widely involved in the responses of plants to biotic and abiotic stresses, as well as plant growth, development, and metabolic regulation. In the present study, a new MYB TF gene, MbMYB108, from Malus baccata (L.) Borkh, was identified and characterized. The open reading frame (ORF) of MbMYB108 was found to be 903 bp, encoding 300 amino acids. Sequence alignment results and predictions of the protein structure indicated that the MbMYB108 protein contained the conserved MYB domain. Subcellular localization showed that MbMYB108 was localized to the nucleus. The expression of MbMYB108 was enriched in young and mature leaves, and was highly affected by cold and drought treatments in M. baccata seedlings. When MbMYB108 was introduced into Arabidopsis thaliana, it greatly increased the cold and drought tolerances in the transgenic plant. Increased expression of MbMYB108 in transgenic A. thaliana also resulted in higher activities of peroxidase (POD) and catalase (CAT), higher contents of proline and chlorophyll, while malondialdehyde (MDA) content and relative conductivity were lower, especially in response to cold and drought stresses. Therefore, these results suggest that MbMYB108 probably plays an important role in the response to cold and drought stresses in A. thaliana by enhancing the scavenging capability for reactive oxygen species (ROS). Malus baccata (L.) Borkh MbMYB108 cold stress drought stress the National Natural Science Foundation of China32172521 the Postdoctoral Scientific Research Development Fund of Heilongjiang ProvinceLBH-Q16020 the Natural Science Fund Joint Guidance Project of Heilongjiang ProvinceLH2019C031 the SIPT Program for Undergraduates of Northeast Agricultural UniversityThis work was supported by the National Natural Science Foundation of China (32172521), the Postdoctoral Scientific Research Development Fund of Heilongjiang Province, China (LBH-Q16020), the Natural Science Fund Joint Guidance Project of Heilongjiang Province (LH2019C031), and the SIPT Program for Undergraduates of Northeast Agricultural University. ==== Body pmc1. Introduction MYB (v-myb avian myeloblastosis viral oncogene homolog) transcription factor (TF) is one of the most numerous and most versatile members of the plant TF family [1]. Its name is based on the fact that the encoded protein contains one or more highly conserved DNA domains—MYB domains [2]. The DNA-binding domains of MYB TFs contain an incompletely repeated R structure (R1, R2, and R3) composed of about 50 amino acid residues. MYB TFs can be divided into four subclasses according to the number of R structures contained in each gene [3]. Among them, the R1/R2-MYB subclass contains only one MYB domain, and the R2R3-MYB subclass contains two MYB domains, the R1R2R3-MYB subclass contains three continuous domains, and the 4R-MYB subclass contains four MYB domains [4,5,6]. A large number of previous studies have shown that MYB TFs regulate the expression of a series of genes by binding the cis-acting elements in the promoter region, and play an important role in the plant stress response [7]. After Paz-Ares cloned the ZmMYBC1 gene related to pigment synthesis from maize [1], a large number of MYB genes were isolated and identified from plants, and it was discovered that MYB genes are widely involved in physiological and biochemical processes such as plant growth and secondary metabolism, and can also regulate many physiological responses of plants under biotic and abiotic stress. Expression of OsMYB3R-2 in transgenic rice can enhance plant resistance to cold, dehydration, and salt stresses, while reducing sensitivity to abscisic acid (ABA) [8]. Overexpression of the rice R2R3-MYB TF gene OsMYB55 enhances the metabolism of amino acids through transcriptional activation to enhance the resistance of rice to high temperatures [9]. In addition, the overexpression of OsMYB4 will also significantly enhance the resistance of transgenic Arabidopsis thaliana to cold or freezing stresses [10]. Through the identification of 156 GmMYB genes in soybeans, it was found that the expression of 43 of them changed under the conditions of ABA, salt, drought, or cold stress [11]. Due to the above, MYB TFs participate in the response of plants to abiotic stress, and have potential application value in transgenic breeding for resistance to abiotic stresses, such as drought, cold, and low-temperature tolerances. According to previous research results, we speculate that there is also a MYB transcription factor gene in Malus baccata (L.) Borkh, which may change the tolerance of transgenic plants to abiotic stress by participating in the physiological and biochemical reactions. Malus baccata is one of the commonly used grafting rootstocks of apple. It has strong grafting affinity between the same genus and a high survival rate. It is widely grown in the northeast and eastern inner Mongolia. Rootstocks are crucial to the morphogenesis of fruit trees and their adaptation to external conditions. Rational use of rootstocks is the most economical and effective measure to control tree crowns and achieve preterm and stable yields, which can further increase yield, improve quality, and improve land utilization [12]. Cold and drought resistances are important indicators for evaluating apple rootstocks [13]. However, most of the current apple rootstocks with high-quality characteristics are weak in cold and drought resistances [14]. From the transcriptome analysis of M. baccata seedlings under cold and/or drought stresses (results not presented here), we found that the MbMYB108 expression level was significantly upregulated under both stresses. More importantly, through NCBI blast (BLAST: Basic Local Alignment Search Tool. Available online: https://blast.ncbi.nlm.nih.gov/Blast.cgi (accessed on 13 March 2021)) of the MbMYB108 gene, we found that the closest A. thaliana MYB gene is AtMYB96, which is a famous MYB TF gene involved in drought stress through the ABA signaling pathway [15]. To better understand the role of MYB TF genes involved in cold and drought stresses, and to provide potential genetic resources for the improvement of the drought resistance of Malus plant, a new MYB TF gene was isolated from M. baccata and designated as MbMYB108. Moreover, it was found that the tolerance of transgenic A. thaliana to cold and drought stresses was increased because of the overexpression of MbMYB108. 2. Results 2.1. Isolation of MbMYB108 Gene from M. baccata The ProtParam analysis (SIB Swiss Institute of Bioinformatics|Expasy. Available online: http://www.expasy.org/tools/protparam.html (accessed on 5 April 2021)) showed that MbMYB108 cDNA was a complete open reading frame of 903 bp, encoding 300 amino acids (Supplementary Figure S1). The predicted theoretical molecular mass of MbMYB108 was 33.939 kDa, with a theoretical isoelectric point of 7.82 and the average hydrophilicity coefficient was −0.753, which indicated that the protein is hydrophilic. Among the amino acids contained in MbMYB108 protein, Asn (10.4%), Ser (10.0%), Ala (8.0%), and Leu (7.0%) were relatively abundant. 2.2. Phylogenetic Relationship and Structural Prediction of MbMYB108 Protein To explore the evolutionary relationship among plant MYB proteins, DNAMAN was used to compare MbMYB108 with 11 other MYB proteins of different species. Inside the red and blue frames was the conserved sequence of the amino acid of MYB proteins (Figure 1A), which was a characteristic sequence of the plant MYB TF family. The amino acid conserved sequence of MbMYB108 had high homology with the amino acid conserved sequences of the other 11 MYB proteins, but the non-conserved sequences had obvious differences, which was consistent with the characteristics of TFs, indicating that MbMYB108 belongs to the MYB TF family. The constructed phylogenetic tree showed that MbMYB108 and MdMYB108 (XP_008340570.2, M. domestica) had the highest homology (Figure 1B). The predicted secondary structure of MbMYB108 contains predominant α-helices (37%) and coil structures (51%) (Figure 2A). The amino acid sequence of MbMYB108 had two SANT conserved domains (Figure 2B), presumably belonging to the R2R3-MYB TF family. The SWISS-MODEL online analysis suggests an overall α-helical structure for the MbMYB108 protein, consistent with the predicted secondary structure (Figure 2C). 2.3. MbMYB108 Was Localized in the Nucleus In order to determine the specific location of MbMYB108 protein in cells, a fusion expression vector of green fluorescent protein (GFP) and the MbMYB108 gene was constructed. As shown in Figure 3, the MbMYB108-GFP fusion protein was targeted into the nucleus (Figure 3E), whereas the control GFP alone was distributed throughout the plasma membrane and the nucleus (Figure 3B). These results showed that MbMYB108 is a nucleus-localized protein. 2.4. Expression Level of MbMYB108 in M. baccata The expression profile of MbMYB108 in various M. baccata tissues under CK was investigated by using the qPCR assay. As shown in Figure 4A, MbMYB108 was expressed at higher levels in young and mature leaves of M. baccata seedlings, but at lower levels in roots and stems. When subjected to low-temperature, high-salt, dehydration, high-temperature, and ABA treatments, the expression level of MbMYB108 in young leaves of M. baccata increased quickly, reached a maximum at 3, 9, 7, 3, and 3 h, respectively, and then decreased (Figure 4B). The expression trend of MbMYB108 in roots was consistent with that in young leaves, reaching a maximum at 7, 3, 5, 7, and 3 h, and then slightly decreasing (Figure 4C). 2.5. Overexpression of MbMYB108 Improves Cold Tolerance in Transgenic A. thaliana To study the effect of overexpression of MbMYB108 on cold and drought stresses in plants, MbMYB108 was introduced into A. thaliana. Using WT and UL lines as controls, the T2 generation transgenic lines were analyzed by qPCR, and the results showed that the target gene had been integrated into the T2 generation transgenic lines (S1, S2, S3, S4, S5, S6) (Figure 5A). The three lines (S2, S4, S6) with a relatively high DNA expression level were selected to continue to cultivate to obtain T3 generation homozygous transgenic A. thaliana, which were used for subsequent research on morphological phenotype and physiological indexes. As shown in Figure 5B, no significant difference in appearance was found among all A. thaliana lines (WT, UL, S2, S4, and S6) under CK (Cold 0 h). However, after dealing with cold (−6 °C) stress for 14 h (Cold 14 h), the transgenic lines (S2, S4, S6) looked much healthier than WT and UL. Under CK, there were no significant difference in the survival rate among all A. thaliana lines (WT, UL, S2, S4, and S6). However, when subjected to cold stress, the survival rates of WT and UL A. thaliana were only 20.5% and 15.8%, while the average survival rate of transgenic (S2, S4, S6) lines reached 84.9%. The survival rates of transgenic (S2, S4, S6) lines were significantly higher than those of WT and UL lines under cold treatment (Figure 5C). To further understand the reasons why transgenic A. thaliana performed better under cold stress, the relevant physiological indexes of all lines (WT, UL, S2, S4, and S6) under normal temperature and cold stress were determined. The results showed that after cold treatment, chlorophyll content, proline content, POD, and CAT activities of transgenic (S2, S4, S6) lines were significantly higher than those of WT and UL, while MDA content and relative conductivity were significantly lower (Figure 5D–I). 2.6. Overexpression of MbMYB108 Promotes the Expression of Cold Stress-Related Genes The low-temperature signal transduction pathway dependent on CBF (CRT/DRE-binding factor) is a very important molecular regulation pathway in A. thaliana in response to cold stress. Therefore, we analyzed the expression level changes of several important genes, AtCBF1, AtCBF3, AtRD29a, and AtCOR15a, downstream of the MYB transcription factor under cold treatment (Figure 6). After 14 h of cold treatment at −6 °C, the expression level of the four genes was upregulated in all A. thaliana lines compared with CK, but the expression level of these four genes in MbMYB108 overexpression transgenic lines (S2, S4, S6) was significantly higher than that in WT and UL lines, indicating that MbMYB108 positively regulates AtCBF1 and AtCBF3, thereby activating the expression of key genes AtRD29a and AtCOR15a under cold stress and improving the cold resistance of plants. 2.7. Overexpression of MbMYB108 Improves Drought Tolerance in Transgenic A. thaliana In order to understand the tolerance of different A. thaliana lines to drought stress, transgenic (S2, S4, S6), UL, and WT A. thaliana with the same growth vigor were not watered for 10 days, and then the phenotype of each line was observed. As shown in Figure 7A, the transgenic (S2, S4, S6), WT, and UL lines all grew well under CK (drought 0 days). However, when watering was stopped for 10 days (drought 10 days), the transgenic (S2, S4, S6) A. thaliana had a better appearance than WT and UL lines. Similarly, under CK, there were no significant differences in the survival rates of all A. thaliana lines (WT, UL, S2, S4, and S6). However, after 10 days of drought stress, the survival rates of WT and UL lines were only 26.8% and 25.5%, while the average survival rate of transgenic (S2, S4, S6) lines was 75.0%. The survival rates of the transgenic (S2, S4, S6) A. thaliana under drought stress were significantly higher than those of WT and UL lines (Figure 7B). Furthermore, the related physiological indexes of each line of A. thaliana treated under CK and drought stress (stopped watering for 10 days) were determined. It was found that under drought stress, overexpression of MbMYB108 resulted in lower MDA content and relative conductivity, higher chlorophyll and proline content, and higher POD and CAT activities in transgenic (S2, S4, S6) A. thaliana relative to WT and UL lines. However, for the above indexes, there were no significant differences among entire test lines (WT, UL, S2, S4, and S6) under CK (Figure 7C–H). 2.8. Overexpression of MbMYB108 Promotes the Expression of Drought Stress-Related Genes Drought stress can induce the increase of ABA levels in plants, and the ABA signal transduction activity is also enhanced. At the same time, studies have shown that ABA can also induce the expression of MbMYB108. Therefore, we selected the ABA synthesis gene AtNCED3 and the ABA signal transduction-related gene AtSnRK2.4, and further analyzed their expression patterns in MbMYB108 transgenic A. thaliana (Figure 8). The results showed that the expression levels of AtNCED3 and AtSnRK2.4 in MbMYB108-overexpressing lines (S2, S4, S6) were significantly higher than those in WT and UL lines under drought stress, indicating that MbMYB108 positively regulates the expression of AtNCED3 and AtSnRK2.4, and enhances drought tolerance in transgenic A. thaliana through ABA synthesis and signal transduction pathways. In addition, the expression levels of other drought stress-responsive genes AtCAT1 and AtP5CS downstream of MbMYB108 were also higher in transgenic (S2, S4, S6) lines under drought stress. 3. Discussion There are many types of MYB family proteins with different functions. In 1987, Paz-Ares et al. cloned the ZmMYBC1 gene by analyzing one genome and two cDNA clone sequences of maize [16], which is the first MYB gene to be identified in plants. In this experiment, M. baccata was used as a test material, taking the nucleic acid sequence of MdMYB108 (XM_008342348.3, M. domestica) as a reference sequence, and primer 5.0 [17] software was used to design specific primers to amplify the target gene MbMYB108. The alignment of protein sequences found that MbMYB108 protein and other species of MYB proteins had high similarity in conserved sequences but large differences in non-conserved regions, which was an obvious feature of TFs. Analysis of the conserved domains revealed that MbMYB108 protein had two SANT conserved domain, which was characteristic of the MYB TF family [18], and it was speculated that MbMYB108 protein belongs to the R2R3-MYB family. Subcellular localization revealed that MbMYB108 is a nuclear-localized protein (Figure 3), which was consistent with other MYB proteins [19,20]. The phylogenetic tree indicated that MbMYB108 is most closely related to MdMYB108 (XP_008340570.2, M. domestica) (Figure 1B). In this study, the expression level of MbMYB108 was the highest in the young leaves of M. baccata, while the expression level was less in the stems, which was only 1/9 of the expression level in the young leaves, indicating that its expression in the stems was inhibited or less relevant. In addition, the expression level of MbMYB108 was unstable in different organs of M. baccata, suggesting that its expression pattern is tissue-specific. The present study showed that abiotic stresses such as low temperature, high salt, dehydration, high temperature, and ABA all induced the expression of MbMYB108, and the expression level of the same TF changed with time under different stress treatments. Among them, under cold and drought stresses, the upregulation of MbMYB108 gene expression in young leaves and roots was the most obvious. Under cold stress, the expression levels of MbMYB108 in young leaves and roots reached the peak at 3 and 7 h after treatment, and were 7.8 and 13.3 times higher than those in the untreated treatment, respectively. Under drought stress, the expression levels of MbMYB108 in young leaves and roots peaked at 7 and 5 h after treatment, respectively, and were 7.2 and 12.9 times higher than those in untreated, respectively. This suggests that this gene may play an important role in the involvement of M. baccata in response to cold and drought stresses. It lays a foundation for further study on the functions of the MbMYB108 gene. After MbMYB108 was introduced into the model plant A. thaliana, WT, UL, and transgenic (S2, S4, S6) A. thaliana were treated with cold and drought, respectively. After stress treatment, all lines showed a certain degree of damage, but the transgenic (S2, S4, S6) A. thaliana emerged with less yellowing and wilting, and statistical analysis showed that the survival rate of the transgenic (S2, S4, S6) A. thaliana was significantly higher than WT and UL lines. The results showed that overexpression of MbMYB108 improves cold and drought resistances in transgenic (S2, S4, S6) A. thaliana. To further study the mechanism of this gene in the process of plant cold and drought resistances, the physiological and biochemical indexes and the expression of downstream stress-related genes of each A. thaliana line were determined and analyzed. When plants are faced with cold stress, their cytoplasmic membrane permeability increases, electrolytes and soluble substances extravasate, and relative conductivity increases. Therefore, the level of relative conductivity can be used as an index for judging the cold resistance of plants [21]. Under cold stress, compared with WT and UL A. thaliana, the relative conductivity of overexpressed transgenic (S2, S4, S6) lines increased less, indicating that their cell membranes are less damaged. Chlorophyll is one of the important indexes for judging the cold resistance of plants. Low temperatures will decompose chlorophyll in plants, causing leaves to turn yellow, and therefore plants cannot perform photosynthesis normally, and even die [22]. Observing the phenotype of each line (WT, UL, S2, S4, and S6) under cold treatment, it was found that the degree of leaf yellowing of the transgenic line is lighter. Cold resistance can also be estimated by osmoregulation and antioxidant capacity [23]. Soluble sugars, proline, and soluble proteins are the main osmotic regulators in plant cytoplasm. They can increase the osmotic concentration of cells, lower the freezing point, buffer dehydration under cold stress, and help maintain the normal metabolism of cells [24]. Under cold stress, plants accumulate osmotic regulators such as proline [25], which enhances the water-holding capacity of cells, promotes protein hydration, and increases the content of soluble proteins [26], thereby maintaining enzyme activity in cold conditions. The proline content of the transgenic (S2, S4, S6) A. thaliana increased more under cold stress, indicating that they suffered less dehydration damage. Studies have shown that the activities of POD and CAT in the roots of cold-acclimated rice are significantly enhanced compared with ordinary rice [27]. Under cold stress, transgenic (S2, S4, S6) A. thaliana overexpressing MbMYB108 had higher CAT and POD activities and lower MDA content compared with WT and UL. It indicated that the transgenic (S2, S4, S6) A. thaliana had a stronger ability to actively remove superoxide ions in the body under cold stress, and the tolerance to cold was also stronger. When plants are subjected to chilling or freezing stress, the concentration of intracellular Ca2+ increases significantly, and the high concentration of intracellular Ca2+ is sensed by downstream signal receptors calmodulin (CAM) [28,29], calcium-dependent protein kinase (CDPK) [30], and calcium-interacting protein kinase (CIPKs) [31], and the signal is further transmitted. Studies have shown that MYB TFs can not only bind to the CBF promoter region to promote the expression of its downstream cold stress-related genes, but can also inhibit the expression of CBFs and negatively regulate the cold resistance of plants [32,33,34]. In addition, CBF can also induce the expression of CRT/DRE cis-acting elements of genes such as CORs, RDs, and LTIs, and improves plant cold resistance [35,36,37]. MYB96 can bind to the promoter of HHP (transmembrane helix structure protein) gene to induce the synthesis of HHP protein, which in turn interacts with ICE1 (CBF gene expression inducer 1), ICE2, and CAMTA3 (calmodulin-binding transcriptional activator 3) to promote the expression of CBF, thereby activating the expression of downstream cold stress-related genes and improving the cold resistance of plants [38,39,40,41]. This present study quantitatively analyzed the expression levels of two key genes, CBF1 and CBF3, in the CBF-dependent pathway, and their downstream cold stress responsive genes COR15a and RD29a under normal conditions and cold treatment, and found that MbMYB108 can positively regulate AtCBF1, AtCBF3, AtCOR15a, and AtRD29a expression in response to cold stress through a CBF-dependent pathway. Drought usually leads to excessive accumulation of reactive oxygen species (ROS), such as O22−, H2O2, OH−, etc., in plants [42,43,44]. Most studies have shown that MYB TFs can participate in plant drought signal transduction and activate the ROS scavenging system, thereby avoiding the peroxidation of cell membrane lipids and the oxidative inactivation of metabolic enzymes, improving drought resistance of transgenic plants [45,46]. The birch BplMYB46 gene binds to the MYBCORE and AC-box motifs and directly activates the expression of SOD, POD, and P5CS genes containing such elements, thereby reducing intracellular ROS levels, increasing proline content, and improving plant drought resistance [47]. The sweet potato IbMYB116 gene was involved in the jasmonic acid (JA) signaling pathway to activate the ROS scavenging system, effectively reducing the oxidative damage of the plasma membrane caused by drought and other stress conditions, thereby improving the drought resistance of transgenic plants [48]. Compared with CK, transgenic (S2, S4, S6) A. thaliana exhibited higher CAT and POD activities, higher proline and chlorophyll content, and lower MDA content and relative conductivity than WT and UL lines under drought stress. The changes of these physiological indexes indicated that under drought stress, the transgenic (S2, S4, S6) lines could quickly scavenge reactive oxygen radicals in the body, reduce the damage to the plasma membrane, and maintain the integrity and stability of cells, and thus improve the drought resistance of plants. Under drought stress, plants can induce stomatal closure by synthesizing ABA, thereby reducing water loss [49,50]. NCED (9-cis-epoxycarotenoid dioxygenase) plays an important role in ABA synthesis [51,52]. A study has shown that AtNCED3 overexpression can increase ABA content in transgenic A. thaliana [53]. Under drought stress, the ABA content in transgenic kidney beans increased due to the expression of NCED1, and drought resistance of the overexpressed line was significantly improved [54]. The ABA signal transduction pathway is also an important pathway for plants to respond to drought stress [55,56]. Under drought stress, related TFs bind to downstream target genes to regulate their expression and generate intercellular signal ABA. ABA receptors sense the accumulation of ABA and bind to it, thereby inhibiting the activity of intracellular PP2C (protein phosphatase 2C) [57,58], resulting in an increase in SnRK (sucrose non-catalytic protein kinase 2) content [59], and downstream TFs are phosphorylated, regulating ABA-related gene expression, response to drought stress, and then stomatal closure [60,61]. In this study, the expression levels of AtNCED3 and AtSnRK2.4 were more significantly increased in transgenic (S2, S4, S6) A. thaliana after drought stress compared with WT and UL lines, indicating that MbMYB108 could positively regulate the expression of AtNCED3 and AtSnRK2.4. It was further shown that MbMYB108 can not only regulate ABA synthesis, but also participates in the regulation of ABA signal transduction genes, and jointly regulates the response of transgenic (S2, S4, S6) A. thaliana to drought stress through these two pathways. In addition, we also analyzed the expression of drought stress-responsive genes AtCAT1 and AtP5CS, and found that the expression levels of both genes were significantly increased in transgenic (S2, S4, S6) A. thaliana under drought stress. These results suggest that MbMYB108 can promote the expression of downstream drought stress-related genes through multiple pathways and improve the drought tolerance of plants. 4. Materials and Methods 4.1. Plant Materials and Growth Conditions Rapid propagation of M. baccata tissue culture seedlings was performed in MS growth medium (MS + 0.6 mg/L of 6-benzylaminopurine (6-BA) + 0.6 mg/L of indolebutyric acid (IBA)) for 30 days. Then, robust tissue culture seedlings were selected and transferred to MS rooting medium (MS + 1.2 mg/L IBA) to continue culturing until white roots grew [62]. Finally, the tissue culture seedlings with new roots were then transferred to Hoagland nutrient solution [63] for cultivation, and the new hydroponic solution was replaced as needed during hydroponics. The temperature of the culture chamber was kept at around 25 °C, and the relative humidity was kept at 80–85%. When the tissue culture seedlings grew 7–9 true leaves and the root system was well-developed, 60 seedlings with good shape and basically the same growth were selected for sampling. First, unexpanded young leaves, completely unfolded mature leaves, phloem at the second and third node stem segments, and newly emerged roots were sampled in 10 of the M. baccata seedlings, respectively. The remaining 50 seedlings were divided into 5 groups for different stress treatments: low-temperature treatment (hydroponic seedlings were cultured at 4 °C), salt treatment (hydroponic seedlings were cultured under high-salt conditions of 200 mM NaCl), dehydration treatment (hydroponic seedlings were cultured in Hoagland nutrient solution with a concentration of 20% PEG6000), high-temperature treatment (hydroponic seedlings were cultured at 38 °C), and ABA treatment (hydroponic seedlings were cultured in hydroponic solution with an ABA concentration of 50 μM), and the seedlings were cultured under normal Hoagland nutrient solution as control, young leaves, and roots of control, and all treatments were sampled. The samples of all control and treated seedlings were sealed after treatments of, respectively, 0, 1, 3, 5, 7, 9, and 12 h, immediately frozen in liquid nitrogen, and stored at −80 °C for RNA extraction. 4.2. Isolation and Cloning of MbMYB108 Extraction of M. baccata total RNA was performed with the OminiPlant RNA Kit (Kangweishiji, Beijing, China). The first-strand cDNA was synthesized using TransScript® First-Strand cDNA Synthesis SuperMix (TransGen Biotech, Beijing, China). A pair of primers (MbMYB-108F and MbMYB-108R, Table S1) were designed based on the homologous regions of MdMYB108 (XM_008342348.3, Malus domestica) to amplify the full-length cDNA sequence. The whole sequence of MbMYB108 was obtained by polymerase chain reaction (PCR) with primers, and the first-strand cDNA of M. baccata was used as a template. The target fragments were separated by agarose gel electrophoresis [64]. The obtained DNA fragment of the MbMYB108 gene was gel-purified and cloned into the pEASY®-T1 vector (TransGen Biotech, Beijing, China) [65] and sequenced (Beijing Genomics Institute, Beijing, China). 4.3. Sequence Analysis and Structure Prediction of MbMYB108 Multiple sequence alignment of MbMYB108 and MYB TFs from other species was performed with DNAMAN 5.2. The phylogenetic tree was constructed by the neighbor-joining method [66] with MEGA 7 (Home. Available online: http://www.megasoftware.net (accessed on 5 April 2021)) [67]. ExPASy (ExPASy-ProtParam tool. Available online: https://web.expasy.org/protparam/ (accessed on 20 April 2021)) was used to predict the primary structure of MbMYB108 protein. The domain of MbMYB108 protein was predicted on the SMART website (SMART: Main page. Available online: http://smart.embl-heidelberg.de/ (accessed on 20 April 2021)), using SWISS-MODEL (SWISS-MODEL. Available online: https://swissmodel.expasy.org/ (accessed on 20 April 2021)) to predict the tertiary structure of MbMYB108 protein. 4.4. Subcellular Localization Analysis of MbMYB108 Protein The MbMYB108 ORF was cloned into the BamHI and SalI sites of the pSAT6-GFP-N1 vector using primers (site-F and site-R, Table S1) with BamHI and SalI restriction sites to construct the MbMYB108-GFP transient expression vector. The recombinant plasmid containing the MbMYB108 gene was injected into onion epidermal cells by particle bombardment [68]. After overnight culture, the fluorescence signals of the control protein GFP and the MbMYB108-GFP fusion protein were observed under confocal microscopy (LSM 510 Meta, Zeiss, Germany). The nucleus was marked using DAPI staining. 4.5. Quantitative Real-Time PCR (qPCR) Expression Analysis of MbMYB108 DNAMAN was used to perform sequence alignment to find out the conserved regions of the MbMYB108 nucleic acid sequence and select the sequences with high specificity to design qPCR [69,70] primers (MbMYB-108qF and MbMYB-108qR, Table S1), and Actin gene (NC_024251.1, M. domestica) primers (Actin-F and Actin-R, Table S1) were sent for synthesis. Using cDNA of pretreated M. baccata material as a template, the expression level of the MbMYB108 gene was detected using TransStart® Green qPCR SuperMix (TransGen Biotech, Beijing, China) according to the manufacturer’s protocol. PCR conditions were as follows: 94 °C for 30 s, 40 cycles of 95 °C for 5 s, 54 °C for 40 s, 72 °C for 30 s, and then 72 °C for 10 min. Analysis of relative transcript level data was carried out using the 2−ΔΔCT method [71]. 4.6. Generation of Transgenic A. thaliana Overexpressing MbMYB108 Primers (HF and HR, Table S1), including target fragment-specific primer sequences, restriction sequences (BamHI and SalI), and overlapping sequences, were designed to amplify target fragments. Then, according to the principle of homologous recombination (ClonExpress®II One Step Cloning Kit, Vazyme, Nanjing, China), the target fragment was ligated into the PCAMBIA2300 vector to construct the PCAMBIA2300-MbMYB108 overexpression vector. The MbMYB108 gene was introduced into columbia-0 ecotype A. thaliana by Agrobacterium-mediated transformation of GV3101 using the inflorescence infusion method [72,73,74,75]. Transgenic A. thaliana was selected on MS medium containing 50 mg/L of Kanamycin until generation T3, and the T3 generation plants were used for further analysis. 4.7. Stress Treatment and Determination of Related Physiological Indexes in A. thaliana Wildtype (WT), empty vector line (UL, line transformed with empty vector only), and T3 transgenic lines (S2, S4, S6) of A. thaliana were sown in 1/2 MS medium, respectively, and after 10 days, seedlings that had cotyledons and had grown well were transferred to nutrient pots (4 plants per pot, the composition of substrate was turfy soil:vermiculite = 2:1). Each line of A. thaliana was divided into two groups: one group was treated with cold stress (−6 °C for 14 h) and the other group was treated with drought stress (stopped watering for 10 days), and then they were returned to normal culture for 7 days to remove stress. Finally, morphological changes were observed, and survival rates were calculated [76]. The materials of each line of A. thaliana under CK and after stress treatment were collected, and their physiological indexes were determined, respectively. Chlorophyll content was measured using the immersion extraction method [77]. According to the method of Huang et al., proline was extracted by the sulfosalicylic acid method and its content was determined [78,79,80]. Relative conductivity was measured using the pump-down method [81]. The guaiacol method was used to determine peroxidase (POD) activity [82,83], the ultraviolet (UV) absorption method was used to measure catalase (CAT) activity [84], and the spectrophotometer color method was used to measure malondialdehyde (MDA) activity [85,86]. 4.8. Expression Analysis of MbMYB108 Downstream Genes Using AtActin as the internal reference, the mRNAs of each line of A. thaliana under CK and after stress were extracted and reverse transcribed into first-strand cDNA. The qPCR experiments were performed on several important regulatory genes downstream of MYB TFs: cold stress response key genes (AtCBF1, AtCBF3, AtCOR15a, AtRD29a) and drought stress response key genes (AtNCED3, AtSnRK2.4., AtCAT1, AtP5CS), using specific primers (Table S1). 4.9. Statistical Analysis SPSS software was used to analyze the differences with Duncan’s multiple range tests. Statistical differences were referred to as significant when * p ≤ 0.05 and ** p ≤ 0.01. 5. Conclusions In the present study, a new MYB gene was isolated from M. baccata and named MbMYB108. Subcellular localization showed that MbMYB108 protein was located in the nucleus. When MbMYB108 was introduced into A. thaliana, it increased the levels of proline and chlorophyll, and improved the activities of POD and CAT, but decreased MDA content and relative conductivity, especially under cold and drought treatments. Taken together, our results suggest that MbMYB108 plays an important role in the response to cold and drought stress by enhancing the capability of scavenging ROS. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/ijms23094846/s1. Click here for additional data file. Author Contributions X.L. (Xingguo Li) and D.H. designed the experiments; C.Y., W.L. (Wenhui Li), X.L. (Xiaoqi Liang) and C.R. performed the experiments; W.L. (Wanda Liu), M.Z. and T.Y. analyzed the data; G.Y. and D.H. wrote the manuscript. All authors have read and agreed to the published version of the manuscript. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Comparison and phylogenetic relationship of MbMYB108 with other species of MYB transcription factor (TF) proteins. (A) Amino acid sequence alignment of MbMYB108 with other species of MYB TF proteins. Conserved domains are shown in red and blue boxes. (B) Phylogenetic tree analysis of MbMYB108 and other species of MYB TF proteins. Figure 2 Structural prediction of MbMYB108 protein. Prediction of the secondary structure (A), domains (B), and tertiary structure (C). Figure 3 Subcellular localization of MbMYB108 protein. Transient expressions of green fluorescent protein (GFP) and MbMYB108-GFP fusion protein in onion epidermal cells was observed by fluorescence microscopy: (A,D) were taken under bright light, (B,E) were taken under dark field, and (C,F) are the results of DAPI staining for 24 h. Scale bar corresponds to 5 μm. Figure 4 Time-course expression patterns of MbMYB108 in Malus baccata. (A) Expression patterns of MbMYB108 in young leaves (partly expanded), mature leaves (fully expanded), roots, and stems in normal condition (room temperature and normal watering). The expression level of young leaves was used as a control. Expression patterns of MbMYB108 in control condition (CK), low-temperature (4 °C), high-salt (200 mM NaCl), dehydration (20% PEG), high-temperature (38 °C), and ABA (50 μM) stress in young leaves (B) and roots (C) at the following time points: 0, 1, 3, 5, 7, 9, and 12 h. Data represent the mean of three replicates. Error bars represent standard deviation. Asterisks above error bars indicate significant differences between treatment and control (0 h) (* p ≤ 0.05, ** p ≤ 0.01). Figure 5 Overexpression of MbMYB108 in Arabidopsis thaliana improved cold tolerance. (A) Transgenic A. thaliana qPCR validation. (B) Phenotypic map of each line (WT, UL, S2, S4, and S6) of A. thaliana under CK (Cold 0 h), cold stress (Cold 14 h), and recovery conditions (stress removed). Scale bar corresponds to 4 cm. (C) The survival rates of each line (WT, UL, S2, S4, and S6) of A. thaliana under CK and cold stress. (D) Catalase (CAT) activity. (E) Chlorophyll content. (F) Peroxidase (POD) activity. (G) Proline content. (H) Malondialdehyde (MDA) content. (I) Relative conductivity. Asterisks above the error bars indicate extremely significant differences between transgenic (S2, S4, S6) and WT A. thaliana (** p ≤ 0.01). The level of each index in the WT line was used as a control. Figure 6 Relative expression levels of cold stress-related genes in WT, UL, and transgenic A. thaliana. Relative expression levels of AtCBF1 (A), AtCBF3 (B), AtCOR15a (C), and AtRD29a (D). Data represent the mean of three replicates. Error bars represent standard deviation. Asterisks above error bars indicate significant difference compared to the WT line (** p ≤ 0.01). Figure 7 Overexpression of MbMYB108 in A. thaliana improved drought tolerance. (A) Phenotypic map of each line (WT, UL, S2, S4, and S6) of A. thaliana under CK (Drought 0 days), drought stress (Drought 10 days), and recovery conditions (stress removed). Scale bar corresponds to 4 cm. (B) The survival rates of each line (WT, UL, S2, S4, and S6) of A. thaliana under CK and drought stress. (C) Catalase (CAT) activity. (D) Chlorophyll content. (E) Peroxidase (POD) activity. (F) Proline content. (G) Malondialdehyde (MDA) content. (H) Relative conductivity. Asterisks above the error bars indicate extremely significant differences between transgenic (S2, S4, S6) and WT A. thaliana (** p ≤ 0.01). The level of each index in the WT line was used as a control. Figure 8 Relative expression levels of drought stress-related genes in WT, UL, and transgenic A. thaliana. Relative expression levels of AtNCED3 (A), AtSnRK2.4 (B), AtCAT1 (C), and AtP5CS (D). Data represent the mean of three replicates. Error bars represent standard deviation. Asterisks above error bars indicate significant difference compared to the WT line (** p ≤ 0.01). Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Feng P. Chen P. Hong W. Zhao X. Liu X. Research Progress of MYB Transcription Factor Family in Arabidopsis thaliana Life Sci. Res. 2016 20 555 560 2. Shi M.Z. Xie D.Y. Biosynthesis and metabolic engineering of anthocyanins in Arabidopsis thaliana Recent Pat. Biotechnol. 2014 8 47 60 10.2174/1872208307666131218123538 24354533 3. Dubos C. Stracke R. Grotewold E. Weisshaar B. Martin C. Lepiniec L. MYB transcription factors in Arabidopsis Trends Plant Sci. 2010 15 573 581 10.1016/j.tplants.2010.06.005 20674465 4. Haga N. Kato K. Murase M. Araki S. Kubo M. Demura T. Suzuki K. Muller I. Voss U. Jurgens G. R1R2R3-Myb proteins positively regulate cytokinesis through activation of KNOLLE transcription in Arabidopsis thaliana Development 2007 134 1101 1110 10.1242/dev.02801 17287251 5. Hou X. Li S. Liu S. Hu C. Zhang J. Moriguchi T. Genome-wide classification and evolutionary and expression analyses of citrus MYB transcription factor families in sweet orange PLoS ONE 2014 9 e112375 10.1371/journal.pone.0112375 25375352 6. Qiao M. Yu Y. Xiang F. The roles of the Arabidopsis R2R3-MYB transcription factors in the stress responses Chin. Bull. Life Sci. 2009 21 145 150 7. Shinozaki Y.K. Shinozaki K. Organization of cis-acting regulatory elements in osmotic- and cold-stressresponsive promoters Trends Plant Sci. 2005 10 88 94 10.1016/j.tplants.2004.12.012 15708346 8. Dai X. Xu Y. Ma Q. Xu W. Wang T. Xue Y. Chong K. Overexpression of an R1R2R3 MYB Gene, OsMYB3R-2, Increases Tolerance to Freezing, Drought, and Salt Stress in Transgenic Arabidopsis Plant Physiol. 2007 143 1739 1751 10.1104/pp.106.094532 17293435 9. Casaretto J.A. El-Kereamy A. Zeng B. Stiegelmeyer S. Chen X. Bi Y. Rothstein S. Expression of OsMYB55 in maize activates stress-responsive genes and enhances heat and drought tolerance BMC Genom. 2016 17 312 10.1186/s12864-016-2659-5 10. Vannini C. Locatelli F. Bracale M. Magnani E. Marsoni M. Osnato M. Mattana M. Baldoni E. Coraggio I. Overexpression of the rice Osmyb4 gene increases chilling and freezing tolerance of Arabidopsis thaliana plants Plant J. 2004 37 115 127 10.1046/j.1365-313X.2003.01938.x 14675437 11. Sun X. Liu J. Yuan X. Pan X. Du W. Ren H. Ma Y. ABE J. Qiu L. Liu B. Cloning and expression analysis of GmMYB gene induced by abiotic stress J. Crops 2012 38 360 368 12. Chen X. Hao Y. Yang H. Mao Z. Jiang Y. Peng F. Shen X. Zhang J. Chen X. Shu H. 10 key technologies for high-quality and efficient development of my country’s apple industry Chin. Fruit Tree 2010 4 65 67 13. Yue Z. Liu H. Ma F. The Malus carotenoid cleavage dioxygenase 7 is involved in stress response and regulated by basic pentacysteine 1 Sci. Hortic. 2015 192 264 270 10.1016/j.scienta.2015.06.027 14. Ding N. Chen J. Feng Y. Sha J. Zhang M. Jiang Y. Effects of dwarf apple loading on nitrogen absorption, distribution and utilization J. Hortic. 2016 43 549 556 15. Liu J. She L. Lan X. Lu Y. Yang J. Liu H. Lu C. Chen Y. Bioinformatics and expression analysis of the MYB family of Himalayan Mirabilis Chin. Agric. Sci. Bull. 2020 36 54 61 16. Xing Z. Cloning and Functional Analysis of Potato Gene StR2R3-MYB1 Master’s Thesis Hunan Agricultural University Haikou, China 2014 2 17. Ren L. Zhu B. Zhang Y. Wang H. Li C. Su Y. Ba C. The research of applying primer premier 5.0 to design PCR primer J. Jinzhou Med. Coll. 2004 25 43 46 18. Boyer L. Latek R. Peterson C. The SANT domain: A unique histone-tail-binding module? Nat. Rev. Mol. Cell Biol. 2004 5 158 163 10.1038/nrm1314 15040448 19. Lv B. Zhang W. Li C. Ming F. MYB family transcription factor OsMYB84 participates in salt stress response through ABA signaling pathway Fudan J. 2015 54 591 600 20. Xie Y.P. Mechanism of Apple MdMYB88 and MdMYB124 Transcription Factors in Low Temperature and Drought Stress Ph.D. Thesis Northwest A&F University Xianyang, China 2018 23 21. Xu C.X. Advances in mechanisms of improving plant cold resistance J. Ecol. 2012 32 7966 7980 22. Ke X. He L. Su Z. Relative chlorophyll index and distribution of four woody plants in southern China J. Cent. South Univ. For. Technol. 2010 30 82 86 23. Pan R.C. Wang X.J. Li N.H. Plant Physiology Version 7 High Education Press Beijing, China 2012 438 455 24. Farhangi-Abriz S. Torabian S. Antioxidant enzyme and osmotic adjustment changes in bean seedlings as affected by biochar under salt stress Ecotoxicol. Environ. Saf. 2017 137 64 70 10.1016/j.ecoenv.2016.11.029 27915144 25. Yao H. Physiological and Biochemical Study on Cold Resistance of Four Cupressaceae Plants Master’s Thesis Inner Mongolia Agricultural University Hohhot, China 2012 2 26. Jiao R. Liu H. Liu G. Wang S. Hou N. Wang Q. Liu Z. Feng X. Hu X. Jin Y. On Proline Accumulation and Plant Resistance to Osmotic Stress Chin. Agric. Sci. Bull. 2011 27 216 221 27. Rezayian M. Niknam V. Ebrahimzadeh H. Penconazole and calcium ameliorate drought stress in canola by upregulating the antioxidative enzymes Funct. Plant Biol. 2020 47 825 839 10.1071/FP19341 32579878 28. Wang Y. Zhou X.Y. Changes of intracellular Ca2+ levels in plant cells under stress Anhui Agric. Sci. Bull. 2010 16 29 31 29. Zhang J. Meng Y. Song Y. Liu S. Zhuang W. Research Developments of Ca2+-CaM Signal System and Its Regulation in Plant J. Chongqing Norm. Univ. 2005 22 49 52 30. Liu G.S. Chen J. The role of calcium-dependent protein kinases (CDPKs) in plant calcium signal transduction Bull. Phys. 2003 20 160 167 31. Verma P. Sanyal S.K. Pandey G.K. Ca2+-CBL-CIPK: A modulator system for efficient nutrient acquisition Plant Cell Rep. 2021 40 2111 2122 10.1007/s00299-021-02772-8 34415375 32. Li M. Lin L. Zhang Y. Sui N. ZmMYB31, a R2R3-MYB transcription factor in maize’ positively regulates expression of CBF genes and enhances resistance to chilling and oxidative stress Mol. Biol. Rep. 2019 46 3937 3944 10.1007/s11033-019-04840-5 31037550 33. Yang A. Dai X. Zhang W.H. A R2R3-type MYB gene, OsMYB2, is involved in salt, cold, and dehydration tolerance in rice J. Exp. Bot. 2012 63 2541 2556 10.1093/jxb/err431 22301384 34. Zhou M.Q. Shen C. Wu L.H. Tang K.X. Lin J. CBF-dependent signaling pathway: A key responder to low temperature stress in plants Crit. Rev. Biotechnol. 2011 31 186 192 10.3109/07388551.2010.505910 20919819 35. Shinozaki K. Yamaguchi-Shinozaki K. Molecular responses to dehydration and low temperature: Differences and cross-talk between two stress signaling pathways Curr. Opin. Plant Biol. 2000 3 217 223 10.1016/S1369-5266(00)80068-0 10837265 36. Thomashow M.F. So what’s new in the field of plant cold acclimation? Lots! Plant Physiol. 2001 125 89 93 10.1104/pp.125.1.89 11154304 37. Baker S.S. Wilhelm K.S. Thomashow M.F. The 5-region of Arabidopsis thaliana cor15a has cis-acting elements that confer cold-, drought- and ABA-regulated gene expression Plant Mol. Biol. 1994 24 701 713 10.1007/BF00029852 8193295 38. Yamaguchi-Shinozaki K. Shinozaki K. A novel cisacting element in an Arabidopsis gene is involved in responsiveness to drought, low-temperature, or high-salt stress Plant Cell 1994 6 251 264 8148648 39. Chinnusamy V. Ohta M. Kanrar S. Lee B.H. Hong X. Agarwal M. Zhu J.K. ICE1: A regulator of cold-induced transcriptome and freezing tolerance in Arabidopsis Genes Dev. 2003 17 1043 1054 10.1101/gad.1077503 12672693 40. Agarwal M. Hao Y. Kapoor A. Dong C.H. Fujii H. Zheng X. Zhu J.K. A R2R3 type MYB transcription factor is involved in the cold regulation of CBF genes and in acquired freezing tolerance J. Biol. Chem. 2006 281 37636 37645 10.1074/jbc.M605895200 17015446 41. Doherty C.J. Van-Buskirk H.A. Myers S.J. Thomashow M.F. Roles for Arabidopsis CAMTA transcription factors in cold-regulated gene expression and freezing tolerance Plant Cell 2009 21 972 984 10.1105/tpc.108.063958 19270186 42. Mhamdi A. Breusegem F. Reactive oxygen species in plant development Development 2018 145 164376 10.1242/dev.164376 43. Mittler R. ROS are good Trends Plant Sci. 2017 22 11 19 10.1016/j.tplants.2016.08.002 27666517 44. Waszczak C. Carmody M. Kangasjarvi J. Reactive oxygen species in plant signaling Annu. Rev. Plant Biol. 2018 69 209 236 10.1146/annurev-arplant-042817-040322 29489394 45. Mahmood T. Khalid S. Abdullah M. Ahmed Z. Shah M.K.N. Ghafoor A. Du X. Insights into Drought Stress Signaling in Plants and the Molecular Genetic Basis of Cotton Drought Tolerance Cells 2020 9 105 10.3390/cells9010105 31906215 46. Noctor G. Veljovic-Jovanovic S. Driscoll S. Novitskaya L. Foyer C.H. Drought and oxidative load in the leaves of C3plants: A predominant role for photorespiration? Ann. Bot. 2002 89 841 850 10.1093/aob/mcf096 12102510 47. Guo H. Wang Y. Wang L. Hu P. Wang Y. Jia Y. Zhang C. Zhang Y. Zhang Y. Wang C. Expression of the MYB transcription factor gene BplMYB46 affects abiotic stress tolerance and secondary cell wall deposition in Betula platyphylla Plant Biotechnol. J. 2016 15 107 121 10.1111/pbi.12595 27368149 48. Zhou Y. Zhu H. He S. Zhai H. Liu Q. A Novel Sweet potato Transcription Factor Gene IbMYB116 Enhances Drought Tolerance in Transgenic Arabidopsis Front. Plant Sci. 2019 10 1025 10.3389/fpls.2019.01025 31475022 49. Seo P. Xiang F. Qiao M. Park J. Lee Y. Kim S. Lee Y. Park W. Park C. The MYB96 Transcription Factor Mediates Abscisic Acid Signaling during Drought Stress Response in Arabidopsis Plant Physiol. 2009 151 275 289 10.1104/pp.109.144220 19625633 50. Tuteja N. Sopory S.K. Plant signaling in stress: G-protein coupled receptors, heterotrimeric G-proteins and signal coupling via phospholipases Plant Signal. Behav. 2008 3 79 86 10.4161/psb.3.2.5303 19516978 51. Gavassi M.A. Silva G.S. da Silva C.D.M.S. Thompson A.J. Macleod K. Oliveira P.M.R. Cavalheiro M.F. Domingues D.S. Habermann G. NCED expression is related to increased ABA biosynthesis and stomatal closure under aluminum stress Environ. Exp. Bot. 2021 185 104404 10.1016/j.envexpbot.2021.104404 52. Taylor I.B. Burbidge A. Thompson A.J. Control of abscisic acid synthesis J. Exp. Bot. 2000 51 1563 1574 10.1093/jexbot/51.350.1563 11006307 53. Yang Y. Tan B. A Distal ABA Responsive Element in AtNCED3 Promoter Is Required for Positive Feedback Regulation of ABA Biosynthesis in Arabidopsis PLoS ONE 2014 9 e87283 24475264 54. Torres G.A. Lelandais-Brière C. Besin E. Jubier M.F. Roche O. Mazubert C. Corre-Menguy F. Hartmann C. Characterization of the expression of Phaseolus vulgaris OCT1, a dehydration-regulated gene that encodes a new type of phloem transporter Plant Mol. Biol. 2003 51 341 349 10.1023/A:1022014229899 12602865 55. Ren X. Chen Z. Liu Y. Zhang H. Zhang M. Liu Q. Hong X. Zhu J.K. Gong Z. ABO3, a WRKY transcription factor, mediates plant responses to abscisic acid and drought tolerance in Arabidopsis Plant J. 2010 63 417 429 10.1111/j.1365-313X.2010.04248.x 20487379 56. Golldack D. Lüking I. Yang O. Plant tolerance to drought and salinity: Stress regulating transcription factors and their functional significance in the cellular transcriptional network Plant Cell Rep. 2011 30 1383 1391 10.1007/s00299-011-1068-0 21476089 57. Leung J. Merlot S. Giraudat J. The Arabidopsis ABSCISIC ACID-INSENSITIVE2 (ABI2) and ABI1 genes encode homologous protein phosphatases 2C involved in abscisic acid signal transduction Plant Cell 1997 9 759 771 9165752 58. Gosti F. Beaudoin N. Serizet C. Webb A.A. Vartanian N. Giraudat J. ABI1 protein phosphatase 2C is a negative regulator of abscisic acid signaling Plant Cell 1999 11 1897 1910 10.1105/tpc.11.10.1897 10521520 59. Nakashima K. Fujita Y. Kanamori N. Katagiri T. Umezawa T. Kidokoro S. Maruyama K. Yoshida T. Ishiyama K. Kobayashi M. Three arabidopsis SnRK2 protein kinases, SRK2D/SnRK2.2, SRK2E/SnRK2.6/OST1 and SRK2I/SnRK2.3, involved in ABA signaling are essential for the control of seed development and dormancy Plant Cell Physiol. 2009 50 1345 1363 10.1093/pcp/pcp083 19541597 60. Muschietti J. McCormick S. Abscisic acid (ABA) receptors: Light at the end of the tunnel F1000 Biol. Rep. 2010 2 15 10.3410/B2-15 20948817 61. Yu X. Han J. Li L. Zhang Q. Yang G. He G. Wheat PP2C-a10 regulates seed germination and drought tolerance in transgenic Arabidopsis Plant Cell Rep. 2020 39 635 651 10.1007/s00299-020-02520-4 32065246 62. Han D. Wang Y. Zhang L. Ma L. Zhang X. Xu X. Han Z. Isolation and functional characterization of MxCS1: A gene encoding a citrate synthase in Malus xiaojinensis Biol. Plant. 2012 56 50 56 10.1007/s10535-012-0015-4 63. Hoagland D.R. Optimum nutrient solutions for plants Science 1920 1354 562 564 10.1126/science.52.1354.562 64. Han D. Zhang Z. Ding H. Chai L. Liu W. Li H. Yang G. Isolation and characterization of MbWRKY2 gene involved in enhanced drought tolerance in transgenic tobacco J. Plant Interact. 2018 13 163 172 10.1080/17429145.2018.1447698 65. Han D. Ding H. Chai L. Liu W. Zhang Z. Hou Y. Yang G. Isolation and characterization of MbWRKY1, a WRKY transcription factor gene from Malus baccata (L.) Borkh involved in drought tolerance Can. J. Plant Sci. 2018 98 1023 1034 10.1139/cjps-2017-0355 66. Saitou N. Nei M. The neighbor-joining method: A new method for reconstructing phylogenetic trees Mol. Biol. Evol. 1987 4 406 425 3447015 67. Kumar S. Stecher G. Tamura K. MEGA7: Molecular Evolutionary Genetics Analysis Version 7.0 for Bigger Datasets Mol. Biol. Evol. 2016 33 1870 1874 10.1093/molbev/msw054 27004904 68. Yang G. Li J. Liu W. Yu Z. Shi Y. Lv B. Wang B. Han D. Molecular cloning and characterization of MxNAS2, a gene encoding nicotianamine synthase in Malus xiaojinensis, with functions in tolerance to iron stress and misshapen flower in transgenic tobacco Sci. Hortic. 2015 183 77 86 10.1016/j.scienta.2014.12.014 69. Radonić A. Thulke S. Mackay I.M. Landt O. Siegert W. Nitsche A. Guideline to reference gene selection for quantitative real-time PCR Biochem. Biophys. Res. Commun. 2004 313 856 862 10.1016/j.bbrc.2003.11.177 14706621 70. Peirson S. Butler J. Foster R. Experimental validation of novel and conventional approaches to quantitative real-time PCR data analysis Nucleic. Acids. Res. 2003 31 e73 10.1093/nar/gng073 12853650 71. Livak K.J. Schmittgen T.D. Analysis of relative gene expression data using real-time quantitative PCR and the 2−∆∆CT method Methods 2001 25 402 408 10.1006/meth.2001.1262 11846609 72. Han D. Han J. Xu T. Li X. Yao C. Li T. Sun X. Wang X. Yang G. Overexpression of MbERF12, an ERF gene from Malus baccata (L.) Borkh increases cold and salt tolerance in Arabidopsis thaliana associated with the ROS scavenging through ethylene signal transduction Vitr. Cell. Dev. Biol.-Plant 2021 57 760 770 10.1007/s11627-021-10199-9 73. Han D. Wang Y. Zhang Z. Pu Q. Ding H. Han J. Fan T. Bai X. Yang G. Isolation and functional analysis of MxCS3: A gene encoding a citrate synthase in Malus xiaojinensis, with functions in tolerance to iron stress and abnormal flower in transgenic Arabidopsis thaliana Plant Growth Regul. 2017 82 479 489 10.1007/s10725-017-0274-3 74. Xu Y.T. Xu Q. Transformation of Arabidopsis thaliana by floral dip Bio-101 2018 e1010203 10.21769/BioProtoc.1010203 75. Ren L. Liu H. Wang H. Zhu Z. Construction of Expression Vectors of Brassica chinensis Myb28 Gene and Transformation to Arabidopsis thaliana Biotechnol. Bull. 2012 3 80 84 76. Yao C. Li X. Li Y. Yang G. Liu W. Shao B. Zhong J. Huang P. Han D. Overexpression of a Malus baccata MYB Transcription FactorGene MbMYB4 Increases Cold and Drought Tolerance in Arabidopsis thaliana Int. J. Mol. Sci. 2022 23 1794 10.3390/ijms23031794 35163716 77. Xu F. Liu S. Liu Y. Xu J. Liu T. Dong S. Effectiveness of lysozyme coatings and 1-MCP treatments on storage and preservation of kiwifruit Food Chem. 2019 288 201 217 10.1016/j.foodchem.2019.03.024 30902282 78. Huang Q. Qian X. Jiang T. Zheng X. Effect of eugenol fumigation treatment on chilling injury and CBF gene expression in eggplant fruit during cold storage Food Chem. 2019 292 143 150 10.1016/j.foodchem.2019.04.048 31054659 79. Bates L.S. Waldren R.P. Teare I.D. Rapid determination of free proline for water stress studies Plant Soil. 1973 39 205 207 10.1007/BF00018060 80. Li P. Zheng X. Liu Y. Zhu Y. Pre-storage application of oxalic acid alleviates chilling injury in mango fruit by modulating proline metabolism and energy status under chilling stress Food Chem. 2014 142 72 78 10.1016/j.foodchem.2013.06.132 24001814 81. Lutts S. Kinet J.M. Bouharmont J. NaCl-induced senescence in leaves of rice (Oryza sativa L.) cultivars differing in salinity resistance Ann. Bot. 1996 78 389 398 10.1006/anbo.1996.0134 82. Sharma A. Yuan H. Kumar V. Ramakrishnan M. Kohli S. Kaur R. Thukral A. Bhardwaj R. Zheng B. Castasterone attenuates insecticide induced phytotoxicity in mustard Ecotoxicol. Environ. Saf. 2019 179 50 61 10.1016/j.ecoenv.2019.03.120 31026750 83. Zhang L. Pei Y. Wang H. Jin Z. Liu Z. Qiao Z. Fang H. Zhang Y. Hydrogen sulfide alleviates cadmium-induced cell death through restraining ROS accumulation in roots of Brassica rapa L. ssp pekinensis. Oxid. Med. Cell Longev. 2014 2015 804603 84. Han Y. Yang H. Wu M. Yi H. Enhanced drought tolerance of foxtail millet seedlings by sulfur dioxide fumigation Ecotoxicol. Environ. Saf. 2019 178 9 16 10.1016/j.ecoenv.2019.04.006 30980964 85. Li L. Lin G. Gu H. Yu L. Ni C. Effects of Srxn1 on growth and Notch signalling of astrocyte induced by hydrogen peroxide Artif. Cells Nanomed. Biotechnol. 2019 47 1917 1923 10.1080/21691401.2019.1614016 31079497 86. Jiang Y. Lin L. Zhong S. Cai Y. Zhang F. Wang X. Miao R. Zhang B. Gao S. Hu X. Overexpression of novel lncRNA NLIPMT inhibits metastasis by reducing phosphorylated glycogen synthase kinase 3β in breast cancer J. Cell Physiol. 2018 234 10698 10708 10.1002/jcp.27738 30417392

PMC009xxxxxx/PMC9099688.txt

==== Front Sensors (Basel) Sensors (Basel) sensors Sensors (Basel, Switzerland) 1424-8220 MDPI 10.3390/s22093138 sensors-22-03138 Article Long-Range Underwater Communication Based on Time Reversal Processing Using Various Diversity Methods https://orcid.org/0000-0002-1839-1666 Kim Donghyeon 1† https://orcid.org/0000-0002-8514-3187 Kim Jeasoo 2*† Hahn Jooyoung 3 Prades Raul Marin Academic Editor 1 Department of Convergence Study on the Ocean Science and Technology, Korea Maritime and Ocean University, Busan 49112, Korea; donghyeon.ual@gmail.com 2 Department of Ocean Engineering, Korea Maritime and Ocean University, Busan 49112, Korea 3 Agency of Defense Development, Changwon-si 51682, Korea; hahnjy@add.re.kr * Correspondence: jskim@kmou.ac.kr † These authors contributed equally to this work. 20 4 2022 5 2022 22 9 313818 3 2022 19 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Time reversal processing (TRP) exploits signal diversity methods, namely, spatial, temporal, beam, and frequency, to mitigate the distortion caused by multipath time delay. Using the same experimental data, this study compares the performance of communication utilizing TRP based on various diversity methods. In October 2018, the biomimetic long-range acoustic communication experiment 2018 (BLAC18) was conducted in the East Sea, east of Pohang, Korea. During the experiment, communication signals modulated by binary phase-shift keying were transmitted over a range of 60 km, and a vertical line array of 16 elements (with an aperture of ∼42 m) was utilized. The BLAC18 analysis showed that the performance of each diversity method depends on the order of diversity. When the order of diversity was one, the beam diversity method with the beamformed signal yielded the best performance. For the maximum order of diversity, however, the spatial diversity method delivered the best performance, owing to the high channel variability and large number of receivers. time reversal diversity deep water communication ==== Body pmc1. Introduction Long-range communication is a challenging task requiring systematic research because of its unstable performance, which is caused by an increase in delay spread, a reduction in channel capacity due to rising transmission loss with range, and channel fluctuation. Recently, in conjunction with the development of autonomous underwater vehicles that cruise over a long range (of the order of hundreds of kilometers), long-range communication technology has attracted growing interest, with numerous experimental studies being conducted to develop a stable long-range underwater communication system. Stojanovic demonstrated the feasibility of long-range communication for the first time using experimental data from the Woods Hole Oceanographic Institution’s long-range (200 km) underwater communication experiment [1,2]. During the experiment, communication signals designed with higher order constellations such as 8-quadrature amplitude modulation (QAM) and 8-phase shift keying (PSK) were transmitted. In 1998, Plaisant demonstrated communication signals designed by spread spectrum techniques as well as PSK modulation with two experimental data sets (50 km: PSK modulation, 20 km: spread spectrum) [3]. To mitigate the intersymbol interference (ISI) caused by multipath time delay, the aforementioned results were analyzed based on a multichannel decision feedback equalizer (M-DFE). This approach is necessarily highly complex computationally, and this complexity is proportional to the length of the equalizer in taps and the number of array elements [4]. To effectively address this problem, time reversal processing (TRP), which enables self-equalization and utilizes diversity to increase the channel capacity, has been utilized as an alternative to M-DFE for long-range communication [4]. TRP is calculated by correlating the received signal with the channel impulse response, which improves the signal-to-noise ratio (SNR) and mitigates the ISI by utilizing multipath time delay that can degrade communication performance. Parvulescu and Clay [5] visually described the improvement of the SNR in a one-transmitter one-receiver environment. The performance of the TRP can be improved using a diversity method [6]. Four types of diversity methods exist, namely, spatial, temporal, beam, and frequency, and they operate on the principle of coherently combining multiple TRP outputs. TRP, which is the precombining process, is a special case of reduced-complexity M-DFE [4]. Since 2008, the Japan Agency for Marine-Earth Science and Technology (JAMSTEC), the institution that has conducted the most long-range underwater acoustic communication experiments to date, has been conducting experiments over various ranges (30–1450 km) and using various modulation methods while investigating multiuser communication [7,8,9,10,11]. JAMSTEC has transmitted communication signals designed up to 16-QAM, which is the highest-order modulation method designed for long-range underwater communication. They analyzed these signals using spatial diversity-based TRP with a vertical line array (VLA). Song presented and compared the results of a basin-scale (∼3250 km) tomography experiment conducted in November 1994 based on spatial and temporal diversity methods [12,13]. Furthermore, unlike other long-range experiments using a VLA, a long-range acoustic communication (LRAC10) experiment was conducted in 2010 at a range of up to 550 km using a horizontal line array [14,15,16,17,18]. A beam diversity-based TRP with an azimuth angle was utilized to evaluate the performance of communication signals that were designed using various modulation methods such as PSK modulation [14,15], OFDM [16], and spread spectrum [18]. The purpose of this study is twofold. First, this paper describes the experiment conducted in the East Sea. The East Sea area is located east of Pohang, Korea, and is a so-called miniature ocean (mini-ocean) because it has an independent deep convection system, which is a characteristic of oceans, and its circulation system and hydrography are similar to that of major ocean basins. The dynamics and oceanographic characteristics of this area can be predicted in advance [19,20,21]. Thus, experiments in the East Sea have oceanographic value. Furthermore, conducting long-range experiments is difficult because they are resource- and time-consuming endeavors, and are performed only by a small number of institutions [1,7,12,14,22]. Therefore, the East Sea long-range communication experiment itself, which is the first experiment conducted using various long-range communication signals in Korea, is very meaningful. The experiment was named the biomimetic long-range acoustic communication experiment 2018 (BLAC18). Second, this paper compares the performance of various diversity methods. During the BLAC18 experiment, communication signals were transmitted several times and obtained using a VLA. Using this experimental data, Park confirmed that combining two diversity methods rather than using a single diversity method improved the communication performance [23]. Kim presented the results of the beam diversity method with a vertical line array, notably the results of a performance comparison based on various beam combinations [24]. The performance of the diversity method depends on the channel variability, and we analyze and compare the communication performance according to the type and order of the diversity method in terms of this variability. However, a study involving a comparison of the results of various diversity methods for the same data based on the order of diversity has not been reported thus far. The contribution of this study is therefore to provide a comprehensive comparison of the communication performance and the dependence thereof on the order of diversity and the type of diversity. The remainder of this paper is organized as follows. Section 2 reviews the TRP formulation using various diversity methods. Section 3 outlines the experiment conducted in the East Sea off the coast of Pohang, Korea, and describes the structure of the communication signal used as the basis for analysis in this study. Section 4 presents and compares the results of our analysis of the long-range communication performance for various diversity methods with the theories introduced in Section 2. Finally, concluding remarks are stated in Section 5. 2. Review of TRP with Various Diversity Methods As mentioned above, TRP is the correlation between a received signal and the channel impulse response and can be extended using a diversity method. The diversity method improves the SNR by coherently combining multiple TRP outputs. For example, the spatial diversity method combines the TRP outputs acquired from different locations. The reader can apply the appropriate diversity method according to the experimental conditions, such as the type of signal, the number of transmissions, and the number of receivers. In this section, we review the theoretical TRP formulation using three diversity methods (spatial, temporal, and beam). As we only transmitted modulated signals with a single carrier frequency, an analysis was not conducted using the frequency diversity method. Equation (1) represents the TRP output using both the spatial and temporal diversity methods [6], where the inside structure of ∑ denotes the TRP output when the order of diversity is one. (1) Y(ω)=∑morn=1MorNRmorn(ω)Hmorn*(ω)=S(ω)∑morn=1MorNHmorn(ω)Hmorn*(ω), where S(ω) and Y(ω) are the source signal and the TRP output using the spatial or temporal diversity method, respectively. R(ω) and H(ω) represent the received signal and the channel impulse response, respectively, and an asterisk (or ()*) denotes a complex conjugate. The order of diversity varies depending on the diversity method. In Equation (1), M and N denote the order of diversity of the spatial and temporal methods, respectively. Because the spatial diversity method-based TRP utilizes the received signals measured from several receivers, the order of spatial diversity (M) is the same as the number of receivers. In the temporal diversity method, which utilizes multiple transmissions measured from a single receiver, the order of temporal diversity (N) is the same as the number of transmissions. Figure 1a shows a schematic of the method used to select the received signals for the spatial and temporal diversity methods. The black boxes indicate multiple transmissions obtained from the array, i.e., all the data acquired during the experiment. The spatial and temporal diversity methods utilize the received signals in the red and blue boxes, respectively. The overall system based on these diversity methods is shown in Figure 1b. Whereas the spatial and temporal diversity methods utilize the received signals, the beam diversity method utilizes the signals obtained by beamforming the received signals. The overall system based on the beam diversity method is shown in Figure 2, and its mathematical formulations are shown in Equations (2)–(6): (2) Bl(ω)=R(ω)WlH(ω)=S(ω)H(ω)WlH(ω)=S(ω)Hl′(ω), (3) R(ω)=R1(ω)…Rm(ω)…RM(ω),H(ω)=H1(ω)…Hm(ω)…HM(ω),Wl(ω)=Wl1(ω)…Wlm(ω)…WlM(ω), (4) Wlm(ω)=exp−iω(m−1)dcsinθl, (5) Y′(ω)=∑l=1LBl(ω)Hl′*(ω)=S(ω)∑l=1LHl′(ω)Hl′*(ω), where d, c, and θl denote the spacing between adjacent receivers, the speed of sound, and the angle of the lth path, respectively; ()H denotes the Hermitian transpose. In Equation (2), Bl(ω) is calculated by steering the received signals to the angle of the lth path and is defined as a beamformed signal [25,26]. H(ω) and R(ω) in Equation (3) represent the channel impulse responses between the source and the array and the received signals obtained from the array, respectively. In this study, as a vertical line array is utilized, R(ω) and H(ω) have M components. Wl(ω) in Equation (3) represents the steering vector in the direction of the angle of the lth path. Bl(ω) can be expressed as the product of S(ω) and Hl′(ω) if R(ω) is separated into S(ω) and H(ω) (Equation (2)). Hl′(ω), the product of H(ω) and Wl(ω), represents the channel impulse response of the beamformed signal along the lth path. In Equation (5), Y′(ω) represents the TRP output using the beam diversity method and has a structure similar to Equation (1). Equations (2) and (5) correspond to the part marked “Beamforming” (blue box) and the part marked “Time reversal processing” (red box) in Figure 2, respectively. A comparison of Figure 1b and Figure 2 reveals two main differences: (1) The first step of the beam diversity method is to beamform (or steer) the received signals (blue box in Figure 2); (2) In the TRP step (red box in Figure 2), these beamformed signals are used instead of the received signals. Therefore, in the beam diversity method, the order of beam diversity is the same as the number of beams. 3. BLAC18 Experiment In October 2018, BLAC18 was conducted in the East Sea, east of Pohang, Korea. At the site of the experiment, the water depth was in the range of 950–1500 m. The communication signals were transmitted over a range of 60 km. The VLA consisted of 16 elements spanning a 42-m aperture with an element spacing of 2.8 m. In the experiment, the source depth was 200 m and the VLA covered a depth range of 179–221 m at a water depth of approximately 950 m. Figure 3a shows the experimental area. The five-pointed star indicates the source location and the magenta circle indicates the VLA location. The depth contours represent the depth in meters. A schematic of the experiment is shown in Figure 3b. The sound-speed profile displayed in Figure 3b was obtained by measuring the conductivity, temperature, and depth (CTD) at the VLA location, which features an underwater sound channel with an acoustic axis at a depth of 250 m. In Figure 3b, the red and blue solid circles indicate the receivers and sources, respectively. A positive angle θ is defined for an upward path. This experiment was conducted in collaboration with the Korea Institute of Ocean Science & Technology (KIOST). The KIOST-operated R/V Ieodo was used for equipment deployment/recovery. Iridium and depth sensors were used to track the location and depth of the VLA, respectively. This paper presents an analysis of the data transmitted during a 2-h period. The detailed structure of the transmitted signal is illustrated in Figure 4. Each block denoted by the letter {A} (280 s long) consists of six data packets denoted by the letter {B} and two types of guard times, which are indicated as {GT#1} and {GT#2} and have lengths of 22.5 and 82.5 s, respectively. Each data packet {B} is 7.5 s long and consists of a linear frequency modulated (LFM) signal as a channel probe and a communication sequence. The 280-s long signal was repeated every 55 min. A total of 18 data packets were transmitted during the 2-h period. During the BLAC18 experiment, many click sounds were recorded. Three data packets were highly contaminated owing to the click sounds and were excluded from the analysis. The probe signal was an LFM chirp with a Hanning window having a duration and frequency of 3 s and 2.2–2.9 kHz, respectively. The communication signal consisted of a total of 1255 symbols (2.47 s transmission duration), modulated by BPSK with a bit rate of 512 bits/s. The shaping pulse was a square-root raised cosine filter with a roll-off factor of beta = 0.25 and the carrier frequency was 2560 Hz. Among the 1255 symbols, the first 255 are m-sequence signals designed for Doppler estimation and synchronization. A 2-s-long guard time, indicated as {GT#3} in Figure 4, is included between the probe and communication signals. 4. Experimental Results 4.1. Order of Beam Diversity Method The orders of the spatial and temporal diversity methods were 16 (the number of receivers) and 15 (=18 − 3, the number of packets), respectively. In the beam diversity method, the order of diversity is the number of dominant paths. In general, this is estimated by beamforming the received signals. However, because the carrier frequency was approximately 10 times the design frequency, the number of dominant paths was not estimated owing to aliasing. Additionally, the beam resolution, i.e., 0.89λD (rad) in a line array [27], determined by the wavelength and array aperture is 0.7∘ under our experimental conditions (fc=2560Hz,D=42m), and adjacent paths lower than the beam resolution cannot be separated. Figure 5a shows the conventional beamforming output according to the frequency, with aliasing occurring as expected. Figure 5b shows the conventional beamforming output averaged over the frequency band. An enlargement of the region from −5∘ to 15∘ of Figure 5b is shown in Figure 5c, in which the red lines correspond to the angles of the path obtained from Figure 6. As is evident from the figure, the angles of all dominant paths cannot be estimated by the sidelobe because of aliasing and beam resolution. Therefore, to obtain the angles of all dominant paths, we used the channel impulse response estimated using the probe signal. The channel impulse response is shown in Figure 6a, from which it is clear that four dominant paths exist. Estimation of the channel impulse response ensures the separation of multiple arrivals in the beam-time domain, referred to as “beam-time migration,” through conventional beamforming. Figure 6b shows the beam-time migration of the channel impulse response. The red dashed lines between Figure 6a,b represent the relative time delays of dominant paths and are shown to extract the angles from the beam-time migration. The process whereby the angles are extracted is as follows: (1) Identify locations with values above a certain threshold in the beam-time domain. At this time, the value of the threshold varies with the selected data. (2) When two or more locations have the same time, one location is selected considering the path direction, that is, the sign of the angle, or the level corresponding to the path from the channel impulse response. The relative arrival time and angle information of the dominant paths are indicated with the red circles in Figure 6b, and these angles were used for the beam diversity method. Figure 6 shows the channel impulse response and beam-time migration estimated from one data packet, and the number of dominant paths was maintained in the same range. 4.2. Impact of Diversity The TRP performance is improved with a diversity method [6]. In this section, we reproduce the work conducted in a previous study [6] to explain the effect of the diversity method. Among the three diversity methods used in this study, the results obtained with the spatial diversity method are presented and described using the 60-km data. Figure 7a shows one of the signals received across the 60-km range. During the experiment, a number of click sounds, presumed to be emitted by dolphins, were captured; these are indicated by red arrows in Figure 7a. More than 30,000 dolphins inhabit the East Sea [28], and the whistle sound made by these dolphins was recorded by the VLA; this suggested the presence of dolphins in the vicinity of the VLA during the experiment. In Figure 7a, the yellow box represents the BPSK signal. Because the click sound is an impulsive signal covering all frequencies, communication performance can be affected if it is included in the communication signal, as shown in Figure 7a. However, the effect of a small number of click sounds can be neglected. In this study, the effect of the click sounds was not considered, and their removal is outside the scope of the work presented herein. Therefore, three data packets that failed to decode owing to contamination by a large number of click sounds were excluded from the analysis. (6) Q−function:q(t)=∑i=1M,L,orNhi(t)∗hi(−t). The Q-function, as defined in Equation (6), is a metric used to assess the performance of the diversity method [6] and is expressed as the sum of the autocorrelation between channel impulse responses. As the order of diversity increases, the result of the Q-function more closely approximates that of the delta function, and the performance of the diversity method improves. The result of the Q-function is shown in Figure 7b for different orders of spatial diversity, where the result of each Q-function is the average over 15 packets. In Figure 7b, the blue, red, and black lines are averaged Q-functions when the order of diversity is 1, 8, and 16, respectively. The mainlobe of the Q-function is similar for the three orders of spatial diversity, but as the order of diversity increases, the sidelobes decrease and converge to zero. The convergence of the Q-function to the delta function means that the TRP output using the diversity method is close to the source signal from Equation (1). In other words, the BER and output SNR can be improved, as shown in Figure 8. The output SNR, SNRo, is defined as the reciprocal of the mean-square error between the information symbols and estimated symbols, as in Equation (7) [6]. (7) SNRo=1/Ee→k2=1/EIk−I^k2, where Ik and I^k are the kth information symbols and estimated symbols, respectively, and E denotes expectation. e→k is the difference between Ik and I^k, which is the noise for the kth symbol. As the source signal used in this study was modulated with binary-phase shift keying (BPSK), the number of bits and symbols are the same. The noise power is equal to the denominator in Equation (7). For BPSK, as the information symbols are on the unit circle, the source power is one and is the same as the numerator in Equation (7). That is, the output SNR is related to the distance between the information symbols and estimated symbols in the scatter plot. If the estimated symbols are close to the information symbols, the output SNR is high. Figure 8a–c show scatter plots for three values of the order of spatial diversity: 1, 8, and 16, respectively. Figure 7 and Figure 8 indicate that, as the order of diversity increases, the BER and the sidelobe of the Q-function decreases and the output SNR increases. These results are consistent with the results reported in [6]. When the maximum order of diversity in the spatial diversity method was used, error-free performance was achieved. 4.3. Performance Comparison Using Various Diversity Methods This section presents and compares the communication performance results according to the diversity method used. Because the diversity method depends on the channel variability, the performance may vary even for the same order of diversity. This implies that the performance would vary depending on which diversity method is selected. Figure 9 shows the variation in communication performance as the order of diversity increases for the three diversity methods. Figure 9a,b are the BER and output SNR results, respectively. Similar to the results presented in Section 4.2, in general, the BER decreased and the output SNR increased as the order of diversity increased for all methods. However, the performance depended on the diversity method. First, when the order of diversity is unity, the beam diversity method outperforms the spatial and temporal diversity methods. As mentioned in Section 2, the beam diversity method combines beamformed signals rather than received signals. Beamforming the received signal to the angle of the path can mitigate the effect of multipath time delay, which is known as spatial filtering [25,26,29]. When the order of diversity is unity, the effect of multipath time delay is not as strong as for the other diversity methods, and thus, the beam diversity method exhibits the best performance. However, because the beam diversity method does not have a large order of diversity, the difference in performance between an order of diversity of unity and the maximum order of diversity, is smaller than that observed for the other methods. In addition, the temporal and beam diversity methods exhibited similar performance in terms of output SNR (Figure 9b) based on when the maximum order of diversity was used, and the difference in output SNR was less than approximately 1 dB. However, in terms of BER (Figure 9a), the BER of the temporal diversity method was twice that of the beam diversity method [temporal diversity: 57/15,200 (=3.75×10−3), beam diversity: 24/14,250 (=1.68×10−3)]. Because the other organizations participating in the experiment designed communication signals with channel coding, we designed a communication signal that did not employ channel coding. Therefore, although the result of channel coding cannot be displayed, the black dashed line in Figure 9a shows the limit (3.8×10−3) of the forward error coding (FEC) scheme, which is a standard practice in undersea systems for evaluating the performance of communication [30]. For the spatial and beam diversity methods, a BER lower than the FEC limit [31] was achieved when the order of the diversity was three or more. At the maximum order, the BER for the temporal diversity method was somewhat lower than the FEC limit. If the FEC scheme can be used when the order of diversity is at its maximum, all diversity methods will provide error-free performance. These results can be interpreted using the co-diversity interference matrix shown in Figure 10. In this study, we defined and utilized the co-diversity interference matrix as a metric representing the channel variability between the two diversities (e.g., two receivers in the spatial diversity method). Each element of this matrix is a correlation coefficient of two channel impulse responses within each diversity method: (8) Co−diversityinterferencematrix:Qij=maxhi(t)∗hi(−t)1≤i,j≤M,L,orN. Because the diagonal element of Qij is itself, it has a maximum value (=1), and the off-diagonal elements represent channel similarity between different receivers (or times, beams) in the diversity method. The closer the value of the co-diversity interference matrix is to unity, the smaller the change in the channel. The off-diagonal term in Figure 10 is related to channel variability; the smaller is the value in the off-diagonal, the greater is the channel variability. In the spatial diversity method, the number of off-diagonal term elements lower than 0.5 is greater than that of the other two diversity methods; therefore, in the BLAC18 environment, the channel variability in terms of space is the greatest. Combining the TRP outputs with high variability increases the extent to which the sidelobes decrease. Therefore, the spatial diversity method was superior to the other methods. Figure 9 shows that the performance of the beam diversity method is best when the order of diversity is unity owing to the effect of the spatial filter. However, from the viewpoint of channel variability, the performance improvement with the beam diversity method is the least because of the small channel variability. According to the channel variability results, the channel variability of the temporal diversity method is between that of the spatial diversity and beam diversity methods, and as a consequence, the range of performance change is the second largest after that of the spatial diversity method. However, despite having the second largest variation, when the order of diversity is at its maximum, the performance of the temporal diversity method is inferior to that of the beam diversity method, indicating that the spatial factor plays a greater role than time. In reality, when operating underwater objects with a short aperture, the temporal diversity method may be more effective than spatial or beam diversity methods, although the data rate will be reduced. In other words, a trade-off exists between the data rate and the order of diversity. Figure 11 shows scatter plots of the temporal and beam diversity methods when the maximum order of diversity is utilized. In Figure 11, the red-, white-, and cyan filled circles represent information symbols, estimation symbols, and error symbols, respectively, and the yellow boxes represent the area around the origin. In Figure 11a,b, more symbols are close to the origin, which is the effect of the lowered SNR. The BPSK modulation technique is determined according to the sign of the real value of the symbol; hence, if many symbols have a real value close to the origin, the probability of an error increases. Because the output SNR of the temporal diversity method is lower than that of the beam diversity method, the estimated symbols were distributed more widely from the information symbols, and the number of error bits increased relatively. Most of the error symbols (i.e., cyan-filled circles) were confirmed to appear around the origin. 5. Conclusions In this paper, we reported the first long-range underwater acoustic communication experiment (60 km range) in the East Sea, which was conducted in 2018, and analyzed the communication performance results for various TRP-based diversity methods (spatial, temporal, and beam). As the performance of the diversity method depends on channel variability, even if the order of diversity is the same, a performance difference occurs according to the type of method. For the minimum order of diversity (=1), the performance of the beam diversity method using beamformed signals, which mitigate the effect of multipath time delay, was the best in terms of the two metrics (BER and output SNR). For the maximum order of diversity, the performance of the spatial diversity method was the best among the three diversity methods in that the output SNR with this method was the highest. Furthermore, from the viewpoint of the BER, the performance was error-free only with the spatial diversity method. This is because the order of diversity is large, and the channel variability in terms of space is the greatest. For BLAC18 data, due to the small channel variability, the performance of the temporal diversity method was lower than that of the spatial diversity method. If the interval between data packets increases, the performance of the temporal diversity method can be further improved, but the data rate reduces, indicating a trade-off between the order the diversity and data rate. In practice, nevertheless, for underwater objects (e.g., submarines) with a short aperture, the temporal diversity method would be more efficient than the spatial and beam diversity methods. The results of the temporal and beam diversity methods are not error-free for the maximum order of diversity, but the BERs of these two methods are less than 1%, demonstrating the feasibility of long-range underwater acoustic communication in the East Sea. Author Contributions Conceptualization, D.K. and J.K.; methodology, D.K.; software, D.K.; validation, D.K., J.K. and J.H.; formal analysis, D.K.; investigation, D.K.; resources, D.K.; data curation, D.K.; writing—original draft preparation, D.K.; writing—review and editing, J.K. and J.H.; visualization, D.K.; supervision, J.K.; project administration, J.K.; funding acquisition, J.K. All authors have read and agreed to the published version of the manuscript. Funding This work was supported by the Agency for Defense Development, South Korea, under Grant UD200010DD. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 (a) Schematic of the methods used to select the received signals. The spatial and temporal diversity methods utilize the received signals in the red and blue boxes, respectively. (b) Block diagram for passive time reversal processing (TRP) using the spatial or temporal diversity method. Figure 2 Block diagram illustrating time reversal processing using the beam diversity method. The beam diversity method has two processes: (1) beamforming the received signals (blue box) and (2) estimating the channel impulse response of the beamformed signal and using this response for time reversal processing (red box). Figure 3 (a) Bathymetry of the experimental area in the East Sea. Depth contours are in meters. The biomimetic long-range acoustic communication 2018 (BLAC18) experiment was conducted in October 2018. The five-point star indicates the source location and the magenta circle indicates the VLA location. (b) Schematic of the BLAC18 experimental setup. The red and blue solid circles indicate the receivers and sources, respectively. A 16-element, 42-m-long VLA moored to the sea floor in water with an approximate depth of 950 m recorded the communication signals. A positive angle denotes an upward ray path. The sound-speed profile was determined by measuring the CTD. Figure 4 Communication signal transmitted by the source during BLAC18. Each block denoted by the letter {A} (280-s long), consisting of six data packets denoted by the letter {B} (7.5-s long), was repeated every 55 min. Each data packet {B} consists of an LFM signal as a channel probe and a communication sequence. Blocks labeled {GT#1∼3} represent guard times that are 22.5-, 82.5-, and 2-s long, in this order. Figure 5 Conventional beamforming output using one data packet; (a) beamformed output as functions of the frequency (Hz) and grazing angle (deg). Aliasing occurred because of the carrier frequency, which is approximately 10 times the design frequency. (b) Beamformed output incoherently summed over the frequency band. (c) Enlargement of (b) from −5∘ to 15∘. The red lines represent the angles of all paths estimated from the channel impulse response (Figure 6). The angles cannot be estimated from beamforming because of aliasing and beam resolution. Figure 6 (a) Channel impulse response estimated at a range of 60 km from a linear frequency modulated (LFM) channel probe after matched-filtering. Four dominant paths are identified. (b) Beam-time migration of the channel impulse response corresponding to (a), respectively. The red dashed lines between the channel impulse response and the beam-time migration represent the relative time delays of the dominant paths, and the red circles on the beam-time migration plot represent the relative time delays and angles of the dominant paths. Figure 7 (a) Signal obtained by one receiver. The click sounds captured during the experiment are identified by red arrows. (b) Averaged Q-function as a function of the order of spatial diversity. The blue, red, and black lines correspond to orders of spatial diversity (M) of 1, 8, and 16, respectively. As the order of diversity increases, the sidelobes become smaller, i.e., the Q-function converges to the delta function. Figure 8 Scatter plots according to the order of spatial diversity (M): (a) M=1, (b) M=8, and (c) M=16. As the order of spatial diversity increases, the communication performance, in terms of BER and output SNR, improves. This is consistent with the sidelobe variation of the averaged Q-function. Figure 9 Variation in performance as a function of the order of diversity: (a,b) present the BER and output SNR results for the three diversity methods, respectively. The red, blue, and magenta solid lines with circles as markers represent the spatial, temporal, and beam diversity methods, respectively. To evaluate communication performance, the FEC limit (3.8×10−3) is indicated by the black dashed line with the BER result. The spatial diversity method produced the best performance. Figure 10 Co-diversity interference, Qij, for the three diversity methods. All plots show the correlation coefficient between the channel impulse responses estimated from two sets of data, i.e., the data obtained by two receivers from the point of view of the spatial diversity. The off-diagonal elements of Qij represent the channel variability within the diversity scheme. (a–c) show the Qij of the spatial, temporal, and beam diversity methods, respectively. Figure 11 Performance comparison of the temporal and beam diversity methods when the order of diversity is maximum; (a,b) show the performance of the beam and temporal diversity methods, respectively. The yellow boxes indicate the area around the origin, and cyan filled circles indicate symbols corresponding to errors. In this experiment, the beam diversity method, which utilizes the features of the spatial filter, outperformed the temporal diversity method. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Stojanovic M. Catipovic J. Proakis J.G. Coherent communications over long range underwater acoustic telemetry channels Acoustic Signal Processing for Ocean Exploration Moura J.M.F. Lourtie I.M.G. Springer Dordrecht, The Netherlands 1993 607 612 2. Stojanovic M. Catipovic J. Proakis J.G. Adaptive multichannel combining and equalization for underwater acoustic communications J. Acoust. Soc. Am. 1993 94 1621 1631 10.1121/1.408135 3. Plaisant A. Long range acoustic communications Proceedings of the IEEE Oceanic Engineering Society, OCEANS’98 Nice, France 28 September–1 October 1998 4. Song H.C. An overview of underwater time-reversal communication IEEE J. Ocean. Eng. 2015 41 644 655 10.1109/JOE.2015.2461712 5. Parvulescu A. Clay C.S. Reproducibility of signal transmissions in the ocean Radio Electron. Eng. 1965 29 223 228 10.1049/ree.1965.0047 6. Song H.C. Hodgkiss W.S. Kuperman W.A. Higley W.J. Raghukumar K. Akal T. Stevenson M. Spatial diversity in passive time reversal communications J. Acoust. Soc. Am. 2006 120 2067 2076 10.1121/1.2338286 7. Shimura T. Watanabe Y. Ochi H. Basic research on time-reversal waves in deep ocean for long acoustic communication Jpn. J. Appl. Phys. 2005 44 4722 10.1143/JJAP.44.4722 8. Shimura T. Ochi H. Watanabe Y. Hattori T. Experiment results of time-reversal communication at the range of 300 km Jpn. J. Appl. Phys. 2010 49 07HG11 10.1143/JJAP.49.07HG11 9. Mosca F. Matte G. Shimura T. Low-frequency source for very long-range underwater communication J. Acoust. Soc. Am. 2013 133 EL61 EL67 10.1121/1.4773199 23298019 10. Shimura T. Watanabe Y. Ochi H. Song H.C. Long-range time reversal communication in deep water: Experimental results J. Acoust. Soc. Am. 2012 132 EL49 EL53 10.1121/1.4730038 22779572 11. Shimura T. Ochi H. Watanabe Y. Hattori T. Demonstration of time-reversal communication combined with spread spectrum at the range of 900 km in deep ocean Acoust. Sci. Technol. 2012 33 113 116 10.1250/ast.33.113 12. Colosi J.A. Baggeroer A.B. Birdsall T.G. Clark C. Cornuelle B.D. Costa D. Dushaw B.D. Dzieciuch M.A. Forbes A.M.G. Howe B.M. A review of recent results on ocean acoustic wave propagation in random media: Basin scales IEEE J. Ocean. Eng. 1999 24 138 155 10.1109/48.757267 13. Song H.C. Kuperman W.A. Hodgkiss W.S. Basin-scale time reversal communications J. Acoust. Soc. Am. 2009 125 212 217 10.1121/1.3021435 19173408 14. Song H.C. Cho S. Kang T. Hodgkiss W.S. Preston J.R. Long-range acoustic communication in deep water using a towed array J. Acoust. Soc. Am. 2011 129 EL71 EL75 10.1121/1.3554707 21428470 15. Song H.C. Acoustic communication in deep water exploiting multiple beams with a horizontal array J. Acoust. Soc. Am. 2012 132 EL81 EL87 10.1121/1.4734242 22894320 16. Kang T. Song H.C. Hodgkiss W.S. Long-range multi-carrier acoustic communication in deep water using a towed horizontal array J. Acoust. Soc. Am. 2012 131 4665 4671 10.1121/1.4711009 22712939 17. Liu Z. Yoo K. Yang T.C. Cho S.E. Song H.C. Ensberg D.E. Long-range double-differentially coded spread-spectrum acoustic communications with a towed array IEEE J. Ocean. Eng. 2013 39 482 490 10.1109/JOE.2013.2264994 18. Song H.C. Hodgkiss W.S. Diversity combining for long-range acoustic communication in deep water J. Acoust. Soc. Am. 2012 132 EL68 EL73 10.1121/1.4731639 22894318 19. Ichiye T. Some problems of circulation and hydrography of the Japan Sea and the Tsushima Current Ocean Hydrodynamics of the Japan and East China Seas Ichiye T. Elsevier Amsterdam, The Netherlands 1984 Volume 39 15 54 0422-9894 20. Gamo T. Chemical Oceanography of the Japan Sea, a Unique Marginal Sea in the Northwestern Pacific Proceedings of the 2013 Asia Oceania Geosciences Society Annual Meeting Brisbane, Australia 24–28 June 2013 21. Kim T. Jo H.-J. Moon J.-H. Occurrence and Evolution of Mesoscale Thermodynamic Phenomena in the Northern Part of the East Sea (Japan Sea) Derived from Satellite Altimeter Data Remote Sens. 2021 13 1071 10.3390/rs13061071 22. Worcester P.F. Dzieciuch M.A. Mercer J.A. Andrew R.K. Dushaw B.D. Baggeroer A.B. Heaney K.D. D’Spain G.L. Colosi J.A. Stephen R.A. The North Pacific Acoustic Laboratory deep-water acoustic propagation experiments in the Philippine Sea J. Acoust. Soc. Am. 2013 134 3359 3375 10.1121/1.4818887 24116529 23. Park H. Kim D. Kim J. Hahn J. Park J. Performance improvement of long-range underwater acoustic communication in deep water using spatiotem-poral diversity J. Acoust. Soc. Korea 2019 38 587 592 24. Kim D. Park H. Kim J. Park J. Hahn J. Performance analysis of underwater acoustic communication based on beam diversity in deep water J. Acoust. Soc. Korea 2019 38 678 686 25. Van Veen B.D. Buckley K.M. Beamforming: A versatile approach to spatial filtering IEEE ASSP Mag. 1988 5 4 24 10.1109/53.665 26. Eom M. Park J. Ji Y. Kim J.S. Mitigation of inter-symbol interference in underwater acoustic communication using spatial filter J. Acoust. Soc. Korea 2014 33 48 53 27. Van Trees H.L. Optimum Array Processing: Part IV of Detection, Estimation, and Modulation Theory John Wiley & Sons Hoboken, NJ, USA 2004 154 196 0471463833 28. Kim J.H. Taxonomic Status of Common Dolphin in the East Sea Master’s Thesis Pukyong National University Busan, Korea 2015 29. Howe G.S. Tarbit P.S.D. Hinton O.R. Sharif B.S. Adams A.E. Sub-seas acoustic remote communications utilising an adaptive receiving beamformer for multipath suppression Proceedings of the IEEE Oceanic Engineering Society, OCEANS’94 Brest, France 13–16 September 1994 30. Cai Y. Ramanujam N. Morris J.M. Adali T. Lenner G. Puc A.B. Pilipetskii A. Performance limit of forward error correction codes in optical fiber communications Proceedings of the Optical Fiber Communication Conference Anaheim, CA, USA 17–22 March 2001 31. Cho J. Xie C. Winzer P.J. Analysis of soft-decision FEC on non-AWGN channels Opt. Express 2012 20 7915 7928 10.1364/OE.20.007915 22453465

PMC009xxxxxx/PMC9099689.txt

==== Front J Clin Med J Clin Med jcm Journal of Clinical Medicine 2077-0383 MDPI 10.3390/jcm11092568 jcm-11-02568 Systematic Review Bracket Transfer Accuracy with the Indirect Bonding Technique—A Systematic Review and Meta-Analysis https://orcid.org/0000-0002-9535-6901 Sabbagh Hisham 1*† Khazaei Yeganeh 2† https://orcid.org/0000-0002-4693-6969 Baumert Uwe 1 Hoffmann Lea 1 Wichelhaus Andrea 1 Janjic Rankovic Mila 1 Nissan Joseph Academic Editor Chaushu Gavriel Academic Editor 1 Department of Orthodontics and Dentofacial Orthopedics, University Hospital, LMU Munich, Goethestrasse 70, 80336 Munich, Germany; uwe.baumert@med.uni-muenchen.de (U.B.); lea.hoffmann@med.uni-muenchen.de (L.H.); kfo.sekretariat@med.uni-muenchen.de (A.W.); mila.janjic@med.uni-muenchen.de (M.J.R.) 2 Statistical Consultation Unit, StaBLab, Department of Statistics, LMU Munich, 80799 Munich, Germany; yeganekhazaei@gmail.com * Correspondence: hisham.sabbagh@med.uni-muenchen.de; Tel.: +49-89-4400-53223 † These authors contributed equally to this work. 04 5 2022 5 2022 11 9 256825 3 2022 30 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Purpose: To investigate the bracket transfer accuracy of the indirect bonding technique (IDB). Methods: Systematic search of the literature was conducted in PubMed MEDLINE, Web of Science, Embase, and Scopus through November 2021. Selection Criteria: In vivo and ex vivo studies investigating bracket transfer accuracy by comparing the planned and achieved bracket positions using the IDB technique were considered. Information concerning patients, samples, and applied methodology was collected. Measured mean transfer errors (MTE) for angular and linear directions were extracted. Risk of bias (RoB) in the studies was assessed using a tailored RoB tool. Meta-analysis of ex vivo studies was performed for overall linear and angular bracket transfer accuracy and for subgroup analyses by type of tray, tooth groups, jaw-related, side-related, and by assessment method. Results: A total of 16 studies met the eligibility criteria for this systematic review. The overall linear mean transfer errors (MTE) in mesiodistal, vertical and buccolingual direction were 0.08 mm (95% CI 0.05; 0.10), 0.09 mm (0.06; 0.11), 0.14 mm (0.10; 0.17), respectively. The overall angular mean transfer errors (MTE) regarding angulation, rotation, torque were 1.13° (0.75; 1.52), 0.93° (0.49; 1.37), and 1.11° (0.68; 1.53), respectively. Silicone trays showed the highest accuracy, followed by vacuum-formed trays and 3D printed trays. Subgroup analyses between tooth groups, right and left sides, and upper and lower jaw showed minor differences. Conclusions and implications: The overall accuracy of the indirect bonding technique can be considered clinically acceptable. Future studies should address the validation of the accuracy assessment methods used. bracket bonding indirect bonding orthodontic brackets transfer accuracy bracket positioning bonding accuracy bonding tray This research received no external funding. ==== Body pmc1. Introduction The straight-wire technique derived from the works of Andrews [1,2] is the most commonly used technique in fixed orthodontic treatment [3]. In this technique, the ideal placement of the brackets is of utmost importance [4,5,6,7]. Positioning errors necessitate the repositioning of brackets or the insertion of additional compensatory bends [4,8,9,10,11,12,13,14] and increase the number of visits and the treatment duration [5], thus compromising treatment efficiency. Clinically, brackets can be positioned directly with an instrument or indirectly with a transfer tray. Indirect bonding (IDB) was first proposed in 1972 [15] and has since been used mainly to improve accuracy through pre-planning the ideal bracket position [7]. Numerous studies have shown that IDB can increase the precision of bracket placement [8,16,17,18,19,20], but neither the direct nor the indirect technique achieves ideal clinical results, and readjustments remain necessary [3,21,22,23,24]. More recently, with the introduction of software for virtual treatment planning and workflows for additive transfer tray manufacturing for IDB, another approach for ideal bracket placement was introduced [25]. By calculating and visualizing the tooth movements resulting from the application of the virtually positioned brackets, adjustments can be made to realize the treatment objectives in the digital setup [13]. Accurate clinical implementation of the planned bracket positions is crucial in this method to achieve the virtually simulated alignment [26]. A growing number of studies have addressed the topic of IDB accuracy [26,27,28]. There is, however, great variability in the reported results between studies, which might be due to underlying methodological or clinical heterogeneity. Thus, the aim of this study was to synthesize the findings and assess the accuracy of the IDB technique, focusing not only on the overall accuracy of the method or different types of indirect bonding trays but also taking into account methodological and clinical aspects such as the method used to evaluate accuracy, and tooth-type-specific and jaw-related differences. 2. Materials and Methods This systematic review was conducted according to the “Preferred Reporting Items for a Systematic Review and Meta-analysis of Diagnostic Test Accuracy Studies” (PRISMA-DTA) statement [29] and registered at the PROSPERO platform (registration number: CRD42021243227). The PICO model (problem/patient, intervention, comparison, outcome) was followed to define the research question and eligibility criteria [30]. Detailed information on how this model influenced the study design, and the definition of each PICO element can be found in Supplementary Table S1. 2.1. Eligibility Criteria Prospective and retrospective in vivo and ex vivo studies investigating bracket transfer accuracy by comparing the planned and achieved bracket positions for buccal bracket bonding were considered. The following eligibility criteria were applied. (1) At least one of the measurements in the linear (mesiodistal, buccolingual, vertical) and/or angular (angulation, rotation, torque) directions was reported. (2) Actual status of the bracket position was confirmed by comparing it to the planned bracket position. Studies assessing lingual bracket bonding accuracy were not considered for inclusion. Only studies published in English were considered, and the last update of the search according to the search strategy was performed on 1 November 2021. 2.2. Literature Search and Study Selection Process Based on the research question and the aforementioned eligibility criteria, a search strategy was developed. Following the Cochrane recommendations for studies dealing with very specific topics, such as indirect bonding, we applied the following concept and broke it into three sub-concepts in order to create our search strategy [31] (Table 1). This template was applied to four bibliographic databases (PubMed, Embase, Web of Science, and Scopus) with specific adaptations for each bibliographic database (Table 2). Sets of records from each database were downloaded to the bibliographic software package EndNote X9 (Clarivate Analytics, Philadelphia, PA, USA) and merged into one core database in order to remove duplicate records. All records identified by the searches were primarily checked on the basis of title and abstract. Full texts of the records identified as relevant were then downloaded and checked for meeting the eligibility criteria. The articles that did not meet the predefined inclusion criteria after the full-text assessment were excluded from further examination. The whole literature screening process was conducted independently in parallel by two of the authors (H.S, M.J.R). The Cohen’s K coefficient for agreement between the two reviewers was 0.89. Any doubts or disagreements were solved by discussion. 2.3. Data Extraction Data from the included studies were extracted by both reviewers in specially prepared data extraction sheets. Any differences in extracted data were resolved through discussion until reaching a consensus. Briefly, the following information was extracted from papers: author and year of publication, study design, number of assessed teeth (incisors, canines, premolars, molars); patient information in case of in vivo studies; IDB technique used (double polyvinyl siloxane (double-PVS); double vacuum-form (double-VF), polyvinyl siloxane vacuum-form (PVS-VF), polyvinyl siloxane putty (PVS-putty), and single vacuum-form (single-VF)); type of brackets used in the study; method for measuring transfer accuracy (digital photography, calipers, CBCT, 3D-scan and superimposition); mean transfer errors (MTE) in linear (mesiodistal, buccolingual, vertical) and angular (angulation, rotation, torque) directions expressed in millimeters (mm) and degrees (°). All corresponding authors of the included studies were contacted to provide the complete data sets or additional data if available. For included studies reporting data only graphically [10,32], data were collected using a data extraction software (WebPlotDigitzer, Version 4.4, Pacifica, CA, USA) as described and validated by Drevon et al. [33]. All data were later transferred to Excel spreadsheets (Excel 2010, Microsoft Corporation, Redmond, WA, USA). The data transfer was checked twice by both reviewers involved before further analysis. 2.4. Risk of Bias Assessment in Included Studies In this review, an adapted risk of bias (RoB) assessment tool was used (Supplementary Table S3) [34,35]. The tool contained four domains (selection bias; reference test bias; verification bias; outcome bias), each of them included items that cover different sources of bias. One of the following three modalities was used to judge the RoB in the primary studies: high, low, or unclear risk of bias. The category “unclear RoB” was applied whenever incomplete details or no information could be found in the study. RoB assessment was performed independently by the two of the authors (H.S., M.J.R). 2.5. Meta-Analysis and Synthesis of Results Meta-analysis was performed using R Statistical Software (Version 4.1.1, R Core Team, Vienna, Austria) according to published procedures [36,37]. To be included in the meta-analysis, the sample size and the mean and standard deviation (SD) of the bracket transfer error expressed in millimeters (mm) or degrees (°) were required. The overall mean transfer errors (MTE) and further subgroup analyses in linear (mesiodistal, buccolingual, vertical) and angular (angulation, rotation, torque) directions were performed in the following categories: overall MTE; tooth group related MTE; jaw-related MTE (left vs. right/upper vs. lower); MTE in relation to accuracy assessment method; MTE in relation to the type of IDB tray. Data wrangling and manipulation were performed using the statistical packages “tidyverse” [37], “dplyr” [38], and “ggplot2” [39]. Meta-analytic syntheses and further investigations were performed by “meta” and “dmetar” in RStudio (Rstudio Inc., Boston, MA, USA) [36,40]. Effect sizes of the overall MTE and subgroup analyses were calculated by the metamean function provided by “meta” and are reported in Table 2. Heterogeneity was assessed using Cochran’s Q and I2-statistics. A random-effects model was retained to pool effect sizes to better account for the differences in design amongst the included studies for both overall category and subgroups analysis. The restricted maximum likelihood estimator was used to calculate the heterogeneity variance τ2 [41]. Knapp–Hartung adjustments were used to calculate the confidence interval around the pooled effect [42]. To investigate publication bias, funnel plots were prepared using the functionalities of the “meta” package. Additionally, drapery plots were produced based on p-value functions. 3. Results 3.1. Literature Search Results The PRISMA workflow illustrating the whole study selection process is summarized in Figure 1. The electronic search resulted in 218 records from PubMed, 187 records from Web of Science, 101 records from EMBASE, and 125 records from Scopus. After duplicate elimination, altogether, 312 studies were identified. Upon checking the titles and abstracts of the identified records, 35 studies were selected for full-text reading. Studies that did not meet the eligibility criteria (n = 19) were excluded from further assessment, and the reasoning is summarized in Supplementary Table S2. Additionally, one more study was selected for inclusion by cross-checking the reference lists of literature selected for inclusion, resulting in a total number of 16 included studies. For two publications [7,28], additional data that was not included in the original manuscripts were provided by the respective authors. 3.2. Results of the Risk of Bias Assessment The overall risk of bias (RoB) of the different domains and items is given in Figure 2. Results of the RoB assessment of the individual studies are available in Supplementary Table S5. In eight studies, indirect bracket placement might have been affected by malocclusions, such as severe crowding or rotations, or no such information was provided [7,18,28,43,44,45,46,47]. Only seven studies provided information on sample size calculation [26,27,28,44,46,47,48]. Three studies considered only specific tooth groups in their investigations [44,45,48]. Nine studies did not report the experience and training of the bonding clinicians or indicated low experience [7,10,26,32,43,44,45,47,48], and in three studies, the bonding clinicians’ experience was unclear based on the provided information [27,46,49]. None of the included studies provided information on calibration, and only two studies provided information on blinding of the examiners [47,48]. Eight studies had a high or unclear risk of bias due to an insufficient method for reproducibility assessment or insufficient reporting [7,10,18,46,47,48,49,50]. 3.3. Study Characteristics and Results of Individual Studies The characteristics of the studies included for quality assessment are illustrated in Table 3. 3.3.1. Study Characteristics and Results of the In Vivo Studies Not Included in the Quantitative Synthesis Four in vivo studies were eligible for quality assessment after full-text reading [7,26,47,48]. Two of the studies investigated the bracket transfer accuracy of 3D printed trays [26,47], one of which compared 3D printed trays to silicone trays [47]. The other two studies investigated silicone trays [46] and vacuum-formed trays [48]. All included in vivo studies investigated the accuracy of bracket transfer with conventional brackets. In these studies, three different methods were used to evaluate accuracy: CBCT and software [7], photography [48], and scans and software [26,47]. Due to the small number of in vivo studies with different study characteristics, they were not included in the quantitative synthesis. The reported linear mean transfer errors ranged from 0.001 to 0.050 mm. The angular mean transfer errors ranged from 0.001 to 1.757°. The full extracted data is available in in Supplementary Tables S4.1–S4.6. 3.3.2. Study Characteristics of the Ex vivo Studies Included in the Quantitative Synthesis A total of 12 ex vivo studies were eligible for quality assessment after full text reading [10,18,27,28,32,43,44,45,46,49,50,51]. Of these, 7 studies investigated 3D printed trays [27,28,32,43,44,45,50], while 5 investigated vacuum-formed trays [10,32,46,49,51], and 4 studies investigated silicone trays [10,18,27,46], with 4 of the 12 included studies comparing more than 1 material group [10,27,32,46]. Three studies used self-ligating brackets for indirect bonding [43,44,50]. The most common method of analysis was the use of scans and software (n = 9) [10,27,28,32,43,44,45,46,49,50], followed by methods using photography (n = 3) [10,18,51]. 3.4. Results of the Meta-Analysis The results of the meta-analysis are summarized in Table 4. The overall linear and angular mean bracket transfer errors are shown in forest plots in Figure 3. The full data sets, including forest plots, drapery plots, and funnel plots for different analysis groups, are available in Supplementary Tables S6.1–S8.12. 3.5. Linear Mean Transfer Errors Overall linear mean transfer errors (MTE) in mesiodistal, buccolingual, and vertical directions were 0.08 mm, 0.09 mm, and 0.14 mm, respectively (Table 4). A comparison of linear MTE between different tooth groups revealed that IDB was less accurate in the incisor group, with an MTE of 0.14 mm in the buccolingual direction and 0.15 mm in the vertical direction. No significant differences could be observed in a comparison of IDB transfer accuracy between left and right sides in all three linear directions. The comparison between the upper and lower jaw showed slightly higher bracket transfer accuracy in the upper jaw in the mesiodistal and vertical directions (MTE 0.10 mm, 0.18 mm), whereas accuracy in the buccolingual direction was lower than in the lower jaw (MTE 0.09 mm). Among the different types of IDB trays, 3D printed trays showed the highest accuracy in the mesiodistal (MTE 0.06 mm) and vertical directions (MTE 0.12 mm) but the lowest accuracy in the buccolingual dimension (MTE 0.10 mm). In studies that used photography as a method to assess accuracy, the MTE was higher in the mesiodistal and vertical directions (MTE 0.12, 0.22) than in studies that used 3D assessment methods (MTE 0.06, 0.11). 3.6. Angular Mean Transfer Errors Overall angular mean transfer errors (MTE) regarding angulation, rotation, and torque were 1.13°, 0.93°, and 1.11°, respectively. Compared to the other tooth groups, molar tubes showed the highest transfer accuracy in rotation (MTE 0.69°) but the lowest in torque (MTE 2.29°). In the premolar group, the highest accuracy was observed for angulation (MTE 0.13°) and torque (0.95°), while rotation (MTE 1.46°) showed the lowest accuracy compared to the other tooth groups. The comparisons between the left and right sides, between upper and lower jaws, and between 3D accuracy assessment and photography could only be partially evaluated on the basis of the available data for angular values. Studies that used photography as a method showed a lower bracket transfer accuracy for angulation (MTE 2.74°) compared to studies that used 3D assessment (MTE 0.95°). IDB showed higher accuracy regarding angulation in the upper jaw (MTE 1.26°) but lower accuracy for rotation (MTE 0.59°) and torque (0.73°). For 3D printed trays, higher torque deviations were observed (MTE 1.42°) than for other types of IDB trays. For silicone trays, the highest accuracy was observed for angulation (MTE 0.66°) and torque (0.79°). 4. Discussion 4.1. Overall In this study, the available literature on the indirect bonding technique was systematically reviewed regarding the accuracy of bracket transfer and differences among available methods to draw conclusions on methodological and clinical aspects. The results of the meta-analysis showed an overall bracket transfer accuracy for the indirect bonding technique between 0.08 and 0.14 mm for linear and 0.93° and 1.13° for angular deviations, respectively. As there are no evidence-based limits for clinically acceptable bracket position deviations in the literature, most studies refer to the professional standards of the American Board of Orthodontics of 0.5 mm for linear and 2° for angular deviations [7,10,26,27,32,45,46,49,52]. However, these limits apply by definition to deviations of tooth positions. As full slot engagement with orthodontic archwires cannot be achieved in the straight-wire technique [27,53,54,55], exceeding these limits cannot be equated with malpositioning of the associated teeth. In view of these considerations and the limitations due to the current reference standard, the overall accuracy of the indirect bonding technique can be considered clinically acceptable. Regarding linear deviations, a higher mean transfer error was observed for the vertical direction than for the mesiodistal and buccolingual directions, which is in line with previous studies [10,46,48,56] and mostly attributed by the authors to misfit phenomena of the indirect bonding trays. Therefore, it has been proposed to increase the distance between the dentition and the transfer tray by adapted designs to improve the fit and reduce vertical deviations [26]. Angular deviations (torque, rotation, and angulation), on the other hand, showed comparable values, although deviations for torque were reported to be highest in previous studies [26,27,28,32,57]. It is possible that the angular deviations are more dependent on the amount of adhesive, tray material, and tray design, and therefore different results are observed in the respective studies depending on the method used [28]. 4.2. Tooth Groups Subgroup analysis by tooth groups showed the lowest angular deviations in the premolar group for all directions but rotation, where transfer was most accurate for molar attachments. Interestingly, linear bracket transfer errors were higher for anterior teeth (incisors and canines) than for posterior teeth (premolars and molars), contrary to previous findings [7,32,57]. The high rotational accuracy of molar attachments could be explained by the larger mesio-distal extension compared to the attachments of other tooth groups. However, the overall differences between the tooth groups in the included ex vivo studies were small and likely to be clinically negligible. 4.3. Side Differences and Differences between Upper and Lower Jaw It is considered that one of the advantages of the indirect bonding technique is that it allows for consistent accuracy in bracket placement, regardless of the practitioner’s handedness or direction of viewing direction and sitting position [32]. However, only a few studies that met the inclusion criteria provided accuracy data separately for the right and left sides, and for the upper and lower jaws, so only limited conclusions can be drawn. Based on data from five studies included in the meta-analysis, no differences in bracket transfer accuracy were found between the right and left sides. In contrast, slightly higher bracket transfer accuracy was found for the upper jaw than for the lower jaw. This result should be interpreted with caution, as it may be biased by the limited number of included studies providing accuracy data for the lower jaw. 4.4. Tray Materials Regarding tray materials, silicone trays represent the reference in terms of accuracy [10,27,46,56]. In previous studies that compared 3D printed trays with other methods, 3D printed trays were found to have a higher bracket transfer accuracy than vacuum-formed trays [32] but lower than silicone trays [27,57]. Interestingly, in this study, 3D printed trays showed lower MTE in the mesiodistal and vertical directions and in angulation compared with the other tray material groups. The use of 3D-printed trays has been suggested to potentially increase treatment efficiency by improving treatment planning through digital setup, treatment simulation, implementation of 3D imaging data such as CBCT or MRI, and by simplifying the laboratory process [11,13,26]. However, further research is necessary to determine the influence of factors like tray design [26,32], material used [57], and manufacturing process. 4.5. Accuracy Assessment Method Included studies using photography as a method for accuracy assessment showed a higher MTE in comparison to studies using 3D assessment. 3D assessment methods for bracket transfer accuracy using scanners or CBCT have been proposed to generally achieve higher accuracy [26,27]. However, most of the included studies did not adequately evaluate the accuracy of the assessment workflow or did not report all relevant reliability data. The sole use of Dahlberg’s formula, intraclass correlation coefficient, or analysis using a paired t-test for reliability reporting in orthodontic research is not adequate [58]. Furthermore, Jungbauer et al. [28] questioned the suitability of intraoral scanners for accurately determining the bracket transfer accuracy because of significant artifacts on scanned brackets and low intra- and inter-rater reliability in their study. The use of photographic methods, on the other hand, has the disadvantage that not all deviation directions can be evaluated. Accurate registration of achieved bracket positions is a technical challenge, which may partly explain why ex vivo studies are predominantly available on this topic. Despite the methodological limitations discussed, scans, photographs, and micro-Ct data appear to be suitable, in principle, for the assessment of IDB accuracy. However, adequate validation of the accuracy assessment method is required to reduce the risk of bias in future studies and to support more targeted research, in which the accuracy values obtained may be useful to practitioners with respect to the clinical protocols. Finally, all relevant data should be made available in future studies to allow for more comprehensive reviews. 5. Strengths and Limitations To the authors’ knowledge, to date, no systematic review has comprehensively addressed the assessment of bracket transfer accuracy, including methodological and clinical aspects of the IDB method. In addition, the number of available studies without standards on the methodological aspects of assessment, validation, and reporting is increasing, which limits the validity and generalizability of the results. However, it should not be neglected that conducting a meta-analysis with a small number of available studies is also subject to limitations. As we anticipated considerable between-study heterogeneity, a random-effects model was used to pool effect sizes. The results of τ2, I2-statistics, and the corresponding p-values indicated that between-study heterogeneity existed in most of the categories and that the use of a random-effects model was appropriate. Nevertheless, the results of the subgroup analyses should be interpreted with caution. The statistical power of small subgroups is limited because the effects are smaller than in the meta-analysis performed for the overall group [36]. Moreover, a controversy around p-value and the sole use of forest plots to visualize results of meta-analyses is rising [59]. Forest plots can only display confidence intervals with the assumption of a fixed significance threshold (p-value < 0.05). Therefore, in this study, we used drapery plots in addition to forest plots. Drapery plots that present the p-value function for all individual studies are suggested as being complementary figures to forest plots for presentation and interpretation of the results of a meta-analysis, specifically with a low number of studies, such as our study [60]. This prevents researchers from solely relying on the p-value < 0.05 significance threshold when interpreting the results. The resulting drapery plots are documented in Supplementary Tables S8.1–S8.12. Due to the low number of in vivo studies (n = 4), with significant differences in the applied methodologies and an extensive and sometimes contradictory range of published results, a meta-analysis could only be carried out for ex vivo studies. The bracket transfer accuracy in in vivo settings could be lower due to limited accessibility of the oral cavity [46], moisture control and soft-tissue interference [32], patient management [45], malocclusion [10], and other factors. Therefore, further methodologically sound in vivo studies are necessary to evaluate the accuracy of the indirect bonding technique in clinical settings. 6. Clinical Implications Accurate bracket placement is essential for effective and efficient treatment with fixed orthodontic appliances [1,7,10,26]. However, due to the complexity of the various clinical and technical aspects of bracket bonding and despite the large number of studies dealing with this topic, there is disagreement on the most appropriate techniques or methods [16]. Objective evidence from well-conducted, prospective, randomized clinical trials is still lacking [16,61]. The findings of this systematic review suggest that indirect bonding as a technique allows achieving planned bracket positions with high overall accuracy, even though the results addressed herein are not sufficient to reflect all of the various clinical aspects. It was shown that using indirect bonding, tooth-type-specific and jaw-related differences appear to have a rather negligible overall influence on accuracy. In contrast to previously published studies [27,47], indirect bracket positioning with 3D printed trays generally appears to be as accurate as silicone trays. Therefore, the selection of one of these techniques could be based on preferences or criteria such as fabrication cost, time, or cost-effectiveness, even though the reduced number of manufacturing steps and further advances in computer-aided technologies will likely favor 3D-printed trays [61]. Indirect bonding remains more time and cost-consuming overall than direct bonding due to the laboratory process required, although it has been shown to reduce clinical chair time [8,61,62]. Further research is needed to evaluate the correlation between the accuracy of bracket placement and the need for compensatory bends, bracket repositioning, and reduction in total treatment time, given the conflicting results to date [5,61,63,64]. 7. Conclusions The results of this meta-analysis indicate a generally precise implementation of planned bracket positions in the indirect bonding technique. Among tray materials, silicone trays and 3D printed trays showed higher accuracy compared to vacuum-formed trays. Subgroup analyses between tooth groups, right and left sides, and upper and lower jaw showed only minor differences. In addition to the main objectives, future studies should address the validation of the accuracy assessment methods and provide complete data sets, including adequate reliability data, to reduce the risk of bias. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/jcm11092568/s1, Supplementary Tables S1–S9. Additional references [65,66,67,68,69,70,71,72,73,74,75,76,77,78,79] are cited in the supplementary content. Click here for additional data file. Author Contributions H.S. Reviewer 1, original draft preparation. M.J.R. Reviewer 2. Y.K. statistical analysis and meta-analysis. A.W. Conceptualisation. L.H. interpretation of the results. U.B. data verification, review, and editing. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Flow diagram of information through the different phases of a systematic review according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (study selection process). Figure 2 Overview of the overall RoB among different domains and items. Figure 3 Forest plots showing overall linear and angular mean bracket transfer errors (MTE) [10,18,27,28,32,43,44,45,46,49,50,51]. jcm-11-02568-t001_Table 1 Table 1 Concept of the search strategy. Domain Search Term Field orthodont* AND Intervention bonding AND Outcome positioning differences OR accuracy OR transfer accuracy OR ideal bracket placement OR accurate bracket positioning OR accurat* jcm-11-02568-t002_Table 2 Table 2 List of adapted search strategies used for different databases and number of identified records. Database Search Strategies Results PubMed orthodont* [All Fields] AND bonding [All Fields] AND ((positioning [All Fields] differences [All Fields]) OR accuracy [All Fields] OR (transfer [All Fields] accuracy [All Fields]) OR (ideal [All Fields] bracket [All Fields] placement [All Fields]) OR (accurate bracket [All Fields] positioning [All Fields]) [All Fields] OR accurat* [All Fields]) 218 Embase orthodont*.mp. AND bonding.mp. AND ((positioning differences).mp. OR accuracy.mp. OR (transfer accuracy).mp. OR (ideal bracket placement).mp. OR (accurate bracket positioning).mp. OR accurat*.mp.) 101 Web of Science orthodont* AND bonding AND (positioning differences OR accuracy OR transfer accuracy OR ideal bracket placement OR accurate bracket positioning OR accurat*) 187 Scopus TITLE-ABS-KEY (orthodont* AND bonding AND (“positioning differences” OR “positioning difference” OR accurac* OR “transfer accuracy” OR “ideal bracket placement” OR “ideal bracket placements” OR “accurate bracket positioning” OR accurat*)) 125 Total 312 jcm-11-02568-t003_Table 3 Table 3 Study characteristics of the included studies. Study Details Sample Details Bonding Procedure (Indirect) Transfer Accuracy Assessment Author (Year) Type of Study Sample Size Calculation/Method No. of Assessed Brackets No. of Bonding Clinicians Type of IDB Tray Bonded Subject (s)/Object (s) Data for Reference Model(s) Tray Construction Type of Brackets No. of Examiners Measuring Method Total/I/C/PM/M Jungbauer et al. [28], 2021 ex vivo Yes 280/80/40/80/80 NR 3D printed (soft) bonding on plaster or printed model impression Virtual model, Rapid prototyping conventional NR Scan + Software 280/80/40/80/80 3D printed (hard) bonding on plaster or printed model impression Virtual model, Rapid prototyping Park et al. [43], 2021 ex vivo No 506/147/79/122/158 1 3D printed bonding on plaster or printed model model scan Virtual model, Rapid prototyping self-ligating 1 Scan + Software Park et al. [44], 2021 ex vivo Yes 225/NR 1 3D printed bonding on plaster or printed model model scan Virtual model, Rapid prototyping self-ligating 1 Scan + Software Faus-Matoses et al. [50], 2021 ex vivo No 335/NR NR 3D printed bonding on plaster or printed model scan Virtual model, Rapid prototyping self-ligating NR Scan + Software Niu et al. [32], 2021 ex vivo Yes 108/37/10 19/32/20 NR 3D printed bonding on plaster or printed model intraoral scan Virtual model, Rapid prototyping conventional NR Scan + Software Yes 104/31/18/35/20 NR Vacuum Form bonding on plaster or printed model intraoral scan Virtual model, Rapid prototyping conventional NR Scan + Software Süpple et al. [49], 2021 ex vivo No 729/210/107/207/205 NR Vacuum Form (group H) bonding on plaster or printed model scan Virtual model, Rapid prototyping conventional NR Scan + Software No 724/209/106/206/203 Vacuum Form bonding on plaster or printed model scan Model and laboratory process conventional NR Scan + Software (group V) Pottier et al. [27], 2020 ex vivo Yes 97/38/20/39/- 1 Silicone bonding on plaster or printed model intraoral scan Virtual model, Rapid prototyping conventional 1 Scan + Software Yes 98/40/19/39/- 3D printed tray bonding on plaster or printed model intraoral scan Virtual model, Rapid prototyping conventional 1 Scan + Software Kalra et al. [51], 2018 ex vivo No 100/20/10/20/0 5 Vacuum Form bonding on plaster or printed model impression Model cast and laboratory process conventional NR Photography Kim et al. [45], 2018 ex vivo No 60/-/-/40/20 1 3D printed tray bonding on plaster or printed model model scan Virtual model, Rapid prototyping conventional NR Scan + Software 30/-/-/20/10 No 60/-/-/40/20 3D printed tray bonding on plaster or printed model model scan Virtual model, Rapid prototyping conventional NR Scan + Software 30/-/-/20/10 Schmid et al. [46], 2018 ex vivo Yes 132/54/24/54/- 1 Silicone bonding on plaster or printed model impression Model cast and laboratory process conventional NR Scan + Software Yes 134/52/29/53/- 1 Vacuum form bonding on plaster or printed model impression Model cast and laboratory process conventional NR Scan + Software Castilla et al. [10], 2014 ex vivo No 296/98/50/98/50 NR Double PVS bonding on plaster or printed model impression Model cast and laboratory process conventional NR Photography, digital caliper 60/20/10/20/10   No 296/98/50/98/50 PVS putty bonding on plaster or printed model impression Model cast and laboratory process conventional NR Photography, digital caliper 60/20/10/20/1 No 296/98/50/98/50 PVS-VF bonding on plaster or printed model impression Model cast and laboratory process conventional NR Photography, digital caliper 60/20/10/20/10 No 296/98/50/98/50 Double Vacuum Form bonding on plaster or printed model impression Model cast and laboratory process conventional NR Photography, digital caliper 58/20/10/18/10 No 296/98/50/98/50 Single Vacuum Form bonding on plaster or printed model impression Model cast and laboratory process conventional NR Photography, digital caliper 58/18/10/20/10 Koo et al. [18], 1999 ex vivo No 180/72/26/72/0 9 Silicone bonding on plaster or printed model impression Model cast and laboratory process conventional NR Photography Chaudhary et al. [47], 2021 in vivo Yes 300/120/60/120/0 NR 3D printed bonding on patient intraoral scan Virtual model, Rapid prototyping conventional NR Scan + Software Yes 300/120/60/120/0 PVS bonding on patient intraoral scan Model cast and laboratory process conventional NR Scan + Software Xue et al. [26], 2020 in vivo Yes 205/71/36/62/36 1 3D printed tray digital or virtual bonding procedure intraoral scan Virtual model, Rapid prototyping conventional NR Scan + Software Grünheid et al. [7], 2016 in vivo No 136/54/26/46/10 4 Silicone Bonding on patient impression Model cast and laboratory process conventional 1 CBCT + Software Hodge et al. [48], 2004 in vivo Yes 156/104/52/0/0 NR Vacuum Form Bonding on patient impression Model cast and laboratory process conventional NR Photography, acetate copies jcm-11-02568-t004_Table 4 Table 4 Summary of the results of the meta-analysis. MTE, mean transfer errors. Analyzed Parameters Mesiodistal Buccolingual Vertical Angulation Rotation Torque Overall accuracy n 23 21 23 20 10 10 MTE (95% CI) 0.08 (0.05; 0.10) 0.09 (0.06; 0.11) 0.14 (0.10; 0.17) 1.13 (0.75; 1.52) 0.93 (0.49; 1.37) 1.11 (0.68; 1.53) Prediction interval [−0.05; 0.20] [−0.04; 0.21] [−0.02; 0.30] [−0.61; 2.87] [−0.88; 2.74] [−0.61; 2.83] Tooth group comparison Incisors n 14 12 14 14 8 12 MTE (95% CI) 0.09 (0.05; 0.12) 0.14 (0.07; 0.21) 0.15 (0.10; 0.20) 1.43 (0.97; 1.89) 0.74 (0.43; 1.05) 1.63 (0.95; 2.32) Prediction interval [−0.04; 0.22] [−0.11; 0.40] [−0.09; 0.39] [−0.32; 3.18] [−0.18; 1.66] [−0.81; 4.08] Canines n 14 12 14 14 8 12 MTE (95% CI) 0.09 (0.05; 0.13) 0.13 (0.07; 0.19) 0.15 (0.09; 0.24) 1.95 (1.15; 2.75) 0.90 (0.47; 1.32) 2.11 (1.13;3.09) Prediction interval [−0.04; 0.22] [−0.09; 0.34] [−0.09; 0.40] [−1.07; 4.97] [−0.35; 2.15] [−1.36; 5.58] Premolars n 16 14 16 16 16 10 MTE (95% CI) 0.09 (0.05; 0.13) 0.10 (0.06; 0.14) 0.13 (0.10;0.17) 0.13 (0.10; 0.17) 1.46 (0.97;1.94) 0.95 (0.37; 1.53) Prediction interval [−0.06; 0.24] [−0.05; 0.24] [−0.01; 0.27] [−0.01; 0.27] [−0.45; 3.36] [−0.81; 2.71] Molars n 10 10 10 10 6 10 MTE (95% CI) 0.06 (0.04; 0.08) 0.09 (−0.04; 0.13) 0.11 (0.04; 0.18) 1.47 (0.70; 2.23) 0.69 (0.32; 1.06) 2.29 (1.20; 3.38) Prediction interval [0.01; 0.11] [−0.04; 0.21] [−0.08; 0.31] [−0.99; 3.92] [−0.26; 1.64] [−1.24; 5.82] Left vs. Right Left n 5 3 5 2 - - MTE (95% CI) 0.14 (0.04; 0.24) 0.11 (0.06; 0.17) 0.22 (0.10; 0.35) 2.91 (−1.59; 7.41) Prediction interval [−0.14; 0.42] [−0.12; 0.35] [−0.13; 0.57] - Right n 5 3 5 2 - - MTE (95% CI) 0.14 (0.05; 0.22) 0.10 (0.02; 0.17) 0.23 (0.04; 0.42) 2.66 (2.59; 2.72) Prediction interval [−0.10: 0.37] [−0.29; 0.48] [−0.30; 0.76] Upper vs. Lower Upper n 9 7 9 6 4 4 MTE (95% CI) 0.10 (0.05; 0.16) 0.09 (0.02; 0.15) 0.18 (0.09; 0.26) 1.26 (0.00; 2.53) 0.59 (−0.49; 1.6) 0.73 (−0.50; 1.96) Prediction interval [−0.08; 0.29] [−0.10; 0.27] [−0.10; 0.45] [−2.34; 4.86] [−2.67; 3.85] [−2.97; 4.43] Lower n 4 2 4 4 2 2 MTE (95% CI) 0.12 (−0.09; 0.33) 0.01 (−0.04; 0.05) 0.22 (−0.00; 0.44) 1.49 (−1.10; 4.08) 0.01 (−0.09;0.10) 0.18 (0.01; 0.35) Prediction interval [−0.52; 0.76] [−0.10; 0.45]- [−6.32; 9.31] 3D accuracy assessment vs. Photography 3D n 18 18 18 18 17 18 MTE (95% CI) 0.06 (0.04; 0.08) 0.09 (0.05; 0.12) 0.11 (0.09; 0.13) 0.95 (0.63; 1.27) 0.93 (0.49; 1.37) 1.11 (0.68; 1.53) Prediction interval [−0.03; 0.15] [−0.05; 0.22] [0.03; 0.18] [−0.42; 2.32] [−0.88; 2.74] [−0.61; 2.83] Photography n 7 5 7 2 - - MTE (95% CI) 0.12 (0.06; 0.18) 0.09 (0.09; 0.10) 0.22 (0.12; 0.31) 2.74 (−1.50; 6.97) Prediction interval [−0.05; 0.30] [ 0.09; 0.10] [−0.07; 0.50] - Type of tray 3D printed n 13 13 4 13 11 13 MTE (95% CI) 0.06 (0.03; 0.09) 0.10 (0.06; 0.13) 0.12 (0.09; 0.15) 1.14 (0.69; 1.60) 0.90 (0.36; 1.45) 1.42 (0.76; 2.09) Prediction interval [−0.05; 0.16] [−0.04; 0.24] [ 0.02; 0.21] [−0.57; 2.86] [−0.94; 2.75] [−1.01; 3.86] Silicone n 4 3 4 3 2 2 MTE (95% CI) 0.10 (0.00; 0.19) 0.08 (−0.01; 0.18) 0.14 (−0.03; 0.32) 1.17 (−1.55; 3.88) 0.66 (−3.82; 5.13) 0.79 (−4.47; 6.05) Prediction interval [−0.20; 0.39] [−0.44; 0.61] [−0.38; 0.67] [−14.17; 17.12] Combined Silicone/Vacuum Form n 1 1 1 - - - MTE (95% CI) 0.09 (0.07; 0.11) 0.09 (0.07; 0.11) 0.14 (0.11; 0.17) Prediction interval - - - Vacuum Form n 6 5 6 5 4 4 MTE (95% CI) 0.10 (0.02; 0.18) 0.08 (−0.03; 0.19) 0.16 (0.03; 0.29) 1.32 (−0.06; 2.71) 1.16 (−0.84; 3.16) 0.86 (0.26;1.46) Prediction interval [−0.13; 0.33] [−0.22; 0.39] [−0.20; 0.52] [−2.52; 5.17] [−4.80; 7.13] [−0.92; 2.63] Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Andrews L.F. The straight-wire appliance, origin, controversy, commentary J. Clin. Orthod. 1976 10 99 114 1074876 2. Andrews L.F. The straight-wire appliance, explained and compared J. Clin. Orthod. 1976 10 174 195 1074877 3. Fukuyo K. Nishii Y. Nojima K. Yamaguchi H. A comparative study in three methods of bracket placement Orthod. Waves 2004 63 63 70 4. Birdsall J. Hunt N.P. Sabbah W. Moseley H.C. Accuracy of positioning three types of self-ligating brackets compared with a conventionally ligating bracket J. Orthod. 2012 39 34 42 10.1179/14653121226806 22433325 5. Brown M.W. Koroluk L. Ko C.C. Zhang K. Chen M. Nguyen T. Effectiveness and efficiency of a CAD/CAM orthodontic bracket system Am. J. Orthod. Dentofac. Orthop. 2015 148 1067 1074 10.1016/j.ajodo.2015.07.029 6. Garino F. Garino G.B. Computer-aided interactive indirect bonding Prog. Orthod. 2005 6 214 223 16276430 7. Grunheid T. Lee M.S. Larson B.E. Transfer accuracy of vinyl polysiloxane trays for indirect bonding Angle Orthod. 2016 86 468 474 10.2319/042415-279.1 26355994 8. Aguirre M.J. King G.J. Waldron J.M. Assessment of bracket placement and bond strength when comparing direct bonding to indirect bonding techniques Am. J. Orthod. 1982 82 269 276 10.1016/0002-9416(82)90461-4 6760721 9. Carlson S.K. Johnson E. Bracket positioning and resets: Five steps to align crowns and roots consistently Am. J. Orthod. Dentofac. Orthop. 2001 119 76 80 10.1067/mod.2001.111220 10. Castilla A.E. Crowe J.J. Moses J.R. Wang M. Ferracane J.L. Covell D.A.J. Measurement and comparison of bracket transfer accuracy of five indirect bonding techniques Angle Orthod. 2014 84 607 614 10.2319/070113-484.1 24555689 11. El-Timamy A.M. El-Sharaby F.A. Eid F.H. Mostafa Y.A. Three-dimensional imaging for indirect-direct bonding Am. J. Orthod. Dentofac. Orthop. 2016 149 928 931 10.1016/j.ajodo.2015.12.009 12. Guenthner T.A. Larson B.E. Indirect bonding: A technique for precision and efficiency Semin. Orthod. 2007 13 58 63 10.1053/j.sodo.2006.11.007 13. Suárez C. Vilar T. The effect of constant height bracket placement on marginal ridge levelling using digitized models Eur. J. Orthod. 2010 32 100 105 10.1093/ejo/cjp029 19487435 14. Israel M. Kusnoto B. Evans C.A. BeGole E. A comparison of traditional and computer-aided bracket placement methods Angle Orthod. 2011 81 828 835 10.2319/072110-425.1 21524241 15. Silverman E. Cohen M. Gianelly A.A. Dietz V.S. A universal direct bonding system for both metal and plastic brackets Am. J. Orthod. 1972 62 236 244 10.1016/S0002-9416(72)90264-3 4559001 16. Yildirim K. Saglam-Aydinatay B. Comparative assessment of treatment efficacy and adverse effects during nonextraction orthodontic treatment of class I malocclusion patients with direct and indirect bonding: A parallel randomized clinical trial Am. J. Orthod. Dentofac. Orthop. 2018 154 26.e1 34.e1 10.1016/j.ajodo.2017.12.009 17. Shpack N. Geron S. Floris I. Davidovitch M. Brosh T. Vardimon A.D. Bracket placement in lingual vs. labial systems and direct vs. indirect bonding Angle Orthod. 2007 77 509 517 10.2319/0003-3219(2007)077[0509:BPILVL]2.0.CO;2 17465662 18. Koo B.C. Chung C.-H. Vanarsdall R.L. Comparison of the accuracy of bracket placement between direct and indirect bonding techniques Am. J. Orthod. Dentofac. Orthop. 1999 116 346 351 10.1016/S0889-5406(99)70248-9 19. Bousema E.J. Koops E.A. van Dijk P. Dijkstra P.U. Association Between Subjective Tinnitus and Cervical Spine or Temporomandibular Disorders: A Systematic Review Trends Heart 2018 22 2331216518800640 10.1177/2331216518800640 20. Nichols D.A. Gardner G. Carballeyra A.D. Reproducibility of bracket positioning in the indirect bonding technique Am. J. Orthod. Dentofac. Orthop. 2013 144 770 776 10.1016/j.ajodo.2013.04.023 21. Dellinger E.L. A scientific assessment of the straight-wire appliance Am. J. Orthod. 1978 73 290 299 10.1016/0002-9416(78)90135-5 274074 22. Miethke R.R. Melsen B. Effect of variation in tooth morphology and bracket position on first and third order correction with preadjusted appliances Am. J. Orthod. Dentofac. Orthop. 1999 116 329 335 10.1016/S0889-5406(99)70246-5 23. Germane N. Bentley B.E.J. Isaacson R.J. Three biologic variables modifying faciolingual tooth angulation by straight-wire appliances Am. J. Orthod. Dentofac. Orthop. 1989 96 312 319 10.1016/0889-5406(89)90350-8 24. Li Y. Mei L. Wei J. Yan X. Zhang X. Zheng W. Li Y. Effectiveness, efficiency and adverse effects of using direct or indirect bonding technique in orthodontic patients: A systematic review and meta-analysis BMC Oral Health 2019 19 137 10.1186/s12903-019-0831-4 31286897 25. Duarte M.E.A. Gribel B.F. Spitz A. Artese F. Miguel J.A.M. Reproducibility of digital indirect bonding technique using three-dimensional (3D) models and 3D-printed transfer trays Angle Orthod. 2019 90 92 99 10.2319/030919-176.1 31411488 26. Xue C. Xu H. Guo Y. Xu L. Dhami Y. Wang H. Liu Z. Ma J. Bai D. Accurate bracket placement using a computer-aided design and computer-aided manufacturing–guided bonding device: An in vivo study Am. J. Orthod. Dentofac. Orthop. 2020 157 269 277 10.1016/j.ajodo.2019.03.022 32005479 27. Pottier T. Brient A. Turpin Y.L. Chauvel B. Meuric V. Sorel O. Brezulier D. Accuracy evaluation of bracket repositioning by indirect bonding: Hard acrylic CAD/CAM versus soft one-layer silicone trays, an in vitro study Clin. Oral Investig. 2020 24 3889 3897 10.1007/s00784-020-03256-x 28. Jungbauer R. Breunig J. Schmid A. Hüfner M. Kerberger R. Rauch N. Proff P. Drescher D. Becker K. Transfer accuracy of two 3D printed trays for indirect bracket bonding—An in vitro pilot study Appl. Sci. 2021 11 6013 10.3390/app11136013 29. Liberati A. Altman D.G. Tetzlaff J. Mulrow C. Gøtzsche P.C. Ioannidis J.P. Clarke M. Devereaux P.J. Kleijnen J. Moher D. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: Explanation and elaboration PLoS Med. 2009 6 e1000100 10.1371/journal.pmed.1000100 19621070 30. Richardson W.S. Wilson M.C. Nishikawa J. Hayward R.S. The well-built clinical question: A key to evidence-based decisions ACP J. Club 1995 123 A12 A13 10.7326/ACPJC-1995-123-3-A12 31. Lefebvre C. Glanville J. Briscoe S. Featherstone R. Littlewood A. Marshall C. Metzendorf M.-I. Noel-Storr A. Paynter R. Rader T. Chapter 4: Searching for and selecting studies Cochrane Handbook for Systematic Reviews of Interventions version 6.3 Updated February 2022 Higgins J.P.T. Thomas J. Chandler J. Cumpston M. Li T. Page M.J. Welch V.A. Cochrane London, UK 2022 32. Niu Y. Zeng Y. Zhang Z. Xu W. Xiao L. Comparison of the transfer accuracy of two digital indirect bonding trays for labial bracket bonding Angle Orthod. 2021 91 67 73 10.2319/013120-70.1 33289807 33. Drevon D. Fursa S.R. Malcolm A.L. Intercoder reliability and validity of webplotdigitizer in extracting graphed data Behav. Modif. 2017 41 323 339 10.1177/0145445516673998 27760807 34. Kühnisch J. Janjic Rankovic M. Kapor S. Schüler I. Krause F. Michou S. Ekstrand K. Eggmann F. Neuhaus K.W. Lussi A. Identifying and avoiding risk of bias in caries diagnostic studies J. Clin. Med. 2021 10 3223 10.3390/jcm10153223 34362007 35. Whiting P.F. Rutjes A.W. Westwood M.E. Mallett S. Deeks J.J. Reitsma J.B. Leeflang M.M. Sterne J.A. Bossuyt P.M. QUADAS-2 Group QUADAS-2: A revised tool for the quality assessment of diagnostic accuracy studies Ann. Intern. Med. 2011 155 529 536 10.7326/0003-4819-155-8-201110180-00009 22007046 36. Harrer M. Cuijpers P. Furukawa T.A. Ebert D.D. Doing Meta-Analysis with R: A Hands-on Guide CRC Press Boca Raton, FL, USA 2021 37. Wickham H. Averick M. Bryan J. Chang W. McGowan L.D.A. François R. Grolemund G. Hayes A. Henry L. Hester J. Welcome to the Tidyverse J. Open Source Softw. 2019 4 1686 10.21105/joss.01686 38. Wickham H. François R. Henry L. Müller K. A Grammar of Data Manipulation [R package dplyr Version 1.0.2] 2020 39. Wilkinson L. ggplot2: Elegant Graphics for Data Analysis by WICKHAM, H Biometrics 2011 67 678 679 10.1111/j.1541-0420.2011.01616.x 40. Balduzzi S. Rücker G. Schwarzer G. How to perform a meta-analysis with R: A practical tutorial Évid. Based Ment. Health 2019 22 153 160 10.1136/ebmental-2019-300117 31563865 41. Veroniki A.A. Jackson D. Viechtbauer W. Bender R. Bowden J. Knapp G. Kuss O. Higgins J.P.T. Langan D. Salanti G. Methods to estimate the between-study variance and its uncertainty in meta-analysis Res. Synth. Methods 2016 7 55 79 10.1002/jrsm.1164 26332144 42. Knapp G. Hartung J. Improved tests for a random effects meta-regression with a single covariate Stat. Med. 2003 22 2693 2710 10.1002/sim.1482 12939780 43. Park J.-H. Choi J.-Y. Oh S.H. Kim S.-H. Three-dimensional digital superimposition of orthodontic bracket position by using a computer-aided transfer jig system: An accuracy analysis Sensors 2021 21 5911 10.3390/s21175911 34502801 44. Park J.H. Choi J.Y. Kim S.H. Kim S.J. Lee K.J. Nelson G. Three-dimensional evaluation of the transfer accuracy of a bracket jig fabricated using computer-aided design and manufacturing to the anterior dentition: An in vitro study Korean J. Orthod. 2021 51 375 386 10.4041/kjod.2021.51.6.375 34803026 45. Kim J. Chun Y.S. Kim M. Accuracy of bracket positions with a CAD/CAM indirect bonding system in posterior teeth with different cusp heights Am. J. Orthod. Dentofac. Orthop. 2018 153 298 307 10.1016/j.ajodo.2017.06.017 29407508 46. Schmid J. Brenner D. Recheis W. Hofer-Picout P. Brenner M. Crismani A.G. Transfer accuracy of two indirect bonding techniques-an in vitro study with 3D scanned models Eur. J. Orthod. 2018 40 549 555 10.1093/ejo/cjy006 29471483 47. Chaudhary V. Batra P. Sharma K. Raghavan S. Gandhi V. Srivastava A. A comparative assessment of transfer accuracy of two indirect bonding techniques in patients undergoing fixed mechanotherapy: A randomised clinical trial J. Orthod. 2021 48 13 23 10.1177/1465312520968571 33843329 48. Hodge T.M. Dhopatkar A.A. Rock W.P. Spary D.J. A randomized clinical trial comparing the accuracy of direct versus indirect bracket placement J. Orthod. 2004 31 132 137 10.1179/146531204225020427 15210929 49. Süpple J. von Glasenapp J. Hofmann E. Jost-Brinkmann P.G. Koch P.J. Accurate bracket placement with an indirect bonding method using digitally designed transfer models printed in different orientations-an in vitro study J. Clin. Med. 2021 10 2002 10.3390/jcm10092002 34066993 50. Faus-Matoses I. Guinot Barona C. Zubizarreta-Macho Á. Paredes-Gallardo V. Faus-Matoses V. A novel digital technique for measuring the accuracy of an indirect bonding technique using fixed buccal multibracket appliances J. Pers. Med. 2021 11 932 10.3390/jpm11090932 34575709 51. Kalra R.K. Mittal S. Gandikota C. Sehgal V. Gupta R. Bali Z. Comparison of accuracy of bracket placement by direct and indirect bonding techniques using digital processing—An in-vitro study J. Clin. Diagn. Res. 2018 12 7 11 10.7860/JCDR/2018/32466.11975 52. Armstrong D. Shen G. Petocz P. Darendeliler M.A. A comparison of accuracy in bracket positioning between two techniques—Localizing the centre of the clinical crown and measuring the distance from the incisal edge Eur. J. Orthod. 2007 29 430 436 10.1093/ejo/cjm037 17660488 53. Archambault A. Lacoursiere R. Badawi H. Major P.W. Carey J. Flores-Mir C. Torque expression in stainless steel orthodontic brackets. A systematic review Angle Orthod. 2010 80 201 210 10.2319/080508-352.1 19852662 54. Dalstra M. Eriksen H. Bergamini C. Melsen B. Actual versus theoretical torsional play in conventional and self-ligating bracket systems J. Orthod. 2015 42 103 113 10.1179/1465313314Y.0000000126 25588827 55. Arreghini A. Lombardo L. Mollica F. Siciliani G. Torque expression capacity of 0.018 and 0.022 bracket slots by changing archwire material and cross section Prog. Orthod. 2014 15 53 10.1186/s40510-014-0053-x 25329505 56. Dörfer S. König M. Jost-Brinkmann P.G. Übertragungsgenauigeit beim indirekten Platzieren von Brackets Kieferorthoädie 2006 20 91 103 57. Hoffmann L. Sabbagh H. Wichelhaus A. Kessler A. Bracket transfer accuracy with two different three-dimensional printed transfer trays vs. silicone transfer trays Angle Orthod. 2022 92 364 371 10.2319/040821-283.1 34982114 58. Donatelli R.E. Lee S.J. How to report reliability in orthodontic research: Part 1 Am. J. Orthod. Dentofac. Orthop. 2013 144 156 161 10.1016/j.ajodo.2013.03.014 59. Wellek S. A critical evaluation of the current “p-value controversy” Biom. J. 2017 59 854 872 10.1002/bimj.201700001 28504870 60. Rücker G. Schwarzer G. Beyond the forest plot: The drapery plot Res. Synth. Methods 2021 12 13 19 10.1002/jrsm.1410 32336044 61. Czolgosz I. Cattaneo P.M. Cornelis M.A. Computer-aided indirect bonding versus traditional direct bonding of orthodontic brackets: Bonding time, immediate bonding failures, and cost-minimization. A randomized controlled trial Eur. J. Orthod. 2021 43 144 151 10.1093/ejo/cjaa045 32780096 62. Bozelli J.V. Bigliazzi R. Barbosa H.A. Ortolani C.L. Bertoz F.A. Faltin Junior K. Comparative study on direct and indirect bracket bonding techniques regarding time length and bracket detachment Dental Press. J. Orthod. 2013 18 51 57 10.1590/S2176-94512013000600009 24351150 63. Weber D.J. II. Koroluk L.D. Phillips C. Nguyen T. Proffit W.R. Clinical effectiveness and efficiency of customized vs. Conventional preadjusted bracket systems J. Clin. Orthod. 2013 47 261 266 23660822 64. Deahl S.T. Salome N. Hatch J.P. Rugh J.D. Practice-based comparison of direct and indirect bonding Am. J. Orthod. Dentofac. Orthop. 2007 132 738 742 10.1016/j.ajodo.2006.01.037 65. Brignardello-Petersen R. There may be no difference in failure and orthodontic bracket placement accuracy when using a direct or indirect bonding technique J. Am. Dent. Assoc. 2020 151 e7 10.1016/j.adaj.2019.08.002 31757358 66. De Oliveira N.S. Rossouw E. Lages E.M.B. Macari S. Pretti H. Influence of clinical experience on accuracy of virtual orthodontic attachment bonding in comparison with the direct procedure Angle Orthod. 2019 89 734 741 10.2319/100618-724.1 30920869 67. El Nigoumi A. Assessing the Accuracy of Indirect Bonding with 3D Scanning Technology J. Clin. Orthod. 2016 50 613 619 27888654 68. Gayake P.V. Chitko S.S. Sutrave N. Gaikwad P.M. The direct way of indirect bonding—the combined effect Int. J. Orthod. Milwaukee 2013 24 15 17 69. Hiro T. Iglesia F. Andreu P. Indirect bonding technique in lingual orthodontics: The HIRO system Prog. Orthod. 2008 9 34 45 19350057 70. Mazzeo F. Marchese E. Assumma V. Sepe J. Perillo L. A new device (FAQ.FIX®) for orthodontic bracket placement in straight wire technique Prog. Orthod. 2013 14 23 10.1186/2196-1042-14-23 24325941 71. Mohlhenrich S.C. Alexandridis C. Peters F. Kniha K. Modabber A. Danesh G. Fritz U. Three-dimensional evaluation of bracket placement accuracy and excess bonding adhesive depending on indirect bonding technique and bracket geometry: An in-vitro study Head Face Med. 2020 16 17 10.1186/s13005-020-00231-5 32741369 72. Mota Júnior S.L. de Andrade Vitral J. Schmitberger C.A. Machado D.B. Avelar J.C. Fraga M.R. da Silva Campos M.J. Vitral R.W. Evaluation of the vertical accuracy of bracket placement with the Boone gauge Am. J. Orthod. Dentofacial Orthop. 2015 148 821 826 10.1016/j.ajodo.2015.05.028 26522043 73. Mota Júnior S.L. Campos M. Schmitberger C.A. Vitral J.A. Fraga M.R. Vitral R.W.F. Evaluation of the prototype of a new bracket-positioning gauge Dental Press J. Orthod. 2018 23 68 74 10.1590/2177-6709.23.2.068-074.oar 29898160 74. Nojima L.I. Araújo A.S. Alves Júnior M. Indirect orthodontic bonding—A modified technique for improved efficiency and precision Dental Press J. Orthod. 2015 20 109 117 10.1590/2176-9451.20.3.109-117.sar 26154464 75. Oliveira N.S. Gribel B.F. Neves L.S. Lages E.M.B. Macari S. Pretti H. Comparison of the accuracy of virtual and direct bonding of orthodontic accessories Dental Press J. Orthod. 2019 24 46 53 10.1590/2177-6709.24.4.046-053.oar 76. Schubert K. Halbich T. Jost-Brinkmann P.G. Müller-Hartwich R. Precision of indirect bonding of lingual brackets using the Quick Modul System (QMS)® J. Orofac. Orthop. 2013 74 6 17 10.1007/s00056-012-0122-z 23299653 77. Shin S.H. Lee K.J. Kim S.J. Yu H.S. Kim K.M. Hwang C.J. Cha J.Y. Accuracy of bracket position using thermoplastic and 3D-printed indirect bonding trays Int. J. Comput. Dent. 2021 24 133 145 34085499 78. Wendl B. Droschl H. Muchitsch P. Indirect bonding--a new transfer method Eur. J. Orthod. 2008 30 100 107 10.1093/ejo/cjm094 18276930 79. Zhang Y. Yang C. Li Y. Xia D. Shi T. Li C. Comparison of three-dimensional printing guides and double-layer guide plates in accurate bracket placement BMC Oral Health 2020 20 127 10.1186/s12903-020-01110-w 32345292

PMC009xxxxxx/PMC9099690.txt

==== Front Cancers (Basel) Cancers (Basel) cancers Cancers 2072-6694 MDPI 10.3390/cancers14092273 cancers-14-02273 Article RR Myelo POINT: A Retrospective Single-Center Study Assessing the Role of Radiotherapy in the Management of Multiple Myeloma and Possible Interactions with Concurrent Systemic Treatment Guerini Andrea Emanuele 1 Tucci Alessandra 2 https://orcid.org/0000-0002-6569-8628 Alongi Filippo 3 Mataj Eneida 1 Belotti Angelo 2 Borghetti Paolo 1* Triggiani Luca 1 Pegurri Ludovica 1 Pedretti Sara 1 Bonù Marco 1 Tomasini Davide 1 Imbrescia Jessica 1 Donofrio Alessandra 1 https://orcid.org/0000-0002-4555-1725 Facheris Giorgio 1 https://orcid.org/0000-0001-9290-6448 Singh Navdeep 1 Volpi Giulia 1 https://orcid.org/0000-0002-3699-8744 Tomasi Cesare 4 https://orcid.org/0000-0002-4126-733X Magrini Stefano Maria 1 Spiazzi Luigi 5 https://orcid.org/0000-0002-5485-1260 Buglione Michela 1 Wong David Academic Editor 1 Department of Radiation Oncology, University of Brescia and Spedali Civili Hospital, Piazzale Spedali Civili 1, 25123 Brescia, Italy; a.e.guerini@gmail.com (A.E.G.); e.mataj@unibs.it (E.M.); luca.triggiani@unibs.it (L.T.); ludovicapegurri@libero.it (L.P.); sara.pedretti@asst-spedalicivili.it (S.P.); marcolorenzo89.mlb@gmail.com (M.B.); tomad88@libero.it (D.T.); imbresciajessica@gmail.com (J.I.); alessandra.donofrio90@gmail.com (A.D.); giorgio.facheris@gmail.com (G.F.); singh.nav92@gmail.com (N.S.); g.volpi50@gmail.com (G.V.); stefano.magrini@unibs.it (S.M.M.); michela.buglione@unibs.it (M.B.) 2 Department of Haematology, ASST-Spedali Civili Hospital, 25123 Brescia, Italy; alessandra.tucci@asst-spedalicivili.it (A.T.); angelo.belotti@asst-spedalicivili.it (A.B.) 3 Advanced Radiation Oncology Department, Sacro Cuore Don Calabria Hospital, IRCCS Ospedale Sacro Cuore Don Calabria, 37024 Negrar Di Valpolicella, Italy; filippo.alongi@unibs.it 4 Department of Medical and Surgical Specialties, Radiological Sciences and Public Health, Section of Public Health and Human Sciences, University of Brescia, 25123 Brescia, Italy; cesare.tomasi@live.com 5 Medical Physics Department, ASST Spedali Civili Hospital, 25123 Brescia, Italy; luigi.spiazzi@unibs.it * Correspondence: paolobor82@yahoo.it; Tel.: +39-0303995272 02 5 2022 5 2022 14 9 227306 4 2022 29 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Simple Summary Currently, few papers have been published regarding the possible interactions between radiotherapy and systemic agents for the treatment of multiple myeloma. In this paper, we retrospectively analyze the data from 312 patients (577 lesions) who received radiotherapy at our institution from 2005 to 2020, with the aim of clarifying the clinical impact of radiotherapy dose and concurrent systemic treatment (CST). The safety profile of the radiotherapy was excellent; high biologically effective doses (BEDs) and CST were associated with higher toxicity rates at the end of radiotherapy, but not after one and three months. The pain control rate was 87.4% at the end of treatment and further increased at three and six months. Radiological progression was reported only for 4.4% of the lesions at six months (based on the data available for 181 lesions) and was significantly more frequent for lesions treated without CST or BED < 15 Gy. Abstract Background and purpose: Although chemotherapy, biological agents, and radiotherapy (RT) are cornerstones of the treatment of multiple myeloma (MM), the literature regarding the possible interactions of concurrent systemic treatment (CST) and RT is limited, and the optimal RT dose is still unclear. Materials and methods: We retrospectively analyzed the records of patients who underwent RT for MM at our institution from 1 January 2005 to 30 June 2020. The data of 312 patients and 577 lesions (treated in 411 accesses) were retrieved. Results: Most of the treated lesions involved the vertebrae (60%) or extremities (18.9%). Radiotherapy was completed in 96.6% of the accesses and, although biologically effective doses assuming an α/β ratio of 10 (BED 10) > 38 Gy and CST were significantly associated with higher rates of toxicity, the safety profile was excellent, with side effects grade ≥2 reported only for 4.1% of the accesses; CST and BED 10 had no impact on the toxicity at one and three months. Radiotherapy resulted in significant improvements in performance status and in a pain control rate of 87.4% at the end of treatment, which further increased to 96.9% at three months and remained at 94% at six months. The radiological response rate at six months (data available for 181 lesions) was 79%, with only 4.4% of lesions in progression. Progression was significantly more frequent in the lesions treated without CST or BED 10 < 15 Gy, while concurrent biological therapy resulted in significantly lower rates of progression. Conclusion: Radiotherapy resulted in optimal pain control rates and fair toxicity, regardless of BED 10 and CST; the treatments with higher BED 10 and CST (remarkably biological agents) improved the already excellent radiological disease control. myeloma radiotherapy concurrent chemotherapy biologic therapy immunotherapy toxicity pain control multiple myeloma disease control This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. ==== Body pmc1. Background Although it is considered a rare cancer, multiple myeloma (MM) is diagnosed in over 100,000 patients each year worldwide, and its prevalence is increasing [1]. Although chemotherapy and systemic agents constitute the cornerstone of its treatment, radiotherapy (RT) also has an established role in its management [2]. Osteolytic lesions are present in up to 80% of cases at diagnosis and often invade surrounding tissue, with possible spinal cord and nerve-root compression [3]. As MM is highly radiosensitive, RT allows a high rate of local response and is thus beneficial in the control of multiple complications of the disease, including pain, structural instability, and spinal cord compression [4]. In the last fifteen years, the approval of innovative systemic treatments (including bortezomib, carfilzomib, ixazomib, lenalidomide, pomalidomide, and daratumumab) has led to an impressive improvement in overall survival [5]. The advent of novel agents has also opened new perspectives for the concomitant administration of RT [6], as possible synergy has been observed in pre-clinical studies [7], but there is also concern over the possible sensitization of normal tissues [8]. Nonetheless, only a few studies, with limited populations, have been published regarding the concurrent use of RT and newly approved drugs [9,10]. This, combined with isolated case reports of severe side effects [11], might lead to the unnecessary discontinuation of systemic treatment or to depriving patients of the benefits of RT, due to fear over unexpected toxicities. We therefore designed a retrospective analysis of our center’s experience to evaluate the effect of RT on the symptomatic and radiological control of the disease and the clinical impact of the concomitant administration of systemic therapy (CST) and RT, focusing on newly approved agents. 2. Methods We retrospectively evaluated records of all the patients who underwent radiation therapy on MM lesions during the period between 1 January 2005 and 30 June 2020 at our institution. Data were retrieved from the electronic medical record systems and picture archiving and communication system used by our Institution. An access was defined as a presentation at our institution for a new radiation therapy prescription for one or more sites of disease. For each patient, information regarding RT and eventual concurrent systemic treatment were reported, as well as information regarding number of previous systemic treatment lines, age, and sex. Systemic treatment was defined as concurrent if the last administration was performed before the start of RT, with an interval of less than five half-lives of the drug, regardless of the line of treatment and the time from the start of treatment. Patients were classified in three categories, according to systemic treatment administered concurrently with RT: (a) no concurrent therapy, (b) concurrent chemotherapy, (c) concurrent biologic treatment (including proteasome inhibitors, immunomodulating agents and/or monoclonal antibodies) given alone or in combination with chemotherapy. For patients treated with combination regimens, only drugs administered with an interval of less than five half-lives from the start of RT were considered as concurrent. Patients were also divided into three groups, according to BED10 (biologically effective dose assuming an α/β ratio of 10) of RT: <15 Gy, 15–38 Gy, >38 Gy. The following outcomes were evaluated:(a) Incidence of treatment toxicity during RT and at one and three months after the end of RT and percentage of RT suspension attributable to toxicity. Toxicities were graded according to National Cancer Institute Common Terminology Criteria for Adverse Events, version 5.0. Only toxicities with a causal relationship with RT were reported. Our analysis included events with the following causal relationships. (a) Possible: Some evidence suggesting a causal relationship, with other factors that might have contributed. (b) Probable: Evidence to suggest a causal relationship, influence of other factors is unlikely. (c) Definite: Clear evidence to suggest a causal relationship, and other possible contributing factors can be ruled out. Side effects with no causal relationship with RT were excluded from the analysis. (b) Pain control at the end of RT and at one, three, and six months after RT; pain control was defined as partial, complete, or absent and possibly assessed by the patient’s self-rated pain verbal numeric scale, VNS, in which patients are asked to verbally state a number between 0 (no pain) and 10 (worst imaginable pain); data regarding reduction or variation in analgesic intake were also included, when available, for the assessment of pain control. (c) Local radiologic in-field response at six months after the end of RT (defined according to RECIST 1.1 or PERCIST 1.0). This protocol was developed, and the study was conducted, according to the criteria established by the 1964 Helsinki declaration and its later amendments, as well as by the ICH Good Clinical Practice, and approved by the Ethics Committee of our Hospital (number of approval NP 4595, received 1 April 2021). Statistical analysis was performed using IBM-SPSS® software ver. 26.0.1 (IBM SPSS Inc. Chicago, IL, USA). The use of Stata® software ver. 16.0 (Stata Corporation, College Station, TX, USA) was also considered for comparisons or implementations of test output. Normality of the distributions was assessed using the Kolmogorov–Smirnov test. Categorical variables were presented as frequencies or percentages and compared with the use of the chi-square test or the Fisher’s exact test, as appropriate. Continuous variables were compared with the use of Wilcoxon paired test. Associations of the crosstabs were verified using standardized adjusted residuals. A two-sided α level of 0.05 was used for all tests. 3. Results The data from 312 patients (158 males and 154 females) were retrieved, with 577 lesions treated in 411 accesses (mean number of treated lesions per access 1.4, median 1). The characteristics of the patients, the diseases, and their treatment are summarized in Table 1. The median age at the start of RT was 69.8 years; the majority of the treated lesions involved the vertebrae (60%) or the extremities (18.9%). Steroid use during RT was reported for 62.9% of the lesions and 55.1% of the accesses. The performance status, according to the Karnofsky score, significantly improved by the last day of RT (data available for 377 accesses: mean 70 before RT vs. 71.7 on the last day of RT, p < 0.001) with no significant impact of CST or BED10. The pain reported, according to the VNS, improved from the beginning of RT to the last day of treatment (data available for 139 accesses: mean 4.8 vs. 1.7, median 5 vs. 0, p < 0.001) with no significant impact of BED10 (p = 0.060) or CST (p = 0.103). The most common side effects were gastro-intestinal (G1 for 13.7% of accesses, G2 for 2%), esophagitis (G1 10.2%, G2 0.7%), pharyngodinia (G1 5.6%), infections (G1 4.9%, G2 1.2%), and mucositis (G1 2.4%, G2 0.2%). Only three G3 toxicity events were reported (one case of esophagitis in a patient administered concurrent chemotherapy and two infections, one in a patient undergoing RT alone, and one in a patient administered concurrent bortezomib), and one G4 infection was reported in a patient administered concurrent chemotherapy; all these severe toxicities were reversible. Pain ‘flare’ during RT was reported only for 8.1% of lesions (47/577) and steroid use during RT had no significant effect on its incidence (p = 0.587). Concomitant systemic treatment had a significant impact on the toxicity reported during RT administration (data available for 410 accesses, p < 0.001), with a lower rate of G2 toxicities for patients receiving RT alone and higher rates of G4 toxicity and lower percentages of patients not reporting toxicity in the concurrent chemotherapy group. The biological dose also had an impact on toxicity: significantly higher rates of grade 1 toxicity during RT were described for treatment with a BED10 > 38 Gy (p < 0.001), while a BED10 ≤ 38 Gy was significantly associated with the absence of toxicity during RT (p < 0.001). Nonetheless, as shown in Table 2, the safety profile was excellent, with no toxicity developed during RT for 59% of the accesses and a toxicity grade ≥2 reported only for 4.1% of the accesses. Moreover, no significant difference in toxicity possibly attributable to RT was reported at one (p = 0.753) or three months (p= 0.677) across the groups receiving no CST, biological agents, or chemotherapy. Similarly, no statistical difference in toxicity was reported at one (p = 0.094) or three (p = 0.358) months after the end of RT between the groups treated with different BED10. Radiotherapy was completed for 96.6% of the accesses (397 of 411), while definitive suspension due to toxicity was reported only for five accesses (1.2%). In all the accesses that required temporary suspension due to toxicity (1.2%) or other causes (0.5%), RT was resumed and completed. The results regarding pain control and radiological control are summarized in Table 3. The overall pain control rate at the end of RT was 87.4%: complete control was obtained for 35.9% of lesions and partial control for 51.5%. There was a statistically significant correlation between BED10 and pain control at the end of RT (p < 0.001), with higher rates of pain control for a BED10 > 38 Gy and lower rates for a BED10 < 15 Gy. Conversely, CST had no impact on the pain control rate at the end of RT (p = 0.244). The data regarding pain control at 1 month were available for 367 lesions. Overall, the complete pain control rate was 66.5%, and the rate of partial control was 30.5%; only 3% of patients had no pain relief on irradiated sites. Radiotherapy alone was significantly linked with higher rates of complete control and lower rates of partial control, and biological treatment was associated with no pain control (p = 0.034). This finding could have been biased by the extremely low rate of absence of pain control. On the other hand, BED10 had no significant effect (p = 0.112) on the pain control rates (p = 0.175). At 3 months (with pain assessed for 355 lesions), the complete pain control rate increased to 73.5%, and the partial and no-control rates were 23.4% and 3.1%, respectively. Radiotherapy alone was significantly linked with higher rates of complete control, concurrent chemotherapy with partial control, and biological treatment with no control (p = 0.001). Again, BED10 had no significant effect (p = 0.112) on the pain control rates at three months. At six months (with data available for 352 lesions), pain control was complete for 77% of the lesions and partial for 17%, while for 6% of the lesions, pain was not controlled. No significant differences in pain control at 6 months were reported for the different systemic-treatment (p = 0.155) or BED10 (p = 0.499) groups. The data regarding the radiological response at 6 months after RT were available for 181 lesions. Overall, local complete response (CR) was observed for 8.8% of the lesions, partial response (PR) for 70.2%, stable disease (SD) for 16.6%, and progressive disease (PD) only for eight lesions (4.4%). Progressive disease was significantly more frequent for the lesions treated without CST, while the administration of concurrent biologic therapy resulted in significantly lower rates of PD (p = 0.044). Due to the multitude of different CSTs, the numbers were not sufficient to allow conclusions regarding the impact of each single drug on tumor response. The impact of BED10 on the radiological response at 6 months was less clear, and it was possibly biased by the small number of some samples (e.g., data available for only 11 lesions treated with a BED10 < 15 Gy). A BED10 > 38 Gy significantly correlated (p = 0.001) with lower rates of SD and CR and higher rates of PR, while a BED10 of 15–38 Gy resulted in higher rates of SD and CR and a BED10 <15 Gy resulted in lower rates of PR and higher rates of PD. 4. Discussion In contrast with solitary plasmocytoma, for which RT is prescribed with curative intent [4,12], radiation treatment for MM is prescribed with palliative intent [4]. Radiotherapy exerts its effect through multiple mechanisms, including the apoptosis of cancer cells, the decompression of nervous structures, the reduction in tumor-associated inflammatory molecules, the inhibition of osteoclasts, and the remineralization of bone [13]. Despite the widespread use of RT for the symptomatic treatment of MM, multiple issues still have to be clarified, including the impact of the RT dose and the possible effect of CST. 4.1. Radiotherapy Dose The definition of the best dose and fractionation of RT for MM is debated, as the current literature is mostly composed of limited retrospective series. While initial analyses revealed an extremely high rate of symptom relief without a dose–response relationship [14], subsequent reports generally described higher pain control and recalcification rates for higher doses [15,16]. Only a few studies normalized the dose as BED. Among these, in a series of 153 patients (based on available information from 81 patients for pain relief and 69 patients for recalcification), the equivalent dose in 2-Gy fractions was significantly associated with increased rates of pain relief and recalcification [13]. Conversely, no impact of BED was reported in another cohort of 149 patients [2]. Another paper, published in 2019, reported no significant differences in pain response in 130 patients treated in the ‘biological era’, but patients who received a dose of 20–30 Gy had a significantly lower risk of pain recurrence compared with those who received lower doses [17]. Similarly, a comparison of short-course (8 Gy single fraction or 20 Gy/5 fr) and longer-course (30–40 Gy in 10–20 fractions) RT for 172 patients with spinal-cord compression from MM revealed better results in terms of motor function from the long-course treatment [18]. Only one randomized clinical trial compared different schedules for the treatment of MM (8 Gy single fraction versus 30 Gy in 10 fr): although the pain response and recalcification rates were similar among the two regimens, the patients in the fractionated treatment group reported better quality of life [19]. The results of these experiences influenced the current ILROG recommendations of 20–30 Gy administered in 5–15 daily fractions for bony sites, with 8 Gy in a single fraction preferred for patients with a dismal prognosis [4]. A study of the largest cohort of MM patients undergoing RT was recently published by Elhammali et al. [20]: 772 patients treated on 1513 sites were analyzed. The majority of the patients was treated with five of more fractions for a total dose of 20–25 Gy. The authors suggested that such doses were sufficient to achieve a low rate of re-irradiation (2.6%) and could provide durable pain relief and high radiological local control. Nonetheless, a BED10 ≤ 28 Gy was associated with a significantly higher risk of re-irradiation, and only 82 treated lesions were assessable for radiological response. Unfortunately, no data were provided regarding CST and toxicity. Our data seem to confirm previous reports, as the RT allowed high rates of pain control at the end of the course (87.4%), which further increased at three months (96.9%) and were substantially maintained at 6 months (94%). To the best of our knowledge, this is the largest cohort assessing radiological response, with promising results. Disease control was obtained for the vast majority of the lesions, with an overall response rate of 86.8%, and only 4.4% of the lesions developed in-field progression at 6 months. Consistently with some of the published series, the impact of BED10 on the radiological response was somehow unclear and possibly biased by the reduced numbers of some groups. With these limits, a BED10 < 15 Gy resulted in significantly higher rates of PD. On the other hand, although higher rates of pain control were achieved by the last day of RT with a BED10 > 38 Gy and lower rates with a BED10 < 15 Gy, BED10 had no impact on pain control at 3 or 6 months. Indeed, the assessment of pain response after RT schedules with a low number of fractions (1–5) is biased by the fact that a few weeks might be required to achieve maximal pain relief after treatment. Thus, as suggested by current guidelines, treatment with schedules characterized by a low BED10 (such as 8 Gy in a single fraction) could be an effective option for pain relief, while a BED10 > 15 Gy could be a better option if the aim is to achieve local disease control. The adoption of schedules with higher BED10 should not raise concerns of increased toxicity as, although a BED10 > 38 Gy resulted in higher rates of side effects, this was limited to reversible grade 1 events, and no difference in toxicity was reported at three or six months after RT between the groups treated with different BED10. 4.2. Concurrent Systemic Treatment Even fewer data, all from retrospective studies, are available regarding the potential synergy, interactions, and toxicities of RT and CST. Although some experiences suggest higher rates of recalcification for patients undergoing concurrent chemotherapy [16], other analyses reported no influence on recalcification [14] or pain response [13]. Moreover, in the majority of the available studies, the characteristics of systemic treatment and timing relative to RT were not clearly specified, and the definition of CST was variable. Only two studies analyzed the combination of biological agents and RT, with no details regarding the effect on tumor response. The data from a cohort of 39 patients (64 sites) who received RT with CST for osteolytic lesions were published in 2014 by Shin et al. [10]. RT was completed by 89.7% of the patients and by 14 of 16 the administered novel agents, the toxicity was fair and mainly hematologic, and the clinical pain response was optimal regardless of CST. Salgado et al. [9] evaluated 130 patients (279 treated sites) who received RT from 2007 to 2017, of whom 91 were administered concurrent (defined as administration from one month before to one month after RT) biological agents, mainly bortezomib, carfilzomib, and daratumumab. No significant difference in the incidence of acute or sub-acute toxicity was reported between the patients who received concurrent biological agents and those who received RT alone, and all the reported toxicities were grade < 3. Talamo et al. [2] retrospectively evaluated a cohort of 149 patients treated on 262 sites during the ‘biological era’. Unfortunately, the timing of the RT and systemic treatments and possible interactions were not reported, but the RT did not decrease the number of peripheral blood stem cells collected for autologous stem-cell transplant. In the absence of solid clinical data, the main concern over administering RT concurrently with systemic treatment is the possible sensitization of normal tissues, leading to increased toxicity. Moreover, two case reports suggest the potential of augmented intestinal toxicity when RT is performed on abdominal or pelvic sites concomitantly with bortezomib [11,21]. This could be explained by the radio-sensitizing effect of bortezomib [11] and the rapid turnover of the intestinal mucosa, and other classes of target therapy, such as cyclin-dependent kinase inhibitors [22], were also linked with unexpected severe radiation-induced enteritis [23]. Nonetheless, no increased toxicity was reported for the patients undergoing regimens including bortezomib during RT in our cohort. To the best of our knowledge, this is the largest series assessing the impact of systemic treatment during RT for MM and the only one to accurately define ‘concurrent’ treatment in the ‘biological era’. The results of our study seem to confirm the safety of RT administered to patients undergoing CST, considering both chemotherapy and target drugs. Although lower rates of G2 toxicities were reported for RT without CST and higher rates of G4 toxicities for concurrent chemotherapy, the safety was excellent overall, and no difference in toxicity was reported at one or three months across the groups receiving no CST, biologic agents, or chemotherapy. Systemic treatment had no impact on pain control at the end of RT. Surprisingly, the patients who underwent RT without CST reported a significantly higher rate of complete pain control at 1 and 3 months after the end of RT. This finding is likely biased by the extremely low rate of absence of pain control, and it should also be considered that pain control at six months was not affected by the CST. On the other hand, the radiological disease progression at six months was significantly more frequent for the lesions treated with RT without CST and less common in cases of the administration of concurrent biologic therapy (p = 0.044), suggesting a potential synergy with ionizing radiation. Nonetheless, systemic treatment in the intervening period between RT and radiological assessment could also have had an impact on the response at imaging, which would have been difficult to quantify; this may also have been due to the large numbers, combinations, and sequences of different systemic therapies. Despite the reassuring reported toxicity profile, when RT is performed on abdominal sites concurrently with potentially radio-sensitizing systemic agents, as a precaution, the treatment plan should be optimized to limit doses to the intestine. In selected cases, conformal treatments, such as intensity-modulated radiation therapy, could aid in reducing doses to the organs at risk, as has already been demonstrated for other hematological malignancies, such as lymphoma [24,25]. If concerns over toxicity persist, relatively low prescription doses could still allow satisfactory disease control [6]. The limits of our analysis must also be acknowledged. The evaluation of pain and its response to treatment is complex, as defining the precise site of pain in relation to RT is not always straightforward for a patient, and several confounding factors (including the use of analgesics and steroids) affect pain control. Although consisting of a large sample, our cohort was retrospective, which could explain the limited proportion of patients for whom complete data regarding the radiological response and the definition of the pain response according to VNS and analgesic drug intake were available. Moreover, some groups (considering both BED10 and the type of systemic treatment) were under-represented, limiting the statistical analysis and the potential to define the impact of each single drug. On the other hand, it should be considered that the main outcomes of this analysis were to assess the tolerability of RT alone or in combination with systemic agents, and the rates of radiological and pain control regardless of BED10. Prospective data from large populations are awaited to confirm our results. The management of MM is still evolving, and the possible impact of new imaging modalities [26,27] and combinations of treatments with new therapies, such as chimeric antigen receptor T-cell therapies [28] and checkpoint inhibitors [29] could open new perspectives. 5. Conclusions Despite a slight increase in the toxicity rates in the case of CST and the increased BED10, the toxicity profile of RT for the treatment of MM was excellent. Irradiation resulted in high pain control rates at the end of treatment, which further improved at three months and were substantially maintained at six months. The radiological disease control was also optimal, with only a few lesions presenting in-field progression. Schedules with a BED10 < 15 Gy allowed satisfactory pain control, but resulted in significantly higher rates of radiologic PD, and higher doses are thus suggested if the aim is not purely analgesic. A potential synergy of ionizing radiation and CST was suggested by the significantly higher rates of radiological PD for the lesions treated with RT without CST, and by the lower rates in case of the administration of concurrent biological therapy. Author Contributions Conceptualization, A.E.G., A.T., F.A., P.B., L.T., L.P., C.T., S.M.M., L.S. and M.B. (Michela Buglione); Data curation, A.E.G., A.T., F.A., E.M., A.B., P.B., L.T., L.P., S.P., M.B. (Marco Bonù), D.T., J.I., A.D., N.S., G.V. and M.B. (Michela Buglione); Formal analysis, A.E.G., A.T., F.A., E.M., A.B., L.T., L.P., S.P., D.T., C.T., S.M.M., L.S. and M.B. (Michela Buglione); Investigation, A.E.G., F.A., E.M., A.B., P.B., L.P., S.P., M.B. (Marco Bonù), D.T., J.I., A.D., G.F., N.S., G.V. and S.M.M.; Methodology, A.E.G., E.M., A.B., P.B., L.T., M.B. (Marco Bonù), D.T., C.T., S.M.M., L.S. and M.B. (Michela Buglione); Project administration, A.E.G., E.M., P.B., L.T., L.P., S.P., M.B. (Marco Bonù), J.I., A.D., G.F., N.S., G.V. and M.B. (Michela Buglione); Supervision, A.E.G., A.T., F.A., A.B., P.B., L.T., L.P., S.P., M.B. (Marco Bonù), D.T., J.I., G.F., N.S., G.V., C.T., S.M.M., L.S. and M.B. (Michela Buglione); Validation, F.A., A.B., L.T., L.P., S.P., M.B. (Marco Bonù), D.T., C.T., S.M.M., L.S. and M.B. (Michela Buglione); Visualization, A.D. and G.F.; Writing—original draft, A.E.G., A.T., E.M., S.M.M. and M.B. (Michela Buglione); Writing—review & editing, A.E.G., A.T., F.A., E.M., A.B., P.B., L.T., L.P., S.P., M.B. (Marco Bonù), D.T., J.I., A.D., G.F., N.S., G.V., C.T., S.M.M., L.S. and M.B. (Michela Buglione). All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement This study was approved by Spedali Civili of Brescia Ethics Committee, number of approval NP 4595, received 1 April 2021. Informed Consent Statement Specific patient consent for the participation in this study was waived due to its retrospective nature; nonetheless, all the enrolled patients signed a generic informed consent for the clinical and scientific treatment of their data. Data Availability Statement The data and material are stored according to our institutional protocols and are available upon request. Conflicts of Interest The authors declare no conflict of interest. cancers-14-02273-t001_Table 1 Table 1 Patients’ and treatment characteristics. BED10 = biologically effective dose (assuming an α/β ratio of 10); 3DCRT = 3-D conformal radiotherapy; 2DRT = 2-D radiotherapy; IMRT = intensity-modulated radiotherapy; VMAT = volumetric modulated arc therapy; Tomo = tomotherapy. Patients’ Characteristics (Per Access) Lines of Systemic Treatment before (Number and Percentage, Data Available for 409 Accesses) Mean 1.65 Range 0–10 One line 42 (10.3%) Two lines 65 (15.9%) Three lines 38 (9.3%) Four lines 28 (6.8%) Five or more lines 43 (10.5%) Number of Treated Lesions Mean 1.41 Range 1–4 One lesion 274 (66.7%) Two lesions 110 (26.8%) Three or more 27 (6.6%) Age at Start of Radiotherapy Median 68.1 years Range 31.2–92.5 years Treatment Characteristics (per Lesion) Systemic Treatment Concurrent with Radiotherapy No 220 (38.1%) Chemotherapy 76 (13.2%) Biologic treatment 281 (48.7%) Type of concurrent biological systemic treatment Lenalidomide or lenamide-containing combinations 76 (13.2%) Bortezomib 67 (11.6%) Bortezomib-containing combinations 117 (20.3%) Other biological agents 21 (3.6%) Site of treated lesions Vertebral 346 (60%) Extremities 108 (18.7%) Pelvic bones 57 (9.9%) Ribs/sternum 36 (6.4%) Skull 15 (2.6%) Other 15 (2.6%) Radiotherapy BED10 Less than 15 Gy 42 (7.3%) 15–38 Gy 165 (28.6%) More than 38 Gy 370 (64.1%) Radiotherapy schedules 30 Gy/10 fr 360 (62.4%) 20 Gy/5 fr 158 (27.4%) 8 Gy/1 fr 34 (5.9%) Radiotherapy technique 3DCRT 440 (76.3%) 2DRT 117 (20.3%) IMRT/VMAT/Tomo 20 (3.5%) cancers-14-02273-t002_Table 2 Table 2 Higher-grade toxicity reported during radiotherapy, at one month and at three months after the end of radiotherapy. RT = radiotherapy; CHT = chemotherapy; BED10 = biologically effective dose (assuming an α/β ratio of 10). Higher-Grade Toxicity (Per n of Accesses) During Radiotherapy (Data Available for 410 Accesses) G0 G1 G2 G3 G4 Total Overall 242 (59%) 151 (36.8%) 13 (3.2%) 3 (0.7%) 1 (0.2%) 410 RT alone 101 (64.3%) 54 (34.4%) 1 (0.6%) 1 (0.6%) 0 (0%) 157 Concurrent CHT 22 (43.1%) 24 (47.1%) 3 (5.9%) 1 (2%) 1 (2%) 51 Concurrent biological agent 119 (58.9%) 73 (36.1%) 9 (4.5%) 1 (0.5%) 0 (0%) 202 BED10 < 15 Gy 28 (90.3%) 2 (6.5%) 0 (0%) 1 (3.2%) 0 (0%) 31 BED10 15–38 Gy 86 (71.7%) 30 (25%) 3 (2.5%) 0 (0%) 1 (0.8%) 120 BED10 > 38 Gy 128 (50.6%) 119 (45.9%) 10 (3.9%) 2 (0.8%) 0 (0%) 259 At 1 Month after Radiotherapy (Data Available for 267 Accesses) G0 G1 G2 G3 G4 Total Overall 251 (94%) 14 (5.2%) 2 (0.7%) 0 (0%) 0 (0%) 267 RT alone 96 (92.3%) 7 (6.7%) 1 (1%) 0 (0%) 0 (0%) 104 Concurrent CHT 21 (100%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 21 Concurrent biological agent 134 (94.4%) 7 (4.9%) 1 (0.7%) 0 (0%) 0 (0%) 142 BED10 < 15 Gy 20 (100%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 20 BED10 15–38 Gy 67 (98.5%) 0 (0%) 1 (1.5%) 0 (0%) 0 (0%) 68 BED10 > 38 Gy 164 (91.6%) 14 (7.8%) 1 (0.6%) 0 (0%) 0 (0%) 179 At 3 Months after Radiotherapy (Data Available for 263 Accesses) G0 G1 G2 G3 G4 Total Overall 254 (96.6%) 8 (3%) 1 (0.4%) 0 (0%) 0 (0%) 263 RT alone 99 (96.1%) 3 (2.9%) 1 (1%) 0 (0%) 0 (0%) 103 Concurrent CHT 20 (100%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 20 Concurrent biological agent 135 (96.4%) 5 (3.6%) 0 (0%) 0 (0%) 0 (0%) 140 BED10 < 15 Gy 18 (100%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 18 BED10 15–38 Gy 66 (100%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 66 BED10 > 38 Gy 160 (94.7%) 8 (4.7%) 1 (0.6%) 0 (0%) 0 (0%) 169 cancers-14-02273-t003_Table 3 Table 3 Data regarding pain control and radiologic response. RT = radiotherapy; CHT = chemotherapy; BED10 = biologically effective dose (assuming an α/β ratio of 10). Pain Control (per n of Lesions) At End of Radiotherapy Complete control Partial control No control Total Overall 193 (35.9%) 277 (51.5%) 68 (12.6%) 538 RT alone 70 (34%) 102 (49.5%) 34 (16.5%) 206 Concurrent CHT 29 (39.7%) 35 (47.9%) 9 (12.3%) 73 Concurrent biological agent 94 (36.3%) 140 (54.1%) 25 (9.7%) 259 BED10 < 15 Gy 11 (39.3%) 6 (21.4%) 11 (39.3%) 28 BED10 15–38 Gy 51 (34.5%) 71 (48%) 26 (17.6%) 148 BED10 > 38 Gy 131 (36.2%) 200 (55.2%) 31 (8.6%) 362 At 1 Month after Radiotherapy Complete control Partial control No control Total Overall 244 (66.5%) 112 (30.5%) 11 (3%) 367 RT alone 105 (75.5%) 31 (22.3%) 3 (2.2%) 139 Concurrent CHT 17 (56.7%) 13 (43.3%) 0 (0%) 30 Concurrent biological agent 122 (61.6%) 68 (34.3%) 8 (4%) 198 BED10 < 15 Gy 14 (56%) 11 (44%) 0 (0%) 25 BED10 15–38 Gy 56 (60.9%) 34 (37%) 2 (2.2%) 92 BED10 > 38 Gy 174 (69.6%) 67 (26.8%) 9 (3.6%) 250 At 3 Months after Radiotherapy Complete control Partial control No control Total Overall 261 (73.5%) 83 (23.4%) 11 (3.1%) 355 RT alone 108 (80.6%) 25 (18.7%) 1 (0.7%) 134 Concurrent CHT 17 (58.6%) 12 (41.4%) 0 (0%) 29 Concurrent biological agent 136 (70.8%) 46 (24%) 10 (5.2%) 192 BED10 < 15 Gy 14 (66.7%) 7 (33.3%) 0 21 BED10 15–38 Gy 58 (64.4%) 28 (31.1%) 4 (4.4%) 90 BED10 > 38 Gy 189 (77.5%) 48 (19.7%) 7 (2.9%) 244 At 6 Months after Radiotherapy Complete control Partial control No control Total Overall 271 (77%) 60 (17%) 21 (6%) 352 RT alone 109 (79.6%) 20 (14.6%) 8 (5.8%) 137 Concurrent CHT 19 (67.9%) 9 (32.1%) 0 (0%) 28 Concurrent biological agent 143 (76.5%) 31 (16.6%) 13 (7%) 187 BED10 < 15 Gy 15 (71.4%) 5 (23.8%) 1 (4.7%) 21 BED10 15–38 Gy 65 (73.9%) 19 (21.6%) 4 (4.5%) 88 BED10 > 38 Gy 192 (78.7%) 36 (14.8%) 16 (6.6%) 244 Radiological Response at 6 Months (n of Lesions) Complete response Partial response Stable disease Disease progression Total Overall 16 (8.8%) 127 (70.2%) 30 (16.6%) 8 (4.4%) 181 RT alone 8 (11.8%) 43 (63.2%) 10 (14.7%) 7 (10.3%) 68 Concurrent CHT 0 (0%) 9 (69.2%) 4 (30.8%) 0 (0%) 13 Concurrent biological agent 8 (8%) 75 (75%) 16 (16%) 1 (1%) 100 BED10 < 15 Gy 2 (18.2%) 3 (27.3%) 4 (36.4%) 2 (18.2%) 11 BED10 15–38 Gy 8 (25.8%) 13 (41.9%) 9 (29%) 1 (3.2%) 31 BED10 > 38 Gy 6 (4.3%) 111 (79.9%) 17 (12.2%) 5 (3.6%) 139 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Cowan J.A. Allen C. Barac A. Basaleem H. Bensenor I. Curado M.P. Foreman K. Gupta R. Harvey J. Hosgood H.D. Global Burden of Multiple Myeloma: A Systematic Analysis for the Global Burden of Disease Study 2016 JAMA Oncol. 2018 4 1221 1227 10.1001/jamaoncol.2018.2128 29800065 2. Talamo G. Dimaio C. Abbi K.K. Pandey M.K. Malysz J. Creer M.H. Zhu J. Mir M.A. Varlotto J.M. Current role of radiation therapy for multiple myeloma Front. Oncol. 2015 5 40 10.3389/fonc.2015.00040 25741475 3. Hameed A. Brady J.J. Dowling P. Clynes M. O ‘Gorman P. Bone Disease in Multiple Myeloma: Pathophysiology and Management Cancer Growth Metastasis 2014 7 33 42 10.4137/CGM.S16817 25187738 4. Tsang R.W. Campbell B.A. Goda J.S. Kelsey C.R. Kirova Y.M. Parikh R.R. Ng A.K. Ricardi U. Suh C.O. Mauch P.M. Radiation Therapy for Solitary Plasmacytoma and Multiple Myeloma: Guidelines from the International Lymphoma Radiation Oncology Group Int. J. Radiat. Oncol. Biol. Phys. 2018 101 794 808 10.1016/j.ijrobp.2018.05.009 29976492 5. Gulla A. Anderson K.C. Multiple myeloma: The (r)evolution of current therapy and a glance into future Haematologica 2020 105 2358 2367 10.3324/haematol.2020.247015 33054076 6. Momm F. Greil C. Schäfer H. Towards individualized radiation therapy in multiple myeloma Haematologica 2020 105 1763 1764 10.3324/haematol.2019.243451 32611577 7. Lee Y.S. Heo W. Nam J. Jeung Y.H. Bae J. The combination of ionizing radiation and proteasomal inhibition by bortezomib enhances the expression of NKG2D ligands in multiple myeloma cells J. Radiat. Res. 2018 59 245 252 10.1093/jrr/rry005 29518205 8. Niyazi M. Maihoefer C. Krause M. Rödel C. Budach W. Belka C. Radiotherapy and “new” drugs-new side effects? Radiat. Oncol. 2011 6 1 19 10.1186/1748-717X-6-177 21208415 9. Salgado L.R. Wang S. Adler A. The Safety Profile of Concurrent Therapy for Multiple Myeloma in the Modern Era Adv. Radiat. Oncol. 2019 4 112 117 10.1016/j.adro.2018.09.009 30706018 10. Shin S.M. Chouake R.J. Sanfilippo N.J. Rapp T.B. Cook P. Formenti S.C. Mazumder A. Silverman J.S. Feasibility and efficacy of local radiotherapy with concurrent novel agents in patients with multiple myeloma Clin. Lymphoma Myeloma Leuk. 2014 14 480 484 10.1016/j.clml.2014.07.010 25176474 11. Mohiuddin M.M. Harmon D.C. Delaney T.F. Severe acute enteritis in a multiple myeloma patient receiving bortezomib and spinal radiotherapy: Case report J. Chemother. 2005 17 343 346 10.1179/joc.2005.17.3.343 16038531 12. Alghisi A. Borghetti P. Maddalo M. Roccaro A.M. Tucci A. Mazzola R. Magrini S.M. Lo Casto A. Bonù M.L. Tomasini D. Radiotherapy for the treatment of solitary plasmacytoma: 7-year outcomes by a mono-institutional experience J. Cancer Res. Clin. Oncol. 2021 147 1773 1779 10.1007/s00432-020-03452-y 33201300 13. Matuschek C. Ochtrop T.A. Bölke E. Ganswindt U. Fenk R. Gripp S. Kröpil P. Gerber P.A. Kammers K. Hamilton J. Effects of Radiotherapy in the treatment of multiple myeloma: A retrospective analysis of a Single Institution Radiat. Oncol. 2015 10 71 10.1186/s13014-015-0374-z 25889851 14. Leigh B.R. Kurtts T.A. Mack C.F. Matzner M.B. Shimm D.S. Radiation therapy for the palliation of multiple myeloma Int. J. Radiat. Oncol. Biol. Phys. 1993 25 801 804 10.1016/0360-3016(93)90308-I 7683017 15. Adamietz I.A. Schober C. Schulte R.W. Peest D. Renner K. Palliative radiotherapy in plasma cell myeloma Radiother. Oncol. 1991 20 111 116 10.1016/0167-8140(91)90144-6 1709508 16. Stölting T. Knauerhase H. Klautke G. Kundt G. Fietkau R. Total and single doses influence the effectiveness of radiotherapy in palliative treatment of plasmacytoma Strahlenther. Onkol. 2008 184 465 472 10.1007/s00066-008-1808-4 19016025 17. Wang S. Salgado L.R. Adler A. Chang S. Ru M. Moshier E. Dharmarajan K. Cho H.J. Bakst R.L. Dose Selection for Multiple Myeloma in Modern Era Pract. Radiat. Oncol. 2019 9 e400 e406 10.1016/j.prro.2019.02.009 30802617 18. Rades D. Hoskin P.J. Stalpers L.J. Schulte R. Poortmans P. Veninga T. Dahm-Daphi J. Obralic N. Wildfang I. Bahrehmand R. Short-course radiotherapy is not optimal for spinal cord compression due to myeloma Int. J. Radiat. Oncol. Biol. Phys. 2006 64 1452 1457 10.1016/j.ijrobp.2005.10.018 16413695 19. Rudzianskiene M. Inciura A. Gerbutavicius R. Rudzianskas V. Macas A. Simoliuniene R. Dambrauskiene R. Kiavialaitis G.E. Juozaityte E. Single vs. multiple fraction regimens for palliative radiotherapy treatment of multiple myeloma: A prospective randomised study Strahlenther. Onkol. 2017 193 742 749 10.1007/s00066-017-1154-5 28573476 20. Elhammali A. Amini B. Ludmir E.B. Gunther J.R. Milgrom S.A. Pinnix C.C. Andraos T. Yoder A. Weber D. Orlowski R. New paradigm for radiation in multiple myeloma: Lower yet effective dose to avoid radiation toxicity Haematologica 2020 105 e355 e357 10.3324/haematol.2019.235804 31919088 21. Berges O. Decaudin D. Servois V. Kirova Y.M. Concurrent radiation therapy and bortezomib in myeloma patient Radiother. Oncol. 2008 86 290 292 10.1016/j.radonc.2007.12.015 18199515 22. Guerini A.E. Pedretti S. Salah E. Simoncini E.L. Maddalo M. Pegurri L. Pedersini R. Vassalli L. Pasinetti N. Peretto G. A single-center retrospective safety analysis of cyclin-dependent kinase 4/6 inhibitors concurrent with radiation therapy in metastatic breast cancer patients Sci. Rep. 2020 10 13589 10.1038/s41598-020-70430-2 32788596 23. Buglione M. Guerini A.E. Filippi A.R. Spiazzi L. Pasinetti N. Magli A. Toraci C. Borghetti P. Triggiani L. Alghisi A. A Systematic Review on Intensity Modulated Radiation Therapy for Mediastinal Hodgkin ‘s Lymphoma Crit. Rev. Oncol. Hematol. 2021 167 103437 10.1016/j.critrevonc.2021.103437 34358649 24. Filippi A.R. Meregalli S. Di Russo A. Levis M. Ciammella P. Buglione M. Guerini A.E. De Marco G. De Sanctis V. Vagge S. Fondazione Italiana Linfomi (FIL) expert consensus on the use of intensity-modulated and image-guided radiotherapy for Hodgkin ‘s lymphoma involving the mediastinum Radiat. Oncol. 2020 15 62 10.1186/s13014-020-01504-8 32164700 25. Albano D. Stecco A. Micci G. Sconfienza L.M. Colagrande S. Reginelli A. Grassi R. Carriero A. Midiri M. Lagalla R. Whole-body magnetic resonance imaging (WB-MRI) in oncology: An Italian survey Radiol. Med. 2021 126 299 305 10.1007/s11547-020-01242-7 32572763 26. Belgioia L. Vagge S. Tagliafico A. Corvò R. How Can Imaging Help the Radiation Oncologist in Multiple Myeloma Treatment Medicina 2020 57 20 10.3390/medicina57010020 33379406 27. Treglia G. Bertagna F. Albano D. The “Undetermined Significance” of 18F-FDG PET/CT or PET/MRI in Patients with Monoclonal Gammopathy of Undetermined Significance (MGUS) Medicina (Kaunas) 2021 57 856 10.3390/medicina57080856 34441062 28. Guerini A.E. Filippi A.R. Tucci A. Simontacchi G. Re A. Guaineri A. Morelli V. Borghetti P. Triggiani L. Pegurri L. ‘Le Roi est mort, vive le Roi‘: New roles of radiotherapy in the treatment of lymphomas in combination with immunotherapy Clin. Lymphoma Myeloma Leuk. 2022 22 e135 e148 10.1016/j.clml.2021.09.005 34728169 29. Kazandjian D. Dew A. Hill E. Ramirez E.G. Morrison C. Mena E. Lindenberg L. Yuan C. Maric I. Wang H.W. Avelumab, a PD-L1 Inhibitor, in Combination with Hypofractionated Radiotherapy and the Abscopal Effect in Relapsed Refractory Multiple Myeloma Oncologist 2021 26 288-e541 10.1002/onco.13712 33554406

PMC009xxxxxx/PMC9099691.txt

==== Front Nutrients Nutrients nutrients Nutrients 2072-6643 MDPI 10.3390/nu14091930 nutrients-14-01930 Systematic Review Acute Effects of Caffeine on Overall Performance in Basketball Players—A Systematic Review Lazić Anja 1 Kocić Miodrag 1 Trajković Nebojša 1 Popa Cristian 2† https://orcid.org/0000-0003-1742-5016 Peyré-Tartaruga Leonardo Alexandre 3† https://orcid.org/0000-0002-4254-3105 Padulo Johnny 4*† Davison Glen Academic Editor 1 Faculty of Sport and Physical Education, University of Niš, 18000 Niš, Serbia; anja.lazic96@hotmail.com (A.L.); miodrag.kocic73@gmail.com (M.K.); nele_trajce@yahoo.com (N.T.) 2 Faculty of Physical Education and Sport, Ovidius University of Constanta, 900470 Constanta, Romania; crispopa2002@yahoo.com 3 Exercise Research Laboratory, Universidade Federal do Rio Grande do Sul, Porto Alegre 90690-200, RS, Brazil; leonardo.tartaruga@ufrgs.br 4 Department of Biomedical Sciences for Health, Università degli Studi di Milano, 20133 Milan, Italy * Correspondence: johnny.padulo@unimi.it † These authors contributed equally to this work. 05 5 2022 5 2022 14 9 193022 3 2022 23 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Caffeine supplementation has become increasingly popular among athletes. The benefits of caffeine include delaying the negative effects of fatigue, maintaining a high level of physical and mental performance, and improving certain abilities necessary for sport success. Given the complex nature of basketball, caffeine could be a legal, ergogenic stimulant substance, which will positively affect overall basketball performance. The purpose of this systematic review was to summarize evidence for the effect of acute caffeine ingestion on variables related to the basketball performance. Web of Science, PubMed, Scopus and ProQuest, MEDLINE, and ERIC databases were searched up to February 2021. Studies that measured the acute effect of caffeine on basketball performance were included and analyzed. Eight studies published between 2000 and 2021 were included in the analysis. Pre-exercise caffeine intake increased vertical jump height, running time at 10 and 20 m without the ball, overall basketball performance (number of body impacts, number of free throws, rebounds, and assists) during simulated games, and reduced the time required to perform a basketball-specific agility test. Equivocal results between caffeine and placebo groups were found for aerobic capacity, free throw and three-point accuracy, and dribbling speed. Pre-exercise caffeine ingestion did not affect RPE, but insomnia and urinary excretion were increased. The pre-exercise ingestion of 3 and 6 mg/kg caffeine was found to be effective in increasing several physical performance variables in basketball players during sport-specific testing and simulated matches. However, considering the intermittent nature and complexity of basketball, and individual differences between players, future studies are needed. supplementation nutrition team sport explosive power agility speed ==== Body pmc1. Introduction Basketball players have been affected by very high internal and external loads during training and competition [1]. Performance implies a large volume of high-intensity activities of short duration—speed, strength, agility, and endurance [2]. These activities represent the basis for the performance of jumping and sprinting movements in different directions and at different angles, and they are combined with sports-specific elements [3,4]. As a multifactorial game, basketball includes shooting and passing accuracy as an integral, crucial, and most frequent technical part of the activity [5,6]. During a basketball match, players perform over 50 jumps [7], 48.7% of all basketball activities include combination of jumping and shooting movements, and 28.5% imply repeated sprint ability [8]. Additionally, a distance of 4000 to 5000 m is covered within 40 min [9]. The previously highlighted facts state the complex nature of the metabolic systems of basketball, mixing aerobic and anaerobic demands. Performing repeated high-intensity activities during training sessions and games induces a great pressure on the central nervous system (CNS) and causes fatigue [10]. Unquestionably, fatigue may decrease basketball performance [11]. In recent years, researchers have become increasingly interested in a variety of ergogenic substances that can delay the effect of fatigue and improve athletic performance [12]. Moreover, by removing caffeine from the list of prohibited substances [13] this supplement has become the most popular ergogenic substance used by athletes [14]. More precisely, data obtained from urine samples of athletes from different sports show that three out of four athletes consume caffeine-based supplements before or during competition [14]. Most researchers working in the area of caffeine effects on athletic performance agree that caffeine has an important role in performance improvement during endurance activities [15,16,17], as well as in power-related activities [18,19]. In addition, there is strong evidence that the consumption of certain doses of caffeine may even improve performance during repeated sub-maximal and maximal activities [18,20,21]. The main physiological symptoms of caffeine use are respiratory rate and minute volume increases, overstimulation of the respiratory centers, intensifying pulmonary blood flow, and the sensitivity of central medullary areas to hypercapnia. These alterations are mediated by the antagonistic action of caffeine on adenosine receptors [22]. Given that many high-intensity activities are present in basketball, and that their manifestation leads to the presence of acute fatigue, and at the same time to great stress to the CNS, caffeine supplementation could be an adequate solution for maintenance or improvement overall performance in this sport. Previous studies [23,24,25,26] have explored the role of caffeine intake in team ball games. Recently, in a systematic review conducted by Chia and colleagues [23], conflicting results have been reported in the most popular ball games. Improvements caused by caffeine intake are shown in jumping and running activities, but authors have failed to report positive effects on specific motor abilities—agility and accuracy. This finding is congruent with a later systematic review with meta-analysis [24], which showed that ingestion of caffeine has small but significant effects on several components of physical performance in team sports. However, previous research in the field of investigation has been restricted to only two studies in basketball and a small sample size. Principally, with this background, it is very difficult to determine whether caffeine is an ergogenic substance for basketball players. In addition to the unclear literature evaluating caffeine’s effect on basketball performance, to the best of our knowledge, there is a limited number of systematic reviews that focused on the acute effects of caffeine on the overall performance in basketball. This paper provides an overview of the effect of acute caffeine ingestion on variables related to basketball performance. 2. Materials and Methods 2.1. Search Strategy The following databases were used to collect adequate literature: Web of Science, PubMed, Scopus and ProQuest, MEDLINE and ERIC up to February 2022. A search of databases was carried out using a combination of the keywords: caffeine, caffeinated, basketball, endurance, performance, strength. The PubMed search is shown below: (Caffeine) OR (caffeinated energy drinks)) OR (caffeinated energy drink)) AND (basketball)) OR (basketball performance)) AND (explosive power)) OR (strength)) OR (speed)) OR (aerobic capacity)) OR (endurance)) OR (flexibility)) OR (agility)) OR (change of direction speed)) OR (precision)) OR (accuracy) A systematic review of the available literature was undertaken in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [27] (Figure 1). All studies that passed the process of previous selection were downloaded to Endnote X9 (Clarivate Analytics, New York, NY, USA) and a subsequent review of the entire text of all studies was performed to remove duplicates. The search strategy was performed by two authors independently (N.T. and A.L.) and study selection was limited to the articles published in English. Disagreements were resolved through discussion of two mentioned authors. The selected studies are presented according to the following parameters in Table 1 and Table 2. Table 1 shows basic information about the study and the participants: (a) reference, (b) participants (gender, number, and age), (c) level of playing, (d) daily dose of caffeine intake through various products, (e) PEDro score rating. Table 2 presents the parameters of the experimental treatment: (a) dose of caffeine, (b) the form in which caffeine is consumed during testing, (c) tested variables, (d) results, (e) results shown by an adequate symbol (↑—improvement, ↓—reduction, ↔—no changes). 2.2. Inclusion Criteria In this systematic review, the following criteria had to be met: (1) studies published in English; (2) studies examining the acute effects of caffeine on basketball performance; (3) studies that were subjected to a randomization process, were controlled, armored and with a cross-over design, and which involved comparing the acute effects of caffeine with the effects of a placebo substance; (4) studies in which the exact dose of caffeine used per kilogram of body mass was precisely explained in the experimental design, as well as studies in which the time period of caffeine/placebo consumption before the experimental program was highlighted; (5) it was necessary that the time interval between tests was at least 24 h; (6) the sample of participants consisted of basketball players of both sexes, regardless of age and competitive rank; (7) studies published in the period from 2000 to 2021; (8) all studies had to be approved by the appropriate Ethics Committee, with the appropriate number and confirmed date of approval. 2.3. Exclusion Criteria The following criteria were applied for studies exclusion from further analysis: (1) studies were not published in English; (2) studies with abstract available only; (3) longitudinal studies with the aim of examining the long-term effects of caffeine supplementation; (4) if the caffeine intervention contained less than 2 mg/kg which showed insufficient ergogenic effects on various forms of exercise; (5) lack of placebo treatment group; (6) studies that did not have basketball players as the participants; (7) All systematic reviews research or systematic reviews with meta-analysis which investigated the acute effects of caffeine in team sports with the ball; (8) the studies which included wheelchair basketball players as the participants. 2.4. Quality Assessment of the Experimental Studies To determine the quality of each experimental program, the Physiotherapy Evidence Database Scale (PEDro) scale was used (Table 1). The PEDro scale is a valid and reliable way to determine the internal validity of randomized measurements [36]. The PEDro scale consists of 11 items that include information on randomization, blinding procedure, statistical analysis data, and presentation of results in the study. Any study that did not meet the criteria, i.e., to which less than six points were assigned during the PEDro scale check, was excluded from further analysis. Two authors (J.P. and M.K.) independently performed methodological quality using the PEDro scale description guidelines, and disagreements were resolved through discussion. 3. Results A total of 1200 studies were extracted by the initial search from the databases search and 7 articles from the list of references. In addition, 1047 studies were excluded due duplicates found, and 160 studies are further screened. We removed 152 studies after full-text review. Finally, in total, eight studies meet the inclusion criteria. The Table 1 and Table 2 highlight general information about the participants and experimental treatment. The eight studies included in this systematic review were published between 2014 and 2021. The total sample consisted of 120 basketball players and the largest number of the participants was n = 21 [32], while the smallest was n = 5 [28]. Of the total number, male participants were included in three studies [28,29,35], while female participants were included in one study only [33]; both male and female participants were included in four studies [30,31,32,34]. The oldest group was 27.9 ± 6.1 years old [30], and the youngest was 14.9 ± 0.8 years old [29]. All participants were low-to-moderate daily caffeine consumers. In seven studies, the caffeine dosage was 3 mg/kg [28,29,30,31,32,33,35] and in one study it was 6 mg/kg [34]. Of the total number of studies, in seven studies, caffeine was taken in the form of a capsule, and in one by consuming an energy drink [29]. The applied dose of caffeine was taken 60 min before testing in all selected studies, and the PEDro scale score ranged from 8 to 10 points and thus, the quality of these studies can be considered to be of moderate to high quality. 4. Discussion The purpose of this systematic review was to summarize the effects of acute caffeine ingestion on variables related to the basketball performance. 4.1. Acute Effects of Caffeine on Physical Performance of Basketball Players The analysis of the results showed that caffeine in doses of 3 mg/kg has a positive effect on improving the height of the vertical jump [29,30,31,33,35] and linear speed at 10 and 20 m without the ball [33]. Subsequently, the dose of 3 mg/kg increased the number of body impacts and overall basketball performance during simulated games [30]. The vertical jump is one of the most important components of the basketball game [37]. The use of 3 mg/kg of caffeine improves the height of the jump in basketball players, which is congruent with previous studies [38,39,40], in which caffeine was the ergogenic substance used for improving vertical jump height. However, it should be mentioned that the increase in the height of the vertical jumping in these studies were obtained by using energy drinks, which contain other stimulant substances in addition to caffeine that may affect performance [41]. The positive effect of caffeine on the vertical jump in basketball was also proven in later studies in which the participants ingested pure caffeine [30,33,35]. The difference is that the effect of the improvement in basketball players was greater than that caused by the energy drinks [29]. Abian-Vicen et al. [29] examined acute caffeine-based drink in young basketball players and found an increase in the vertical jump of 2.1%, which is lower when compared to the increase of 4.6% in countermovement jump, the 3.8% in countermovement jump with arm swing, and the 4.8% in squat jump reported by using pure caffeine [33]. Also, professional players are more susceptible to the ergogenic effects of caffeine when compared to younger players [42], which is one of the reasons for this difference. Since change of direction speed in basketball is a frequent activity and it is related to all other elements of the game and to the CNS, only two studies [30,33] tried to determine the effects of 3 mg/kg of caffeine on this ability. While Puente et al. [30] did not find statistically significant improvements on the change of direction agility test, Stojanović et al. [33] came to the conclusion that caffeine can reduce the time required to perform this activity. The possible reason for the discrepancy in the obtained results can be explained by the fact that the change of direction agility test cannot be applied to a specific basketball performance [33]. Conversely, the work of Stojanović et al. [33] showed a small improvement in time on the agility test and this result should be considered more relevant because a specific basketball test was used in the testing that reflects the movements present in the game [43]. In favor of this finding, Davis and Green [42] reported that the efficacy of caffeine on anaerobic performance is more effective when the assessment protocol reflects sports demands. 4.2. Acute Effects of Caffeine on Specific Basketball Performance Caffeine as the ergogenic substance can affect the improvement of psychomotor and cognitive functions during fatigue that occurs because of high-intensity activities and thus has an impact on the adequate performance of technical elements [44]. However, research in this area of investigation has failed to explore the acute effects of 3 mg/kg [29,30] and 6 mg/kg [34] on overall accuracy and dribbling speed [32] in basketball players. Even though there was an increased number of free throws, the use of 3 mg/kg caffeine did not result in the improvement of shot accuracy. Specifically, Tan et al. [34] were unable to identify changes in accuracy by using 6 mg/kg of caffeine. More precisely, Tan et al. [34] found that in the state of acute fatigue, considering individual variations, there was no overall improvement in accuracy when performing free throws. Accuracy is a complex ability and depends on the interaction of numerous factors such as the trajectory of the ball after ejection, the phase of the shooting activity, the movements of which each phase consists and postural balance of the player [45], as well as a limited visual field and tactile sensation of the fingertips [46]. In addition to physical performance, caffeine has effects on hormonal, metabolic, and psychological status [47]. Some authors [24] do not recommend the use of caffeine in team sports with the ball, in which technical and tactical elements are a key factor [5,6]. Namely, the use of caffeine leads to insomnia, increased nervousness, and tremor [48,49], which can negatively affect the mentioned parameters. However, although the use of caffeine initiated increased insomnia [30] after the experimental treatment, caffeine also increased the number of rebounds, assists, performance index, and total body impacts [30] by increasing awareness and alertness during testing basketball players. To avoid the negative effects of caffeine such as insomnia, it is better to use it in the morning, which was confirmed in the study by Stojanović et al. [35]. It was concluded that 3 mg/kg of caffeine had effects on improved performance in the morning, compared to afternoon testing. However, it is difficult to apply these results in practice, due to most games and training being performed in the evening. 4.3. Dosing, Timing and Individual Responses to Caffeine The dose of caffeine used in the studies were above 3 mg/kg, suggesting that doses below 2 mg/kg may not be effective for performance in team sport [24,32]. Caffeine, or an adequate placebo, was ingested 60 min before the experimental program, since is it claimed that caffeine needs 30 to 60 min to be absorbed and to reach the plasma concentration peak [50]. The daily use of caffeine in all studies is labeled as low to moderate. However, daily use was measured based on the mean value of the group and it is necessary to emphasize the individual amount of consumption, because caffeine is strongly individually determined and may not have the ergogenic effect in all users [51]. Also, training status, tolerance to caffeine, or genotype variation have been defined as factors that can influence acute caffeine effects [52]. While some limiting factors can be reversed by increasing the dose of caffeine, numerous studies [53,54,55,56] have drawn parallels between an adequate response to caffeine intake and the presence of different genetic variations. However, a study carried out with 21 male and female basketball players found that both AA homozygotes and C-allele carriers similarly increased their physical performance after ingestion of 3 mg/kg of caffeine [30]. Future studies will have to identify mechanisms through which genetic variations correlate with caffeine intake. 4.4. Study Limitations This systematic review presents some limitations related to the variety of performance tests and variables used in the studies included, as well as the lack of studies based on caffeine effects on basketball performance. Secondly, our findings are not generalizable to the entire basketball population due to the small number of the participants. Previously highlighted facts resulted in the impossibility of meta-analysis and more relevant insights into the acute effects of caffeine. However, given that basketball is a complex sport in which physical performance is only one of several factors necessary for succeeding, our findings should not be over-interpreted. Therefore, several findings of this systematic review warrant further discussion, such as the acute effects of caffeine supplementation on sport-specific situations, especially decision-making situations when players are fresh, and during acute fatigue. Moreover, future studies in this area of investigation should test mechanisms involved in the generation of post-activation potentiation with concomitant caffeine ingestion. Therefore, these findings are still incomplete and still not enough to determine whether caffeine improves overall basketball performance. 5. Conclusions and Practical Applications In summary, the pre-exercise ingestion of 3 and 6 mg/kg of caffeine was found to be effective in increasing several physical performance variables in basketball players during sport-specific testing and simulated matches. Caffeine significantly increases vertical jump performance, sprint performance without the ball, planned agility, number of three throws, rebounds, assists, and body impacts during simulated matches. Equivocal results were found for endurance, accuracy, and dribbling speed. This study provides data that will help coaches, nutritionists, and basketball players in resolving doubts about the use of caffeine as a stimulant pre-game or during the matches. Finally, future studies will have to continue to explore the effects of caffeine in situations such as those during the games and, during acute fatigue taking individual differences between players into account. Acknowledgments The authors want to acknowledge all researchers who participated in the investigations cited in this manuscript. Author Contributions Conceptualization, N.T. and J.P.; methodology, C.P.; formal analysis, C.P. and M.K.; investigation, N.T.; writing—original draft preparation, A.L.; writing—review and editing, A.L. and L.A.P.-T.; visualization, L.A.P.-T. and M.K.; supervision, J.P. and N.T. All authors have read and agreed to the published version of the manuscript. Funding This research received no external funding. Institutional Review Board Statement No applicable. Informed Consent Statement No applicable. Data Availability Statement No new data were created or analyzed in this study. Conflicts of Interest The authors declare no conflict of interest. Figure 1 PRISMA preferred reporting items for systematic reviews and meta-analysis. nutrients-14-01930-t001_Table 1 Table 1 Descriptive statistics of the participants and PEDro scale for assessing quality of included studies. Study Participants Level of Playing Daily Caffeine Intake (mg/day, week) PEDro Score M/F N Age (Years) Tucker et al. [28] M 5 22 ± 1 professional <500 8 Abian-Vicen et al. [29] M 16 14.9 ± 0.8 professional <60 9 Puente et al. [30] M/F 10/10 27.1 ± 4.0/27.9 ± 6.1 professional <100 9 Puente et al. [31] M/F (AA/CC) 10/9 26.5 ± 2.4/27.0 ± 5.3 professional <100 10 Scanlan et al. [32] M/F 11/10 18.3 ± 3.3 professional <100 10 Stojanović et al. [33] F 10 20.2 ± 3.9 professional <100 10 Tan et al. [34] M/F 12/6 23.1 ± 1.9/22.0 ± 1.3 college <200 10 Stojanović et al. [35] M 11 16.5 ± 1.0 juniors 310 ± 76 10 M/F—male/female; N—number; AA—AA homozygotes; CC—allele carriers, PEDro—Physiotherapy Evidence Database Scale. nutrients-14-01930-t002_Table 2 Table 2 Experimental program of acute caffeine ingestion in basketball players. Study Form of Caffeine Dose (mg/kg) Timing (min) Variables (Unit) Placebo Caffeine Results Tucker et al. [28] Capsule 3 60 VO2max (rep) 10VJ (cm) 118/126/109/122/83 ± 13/14/6/14/4 124/117/119/111/87 ± 5/13/9/9/4 ↔ ↔ Abian-Vicen et al. [29] Energy drink 3 60 FTA (%) TPA (%) CMJ (cm) 15—RJ (cm) YoYo IR 1 (m) UE (μg/mL) 70.7 ± 11.8 39.9 ± 11.8 37.5 ± 4.4 28.8 ± 3.4 1.925 ± 702 0.1± 0.1 70.3 ± 11.0 38.1 ± 12.8 38.3 ± 4.4 30.2 ± 3.6 2.000 ± 706 1.2 ± 0.7 ↔ ↔ ↑ ↑ ↔ ↑ Puente et al. [30] Capsule 3 60 VJ (cm) CODAT (s) CODATwb (s) FTA (%) FT (n) Reb (n) Ass (n) Imp (imp/min) PIR (%) POI (%) 37.3 ± 6.8 5.96 ± 0.29 6.20 ± 0.29 15.4 ± 1.6 0.6 ± 0.8 2.5 ± 2.0 1.1 ± 0.9 396 ± 43 7.2 ± 8.6 19.0 38.2 ± 7.4 5.95 ± 0.3 6.14 ± 0.32 15.6 ± 2.3 1.1 ± 1.1 3.7 ± 2.6 2.1 ± 1.6 410 ± 41 10.6 ± 7.1 54.4 ↑ ↔ ↔ ↔ ↑ ↑ ↑ ↑ ↑ ↑ Puente et al. [31] Capsule 3 60 VJ (cm) CODAT (s) CODATwb (s) Mhr (bpm) Phr (bpm) Imp (imp/min) 39.6 ± 7.2/36.3 ± 5.9 5.91 ± 0.25/5.95 ± 0.33 6.19 ± 0.21/6.14 ± 0.35 158 ± 9/161 ± 13 187 ± 12/182 ± 7 385 ± 48/401 ± 36 40.7 ± 7.3/37.2 ± 6.9 5.88 ± 0.27/5.97 ± 0.38 6.09 ± 0.24/5.97 ± 0.38 160 ± 10/163 ± 9 188 ± 13/185 ± 6 401 ± 36/415 ± 35 ↑/↔ ↔ ↔ ↔ ↔ ↑/↑ Scanlan et al. [32] Capsule 3 60 TDT20m (s) DD20m (s) 3.560 ± 0.184 0.145 ± 0.138 3.528 ± 0.208 0.150 ± 0.129 ↔ ↔ Stojanović et al. [33] Capsule 3 60 CMJ (cm) CMJa (cm) SJ (cm) LAT (s) 5 m sprint (s) 10 m sprint (s) 20 m sprint (s) 5 m sprint wb (s) 10 m sprint wb (s) 20 m sprint wb (s) RSA (s) RPE (AU) PP (AU) 27.92 ± 4.24 33.85 ± 3.92 25.97 ± 3.16 13.22 ± 0.87 1.24 ± 0.15 2.11 ± 0.18 3.59 ± 0.25 1.22 ± 0.08 2.07 ± 0.11 3.65 ± 0.15 32.20 ± 1.74 7.8 ± 1.2 3.6 ± 2.8 29.20 ± 4.39 35.14 ± 5.08 27.22 ± 4.37 12.99 ± 0.86 1.18 ± 0.11 2.01 ± 0.13 3.49 ± 0.23 1.20 ± 0.05 2.05 ± 0.12 3.56 ± 0.25 31.80 ± 1.62 5.6 ± 2.5 3.6 ± 2.8 ↑ ↑ ↑ ↔ ↔ ↑ ↑ ↔ ↔ ↔ ↔ ↓ ↔ Tan et al. [34] Capsule 6 60 FTA (%) HR (bpm) RPE (AU) 5.5 ± 2.0 163 ± 12.1 15.7 ± 2.1 6.1 ± 1.7 166 ± 9.2 15.8 ± 2.1 ↔ ↑ ↔ Stojanović et al. [35] Capsule 3 60 CMJ am/pm (cm) CMJa am/pm (cm) SJ am/pm (cm) LAT am/pm (s) 20 m sprint (s) RSA (s) RSAwb (s) 31.03 ± 4.98 (am) 39.98 ± 5.23 (am) 30.55 ± 4.89 (am) 12.61 ± 0.84 (pmc) X 29.91 ± 1.31 X 33.90 ± 5.38 (am) 42.32 ± 5.69 (am) 33.20 ± 4.71(am) 11.98 ± 0.70 (amc) X 26.49 ± 1.62 (amc) X ↑ ↑ ↑ ↑ ↑ VO2max—maximal oxygen uptake; VJ—vertical jumps; FTA—free throw accuracy; TPA—three point accuracy; CMJ—countermovement jump; RJ—repeated jumps; YoYo IR 1—yoyo intermittent recovery test; UE—urinary excretion; CODAT (wb)—change of direction agility test (with ball); Reb—rebounds; Ass—assists; Imp—body impacts; PIR—performance index; POI—prevalence of insomnia; Mhr—mean heart rate; Phr—peak heart rate; TDT—total dribbling time; DD—dribbling deficit; CMJa—countermovement jump with arm swing; SJ—squat jump; LAT—lane agility drill test; RSA (wb)—repeated sprint ability (with ball); RPE—perceived exertion; PP—performance; HR—heart rate; am—morning; pm—evening; amc—morning caffeine group; pmc—evening caffeine group; ↑—improvement; ↓—reduction; ↔—unchanged. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Hůlka K. Cuberek R. Bělka J.J.A.G. Heart rate and time-motion analyses in top junior players during basketball matches Acta Gymnica 2013 43 27 35 10.5507/ag.2013.015 2. Stojanović E. Stojiljković N. Scanlan A.T. Dalbo V.J. Berkelmans D.M. Milanović Z. The activity demands and physiological responses encountered during basketball match-play: A systematic review Sports Med. 2018 48 111 135 10.1007/s40279-017-0794-z 29039018 3. Arendt E. Dick R. Knee injury patterns among men and women in collegiate basketball and soccer: NCAA data and review of literature Am. J. Sports Med. 1995 23 694 701 10.1177/036354659502300611 8600737 4. Piasecki D.P. Spindler K.P. Warren T.A. Andrish J.T. Parker R.D. Intraarticular injuries associated with anterior cruciate ligament tear: Findings at ligament reconstruction in high school and recreational athletes: An analysis of sex-based differences Am. J. Sports Med. 2003 31 601 605 10.1177/03635465030310042101 12860552 5. Nunes H. Iglelias X. Daza G. Irurtia A. Caparrós T. Anguera M. The influence of pick and roll in attacking play in top-level basketball Cuad. Psicol. Deporte 2016 16 129 142 6. Gómez M.-Á. Battaglia O. Lorenzo A. Lorenzo J. Jimenez S.L. Sampaio J. Effectiveness during ball screens in elite basketball games J. Sports Sci. 2015 33 1844 1852 10.1080/02640414.2015.1014829 25703612 7. Montgomery P.G. Pyne D.B. Minahan C.L. The physical and physiological demands of basketball training and competition Int. J. Sports Physiol. Perform. 2010 5 75 86 10.1123/ijspp.5.1.75 20308698 8. Abdelkrim N.B. Castagna C. El Fazaa S. Tabka Z. El Ati J. Blood metabolites during basketball competitions J. Strength Cond. Res. 2009 23 765 773 10.1519/JSC.0b013e3181a2d8fc 19387403 9. Crisafulli A. Melis F. Tocco F. Laconi P. Lai C. Concu A. External mechanical work versus oxidative energy consumption ratio during a basketball field test J. Sports Med. Phys. Fit. 2002 42 409 10. Rojas-Valverde D. Montoya-Rodríguez J. Azofeifa-Mora C. Sanchez-Urena B. Effectiveness of beetroot juice derived nitrates supplementation on fatigue resistance during repeated-sprints: A systematic review Crit. Rev. Food Sci. Nutr. 2021 61 3395 3406 10.1080/10408398.2020.1798351 32715742 11. Mulazimoglu O. Yanar S. Tunca Evcil A. Duvan A. Examining the effect of fatigue on shooting accuracy in young basketball players Anthropologist 2017 27 77 80 10.1080/09720073.2017.1311671 12. Slater G.J. Sygo J. Jorgensen M. SPRINTING... Dietary approaches to optimize training adaptation and performance Int. J. Sport Nutr. Exerc. Metab. 2019 29 85 94 10.1123/ijsnem.2018-0273 30943814 13. Desbrow B. Leveritt M. Well-trained endurance athletes’ knowledge, insight, and experience of caffeine use Int. J. Sport Nutr. Exerc. Metab. 2007 17 328 339 10.1123/ijsnem.17.4.328 17962708 14. Del Coso J. Muñoz G. Muñoz-Guerra J. Prevalence of caffeine use in elite athletes following its removal from the World Anti-Doping Agency list of banned substances Appl. Physiol. Nutr. Metab. 2011 36 555 561 10.1139/h11-052 21854160 15. Bell D.G. McLellan T.M. Exercise endurance 1, 3, and 6 h after caffeine ingestion in caffeine users and nonusers J. Appl. Physiol. 2002 93 1227 1234 10.1152/japplphysiol.00187.2002 12235019 16. Graham T. Hibbert E. Sathasivam P. Metabolic and exercise endurance effects of coffee and caffeine ingestion J. Appl. Physiol. 1998 85 883 889 10.1152/jappl.1998.85.3.883 9729561 17. Jenkins N.T. Trilk J.L. Singhal A. O’Connor P.J. Cureton K. Ergogenic effects of low doses of caffeine on cycling performance Int. J. Sport Nutr. Exerc. Metab. 2008 18 328 342 10.1123/ijsnem.18.3.328 18562777 18. Meyers B. Cafarelli E. Caffeine increases time to fatigue by maintaining force and not by altering firing rates during submaximal isometric contractions J. Appl. Physiol. 2005 99 1056 1063 10.1152/japplphysiol.00937.2004 15879163 19. Wiles J.D. Coleman D. Tegerdine M. Swaine I.L. The effects of caffeine ingestion on performance time, speed and power during a laboratory-based 1 km cycling time-trial J. Sports Sci. 2006 24 1165 1171 10.1080/02640410500457687 17035165 20. Gliottoni R.C. Meyers J.R. Arngrímsson S.Á. Broglio S.P. Motl R.W. Effect of caffeine on quadriceps muscle pain during acute cycling exercise in low versus high caffeine consumers Int. J. Sport Nutr. Exerc. Metab. 2009 19 150 161 10.1123/ijsnem.19.2.150 19478340 21. Woolf K. Bidwell W.K. Carlson A.G. The effect of caffeine as an ergogenic aid in anaerobic exercise Int. J. Sport Nutr. Exerc. Metab. 2008 18 412 429 10.1123/ijsnem.18.4.412 18708685 22. Barreto G. Grecco B. Merola P. Reis C.E.G. Gualano B. Saunders B. Novel insights on caffeine supplementation, CYP1A2 genotype, physiological responses and exercise performance Eur. J. Appl. Physiol. 2021 121 749 769 10.1007/s00421-020-04571-7 33403509 23. Chia J.S. Barrett L.A. Chow J.Y. Burns S.F. Effects of caffeine supplementation on performance in ball games Sports Med. 2017 47 2453 2471 10.1007/s40279-017-0763-6 28741186 24. Salinero J.J. Lara B. Del Coso J. Effects of acute ingestion of caffeine on team sports performance: A systematic review and meta-analysis Res. Sports Med. 2019 27 238 256 10.1080/15438627.2018.1552146 30518253 25. Zbinden-Foncea H. Rada I. Gomez J. Kokaly M. Stellingwerff T. Deldicque L. Peñailillo L. Effects of caffeine on countermovement-jump performance variables in elite male volleyball players Int. J. Sports Physiol. Perform. 2018 13 145 150 10.1123/ijspp.2016-0705 28488924 26. Munoz A. López-Samanes Á. Pérez-López A. Aguilar-Navarro M. Moreno-Heredero B. Rivilla-García J. González-Frutos P. Pino-Ortega J. Morencos E. Del Coso J. Effects of caffeine ingestion on physical performance in elite women handball players: A randomized, controlled study Int. J. Sports Physiol. Perform. 2020 15 1406 1413 10.1123/ijspp.2019-0847 32335531 27. Moher D. Liberati A. Tetzlaff J. Altman D.G. The PRISMA Group Reprint—Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement Phys. Ther. 2009 89 873 880 10.1093/ptj/89.9.873 19723669 28. Tucker M.A. Hargreaves J.M. Clarke J.C. Dale D.L. Blackwell G.J. The effect of caffeine on maximal oxygen uptake and vertical jump performance in male basketball players J. Strength Cond. Res. 2013 27 382 387 10.1519/JSC.0b013e31825922aa 22561972 29. Abian-Vicen J. Puente C. Salinero J.J. González-Millán C. Areces F. Muñoz G. Muñoz-Guerra J. Del Coso J. A caffeinated energy drink improves jump performance in adolescent basketball players Amino Acids 2014 46 1333 1341 10.1007/s00726-014-1702-6 24599611 30. Puente C. Abián-Vicén J. Salinero J.J. Lara B. Areces F. Del Coso J. Caffeine improves basketball performance in experienced basketball players Nutrients 2017 9 1033 10.3390/nu9091033 31. Puente C. Abián-Vicén J. Del Coso J. Lara B. Salinero J.J. The CYP1A2-163C> A polymorphism does not alter the effects of caffeine on basketball performance PLoS ONE 2018 13 e0195943 10.1371/journal.pone.0195943 29668752 32. Scanlan A.T. Dalbo V.J. Conte D. Stojanović E. Stojiljković N. Stanković R. Antić V. Milanović Z. No effect of caffeine supplementation on dribbling speed in elite basketball players Int. J. Sports Physiol. Perform. 2019 14 997 1000 10.1123/ijspp.2018-0871 30569790 33. Stojanović E. Stojiljković N. Scanlan A.T. Dalbo V.J. Stanković R. Antić V. Milanović Z. Acute caffeine supplementation promotes small to moderate improvements in performance tests indicative of in-game success in professional female basketball players Appl. Physiol. Nutr. Metab. 2019 44 849 856 10.1139/apnm-2018-0671 30633542 34. Tan Z.S. Burns S.F. Pan J.W. Kong P.W. Effect of caffeine ingestion on free-throw performance in college basketball players J. Exerc. Sci. Fit. 2020 18 62 67 10.1016/j.jesf.2019.12.002 31908649 35. Stojanović E. Scanlan A.T. Milanović Z. Fox J.L. Stanković R. Dalbo V.J. Acute caffeine supplementation improves jumping, sprinting, and change-of-direction performance in basketball players when ingested in the morning but not evening Eur. J. Sport Sci. 2021 22 360 370 10.1080/17461391.2021.1874059 33413049 36. Maher C.G. Sherrington C. Herbert R.D. Moseley A.M. Elkins M. Reliability of the PEDro scale for rating quality of randomized controlled trials Phys. Ther. 2003 83 713 721 10.1093/ptj/83.8.713 12882612 37. Ziv G. Lidor R. Vertical jump in female and male basketball players—A review of observational and experimental studies J. Sci. Med. Sport 2010 13 332 339 10.1016/j.jsams.2009.02.009 19443269 38. Abian P. Del Coso J. José Salinero J. Gallo-Salazar C. The ingestion of a caffeinated energy drink improves jump performance and activity patterns in elite badminton players J. Sports Sci. 2015 33 1042 1050 10.1080/02640414.2014.981849 25530454 39. Perez-Lopez A. Salinero J.J. Abián-Vicén J. Valadés D. Lara B. Hernandez C. Areces F. González C. Del Coso J. Caffeinated energy drinks improve volleyball performance in elite female players Med. Sci. Sports Exerc. 2015 47 850 856 10.1249/MSS.0000000000000455 25051390 40. Del Coso J. Pérez-López A. Abián-Vicén J. Salinero J.J. Lara B. Valadés D. Enhancing physical performance in male volleyball players with a caffeine-containing energy drink Int. J. Sports Physiol. Perform. 2014 9 1013 1018 10.1123/ijspp.2013-0448 24664858 41. Souza D.B. Del Coso J. Casonatto J. Polito M.D. Acute effects of caffeine-containing energy drinks on physical performance: A systematic review and meta-analysis Eur. J. Nutr. 2017 56 13 27 10.1007/s00394-016-1331-9 27757591 42. Davis J.K. Green J.M. Caffeine and Anaerobic Performance Sports Med. 2009 39 813 832 10.2165/11317770-000000000-00000 19757860 43. Delextrat A. Badiella A. Saavedra V. Matthew D. Schelling X. Torres-Ronda L. Match activity demands of elite Spanish female basketball players by playing position Int. J. Perform. Anal. Sport 2015 15 687 703 10.1080/24748668.2015.11868824 44. Hogervorst E. Bandelow S. Schmitt J. Jentjens R. Oliveira M. Allgrove J. Carter T. Gleeson M. Caffeine improves physical and cognitive performance during exhaustive exercise Med. Sci. Sports Exerc. 2008 40 1841 1851 10.1249/MSS.0b013e31817bb8b7 18799996 45. Okazaki V.H. Rodacki A. Satern M.N. A review on the basketball jump shot Sports Biomech. 2015 14 190 205 10.1080/14763141.2015.1052541 26102462 46. Kasabova L. Study of Side Factors with Negative Impact on the Accuracy of Shooting in Basketball Гoдишник УНСС 2018 1 247 253 47. Sökmen B. Armstrong L.E. Kraemer W.J. Casa D.J. Dias J.C. Judelson D.A. Maresh C.M. Caffeine use in sports: Considerations for the athlete J. Strength Cond. Res. 2008 22 978 986 10.1519/JSC.0b013e3181660cec 18438212 48. Pallarés J.G. Fernández-Elías V.E. Ortega J.F. Muñoz G. Munoz-Guerra J. Mora-Rodríguez R. Neuromuscular responses to incremental caffeine doses: Performance and side effects Med. Sci. Sports Exerc. 2013 45 2184 2192 10.1249/MSS.0b013e31829a6672 23669879 49. Salinero J.J. Lara B. Abián-Vicén J. González-Millán C. Areces F. Gallo-Salazar C. Ruiz-Vicente D. Del Coso J. The use of energy drinks in sport: Perceived ergogenicity and side effects in male and female athletes Br. J. Nutr. 2014 112 1494 1502 10.1017/S0007114514002189 25212095 50. Magkos F. Kavouras S.A. Caffeine use in sports, pharmacokinetics in man, and cellular mechanisms of action Crit. Rev. Food Sci. Nutr. 2005 45 535 562 10.1080/1040-830491379245 16371327 51. Grgic J. Trexler E.T. Lazinica B. Pedisic Z. Effects of caffeine intake on muscle strength and power: A systematic review and meta-analysis J. Int. Soc. Sports Nutr. 2018 15 1 10 10.1186/s12970-018-0216-0 29311763 52. Pickering C. Kiely J. Are the current guidelines on caffeine use in sport optimal for everyone? Inter-individual variation in caffeine ergogenicity, and a move towards personalised sports nutrition Sports Med. 2018 48 7 16 10.1007/s40279-017-0776-1 28853006 53. Guest N. Corey P. Vescovi J. El-Sohemy A. Caffeine, CYP1A2 genotype, and endurance performance in athletes Med. Sci. Sports Exerc. 2018 50 1570 1578 10.1249/MSS.0000000000001596 29509641 54. Rahimi R. The effect of CYP1A2 genotype on the ergogenic properties of caffeine during resistance exercise: A randomized, double-blind, placebo-controlled, crossover study Ir. J. Med. Sci. 2019 188 337 345 10.1007/s11845-018-1780-7 29532291 55. Salinero J.J. Lara B. Ruiz-Vicente D. Areces F. Puente-Torres C. Gallo-Salazar C. Pascual T. Del Coso J. CYP1A2 genotype variations do not modify the benefits and drawbacks of caffeine during exercise: A pilot study Nutrients 2017 9 269 10.3390/nu9030269 56. Algrain H.A. Thomas R.M. Carrillo A.E. Ryan E.J. Kim C.-H. Lettan R.B. Ryan E.J. The effects of a polymorphism in the cytochrome P450 CYP1A2 gene on performance enhancement with caffeine in recreational cyclists J. Caffeine Res. 2016 6 34 39 10.1089/jcr.2015.0029

PMC009xxxxxx/PMC9099692.txt

==== Front Polymers (Basel) Polymers (Basel) polymers Polymers 2073-4360 MDPI 10.3390/polym14091927 polymers-14-01927 Article Effect of Cyclic Shear Fatigue under Magnetic Field on Natural Rubber Composite as Anisotropic Magnetorheological Elastomers https://orcid.org/0000-0002-9607-4770 Yoon Jeong-Hwan 1* Lee Seung-Won 1 Bae Seok-Hu 1 Kim Nam-Il 1 Yun Ju-Ho 1 Jung Jae-Hum 2 Kim Young-Gil 2 Guzmán Eduardo Academic Editor 1 Energy Materials R&D Centre, Materials Technology R&D Division, Korea Automotive Technology Institute, 303 Pungse-ro, Pungse-myeon, Dongnam-gu, Cheonan-si 31214, Chungnam, Korea; swlee1@katech.re.kr (S.-W.L.); shbae@katech.re.kr (S.-H.B.); nikim@katech.re.kr (N.-I.K.); jhyun@katech.re.kr (J.-H.Y.) 2 Material Development Team, Deaheung Rubber & Technology Co., Ltd., 436 Seobu-ro, Jillye-myeon, Gimhae-si 50872, Gyeongnam, Korea; jaehum.jung@dhrnt.com (J.-H.J.); younggil.kim@dhrnt.com (Y.-G.K.) * Correspondence: jhyoon@katech.re.kr; Tel.: +82-41-559-3128 09 5 2022 5 2022 14 9 192729 3 2022 06 5 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). With the development and wide applicability of rubber materials, it is imperative to determine their performance under various conditions. In this study, the effect of cyclic shear fatigue on natural-rubber-based anisotropic magnetorheological elastomer (MRE) with carbonyl iron particles (CIPs) was investigated under a magnetic field. An anisotropic MRE sample was prepared by moulding under a magnetic field. Cyclic shear fatigue tests were performed using a modified electromechanical fatigue system with an electromagnet. The storage modulus (G′) and loss factor in the absence or presence of a magnetic field were measured using a modified dynamic mechanical analysis system. Under a magnetic field, fatigue exhibited considerable effects to the MRE, such as migration and loss of magnetised CIPs and suppressed increase in stiffness by reducing the energy loss in the strain cycle. Therefore, the G′ of the MRE after fatigue under a magnetic field was lower than that after fatigue in the zero field. The performance of the MRE, such as absolute and relative magnetorheological effects, decreased after subjecting to cyclic shear fatigue. In addition, all measured results exhibited strain-dependent behaviour owing to the Payne effect. natural rubber carbonyl iron particle anisotropic magnetorheology shear fatigue Ministry of Trade, Industry, and Energy of the South Korean government10047791 This research was funded by the Ministry of Trade, Industry, and Energy of the South Korean government, project number 10047791. ==== Body pmc1. Introduction Rubbery materials with viscoelastic properties exhibit complex behaviour owing to their deformability, stress softening, and time-dependent attributes; thus, they are widely used in various industrial applications such as seals, dampers, bushings, and tires, among others [1,2,3,4,5]. As rubber products are exposed to different environmental conditions and subjected to cyclic loading, they often fail owing to nucleation and the presence of defects [6,7]. In addition, the stress–strain curve of a rubber material subjected to cyclic loading exhibits a hysteresis loop, indicating energy loss owing to its viscoelasticity. Most of this energy loss is eventually dissipated into heat. When the heat generated by cyclic loading does not exit the material and accumulates, the temperature of the material increases, resulting in fatigue failure or changes in the material properties [8,9,10,11,12,13]. The unique rheological properties of magnetorheological elastomers (MREs) can be easily tuned when exposed to an external magnetic field, making them suitable for advanced rubber products, such as adaptive dampers and stiffness-tuneable mounts. Natural rubber (NR) exhibiting high mechanical properties can contain high content of magnetic particle to improve the MRE’s performances, while soft carbonyl iron particles (CIPs) are widely used because of their high permeability, high magnetic saturation, and low remnant magnetization. On the other hand, MRE containing a large amount of filler has a relatively low fatigue resistance. There have been ample studies on various approaches to improve the performance of MREs. Recently, researchers have reported the fatigue behaviour of MREs under a magnetic field [14,15,16,17,18,19,20,21]. Zhou et al. [22] reported that the fatigue life of an MRE increases when fatigue is induced under an external magnetic field. Lian et al. [23] demonstrated the effect of repetition of a magnetic field on the fatigue behaviour, hysteresis loss, and storage modulus of MREs. Meanwhile, most studies on the durability of MRE have focused on monitoring the change of tensile strength and compressive load [24,25,26]. However, the effect of an external magnetic field on the property changes of MRE during cyclic fatigue needs to be further investigated for the actual application as engine mount. In this study, an anisotropic MRE was fabricated using NR, which is widely used as the matrix for MRE, and CIPs. Cyclic shear fatigue and strain amplitude sweep tests were performed to investigate the effect of the magnetic field on the prepared samples. 2. Materials and Methods NR (CV-60, Standard Vietnam Rubber) was used as the rubber matrix. Carbon black (N990, Cancarb, Alberta, Canada) was used as the reinforcement to improve the mechanical properties. Spherical CIPs (CC grade, BASF, Ludwigshafen, Germany) with a mean particle size of 3–5 μm were used as the soft magnetic particles. Processing oil (N-2, Michang, Busan, Korea), sulphur (MIDAS SP 325, Miwon, Anyang, Korea), thiuram disulphide (ORICELL TT, OCI, Seoul, Korea), and sulphonamide (ORICELL CZ, OCI, Seoul, Korea) were used as the curing agents and accelerators. A Banbury mixer (HYB-3L, Hyupyoung Machinery, Kimpo, Korea) was used to prepare the MRE samples. First, 150 parts per hundred rubber (phr) CIPs, 15 phr carbon black, 20 phr processing oil, 1.5 phr sulphur, 1.5 phr thiuram disulphide, and 2.0 phr sulphonamide were compounded with NR at 50 °C for 15 min. The resulting mixture was moulded into disks with a thickness of 3.0 mm and diameter of 7.0 mm under approximately 10 MPa in a magnetic field with an intensity of 1000 mT at 160 °C for 400 s. As a result, an anisotropic MRE sample with CIPs aligned out-of-plane along the magnetic field direction in the matrix was obtained. Cyclic shear fatigue tests were performed using a modified electromechanical fatigue system (Fatigue tester, Daekyung engineering, Bucheon, Korea) equipped with an electromagnet that can generate a magnetic field under uniaxial shear loading (Figure 1a). The shear direction was perpendicular to the direction of the magnetic field and aligned CIPs. The system was designed to maintain a constant magnetic flux density during fatigue tests. A relatively weak magnetic flux density of 300 mT was applied to confirm the effect of the magnetic field on the MRE and generate a constant magnetic field for up to 500,000 fatigue cycles. Based on the test conditions, such as engine mounts where MRE can be applied, a strain amplitude of 50% and frequency of 5 Hz at an ambient temperature of 23 °C was set. The maximum load applied to the sample was recorded. The changes in the morphology, specific gravity, and dynamic viscoelastic properties of the anisotropic MRE samples were measured after subjecting them to cyclic shear fatigue. The cross-sectional morphologies of the MREs were observed using a laser confocal microscope (OLS5000, Olympus, Shinjuku, Japan). An analytical balance (MS204TS, Mettler Toledo, Greifensee, Switzerland) was used to obtain the specific gravity according to ISO 1183-1. The rheological properties of the samples were examined using the modified dynamic mechanical analysis system presented in our previous work [27]. Three types of properties were measured: storage modulus, loss modulus, and loss factor. When an external magnetic field was applied, the direction of the magnetic field was perpendicular to the surface of the tested sample and parallel to the aligned structure of the CIPs within the anisotropic MRE. Thus, the modulus measured under the magnetic field detected the response of the sample in the perpendicular direction. The strain amplitude sweep test was performed by varying the strain from 0.1 to 5.0% with 1 Hz at an ambient temperature of 23 °C. In the on-state measurement, a magnetic flux density of 1200 mT was applied for the sufficient saturated magnetisation of the magnetic particles. 3. Results and Discussion 3.1. Cyclic Shear Fatigue Test Figure 1b shows the maximum recorded load with the increasing number of cycles for the NR and MRE samples. The NR sample had the same composition as that of the MRE sample, except for the presence of CIPs in the latter. The MRE sample was subjected to zero and external magnetic fields. The load of the samples increased with the increase in the number of cycles. In particular, the load of the MRE sample was higher than that of the NR sample. Owing to the higher stiffness of the MRE sample, a larger maximum load is required for its deformation. Moreover, the maximum load increases further under an external magnetic field, consistent with the known effect of an external magnetic field on the MRE sample [1,27,28,29,30]. Meanwhile, the material exhibited variable response to the fatigue cycles at a fixed strain. The load deviations increased as the number of fatigue cycles increased. This phenomenon is associated to the stress softening due to the sliding between the matrix chains and filler particles, hysteretic heating of the rubber materials by the energy loss in the strain cycle, and accumulation of crystallisation under non-relaxing conditions. Stress softening under cyclic loading, especially for filled rubber composites, has been described by the Mullins and Payne effects [8,9,10,11,12,13,31,32,33,34,35,36,37]. Softening and hardening are complex events that occur in samples during the cyclic shear fatigue test, and these changes may affect the rheological properties of the MRE. Interestingly, the CIPs embedded in the MRE were attracted to the magnetic field and were released during the fatigue test under an external magnetic field. The detachment of the CIPs weakens physical bonds between the NR matrix and CIPs owing to the relative movement of magnetised CIPs and the large strain amplitude applied to the MRE [18,38,39,40]. This can explain the slightly lower increase in the maximum recorded load under a magnetic field that in a zero field as the number of fatigue cycles increased. 3.2. Morphology The laser confocal microscope images of the cross-sectional morphologies of the initial MRE samples and after the fatigue tests are shown in Figure 2. Initially, the CIPs were aligned and embedded in NR matrix with some of the agglomerations in the magnetic field direction. However, in the samples subjected to fatigue, most of the particles were separated from the surrounding NR matrix, as can be seen in the dark area around the particles in Figure 2c–f. Different morphologies were observed based on the presence or absence of a magnetic field. Figure 2c shows that the separation regions are mainly arranged in the shear strain direction. Meanwhile, most of the CIPs are dissociated from the surrounding NR matrix in this direction. However, when shear fatigue was applied under a magnetic field, separation occurred in the magnetic field direction, as shown in Figure 2e,f. Although the CIPs are separated from the surrounding NR matrix by the shear strain, interparticle forces between the magnetised CIPs maintained the physical bond with the NR matrix. However, the shear strain continuously separated the magnetised CIPs and NR matrix, along with the interparticle forces generated the migration regions by shifting the CIPs, as shown in Figure 2f. Therefore, the separation regions were observed in the shear strain and magnetic field directions as in zig-zag arrows. Some CIPs exhibited agglomerated morphologies in the direction of the interparticle forces. In addition, the magnetised CIPs on the surface of the MRE can be attracted to and lost to an external electromagnet when separated from the NR matrix. 3.3. Specific Gravity Table 1 lists the specific gravities of the MRE samples before and after subjecting them to cyclic shear fatigue. The specific gravity of the MRE sample prior to fatigue testing was 2.586 g/cm3. After subjecting the MRE sample to fatigue for up to 500,000 cycles in the zero field, the specific gravity of the sample remained constant. Meanwhile, the specific gravity of the MRE sample subjected to fatigue under a magnetic field decreased with the increase in the fatigue cycles. This indicates changes in the MRE composition. In particular, the change in specific gravity of the MRE, as confirmed during the cyclic shear fatigue test, can be ascribed to the loss of the CIPs with a specific gravity of 7.86 g/cm3. 3.4. Rheological Properties Strain amplitude sweep tests were performed to investigate the change in the dynamic viscoelastic modulus of the MREs under cyclic shear fatigue. Figure 3 shows the storage modulus (G′0) and loss factor (tan(δ0)) of the NR sample as a function of the strain amplitude in the zero-field after 0, 300,000, and 500,000 fatigue cycles. As shown in Figure 3a, the G′0 of the NR sample decreased with increasing strain amplitude owing to the Payne effect and decreased with fatigue. In addition, tan(δ0) increased as the number of fatigue cycles increased, as shown in Figure 3b. These results are consistent with the generally known behaviour, that is, the remarkable decrease in the storage modulus. As the number of cycles increased, damage gradually propagated, thereby decreasing G′0 and increasing tan(δ0) because the friction inside the material causes more energy dissipation [41,42]. Figure 4 shows the rheological properties of the MRE sample in the absence (off-state) and presence (on-state) of the magnetic field under cyclic shear fatigue. After 300,000 cycles, the G′0 of the MRE sample was lower than that of the initial sample, exhibiting a behaviour similar to the NR sample. After 500,000 cycles, G′0 increased again. This trend can be attributed to the increase in the maximum load during the cyclic shear fatigue test. The MRE sample has more pronounced hysteretic heating and crystallisation accumulation, resulting in its higher stiffness. In contrast, the G′0 value of the MRE sample before and after being subjected to fatigue under a magnetic field exhibits a slightly different trend, as shown in Figure 4b. After 500,000 cycles under a magnetic field, the G′0 of the sample decreased more than that of the sample after 300,000 cycles. This trend can be ascribed to the relative movement of the CIPs limited by the magnetic field and loss of CIPs. In particular, the limited relative movement of the CIPs can suppress the increase in stiffness of the MRE by reducing the energy loss in the strain cycle. In addition, this decreases the softness of the MRE because of the CIPs lost during the cyclic shear fatigue test. The on-state G′ trend of the MRE before and after being subjected to cyclic shear, as shown in Figure 4c,d, is similar to that of the off-state G′0 in Figure 4a,b, respectively. However, the difference of the G′0 values before and after being subjected to fatigue is larger. This suggests the effect of fatigue on the modulus stored in the restrained matrix generated by the interparticle forces. Therefore, the magnetorheological (MR) performance of an MRE can be altered by cyclic shear fatigue as it is dependent on the capacity of the modulus stored in the restrained matrix [43]. When cyclic shear fatigue is applied under a magnetic field, the decrease in G′ is more pronounced owing to agglomeration caused by the migration of the CIPs, as shown in Figure 2f. This is ascribed to the increased distance between the CIP aggregates as the CIPs move, which can weaken the interparticle forces or reduce the area of the restrained matrix around the magnetised CIPs. The performance of the MRE sample is usually evaluated by the absolute and relative MR effects, which represent the change in the storage modulus under a magnetic field [44,45,46,47]. The absolute MR effect (ΔG′) represents the difference between G′ and G′0. The relative MR effect is the percentage of ΔG′ and G′0, as:(1) Relative MR effect =ΔG’G’0×100%. Figure 4e,f shows the absolute and relative MR effects of the MRE sample before and after subjecting it to fatigue, respectively. Except for the sample after 500,000 cycles in the zero field, which has the highest G′, the ΔG′ values of all samples after fatigue were lower than that of the initial sample. Similar to the results of the G′ values, lower ΔG′ values were obtained for the samples after fatigue. As G′0 decreased with increasing strain amplitude owing to the Payne effect, the relative MR effect was strain-dependent, that is, it increased with increasing strain amplitude. The tan(δ0) of the MRE increased with increasing energy dissipation after subjecting it to cyclic shear fatigue, as shown in Figure 5a,b, which is consistent with the results of the NR sample in Figure 3. As shown in Figure 5a, the tan(δ0) value of the sample after fatigue in the zero field at a lower strain amplitude was slightly lower than that of the initial sample. In contrast, the tan(δ0) value of the sample at a lower strain amplitude after 300,000 cycles under a magnetic field was higher than that of the initial sample, as shown in Figure 5b, which can be ascribed to the higher energy dissipation owing to the migration of magnetised CIPs. The tan(δ0) behaviour with the strain amplitude of the sample after 500,000 cycles under a magnetic field is expected to vary under the influence of further migration and CIP losses. Figure 5c,d shows the tan(δ) value measured in the on-state condition, which exhibits a different behaviour from that of tan(δ0). The loss factor based on the relative movement between the CIPs and matrix in the on-state was lower than that in the off-state because the interparticle forces between the CIPs inside the rubber matrix limit the relative movement of the CIPs, thereby reducing the energy dissipation [18,38,39,40]. Therefore, the tan(δ) values of all samples were lower than of tan(δ0). Moreover, the tan(δ) of all samples, except for that after 500,000 cycles in the zero field, was higher than that of the initial sample. The lower tan(δ) of the sample after 500,000 cycles in the zero field is attributed to the reduced sliding between the matrix chains and particles due to the expansion of the separation region and limited relative movement of the magnetised CIPs [31]. The highest tan(δ) was obtained after 500,000 cycles under a magnetic field, as shown in Figure 5d. This can be explained by the agglomerate formation and particle loss due to the migration of the CIPs after cyclic shear fatigue owing to the low G′. The internal changes in the MRE after fatigue increased the distance between the CIPs, thereby weakening the interparticle forces. This reduces the force limiting the relative movement of the magnetised CIPs, thereby increasing energy dissipation. In addition, the cyclic shear fatigue under a magnetic field reduced the area of the separation region between the particles and matrix, and some CIPs maintained the physical bonds with the surrounding matrix even after fatigue, as shown in Figure 2f. Consequently, tan(δ) tends to increase as the number of fatigue cycles increases under a magnetic field. 4. Conclusions In this study, cyclic shear fatigue and strain amplitude sweep tests were performed on NR-based MRE samples to clarify the changes in the dynamic viscoelastic properties under fatigue. When the cyclic shear fatigue tests were performed in the zero field and an external magnetic field, the material response to the fatigue cycles at a fixed strain was not constant. In addition, the maximum recorded load increased with the increase in the number of fatigue cycles. Under a magnetic field, the CIPs on the MRE surface were extracted with reduced specific gravity from 2.586 to 2.565 g/cm3, where the migration of the CIPs was observed in the cross-sectional morphologies. The absence or presence of a magnetic field during fatigue affected the changes in the G′0 of the MRE. In addition, the magnetic field suppressed the increase in stiffness of the MRE by reducing the energy loss in the strain cycle. However, the magnetic field induced the migration and loss of CIPs, and consequently, decreased the absolute and relative MR effects of the MRE after fatigue under a magnetic field. As the relative movement of the CIPs in the MRE is restricted by the magnetic field, tan(δ) is lower than tan(δ0). However, both tan(δ0) and tan(δ) increased after fatigue. After 500,000 fatigue cycles in the zero field, the sliding between the matrix chains and particles decreased owing to the expansion of the separation region, resulting in a lower tan(δ) than the initial sample. After fatigue test, the maximum G′ decreased slightly from 0.53 MPa to 0.51 MPa and 0.45 MPa over the whole strain amplitude range investigated in absence or presence of a magnetic field, respectively. No significant change was also observed in the maximum tan(δ) value. This indicates the effect of the cyclic shear fatigue accumulated in the absence or presence of a magnetic field on the magnetorheological properties of the MRE sample. This is also an important factor to be considered in the application of MRE for further research. Author Contributions Conceptualisation, J.-H.Y. (Jeong-Hwan Yoon) and J.-H.Y. (Ju-Ho Yun); methodology, J.-H.Y. (Jeong-Hwan Yoon), S.-W.L., J.-H.J. and Y.-G.K.; validation, J.-H.Y. (Jeong-Hwan Yoon), S.-H.B. and N.-I.K.; investigation, J.-H.Y. (Jeong-Hwan Yoon); resources, J.-H.Y. (Jeong-Hwan Yoon), S.-W.L., J.-H.J. and Y.-G.K.; data curation, J.-H.Y. (Jeong-Hwan Yoon) and S.-W.L.; writing—original draft preparation, J.-H.Y. (Jeong-Hwan Yoon); writing—review and editing, S.-H.B., N.-I.K. and J.-H.Y. (Ju-Ho Yun); supervision, J.-H.Y. (Jeong-Hwan Yoon); project administration, J.-H.Y. (Jeong-Hwan Yoon); funding acquisition, J.-H.Y. (Ju-Ho Yun) All authors have read and agreed to the published version of the manuscript. Data Availability Statement The data presented in this study are available on request from the corresponding author. Conflicts of Interest The authors declare no conflict of interest. Figure 1 (a) Photos of the modified electromechanical fatigue system equipped with an electromagnet. (b) Maximum recorded load during cyclic shear fatigue of the NR and anisotropic MRE samples. Figure 2 Microscope images with (a) 20× and (b) 100× magnification of the initial MRE cross-section; and (c) 20× and (d) 100× magnification of the MRE cross-section after the cyclic shear fatigue in the zero field; and (e) 20× and (f) 100× magnification of the MRE cross-section after 500,000 shear fatigue cycles under magnetic field. The bright parts represent the magnetic particles embedded in the NR. The white arrows represent the direction of separation or migration regions. Figure 3 (a) G′0 and (b) tan(δ0) of the NR before and after cyclic shear fatigue. Figure 4 Off-state G′0 of the MRE sample before and after being subjected to cyclic shear fatigue in the (a) absence and (b) presence of a magnetic field. On-state G′ of the MRE sample before and after cyclic shear fatigue in the (c) absence and (d) presence of magnetic field. (e) Absolute and (f) relative MR effects of the MRE samples. Figure 5 Off-state tan(δ0) of the MRE sample before and after cyclic shear fatigue in the (a) absence and (b) presence of a magnetic field. On-state tan(δ) of the MRE sample before and after cyclic shear fatigue in the (c) absence and (d) presence of magnetic field. polymers-14-01927-t001_Table 1 Table 1 Specific gravity of the MRE sample before and after cyclic shear fatigue in the absence or presence of a magnetic field. Number of Fatigue Cycles 0 Zero-Field Magnetic Field (300 mT) 300,000 500,000 300,000 500,000 Specific gravity (g/cm3) 2.586 2.581 2.583 2.577 2.565 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Hosseini S.M. Shojaeefard M.H. Saeidi G.H. Fatigue life prediction of magneto-rheological elastomers in magnetic field Mater. Res. Express. 2021 8 025304 10.1088/2053-1591/abe520 2. Laraba-Abbes F. Ienny P. Piques R. A new ‘Tailor-made’ methodology for the mechanical behaviour analysis of rubber-like materials: II. Application to the hyperelastic behaviour characterization of a carbon-black filled natural rubber vulcanizate Polymer 2003 44 821 840 3. Wang X.L. Liao M.Y. Xu Y. Liu X.A. Fatigue crack propagation characteristics of rubbery materials under variable amplitude loading Results Phys. 2018 10 233 240 10.1016/j.rinp.2018.05.043 4. Tee Y.L. Loo M.S. Andriyana A. Recent advances on fatigue of rubber after the literature survey by Mars and Fatemi in 2002 and 2004 Int. J. Fatigue 2018 110 115 129 10.1016/j.ijfatigue.2018.01.007 5. Shangguan W.B. Wang X.L. Deng J.X. Rakheja S. Pan X.Y. Yu B. Experiment and modeling of uniaxial tension fatigue performances for filled natural rubbers Mater. Des. 2014 58 65 73 10.1016/j.matdes.2014.01.035 6. Gent A.N. Lindley P.B. Thomas A.G. Cut growth and fatigue of rubbers. I. The relationship between cut growth and fatigue J. Appl. Polym. Sci. 1964 8 455 466 7. Yeoh O.H. Analysis of deformation and fracture of ‘pure shear’ rubber testpiece Plast. Rubber. Compos. 2001 30 389 397 10.1179/146580101101541787 8. Luo W. Huang Y. Yin B. Jiang X. Hu X. Fatigue life assessment of filled rubber by hysteresis induced self-heating temperature Polymers 2020 12 846 10.3390/polym12040846 9. Le S.V. Marco Y. Calloch S. Doudard C. Charrier P. Fast evaluation of the fatigue lifetime of rubber-like materials based on a heat build-up protocol and micro-tomography measurements Int. J. Fatigue 2010 32 1582 1590 10. Luo W. Yin B. Hu X. Zhou Z. Deng Y. Song K. Modeling of the heat build-up of carbon black filled rubber Polym. Test. 2018 69 116 124 10.1016/j.polymertesting.2018.05.017 11. Zhi J. Wang S. Zhang M. Wang H. Lu H. Lin W. Numerical analysis of the dependence of rubber hysteresis loss and heat generation on temperature and frequency Mech. Time-Depend. Mat. 2019 23 427 442 10.1007/s11043-018-9398-8 12. Le S.V. Marco Y. Calloch S. Charrier P. Taveau D. Heat build-up of rubber under cyclic loadings: Validation of an efficient demarch to predict the temperature fields Rubber Chem. Technol. 2013 86 38 56 13. Katunin A. Criticality of the self-heating effect in polymers and polymer matrix composites during fatigue, and their application in non-destructive testing Polymers 2019 11 19 10.3390/polym11010019 14. Schümann M. Odenbach S. In-situ observation of the particle microstructure of magnetorheological elastomers in presence of mechanical strain and magnetic fields J. Magn. Magn. Mater. 2017 441 88 92 10.1016/j.jmmm.2017.05.024 15. Perales-Martínez I.A. Palacios-Pineda L.M. Lozano-Sánchez L.M. Martínez-Romero O. Puente-Cordova J.G. Elías-Zúñiga A. Enhancement of a magnetorheological PDMS elastomer with carbonyl iron particles Polym. Test. 2017 57 78 86 10.1016/j.polymertesting.2016.10.029 16. Burgaz E. Goksuzoglu M. Effects of magnetic particles and carbon black on structure and properties of magnetorheological elastomers Polym. Test. 2020 81 106233 10.1016/j.polymertesting.2019.106233 17. Cvek M. Kracalik M. Sedlacik M. Mrlik M. Sedlarik V. Reprocessing of injection-molded magnetorheological elastomers based on TPE matrix Compos. B. Eng. 2019 172 253 261 10.1016/j.compositesb.2019.05.090 18. Yu M. Qi S. Fu J. Zhu M. Chen D. Understanding the reinforcing behaviors of polyaniline-modified carbonyl iron particles in magnetorheological elastomer based on polyurethane/epoxy resin IPNs matrix Compos. Sci. Technol. 2017 139 36 46 10.1016/j.compscitech.2016.12.010 19. Dargahi A. Sedaghati R. Rakheja S. On the properties of magnetorheological elastomers in shear mode: Design, fabrication and characterization Compos. B. Eng. 2019 159 269 283 10.1016/j.compositesb.2018.09.080 20. Na B.G. Chung K.H. Effect of precured EPDM on the property of magneto-rheological elastomer based on NR/EPDM blend Elast. Compos. 2018 53 67 74 21. Kim T.W. Choi Y.J. Kim N.Y. Chung K.H. A study on the fatigue property of magneto-rheological elastomers Elast. Compos. 2018 53 150 157 22. Zhou Y. Johnson M. Wen S. Betts A. Jerrams S. Equi-biaxial fatigue behaviour of magnetorheological elastomers in magnetic fields J. Intell. Mater. Syst. Struct. 2016 28 687 696 10.1177/1045389X16657204 23. Lian C. Lee K.H. Choi S.B. Lee C.H. A study of the magnetic fatigue properties of a magnetorheological elastomer J. Intell. Mater. Syst. Struct. 2018 30 749 754 10.1177/1045389X18799495 24. Wang Y. Gong X. Yang J. Xuan S. Improving the dynamic properties of MRE under cyclic loading by incorporating silicon carbide nanoparticles Ind. Eng. Chem. Res. 2014 53 3065 3072 10.1021/ie402395e 25. Gorman D. Murphy N. Ekins R. Jerrams S. The evaluation and implementation of magnetic fields for large strain uniaxial and biaxial cyclic testing of magnetorheological elastomers Polym. Test. 2016 51 74 81 10.1016/j.polymertesting.2016.02.002 26. Johari M.A.F. Mazlan S.A. Nasef M.M. Ubaidillah U. Nordin N.A. Aziz S.A.A. Johari N. Nazmi N. Microstructural behavior of magnetorheological elastomer undergoing durability evaluation by stress relaxation Sci. Rep. 2021 11 10936 10.1038/s41598-021-90484-0 34035434 27. Yoon J.H. Lee S.W. Bae S.H. Kim N.I. Yun J.H. Ryu S.H. Effect of alignment of magnetic particles on the rheological properties of natural rubber composite J. Polym. Res. 2021 28 367 10.1007/s10965-021-02735-3 28. Ubaidillah Sutrisno J. Purwanto A. Mazlan S.A. Recent progress on magnetorheological solids: Materials, fabrication, testing, and applications Adv. Eng. Mater. 2015 17 563 597 10.1002/adem.201400258 29. Bodelot L. Voropaieff J.P. Pössinger T. Experimental investigation of the coupled magneto-mechanical response in magnetorheological elastomers Exp. Mech. 2018 58 207 221 10.1007/s11340-017-0334-7 30. Syam T. Muthalif A. Salem A. Hejazi A. 3D numerical modelling and analysis of a magnetorheological elastomer (MRE) J. Vibroengineering 2020 22 1251 1265 31. Cantournet S. Desmorat R. Besson J. Mullins effect and cyclic stress softening of filled elastomers by internal sliding and friction thermodynamics model Int. J. Solids Struct. 2009 46 2255 2264 10.1016/j.ijsolstr.2008.12.025 32. Harwood J.A.C. Payne A.R. Stress softening in natural rubber vulcanizates. Part III. Carbon black-filled vulcanizates J. Appl. Polym. Sci. 1966 10 315 324 33. Harwood J.A.C. Payne A.R. Hysteresis and strength of rubbers J. Appl. Polym. Sci. 1968 12 889 901 10.1002/app.1968.070120423 34. Mullins L. Tobin N.R. Theoretical model for the elastic behavior of filler-reinforced vulcanized rubbers Rubber Chem. Technol. 1957 30 555 571 10.5254/1.3542705 35. Mullins L. Tobin N.R. Stress softening in rubber vulcanizates. Part I. Use of a strain amplification factor to describe the elastic behavior of filler-reinforced vulcanized rubber J. Appl. Polym. Sci. 1965 9 2993 3009 10.1002/app.1965.070090906 36. Qi H.J. Boyce M.C. Constitutive model for stretch-induced softening of the stress–stretch behavior of elastomeric materials J. Mech. Phys. Solids 2004 52 2187 2205 10.1016/j.jmps.2004.04.008 37. Saintier N. Cailletaud G. Piques R. Cyclic loadings and crystallization of natural rubber: An explanation of fatigue crack propagation reinforcement under a positive loading ratio Mater. Sci. Eng. A 2011 528 1078 1086 10.1016/j.msea.2010.09.079 38. Yu M. Fu J. Ju B.X. Zheng X. Choi S.B. Influence of x-ray radiation on the properties of magnetorheological elastomers Smart. Mater. Struct. 2013 22 125010 10.1088/0964-1726/22/12/125010 39. Fan Y. Gong X. Xuan S. Qin L. Li X. Effect of cross-link density of the matrix on the damping properties of magnetorheological elastomers Ind. Eng. Chem. Res. 2013 52 771 778 10.1021/ie302536e 40. Yang J. Gong X. Deng H. Qin L. Xuan S. Investigation on the mechanism of damping behavior of magnetorheological elastomers Smart. Mater. Struct. 2012 21 125015 10.1088/0964-1726/21/12/125015 41. Movahedi-Rad A.V. Keller T. Vassilopoulos A.P. Modeling of fatigue behavior based on interaction between time- and cyclic-dependent mechanical properties Compos. Part A Appl. Sci. Manuf. 2019 124 105469 10.1016/j.compositesa.2019.05.037 42. Movahedi-Rad A. Keller T. Vassilopoulos A. Interrupted tension-tension fatigue behavior of angle-ply GFRP composite laminates Int. J. Fatigue 2018 113 377 388 10.1016/j.ijfatigue.2018.05.001 43. Burhannuddin N.L. Nordin N.A. Mazlan S.A. Aziz S.A.A. Kuwano N. Jamari S.K.M. Ubaidillah Physicochemical characterization and rheological properties of magnetic elastomers containing different shapes of corroded carbonyl iron particles Sci. Rep. 2021 11 868 10.1038/s41598-020-80539-z 33441824 44. Esmaeilnezhad E. Choi H.J. Schaffie M. Gholizadeh M. Ranjbar M. Polymer coated magnetite-based magnetorheological fluid and its potential clean procedure applications to oil production J. Clean. Prod. 2018 171 45 56 10.1016/j.jclepro.2017.10.004 45. Wang Y. Zhang X. Chung K. Liu C. Choi S.B. Choi H.J. Formation of core–shell structured complex microparticles during fabrication of magnetorheological elastomers and their magnetorheological behavior Smart Mater. Struct. 2016 25 115028 10.1088/0964-1726/25/11/115028 46. Dong X. Ma N. Qi M. Li J. Chen R. Ou J. The pressure-dependent MR effect of magnetorheological elastomers Smart Mater. Struct. 2012 21 075014 10.1088/0964-1726/21/7/075014 47. Jiang W. Yao J. Gong X. Chen L. Enhancement in magnetorheological effect of magnetorheological elastomers by surface modification of iron particles Chinese J. Chem. Phys. 2008 21 87 92 10.1088/1674-0068/21/01/87-92

PMC009xxxxxx/PMC9099693.txt

==== Front Polymers (Basel) Polymers (Basel) polymers Polymers 2073-4360 MDPI 10.3390/polym14091791 polymers-14-01791 Article Polyvinyl Alcohol/Polyaniline/Carboxylated Graphene Oxide Nanocomposites for Coating Protection of Cast Iron in Simulated Seawater Elessawy Noha A. 1* https://orcid.org/0000-0001-5317-845X Gouda Marwa H. 2* Elnouby Mohamed 3 Taha Nahla A. 4 https://orcid.org/0000-0003-4362-4801 Youssef M. Elsayed 1 https://orcid.org/0000-0002-7920-2638 Santos Diogo M. F. 5 Santamaria Antxon Academic Editor 1 Computer Based Engineering Applications Department, Informatics Research Institute IRI, City of Scientific Research and Technological Applications (SRTA-City), Alexandria 21934, Egypt; elsayed168@gmail.com 2 Polymer Materials Research Department, Advanced Technology and New Materials Research Institute, City of Scientific Research and Technological Applications (SRTA-City), Alexandria 21934, Egypt 3 Nanomaterials and Composites Research Department, Advanced Technology and New Materials Research Institute, City of Scientific Research and Technological Applications (SRTA-City), Alexandria 21934, Egypt; mnano2050@yahoo.com 4 Modelling and Simulation Research Department, Advanced Technology and New Materials Research Institute (ATNMRI), City of Scientific Research and Technological Applications (SRTA-City), Alexandria 21934, Egypt; nahlataha_1982@yahoo.com 5 Center of Physics and Engineering of Advanced Materials, Laboratory for Physics of Materials and Emerging Technologies, Chemical Engineering Department, Instituto Superior Técnico, Universidade de Lisboa, 1049-001 Lisbon, Portugal; diogosantos@tecnico.ulisboa.pt * Correspondence: nony_essawy@yahoo.com (N.A.E.); marwagouda777@yahoo.com (M.H.G.) 27 4 2022 5 2022 14 9 179104 3 2022 25 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). In our daily lives and product manufacturing, metal corrosion causes significant economic losses. Numerous polymeric composite coatings have been shown to be resistant to harsh environments, such as those found in marine environments. In this study, a composite of polyvinyl alcohol/polyaniline blend loaded with carboxylated graphene was explored in the search for long-lasting coatings to resist electrochemical deterioration of cast iron in desalination systems of saltwater. Polyvinyl alcohol/polyaniline/carboxylated graphene oxide nanocomposite was spin-coated onto cast iron samples. Electrochemical impedance spectroscopy (EIS) and electrochemical DC corrosion testing with a three-electrode system were used to study corrosion resistance in uncoated and coated cast iron samples. The results exhibit effective corrosion protection properties. The EIS data indicated better capacitance and higher impedance values for coated samples than bare metal, depicting enhanced corrosion resistance against the saline environment. Tafel analysis confirmed a significant decrease in the corrosion rate of the PVA/PANI/GO-COOH coated sample. polyvinyl alcohol polyaniline carboxylated graphene saline environment corrosion inhibition This research received no external funding. ==== Body pmc1. Introduction Seawater desalination develops rapidly, as it can solve water scarcity efficiently, whereas desalination, cooling, and full utilization of seawater are all effective ways to meet human water needs. However, corrosion problems for most metals and alloys in seawater desalination systems are more severe than in regular water because seawater is rich in natural electrolytes and highly corrosive substances, which cause general and localized corrosion. Both types of corrosion harm the service life of seawater desalination equipment and the safe operation of the system. Corrosion in seawater also causes high economic costs, such as lost production, product loss, efficiency loss, and product contamination. Even worse, it may cause catastrophic mishaps, such as harmful chemical leaks, resulting in pollution and putting people’s health at risk [1]. As a result, it is critical to pay attention to corrosion inhibition for metals in a saline environment by investigating corrosion behavior and taking appropriate anti-corrosion measures to use marine resources efficiently and actively respond to sustainable development. Standard coating systems for metal protection incorporate more than one type of coating system. However, protective coatings [2], corrosion inhibitors [3], nanocomposites [4], and electrochemical protection [5,6] are now the most widely utilized ways of improving corrosion resistance. Furthermore, various coating systems have been proposed by many researchers that effectively prevent metals from corrosion. Still, research related to non-toxic and cost-effective coating system that exhibits excellent corrosion resistance behavior is limited. Furthermore, in the corrosion protection of metals, polymer-based coatings are widely used since they possess strong binding capability with the metal surface that acts as a barrier. Through their functional groups, they form complexes with metal ions. These complexes occupy a large surface area on the metal surface and act as a blanket to the surface, protecting the metal from corrosive agents present in the solution [2,4]. However, the presence of micropores and microcracks generated during the coating formation leads to the failure of the coating due to the permeation of corrosive ions. According to recent research, loading the polymeric matrix with conductive polymers and nanoparticles can increase the active surface area of the coating, improve its electrostatic conductivity, and stimulate the creation of a passivation layer at the interface between metal and polymer [4,7,8]. Polyvinyl alcohol (PVA) is a non-toxic and water-soluble synthetic polymer that shows film-forming characteristics. PVA is also used as an anti-corrosion coating because it acts as a high oxygen barrier and adheres firmly to the metal surface [9,10,11]. On the other hand, coatings containing conducting polymers such as polyaniline (PANI) can protect pinholes and defects due to their passivating ability [7,8]. The PANI protection mechanism is electronic in nature, similar to the mechanism seen in metals. Furthermore, there are various reports about the improvement of coatings performance using nanoparticles as reinforcement, such as graphene (G) and graphene oxide (GO, the oxide of graphene) [4,12,13,14]. GO has several unique properties, such as good dispersion capability, and some of graphene’s basic characteristics, such as a large specific surface area, and excellent electrical and thermal conductivity [4,14]. However, by combining GO with PVA, the oxygen functional groups present in GO form a strong molecular interaction with PVA [15,16] and enhance the strength of the coating layer. Therefore, the development of hybrid materials with multiple combinations of polymers and GO is gaining attraction and can bring stability to the coating. It has been demonstrated that PVA has low water absorption behavior. It behaves as a corrosion inhibitor by forming complexes and covering surfaces to protect the metallic material from corrosion [9,10,17]. The dielectric properties of PANI indicate its capability to provide reasonably good corrosion protection [18]. Additionally, GO, which was used as reinforcement in both coatings, has also demonstrated significant anti-corrosion properties due to its hydrophobic nature, as a result of its non-polar covalent double bond [4,14]. However, GO cannot act as an active corrosion protection coating for long-term applications. The pores-deformations formed on the GO coating may increase the corrosion rate compared to the uncoated substrate [19]. To the best of our knowledge, there were no reports in the literature dealing with the direct deposit of PVA/PANI/GO-COOH nanocomposite coating on cast iron from an aqueous solution. The present study deals with the preparation of PVA/PANI blended with carboxylated graphene oxide (GO-COOH) prepared from upcycled plastic waste to form a novel nanocomposite coating. The characteristics of different types of prepared PVA, PVA/PANI, PVA/GO-COOH, and PVA/PANI/GO-COOH nanocomposite coatings were explored using FTIR analysis. In addition, these coatings were characterized by using weight loss measurements, potentiodynamic polarization studies, and electrochemical impedance analysis. Finally, the possibility of utilizing the PVA/PANI/GO-COOH nanocomposite coatings for corrosion protection of cast iron in an aqueous 3.5 wt.% NaCl solution was examined and optimized using the response surface methodology model. An excellent corrosion-resistant coating is expected by reducing the pores, with each constituent of the coating being believed to improve corrosion resistance. Furthermore, its cost of production is low, and large-scale implementation in commercial applications is comparatively convenient. 2. Materials and Methods 2.1. Materials PVA (MW 89,000–98,000, +99% hydrolyzed), aniline monomer, N-methyl-2-pyrrolidone (NMP), ammonium persulfate (APS), HCl, methanol, benzimidazole (98%), sodium chloride, sodium hydroxide, chloroacetic acid (ClCH2COOH, purum ≥ 97.0%) were purchased from Sigma-Aldrich (St. Louis, MI, USA). The chemical composition of the cast iron sample is shown in Table 1. Cast iron specimens of dimensions 3.7 × 1.2 × 0.2 cm3 were polished to a mirror finish, degreased with trichloroethylene, and used for weight loss and surface examination studies. 2.2. Preparation of Composite Coatings 2.2.1. Preparation of Carboxylated Graphene Oxide (GO-COOH) 0.1 g of reduced GO prepared from plastic waste using the previously mentioned procedure [20,21] was subjected to 1 h sonication. The resultant suspension (50 mL) was mixed with NaOH (0.6 g) and chloroacetic acid ClCH2COOH (0.5 g). The mixture was sonicated for 2 h to convert the –OH groups of GO into –COOH via conjugation of acetic acid moieties to obtain GO-COOH [22]. The resulting GO-COOH solution was purified by repeated rinsing with deionized water and filtrations until the product was well dispersed in water. 2.2.2. Preparation of PANI For the polymerizations in the liquid state, aniline (0.2 M) in 100 mL 1 M HCl was oxidized with ammonium peroxydisulfate (0.25 M) in 0.1 M HCl. Solutions of a monomer and an oxidant were mixed after pre-cooling at 0 °C to start the oxidation. The appearance of green color indicated that the polymerization was completed. The solution was washed several times with acetone to remove the unreacted monomer and then dried to obtain the required PANI powder. 2.2.3. Preparation of PVA Solution 10 wt.% PVA was dissolved in deionized water at 90 °C for 2 h until a completely clear solution was obtained. 2.2.4. Preparation of PVA/PANI Blended GO-COOH Nanocomposite Coating In the next step, the specific amount of PANI and GO-COOH were dissolved in 5 mL of NMP and kept on a sonication for 1 h to disperse, then added in PVA solution and mixed for 2 h. The PVA/PANI/GO-COOH nanocomposite composition was set as 98/0.5/1.5 wt.%. The complete process is schematically described in Figure 1. For coating the cast iron sample surfaces, solutions covered the surface using POLOS 300 Advanced PTFE spin coater. For comparison, four coatings were investigated: PVA, PVA/PANI, PVA/GO-COOH, and PVA/PANI/GO-COOH nanocomposite coatings. 2.3. Characterization Fourier transform infrared spectroscopy (FTIR) measurements of GO-COOH, PANI, PVA, and PVA/PANI/GO-COOH nanocomposites were obtained on a Nicolet 6700 spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) at room temperature in the 4000–400 cm−1 range. Weight loss experiments were conducted in 3.5 wt.% NaCl for 7 days at 25 °C. After the immersion period, the specimens were cleaned and reweighed to determine the corrosion rate [23]. Triple experiments were performed in each case, and the mean value of the weight loss is reported. Weight loss allows calculating the mean corrosion rate, as expressed in mg cm−2 h−1. The resulting quantity, corrosion rate (ωcorr) is thereby the fundamental measurement of corrosion. ωcorr can be determined either by chemical analysis of dissolved metal in solution or by gravimetric method measuring the weight of specimen before and after exposure in the aggressive solution applying the following equation: ωcorr = (mi − mf)/(S.t)(1) where mi, mf, S, and t denote initial weight, final weight, specimen surface, and immersion time, respectively. The inhibition efficiency, ηω%, was determined as follows:ηω% = (ω0corr − ωcorr)/ω0corr × 100(2) where ω0corr and ωcorr are the corrosion rates in the absence and presence of inhibition coating, respectively. Tafel polarization, linear polarization resistance, and electrochemical impedance spectroscopy experiments have been performed for further corrosion investigation by using a three-electrode assembly cell having a cast iron working electrode, a graphite rod counter electrode, and a saturated calomel electrode as reference electrode, and using Potentiostat/Galvanostatic Metrohm Autolab (Utrecht, The Netherlands), AUT85664). To test the stability of the prepared PVA/PANI/GO-COOH nanocomposite coating in the variable time interval with the solution temperature and saline solution concentration, the experimental data were evaluated and optimized using the response surface methodology (RSM) models. Data analysis and optimization were conducted using “Design Expert” 13.0.9.0 StatEase software. The results were analyzed using analysis of variance (ANOVA), standard residuals versus predicted values, and three-dimensional surface maps. 3. Results and Discussion 3.1. Characterization of Nanocomposite Coating Using FTIR Spectroscopy Fourier transform infrared spectroscopy was used to investigate the functional groups of the prepared materials as shown in Figure 2. The FTIR spectrum of PVA consisted of a strong broad band at 3765 cm−1 corresponding to OH stretching vibration of hydroxyl groups of PVA. The band corresponding to C–H asymmetric stretching vibration occurred at 2428 cm−1, and C–H symmetric stretching vibration was observed at 2159 cm−1. However, the spectrum of PANI showed bands at 1580, 1460, 1300, and 1130 cm−1, which could be assigned to the C=C stretching of the quinoid ring, C=C stretching of the benzenoid ring, C–N stretching of the benzenoid unit, and C–N stretching of the quinoid unit, respectively [24]. FTIR spectrum of GO-COOH consisted of a strong broad band at 3446 cm−1, corresponding to O-H stretching, while the band at 1636 cm−1 corresponded to C=O stretch in carboxylic acids, and the band at 1393 cm−1 corresponded to O-H bend [25]. In the composite spectrum, esterification between the –OH group of PVA and –COOH of the GO-COOH appeared at 1716 cm−1, and the band at 1427 cm−1 could be assigned to the C=C stretching of the quinoid ring of PANI. 3.2. Corrosion Tests 3.2.1. Corrosion Protection Performance of Prepared Coatings Prior to assessing the corrosion performance of composite coatings, the corrosion inhibition property of PVA/PANI/GO-COOH nanocomposite for cast iron coupon in 3.5 wt.% NaCl aqueous solution was examined as shown in Figure 3. After immersion in NaCl solution for 7 days, the surface of the cast iron coupon was covered with brown corrosion products (Figure 3a), while the sample coated with PVA/PANI/GO-COOH had no change in color (Figure 3e). Pure PVA coating, PVA blended PANI coating, and PVA doped with GO-COOH coating were also prepared for comparison. These results show that PVA/PANI/GO-COOH nanocomposite provided effective inhibition against corrosion. In addition, PVA, PVA/PANI, and PVA/GO-COOH composites also inhibited corrosion to some extent, as shown in Figure 3b–d. That may reveal that the interaction mechanism between the coating and the metal surface generally includes an adsorption phenomenon. The strength of the adsorption on a metallic surface depends largely on the electrostatic forces of attraction between the polar head of the polymer coating molecule with the iron atoms on the metal surface. The adsorption of the coating molecules on the metal surface is not only a precursor to surface adhesion, which forms a physical barrier to preclude the surface from chemical reactions but also a means to induce non-wetting. Therefore, in the case of PVA coating, the corrosion inhibition occurred to the presence of electron donor atoms of nitrogen in its molecular structure. So, the electron lone pair on the nitrogen will coordinate with the metal atoms of the active sites and increase adsorption to hence higher inhibition efficiency, which can be stabilized by the participation of the two adsorption modes, physisorption and chemisorption [23]. PVA/PANI coating presents the advantages of physical barrier protection with PVA and the redox features of PANI [26]. On the other hand, the composite of PVA with GO-COOH, a nanofiller reinforced PVA polymeric coating, acts as an excellent barrier to corrosive solutions and makes its diffusion path available in the tortuous path [27,28]. On the other hand, the carboxylic group on graphene surface nanofillers improve the quality and adherence of the PVA coating, reducing the porosity of the coating matrix and altering the physicochemical properties of the coating–cast iron interface, increasing the anti-corrosion properties [28]. 3.2.2. Weight Loss Studies Figure 4a displays weight loss plots of uncoated and coated samples with PVA, PVA/PANI, PVA/GO-COOH, and PVA/PANI/GO-COOH nanocomposite coating in 3.5 wt.% NaCl aqueous solution. As revealed by the graph, all the coatings showed significant improvement in corrosion resistance. Between all coatings, the PVA/PANI/GO-COOH nanocomposite coating showed a higher resistance in saline solution compared with PVA, PVA/PANI, and PVA/GO-COOH, as displayed by the lower value of weight loss. The data suggested that in the saline environment, the PVA/PANI/GO-COOH nanocomposite coating was more resistant to corrosive ions by 74% more than the PVA coating and resisted the ionic attack. However, PVA coating displayed relatively low resistance to corrosion and provided lower protection, as evident from its comparison with the other coatings. When PANI comes in contact with moisture or a corrosive medium, the nitrogen atoms in the polymer chain meet with hydrogen atoms and are reduced to the emeraldine (partially oxidized state) or nigeraniline (75% oxidized state) [27], resulting in the chain becoming twisted. This induces porosity in the coating, allowing the solvent molecules to reach other polymer chains beneath the top surface, propagating the reduction reaction and contributing to the growth of a passive oxide layer on the metal surface, protecting them [26]. Still, that probability is minimal due to the hydrogen gas evolution from redox reactions involved in the corrosion process. In the case of GO-COOH present in the composite structure, the approach of protons to the PANI nitrogen atoms is supposed to be hindered because GO-COOH remains essentially inert, permeable only to water molecules, and probably not to the solvated protons or hydronium ions. This suggests that the spread of the PVA/PANI/GO-COOH composite over the cast iron surface is most probably in the form of a network. 3.2.3. Morphological Characterization Cast iron coupons coated with PVA, PVA/PANI, PVA/GO-COOH, PVA/PANI/GO-COOH show (Figure 5a,c,e,g) a dense and compactly packed adhesive layer. The EDX results (Figure 5b,d,f,h) for coated coupons with PVA, PVA/PANI, PVA/GO-COOH, PVA/PANI/GO-COOH, respectively, show peaks at 0.25 and 0.50 keV due to the presence of C and O. In PVA/PANI and PVA/PANI/GO-COOH coated coupons, an additional peak of N at 0.30 keV was observed, confirming the doping of PAN into PVA matrix. The morphology images of rust coupons coated with PVA, PVA/PANI, PVA/GO-COOH, and PVA/PANI/GO-COOH are shown in Figure 5i,k,m,o. Compared with the large area of damaged coated layer on cast iron coupons with PVA, a circular cake shape corrosion product was observed for PVA/PANI. Less amount was observed for PVA/GO-COOH, and that was not observed for PVA/PANI/GO-COOH coated sample, indicating that the cast iron coupon coated with PVA/PANI/GO-COOH had the weakest degree of corrosion in the different samples, as confirmed by EDX (Figure 5j,l,n,p). 3.2.4. Electrochemical Characterization The corrosion resistance of the uncoated and PVA, PVA/PANI, PVA/GO-COOH, and PVA/PANI/GO-COOH coated samples was evaluated using Tafel analysis. “E-log I” curves for uncoated and coated samples exposed to 3.5 wt.% NaCl solution are represented in Figure 6a. In general, a higher Ecorr and a lower Icorr indicate greater inhibition efficiency [4]. The corrosion potentials of coated samples were positively shifted, and the corrosion current densities were significantly reduced compared to the uncoated cast iron bar. However, PVA/PANI/GO-COOH coated sample showed the highest right shift compared to PVA/PANI and PVA/GO-COOH coatings, and that increase in Ecorr was an indication of good anodic protection. This may be due to the good compatibility of PVA with PANI and GO-COOH additives and its effect as multiple barriers that effectively promote corrosion resistance by preventing the penetration of corrosive particles to the metal surface. The Nyquist plot in Figure 6b represents the corrosion behavior of uncoated and coated cast iron samples. All the experiments have been done for uncoated and coated cast iron samples at 35 °C and 0.5 M NaCl solution. It can be observed from the Nyquist plot that the semicircle diameter increases with different coatings, in the order PVA, PVA/PANI, PVA/GO-COOH, and PVA/PANI/GO-COOH. The larger arc radius reflects increasing corrosion resistance, the longer the electrolyte solution takes to corrode the coating, thus enhancing the corrosion resistance. The electrochemical impedance data for PVA/PANI/GO-COOH, which shows the best anti-corrosion performance, was interpreted in terms of the equivalent circuit composed of resistors and capacitors, such as the equivalent circuit shown in Figure S1 of supplementary material, including a solution resistance (Rs), coating resistance (Rf), coating capacitance (Cc), double layer capacitance (Cdl), and charge transfer resistance (Rct). However, the charge transfer resistance value is inversely proportional to the corrosion rate. A high charge transfer resistance value correlates to a low corrosion rate [29], and that was observed with PVA/PANI/GO-COOH coating, which showed the highest Rct value and lowest corrosion rate, confirming the PVA/PANI/GO-COOH coating anti-corrosion effect. It was concluded that the PVA/PANI/GO-COOH coating was more effective than other coatings. That can be explained by the fact that the PVA coating can isolate and protect the cast iron surface from the external corrosive environment, but during the curing process, some micropores will be formed on the metal surface due to the evaporation of the solvent. Consequently, the oxidizing medium can easily pass through the pores and cause corrosion. However, adding PANI and GO-COOH, which act as filler materials that fill the voids, improves the compactness and provides an extra barrier layer against the corrosive. In addition, PANI dispersed in PVA exerts a unique anti-corrosion effect that can passivate the iron substrate and delay corrosion. At the same time, the conductive properties of GO-COOH prevent the formation of rust. 3.2.5. Optimization of Corrosion Conditions for PVA/PANI/GO-COOH Coating The influence of time, temperature, and saline solution concentration on the inhibition efficiency, ηω%, of PVA/PANI/GO-COOH nanocomposite coating over the cast iron surface was investigated and optimized by using a design matrix following the Box-Behnken design [30] with 17 trials, as illustrated in Tables S2 and S3 of supplementary file. The 3D surface plots, as shown in Figure 7, exhibit the findings of the Box–Behnken design, demonstrating the type of interaction between the tested variables and the optimum conditions. The ANOVA analysis of variance is well-known for determining the statistical significance of the quadratic response surface model. As shown in Table S4, the quadratic model is very suitable for a high coefficient of determination R2 (0.9606) and low p-value (0.0004) of the model, which indicates the model was significant. In addition, p-values lower than 0.0500 indicate A, B, and C model terms were significant. However, the model F-value of 18.96 implies the model is significant, and there is only a 0.04% chance that an F-value this large could occur due to noise. According to the model’s calculations, the numerical relationship between the independent variable and the responses Y (inhibition efficiency) is as follows:Y= 78 − 7.54A − 3.25B − 7.79C − 2.09A2 − 0.6625B2 − 1.09C2 − 1.25AB + 2.32 AC + 0.75BC(3) The optimal levels of the three components at the maximum point of the polynomial model were 4 days at 28.7 °C and 0.53 M NaCl solution concentration to obtain maximum inhibition efficiency (about 93%) for PVA/PANI/GO-COOH coating. 4. Conclusions A greater number of pores mean a greater number of corrosive particles pass through the coating, which exposes more surface area of the metal to corrosion reaction. Therefore, the porosity of the coating layer is critical in achieving good corrosion resistance. As a result, the coating with the higher pore resistance in any given environment performed better than the one with the lower coating resistance value. According to the proposed mechanism, PVA/PANI/GO-COOH composite molecules arrange to form an optimized network structure when coated on the cast iron surface and thus provide 94% inhibition efficiency compared to the uncoated sample. The excellent corrosion protection of PVA/PANI/GO-COOH was optimized to 93.3% for 4 days at 0.53 M NaCl solution concentration and 28.7 °C. PVA/PANI/GO-COOH nanocomposite has many advantages such as the eco-friendly component, low cost of production, and large-scale implementation in commercial applications is comparatively convenient. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/polym14091791/s1, Figure S1: Suggested equivalent circuit for PVA/PANI/GO-COOH coating, Table S1: Level of various independent variables at coded values of response surface methodology experimental design; Table S2: The Box-Behnken design matrix and results for the three variables that influenced on Inhibition efficiency (%) of PVA/PANI/GO-COOH nanocomposite coating; Table S3: The Box-Behnken design matrix and results for the three variables that influenced on Inhibition efficiency (%) of PVA/PANI/GO-COOH nanocomposite coating; Table S4: ANOVA analysis for response function Y (inhibition efficiency (%)). Click here for additional data file. Author Contributions Conceptualization, N.A.E. and M.H.G.; methodology, M.H.G., M.E. and N.A.T.; software, N.A.E. and M.E.Y.; investigation, N.A.E. and M.H.G.; writing—original draft preparation, N.A.E.; writing—review and editing, M.E.Y. and D.M.F.S.; visualization, N.A.E.; supervision, M.E.Y. and D.M.F.S.; project administration, M.E.Y. and D.M.F.S. All authors have read and agreed to the published version of the manuscript. Data Availability Statement All data generated or analysed during this study are included in this published article and its Supplementary Information files. Conflicts of Interest The authors declare no conflict of interest. Figure 1 Schematic illustration of preparation process of the PVA/PANI blended GO-COOH nanocomposite coatings. Figure 2 FTIR spectra of PVA, PANI, GO-COOH, PVA/PANI/GO-COOH composite coating. Figure 3 Optical images of cast iron coupons after 7 days of exposure to 3.5 wt.% NaCl (a) uncoated, (b) coated with PVA, (c) coated with PVA/PANI, (d) coated with PVA/GO-COOH, (e) coated with PVA/PANI/GO-COOH. Figure 4 (a) Weight loss curves for cast iron in 3.5 wt.% NaCl solution, (b) inhibition efficiency, ηω%, in the absence and presence of various coatings. Figure 5 Typical SEM images and EDX spectra of cast iron coupons coated with (a,b) PVA, (c,d) PVA/PANI, (e,f) PVA/GO-COOH, and (g,h) PVA/PANI/GO-COOH. SEM images and EDX spectra of cast iron coupons coated with (i,j) PVA, (k,l) PVA/PANI, (m,n) PVA/GO-COOH, and (o,p) PVA/PANI/GO-COOH after 7 days of exposure to 3.5 wt.% NaCl. Figure 6 (a) Tafel polarization curves and (b) Nyquist plot of the coating samples after being immersed in 3.5 wt.% NaCl solution for 7 days. Figure 7 (a,c,e) The 3D surface plots and (b,d,f) 2D surface plots of the inhibition efficiency, ηω%, response of PVA/PANI/GO-COOH nanocomposite coating over the cast iron surface. polymers-14-01791-t001_Table 1 Table 1 Chemical composition of cast iron specimens used. Element Composition, wt.% C Si Mn S P Fe 3.17 2.84 0.37 0.12 0.09 93.41 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Hou X. Gao L. Cui Z. Yin J. Corrosion and protection of metal in the seawater desalination IOP Conf. Ser. Earth Environ. Sci. 2018 108 022037 10.1088/1755-1315/108/2/022037 2. Shen L. Zhao W. Miao L. Designed a novel EP + GO/ZRC + GO coating with bilayered structure for enhancing corrosion resistance of steel substrate J. Hazard. Mater. 2021 403 123670 10.1016/j.jhazmat.2020.123670 33264874 3. Aghzzaf A.A. Rhouta B. Rocca E. Khalil A. Steinmetz J. Corrosion inhibition of zinc by calcium exchanged beidellite clay mineral: A new smart corrosion inhibitor Corros. Sci. 2014 80 46 52 10.1016/j.corsci.2013.10.037 4. Chen S. Li B. Xiao R. Luo H. Yu S. He J. Liao X. Design an epoxy coating with TiO2/GO/PANI nanocomposites for enhancing corrosion resistance of Q235 carbon steel Materials 2021 14 2629 10.3390/ma14102629 34069811 5. Zheng H. Liu Y. Zhou Y. Zhao D. Wang D. Yun L. Zhang D. Zhang L. Improved photocathodic protection performance of g-C3N4/rGO/ZnS for 304 stainless steel J. Phys. Chem. Solids 2021 148 109672 10.1016/j.jpcs.2020.109672 6. Xu D. Liu Y. Liu Y. Chen F. Zhang C. Liu B. A review on recent progress in the development of photoelectrodes for photocathodic protection: Design, properties, and prospects Mater. Des. 2021 197 109235 10.1016/j.matdes.2020.109235 7. Zhang Y. Shao Y. Liu X. Shi C. Wang Y. Meng G. Zeng X. Yang Y. A study on corrosion protection of different polyaniline coatings for mild steel Prog. Org. Coat. 2017 111 240 247 10.1016/j.porgcoat.2017.06.015 8. Gao F. Mu J. Bi Z. Wang S. Li Z. Recent advances of polyaniline composites in anticorrosive coatings: A review Prog. Org. Coat. 2021 151 106071 10.1016/j.porgcoat.2020.106071 9. Ramezanzadeh B. Vakili H. Amini R. The effects of addition of poly (vinyl) alcohol (PVA) as a green corrosion inhibitor to the phosphate conversion coating on the anti-corrosion and adhesion properties of the epoxy coating on the steel substrate Appl. Surf. Sci. 2015 327 174 181 10.1016/j.apsusc.2014.11.167 10. Dwivedi A. Bharti P. Shukla S. Surface assimilation and corrosion inhibition characteristic of water soluble polyvinyl alcohol on mild steel surface in 0.5M HCl solution J. Turk. Chem. Soc. 2021 8 217 228 10.18596/jotcsa.794721 11. Peng L. Wang D. Qi J. Study on anti-corrosion of PVA-treated wheat straw and its application in reinforcement of a silty soil Constr. Build. Mater. 2021 291 123305 10.1016/j.conbuildmat.2021.123305 12. Jafari Y. Ghoreishi S.M. Shabani-Nooshabadi M. Polyaniline/graphene nanocomposite coatings on copper: Electropolymerization, characterization, and evaluation of corrosion protection performance Synth. Met. 2016 217 220 230 10.1016/j.synthmet.2016.04.001 13. Hikku G.S. Jeyasubramanian K. Venugopal A. Ghosh R. Corrosion resistance behaviour of graphene/polyvinyl alcohol nanocomposite coating for aluminium-2219 alloy J. Alloy. Compd. 2017 716 259 269 10.1016/j.jallcom.2017.04.324 14. Yang S. Zhu S. Hong R. Graphene oxide/polyaniline nanocomposites used in anticorrosive coatings for environmental protection Coatings 2020 10 1215 10.3390/coatings10121215 15. Li J. Shao L. Zhou X. Wang Y. Fabrication of high strength PVA/rGO composite fibers by gel spinning RSC Adv. 2014 4 43612 43618 10.1039/C4RA07295B 16. Gouda M.H. Ali S.M. Othman S.S. Al-Aziz S.A.A. Abu-Serie M.M. Elsokary N.A. Elessawy N.A. Novel scaffold based graphene oxide doped electrospun iota carrageenan/polyvinyl alcohol for wound healing and pathogen reduction: In-vitro and in-vivo study Sci. Rep. 2021 11 20456 10.1038/s41598-021-00069-0 34650075 17. Srimathi M. Rajalakshmi R. Subhashini S. Polyvinyl alcohol–sulphanilic acid water soluble composite as corrosion inhibitor for mild steel in hydrochloric acid medium Arab. J. Chem. 2014 7 647 656 10.1016/j.arabjc.2010.11.013 18. Abbes I. Srasra E. Electrical and dielectric properties of polyaniline and polyaniline/montmorillonite nanocomposite prepared by solid reaction using spectroscopy impedance J. Nanomater. 2015 2015 516902 19. Schriver M. Regan W. Gannett W.J. Zaniewski A.M. Crommie M.F. Zettl A. Graphene as a long-term metal oxidation barrier: Worse than nothing ACS Nano 2013 7 5763 5768 10.1021/nn4014356 23755733 20. El Essawy N.A. Konsowa A.H. Elnouby M. Farag H.F. A novel one-step synthesis for carbon-based nanomaterials from polyethylene terephthalate (PET) bottles waste J. Air Waste Manag. Assoc. 2017 67 358 370 10.1080/10962247.2016.1242517 27700617 21. El Essawy N.A. Ali S.M. Farag H.A. Konsowa A.H. Elnouby M. Hamad H.A. Green synthesis of graphene from recycled PET bottle wastes for use in the adsorption of dyes in aqueous solution Ecotoxicol. Environ. Saf. 2017 145 57 68 10.1016/j.ecoenv.2017.07.014 28708982 22. Ziółkowski R. Górski Ł. Malinowska E. Carboxylated graphene as a sensing material for electrochemical uranyl ion detection Sens. Actuators B Chem. 2017 238 540 547 10.1016/j.snb.2016.07.119 23. Alaoui K. El Kacimi Y. Galai M. Touir R. Dahmani K. Harfi A. Ebn Touhami M. Anti-corrosive properties of polyvinyl-alcohol for carbon steel in hydrochloric acid media: Electrochemical and thermodynamic investigation J. Mater. Environ. Sci. 2016 7 2389 2403 24. Dong Y. Zhou Y. Ding Y. Chu X. Wang C. Sensitive detection of Pb(II) at gold nanoparticle/polyaniline/graphene modified electrode using differential pulse anodic stripping voltammetry Anal. Methods 2014 6 9367 9374 10.1039/C4AY01908C 25. Chhabra V.A. Deep A. Kaur R. Kumar R. functionalization of graphene using carboxylation process Int. J. Sci. Emerg. Technol. Latest Trends 2012 4 13 19 26. Bandeira R.M. Drunen J. Ferreira F.A. Rodrigues-Filho U.P. Tremiliosi-Filho G. Chemically synthesized polyaniline/polyvinyl chloride blended coatings for the corrosion protection of AA7075 aluminum alloy Corros. Sci. 2018 139 35 46 10.1016/j.corsci.2018.04.031 27. Mahato N. Cho M.H. Graphene integrated polyaniline nanostructured composite coating for protecting steels from corrosion: Synthesis, characterization, and protection mechanism of the coating material in acidic environment Constr. Build. Mater. 2016 115 618 633 10.1016/j.conbuildmat.2016.04.073 28. Liu D. Zhao W. Liu S. Cen Q. Xue Q. Comparative tribological and corrosion resistance properties of epoxy composite coatings reinforced with functionalized fullerene C60 and graphene Surf. Coat. Technol. 2016 286 354 364 10.1016/j.surfcoat.2015.12.056 29. Ammar A.U. Shahid M. Ahmed M.K. Khan M. Khalid A. Khan Z.A. Electrochemical study of polymer and ceramic-based nanocomposite coatings for corrosion protection of cast iron pipeline Materials 2018 11 332 10.3390/ma11030332 30. Box G. Behnken D.W. Some new three level designs for the study of quantitative variables Technometrics 1960 2 455 475 10.1080/00401706.1960.10489912

PMC009xxxxxx/PMC9099694.txt

==== Front Sensors (Basel) Sensors (Basel) sensors Sensors (Basel, Switzerland) 1424-8220 MDPI 10.3390/s22093411 sensors-22-03411 Article Evaluation of Plant Stress Monitoring Capabilities Using a Portable Spectrometer and Blue-Red Grow Light https://orcid.org/0000-0002-0963-4216 Merrick Trina 1* Bennartz Ralf 23 Jorge Maria Luisa S. P. 2 https://orcid.org/0000-0001-8135-9266 Pau Stephanie 4 https://orcid.org/0000-0003-3368-5408 Rausch John 2 Chakraborty Somsubhra Academic Editor 1 Naval Research Laboratory, Remote Sensing Division, 4555 Overlook Ave. SW, Washington, DC 20375, USA 2 Department of Earth and Environmental Science, Vanderbilt University, 5726 Stevenson Center, Nashville, TN 37240, USA; ralf.bennartz@vanderbilt.edu (R.B.); malu.jorge@vanderbilt.edu (M.L.S.P.J.); john.rausch@vanderbilt.edu (J.R.) 3 Space Science and Engineering Center, University of Wisconsin—Madison, 1225 W Dayton St., Madison, WI 53706, USA 4 Department of Geography, Florida State University, 113 Collegiate Loop, Tallahassee, FL 32306, USA; spau@fsu.edu * Correspondence: trina.merrick@nrl.navy.mil 29 4 2022 5 2022 22 9 341124 3 2022 21 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Remote sensing offers a non-destructive method to detect plant physiological response to the environment by measuring chlorophyll fluorescence (CF). Most methods to estimate CF require relatively complex retrieval, spectral fitting, or modelling methods. An investigation was undertaken to evaluate measurements of CF using a relatively straightforward technique to detect and monitor plant stress with a spectroradiometer and blue-red light emitting diode (LED). CF spectral response of tomato plants treated with a photosystem inhibitor were assessed and compared to traditional reflectance-based indices: normalized difference vegetation index (NDVI) and photochemical reflectance index (PRI). The blue-red LEDs provided input irradiance and a “window” in the CF emission range of plants (~650 to 850 nm) sufficient to capture distinctive “two-peak” spectra and to distinguish plant health from day to day of the experiment, while within day differences were noisy. CF-based metrics calculated from CF spectra clearly captured signs of vegetation stress earlier than reflectance-based indices and by visual inspection. This CF monitoring technique is a flexible and scalable option for collecting plant function data, especially for indicating early signs of stress. The technique can be applied to a single plant or larger canopies using LED in dark conditions by an individual, or a manned or unmanned vehicle for agricultural or military purposes. spectroscopy chlorophyll fluorescence vegetation indices NDVI PRI photosynthesis photosystem inhibition Goetz Instrument Loan ProgramThis work was supported by the Goetz Instrument Loan Program (2015) through ASD (Analytical Spectral Devices), as well as a partnership between the São Paulo Research Foundation (Fundação de Amparo A Pesquisa do Estado de São Paulo, FAPESP), Vanderbilt University, Nashville, TN, USA and Universidade Estadual Paulista (UNESP), Rio Claro, Brazil (FAPESP #2013/50421-2 and 20599-00-5). ==== Body pmc1. Introduction Spectrometer measurements of chlorophyll fluorescence (CF) capture photosynthetic activity and plant function information, have been shown related to gross primary production (GPP) and carbon uptake, and can capture signals of plant stress, e.g., [1,2,3]. CF is a physiological process undergone by plants to dissipate excess energy from photosynthesis to protect tissues by emitting radiation mostly in the red and far-red region of the spectrum (approximately 650 to 850 nm) and, as a by-product of photosynthesis, provides a mechanistic link to plant function [4,5]. Thus, CF offers opportunities to assess vegetation status at leaf, plant, canopy, ecosystem, and global levels across spatial and temporal scales, e.g., [6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23]. Two predominant categories of CF measurement techniques are active and passive techniques. Widely used active techniques employ lasers, light emitting diode (LED) illumination, or lamps to excite chlorophyll and instruments, such as cameras or spectrometers, to record CF promptly or delayed and include techniques, such as pulse-amplitude-modulation (PAM) [4,7,12,16,24,25,26,27,28,29,30,31,32,33,34,35]. By activating photosynthesis with a light of a known, narrow wavelength, the PAM technique allows measurements of several aspects of photosynthetic activity in ambient light conditions in great detail. Such active techniques, however, are restricted mainly to leaf level and have intricate protocols to follow for experiment design and are not easily scaled up. The other category is measuring CF in ambient conditions, mostly in sunlight. In contrast to PAM, CF measured with the so-called passive techniques can be made at leaf to ecosystem scales using spectrometers mounted on the laboratory bench, towers, drones, or on satellites. In addition to allowing data collection at multiple scales, passive CF measurements have also been shown to scale effectively to larger levels, such as globally, and are more generalizable because the light activating photosynthesis is usually sunlight, unlike the active technique [23,28,36,37,38,39,40,41]. CF can be challenging to measure because the CF signal is small compared to the amount of sunlight, or potentially alternative light, the plant requires to drive photosynthesis. For this reason, most methods to examine CF have remained complex, time intensive, require extensive knowledge of the instrument and illumination source specifications, uncertainties, protocols, and results are difficult to interpret. For example, current passive CF measurement methods include retrievals under sunlight requiring one of several algorithms to exploit the oxygen absorption bands that coincide with the CF emission spectrum centered at approximately 687 and 761 nm. The retrieved CF values are used for estimates or as model inputs to estimate the CF spectra, e.g., [22,42,43,44,45,46]. Another option is to use an actinic light source having a “window” coincident with the CF spectrum can allow CF to be captured with a spectroradiometer. In many cases, a filter is placed within a chamber, such as within the source chamber for a leaf clip, to block incoming light in the CF wavelength range and then record the plant CF spectra in that “window”. This filtering and measurement process can be tedious, require expensive components, and, in the case of using the leaf clip, restrict measurements to leaf scale similar to the PAM method [42,47,48,49,50]. Therefore, despite advantages over active techniques, passive measurements of CF have remained complex. With the goal of identifying a potentially more straightforward CF-based technique to detect plant stress, an initial investigation was carried out to address the following: (1) what level of sensitivity to plant stress can be measured using a spectrometer and blue-red light emitting diode grow light, and (2) are CF spectra or CF metrics derived from these spectra indicative of stress earlier than visual inspection or traditional vegetation indices? Specifically, an Analytical Spectral Devices (ASD)/Malvern FieldSpec® HandHeld 2™ Spectroradiometer (HH2) with a blue-red grow light emitting diode lamp (MiracleLED™ 2.2 Watt; LED) was used to make repeated measurements of a target plant treated with a photosystem inhibiting poison. The ability to detect CF, the sensitivity of the CF region of the spectra, and metrics derived from the CF region were tested and metrics were compared to measurements of the normalized vegetation index (NDVI) and the photochemical reflectance index (PRI), which are often used to assess vegetation function. 2. Materials and Methods Healthy, similar-sized tomato plants (Solanum lycopersicum), chosen for applicability for follow up studies, were placed in two chambers (target and control) of a dark tent in the laboratory (Figure 1a). On Day 1 of the experiment, a dose of a photosystem II (PSII) inhibitor, an algicide/herbicide called 3-(3′,4′-dichlorophenyl)-1,1-dimethylurea (DCMU) mixed with 300 mL water was applied to the soil of the target plant. DCMU blocks the electron flow from photosystem II (PSII) during photosynthesis, thus inducing stress when plants cannot effectively convert incoming energy to useful forms for carrying out proper functioning. For this reason, DCMU makes an efficient weed killer, for instance. The large application amount relative to what might be used for weed control in a natural setting was chosen to illicit an intentional stress response due to photosynthetic system shutdown in order to test the capability to detect and track [51]. In this scenario, CF would be expected to rise dramatically in the beginning stages because the energy supplied by incoming illumination would be blocked from entering PSII and a greater amount of excess energy would be given off as CF. After this initial increase, a decline in function results in a decline in CF. The HH2 was mounted on a standard tripod above the target tomato plant canopy in the dark tent enabling a field of view of approximately 29 cm diameter (Figure 1a). An LED was positioned in each chamber using flexible bulb holders from a single lamp base to provide incoming photosynthetically active radiation (PAR; ~400–680 nm), one illuminating the target tomato plant and one illuminating the control. The LED emits light only in the blue and red regions, leaving a “window” in the upper red and far-red regions where plants emit CF as a by-product of photosynthesis (Figure 1b). Using a timer, the LEDs were turned on at 7:00 a.m. CDST each day and turned off at 7:00 p.m. CDST each evening, during which the HH2 was set to record a spectral measurement every 30 min. Figure 1c shows examples CF spectra on Day 1 of the experiment and the typical two-peak feature of a CF spectrum are distinguishable. For reference, an incoming (LED) spectrum (measured using a Spectralon® white reference panel (WR)) is overplotted to illustrate the CF and LED spectra are clearly distinguishable from one another. However, it should be noted that the magnitude of the LED spectrum is lower when illuminating the plant than when illuminating the WR (albedo of ~1, while the plant albedo ~0.4). Once daily, a Spectralon® white reference panel (WR) was placed above the target plant canopy and an irradiance measurement of the LED was made. In addition, a measurement of the target and control plant and the WR were made under an ASD/Malvern Illuminator Reflectance Lamp (IRL, 70 W stable quartz-tungsten-halogen calibrated light source, ASD/Malvern, Boulder, CO, USA). The spectra under IRL were used to calculate daily normalized difference vegetation index (NDVI) and photochemical reflectance index (PRI) measurements. The daily reflectance of the control plant did not change throughout the experiment (not shown). The CF portion (650–800 nm region) was extracted from each observation made between 8:00 a.m. CDST and 6:00 p.m. CDST. The first and last 1–2 measurements of the day were omitted to avoid erroneous measurements due to mismatch of the lamp timer and HH2 internal clock/auto-timer setting or any drift in either timer. A daily average CF spectrum and standard deviation at each wavelength was calculated. The resulting CF spectra were smoothed (central-moving average, window = 5). In addition to the full CF spectrum, CF values at 685 nm and 740 nm were extracted (Fred, Ffar, respectively) and then the Fred/Ffar ratio calculated [21,51,52,53]. Daily NDVI [54] was calculated by calculating reflectance from the target plant spectrum and the IRL reference spectrum taken once each day and using:(1) NDVI=R800−R680R800+R680 and PRI [55] as (2) PRI=R531−R570R531+R570 where Rλ is the reflectance and λ is the indicated wavelength in nm. To examine the relative capabilities of the measurements to capture the dynamics of photosynthetic function, we also calculated the proportion of maximum NDVI, PRI, Fred, and Ffar as a time series. Custom programs to read, write, and batch process the spectra from their proprietary format as well as specifically process and make figures from the experiment were written in Interactive Data Language (IDL; L3Harris Geospatial, Boulder, CO, USA). 3. Results Days 0 and 1 show the near and far-red peaks expected for the fluorescence range, which become less distinguishable on Days 2–4 as the variability for each increase (Figure 2a and Figure A1a–e). While the plant canopy is visibly unchanged from Day 0 to Day 3 (Figure 2a and Figure A1a–d), the red peak increases by Day 3 compared to Day 0 by approximately 4 times, while the far-red peak increases by 3.5 times, likely due to the increase in CF corresponding to photosynthetic inhibition. After the Day 3 peak in the fluorescence values, both peaks are decreased by Day 4, when the first signs of leaf discoloration and some wilting appear in the canopy, yet CF is still elevated compared to before DCMU application (Figure 2c and Figure A1e). By Day 5 of the experiment, the two peaks of fluorescence are indistinguishable, eventually reducing to noise around zero by Day 7 (Figure A1f–h) and most leaves on Day 5 show some withering, discoloration, spots, or slightly less green at a minimum and some portions of soil began to show through the canopy when viewed from above continuing to degrade through Day 8, the final day of measurements, the recorded CF spectra and photos show little to no green vegetation being measured by the HH2 (Figure 2d and Figure A1a–i). From Day 0–3, Fred and Ffar increase dramatically (333% and 166%, respectively, Figure 3a) going from approximately 0.2 and 0.3 of maximum to 1.0 (maximum values) in this time period, while PRI increased from 0.8 of maximum to 1.0, an increase of 35%, only subtly capturing photosynthetic response (Figure 3d,e). On Days 3–5, Fred and Ffar decline close to Day 2 levels (183% and 32% of Day 0 values, respectively) reaching approximately 0.2 and 0.3 of maximum (Figure 3a,e) and PRI performs more similarly going from maximum (reached on Day 2) to 0.4 of maximum, indicating the pigment changes occurring in the plant leaves. Day 5 and beyond Fred and Ffar are below the Day 0 values for Fred and Ffar, reaching values on Day 8 of 84% and 80% decrease from Day 0, while PRI reaches a value of −0.05 by Day 8 capturing pigment changes and structural changes ion with structural changes in the plant, e.g., [1,13] (Figure 3d,e). Fred/Ffar increases to a maximum value of 1.2 on Day 4, when Fred surpasses the value of Ffar, then declines (Figure 3b inset shows Day 0–Day 5 detail) which highlights the capability of the instrument to record CF signals that increase when PSII is initially blocked and excess energy dissipation as CF increases, capturing initial stress reactions. The Fred/Ffar (Figure 3b) shows variability so large after Day 5 (due to increasingly low values of Fred and Ffar compared to instrument noise) that the trend over the first five days is obscured by the scale of the graph. The decrease in signal for the Ffar region of the spectrum results in dramatic increase in noise and extreme variability in Fred/Ffar on Days 6–8. In contrast, NDVI decreases throughout the experiment from approximately 0.87 to 0.69 at an average of 0.03 per day (a total decrease of only 21%, and approximately 0.8 of maximum value) and does not capture dynamic changes during photosystem inhibition and does not show a response to the DCMU poisoning of the plant prior to visible signs of stress on the plant on Day 4 (Figure 3c,e). Additionally, the value of 0.67 for NDVI at the end of the experiment indicates NDVI does not fully capture the reduced functioning of the plant (Figure 3c). A comparison of the daily NDVI, PRI, Fred, and Ffar using the proportion of maximum of each (Figure 3e) highlights the relative degree to which these metrics capture changes in the response of the plants to stress. In this manner we show the large changes in PRI, Fred, and Ffar as compared to the change in NDVI. Figure 3e also illustrates the difference in timing of the maximum values, which is the beginning of the experiment for NDVI, Day 2 for PRI, and Day 3 for Fred and Ffar, further supporting that NDVI does not respond clearly to the initial stress, PRI has a potential moderate response, but Fred and Ffar capture both the increase in dissipated energy during the subsequent onset of stress and the decline in function. Pearson correlations (r values) among the metrics are shown in Figure 3f for comparison, especially between the CF metrics and RIs (NDVI and PRI). No significant correlation exists between Fred and NDVI, Fred and Fred/Ffar, as well as Ffar and Fred/Ffar. However, a positive relationship exists between PRI and Ffar (r = 0.88, p < 0.01) for PRI and Fred (r = 0.7, p < 0.05) as well as for NDVI and Ffar (r = 0.75, p < 0.05). The Fred/Ffar is negatively correlated with both NDVI (r = −0.83, p < 0.01) and PRI (r = −0.79, p < 0.05). Fred and Ffar are highly positively correlated and are likely autocorrelated which mathematically explains the lack of relationship to their ratio (Fred/Ffar). PRI and NDVI are also strongly positively correlated (r = 0.96, p < 0.01). 4. Discussion An overarching goal of this study was to confirm that CF at canopy level could be detected and monitored with a relatively straightforward measurement technique. The general expectation that the increase in CF emission that would coincide with photosystem inhibition would manifest in an increase in the far-red peak of the spectra was confirmed. While individual measurements were too noisy to discriminate half-hourly or hourly decreases in plant function, likely due to the combination of an uncalibrated illumination source (LED) and dynamic and complex response of fluorescence, it was found that smoothed signals on a daily basis yielded CF information that could be used for daily plant function information, especially for indication of stress. Based on the results, CF measurements with this technique promise to evaluate more precisely the function and the stress than NDVI and PRI, which are both reflectance-based indices (RI). CF spectra and CF metrics, especially Fred and Ffar, capture the initial photosynthesis stress response and the decline of plant function to a greater degree than PRI and NDVI highlighting the capability of CF to better monitor dynamics of plant status. Because CF originates from photosynthetic machinery of plants, it was anticipated that if distinct CF signals were captured and distinguishable, the CF spectra and CF metrics would detect stress earlier and be more sensitive in monitoring it than RIs. NDVI is the most used vegetation index and is often referred to as a “greenness index” due to its widespread use to track large scale seasonality. NDVI is calculated from reflectance in the red and near-infrared portion of the spectrum, is sensitive to changes in biomass and leaf area index (LAI), is a good indicator of absorbed photosynthetically active radiation (APAR) of a canopy and mainly captures structural changes (biomass and LAI) seasonally [1,13,55]. However, the structural changes in the target plant lagged behind the functional response, i.e., conversion of short wavelength energy from the PAR range to longer wavelength emission byproduct of the photosystem in the CF range. PRI is significantly positively correlated with both Fred and Ffar, while NDVI is only significantly correlated with Ffar. While NDVI and PRI can explain a good deal of variation in CF, RI sensitivity to plant function is lower than measurements of CF. PRI is also an RI, but unlike NDVI, has been shown to detect changes in photosynthetic activity on two timescales: (1) diurnally capturing responses of the xanthophyll cycle to changing illumination and (2) pH of thylakoid lumen and seasonally capturing changes in the chlorophyll-carotenoid ratios, i.e., pigment content of leaves. PRI has also been shown to correlate with light-use efficiency (LUE) of some vegetation [1,13,56,57,58,59]. In this experiment, CF response was more sensitive to the plant stress than either NDVI or PRI, but the decline in PRI over the last days of the experiment showed the changes in leaf pigment present were detectible with the PRI and was more sensitive to changes in the canopy than NDVI. This is consistent with several studies showing PRI as a good indicator of plant function, albeit the precise responses are still being studied for temporal and spatial interpretation in some ecosystems, e.g., [13,59]. The comparison of the portions of the CF spectrum, Fred and Ffar, and their ratio, Fred/Ffar, revealed dynamic responses to plant stress and show promise to make distinctions regarding physiology of the target plant. The shape of CF peaks for a plant at room temperature depend on leaf chlorophyll-a concentration, structure, and constituents, and the optical properties of the leaf determine the penetration depth of incident light and remission of CF from these depths [7,21,51,60,61,62,63,64]. As was the case in this investigation, studies show the Fred region suffers from reabsorption of CF photons to a greater degree than the Ffar region, explaining the lower peak of Fred than Ffar, e.g., [3,65]. Additionally, Fred increased at a relatively higher rate than Ffar in response to DCMU poisoning, which was expected from this particular herbicide. DCMU as a photosystem inhibitor most affects PSII; thus, it is expected that the PSII inhibition of photosynthesis would manifest as a dramatic increase in Fred, which is the case in this study [28,62,66]. While there is a PSI and PSII contribution to both Fred and Ffar, Fred is produced dominantly by PSII and is more variable compared to Ffar. Furthermore, in healthy leaves, Fred is close to or less than Ffar, but in stressed leaves, Fred increases and Ffar decreases [5,60,67]. It is for the reasons mentioned that the red peak (Fred) and far-red peak (Ffar) and the red-far-red ratio (Fred/Ffar) are metrics that have been utilized to detect stressed vegetation and estimate plant functioning status, and all three were indicative of changing function of the target plant in this experiment [53,62,68,69,70]. In this study, we demonstrate that a system using the HH2 and blue-red LED grow lamp to detect and monitor measure plant canopy level function, especially in the context of stress, has potential in a variety of applications. The capability to capture distinct signals of plant canopy stress earlier than traditional reflectance-based indices, NDVI and PRI, plus the capability to scale from small to larger canopies holds promise for both research and applications in the lab or field, at night for instance. Recently, two larger scale studies have employed LEDs with spectroscopy at night to study fluorescence responses, one in a forested area examined steady state responses of canopy and understory in a scots pine forest [71] and another used tractor mounted spectrometer and LEDs to measure fluorescence and compare to aerial net primary production among varieties of soybean, both rainfed and irrigated [72]. In the study of soybeans, the authors also show that the Fred/Ffar revealed differences in plant function among cultivars, while traditional RI’s did not. Our results support findings in these studies, albeit on a smaller scale, such that might be found in a laboratory or greenhouse. Taken together, the results of this study show the potential of the HH2 sensor to provide robust plant function information with this technique. 5. Conclusions This work presents a process to measure CF using relatively straightforward methods and interpretation compared to other approaches aiming to capture fluorescence responses. Additionally, the technique can be applied to a single plant or larger canopies using LED, in dark conditions. The ability to collect data at leaf, plant, and canopy levels reliably and scale between these levels effectively with measurable uncertainties could also be applied, for instance, using manned or unmanned vehicles in agricultural or military applications. This scaling cannot be accomplished as easily or at all with other photosynthesis measurement techniques, such as pulse amplitude modulated fluorimetry (PAM). These procedures can be effectively employed as-is or with modifications in vegetation research and applications in multiple ways. For instance, plant function on multiple timescales could be further investigated by employing calibration for LEDs and examining timescales where signals would be statistically significantly different. In addition, CF measurements with our technique could be used in conjunction with PRI and NDVI measurements to inform plant studies along multiple timescales, i.e., CF is sensitive on the finer scale to track photosynthetic function and PRI on longer timescales to inform changes in leaf pigment, and add NDVI for seasonal structure changes. Therefore, the HH2 and LED system presented has broad appeal to multiple vegetation research areas. Author Contributions Conceptualization, T.M.; methodology, T.M., R.B., M.L.S.P.J. and J.R.; data curation and analysis T.M., R.B., J.R. and S.P.; writing—original draft preparation, T.M.; writing—review and editing, T.M., R.B., S.P., M.L.S.P.J. and J.R. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement All data and code are freely available. Location and access provided upon request to the primary author. Conflicts of Interest The authors declare no conflict of interest. Appendix A Daily fluorescence spectra, images and visual descriptions of experiment. Figure A1 Daily average smoothed fluorescence and photographs of target tomato plant for all days of experiment (Days 0 to 8 spectra and photographs labelled (a–i), respectively). Figure 1 (a) The experimental instrument setup included the handheld ASD/Malvern HandHeld-2 Pro spectroradiometer (Boulder, CO, USA; HH2) and a blue-red LED grow light. A full spectrum lamp is added here for the purposes of a clear photo only. (b) Plot of blue-red light emitting diode grow light (MiracleLED™ 2.2 Watt; LED) spectrum, measured by recording the spectrum of the blue-red LED incident on a Spectralon® white reference panel (Boulder, CO, USA; WR). Inset: zoom in to a portion of the fluorescence range from 670–800 nm. (c) Two sample spectra of the tomato plant within the fluorescence range (approximately 670 nm–780 nm with the spectrum of the WR measurement of the LED for reference. Figure 2 CF spectra, images and descriptions of selected days of experiment: (a) Day 0, (b) Day 3, (c) Day 4, (d) Day 8. Left column: Daily average smoothed fluorescence (Fsm), blue lines indicate Fsm spectra, shaded pink region indicates standard deviation for the day. Middle column: photographs of tomato plant each day. Right column: description of visual inspection of plant for each day. Daily fluorescence spectra, pictures and visual descriptions are included in Appendix A. Figure 3 Results of plant stress experiments. CF metrics, NDVI and PRI measurement time series for the experiment. (a) Daily mean, Fred and Ffar, (b) Daily mean Fred/Ffar, Inset plot of the first five days of the experiment highlighting the trend of increasing Fred/Ffar, (c) The normalized difference vegetation index (NDVI) taken one time daily for the experiment. (d) The photochemical reflectance index taken one time daily for the experiment (e) Fred, Ffar, NDVI, and PRI plotted as the proportion of maximum over the course of the experiment. (f) Pearson Correlations among the mean values. Only significant correlations (p-values < 0.05) are shown along the lower right half. Blank areas in the lower right half indicate insignificant correlations (p-values > 0.05). Error bars indicate the standard deviation for each day. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Campbell P.K.E. Huemmrich K.F. Middleton E.M. Ward L.A. Julitta T. Daughtry C.S.T. Burkart A. Russ A.L. Kustas W.P. Diurnal and Seasonal Variations in Chlorophyll Fluorescence Associated with Photosynthesis at Leaf and Canopy Scales Remote Sens. 2019 11 488 10.3390/rs11050488 2. Meroni M. Picchi V. Rossini M. Cogliati S. Panigada C. Nali C. Lorenzini G. Colombo R. Leaf level early assessment of ozone injuries by passive fluorescence and photochemical reflectance index Int. J. Remote Sens. 2008 29 5409 5422 10.1080/01431160802036292 3. Meroni M. Rossini M. Picchi V. Panigada C. Cogliati S. Nali C. Colombo R. Assessing Steady-state Fluorescence and PRI from Hyperspectral Proximal Sensing as Early Indicators of Plant Stress: The Case of Ozone Exposure Sensors 2008 8 1740 1754 10.3390/s8031740 27879790 4. Baker N.R. Chlorophyll fluorescence: A probe of photosynthesis in vivo Annu. Rev. Plant Biol. 2008 59 89 113 10.1146/annurev.arplant.59.032607.092759 18444897 5. Genty B. Briantais J.-M. Baker N.R. The relationship between the quantum yield of photosynthetic electron transport and quenching of chlorophyll fluorescence Biochim. Biophys. Acta 1989 990 87 92 10.1016/S0304-4165(89)80016-9 6. Gitelson A.A. Gritz Y. Merzlyak M.N. Relationships between leaf chlorophyll content and spectral reflectance and algorithms for non-destructive chlorophyll assessment in higher plant leaves J. Plant Physiol. 2003 160 271 282 10.1078/0176-1617-00887 12749084 7. Porcar-Castell A. Tyystjärvi E. Atherton J. van der Tol C. Flexas J. Pfündel E.E. Moreno J. Frankenberg C. Berry J.A. Linking chlorophyll a fluorescence to photosynthesis for remote sensing applications: Mechanisms and challenges J. Exp. Bot. 2014 65 4065 4095 10.1093/jxb/eru191 24868038 8. Hao D. Asrar G.R. Zeng Y. Yang X. Li X. Xiao J. Guan K. Wen J. Xiao Q. Berry J.A. Potential of hotspot solar-induced chlorophyll fluorescence for better tracking terrestrial photosynthesis Glob. Chang. Biol. 2021 27 2144 2158 10.1111/gcb.15554 33560585 9. Dechant B. Ryu Y. Badgley G. Zeng Y. Berry J.A. Zhang Y. Goulas Y. Li Z. Zhang Q. Kang M. Canopy structure explains the relationship between photosynthesis and sun-induced chlorophyll fluorescence in crops Remote Sens. Environ. 2020 241 111733 10.1016/j.rse.2020.111733 10. Maguire A.J. Eitel J.U.H. Griffin K.L. Magney T.S. Long R.A. Vierling L.A. Schmiege S.C. Jennewein J.S. Weygint W.A. Boelman N.T. On the Functional Relationship Between Fluorescence and Photochemical Yields in Complex Evergreen Needleleaf Canopies Geophys. Res. Lett. 2020 47 e2020GL087858 10.1029/2020GL087858 11. Li X. Xiao J. He B. Arain M.A. Beringer J. Desai A.R. Emmel C. Hollinger D.Y. Krasnova A. Mammarella I. Solar-induced chlorophyll fluorescence is strongly correlated with terrestrial photosynthesis for a wide variety of biomes: First global analysis based on OCO-2 and flux tower observations Glob. Chang. Biol. 2018 24 3990 4008 10.1111/gcb.14297 29733483 12. Magney T. Frankenberg C. Fisher J.B. Sun Y. North G.B. Davis T.S. Kornfeld A. Siebke K. Connecting active to passive fluorescence with photosynthesis: A method for evaluating remote sensing measurements of Chl fluorescence New Phytol. 2017 215 1594 1608 10.1111/nph.14662 28664542 13. Springer K.R. Wang R. Gamon J.A. Parallel Seasonal Patterns of Photosynthesis, Fluorescence, and Reflectance Indices in Boreal Trees Remote Sens. 2017 9 691 10.3390/rs9070691 14. Wagle P. Zhang Y. Jin C. Xiao X. Comparison of solar-induced chlorophyll fluorescence, light-use efficiency, and process-based GPP models in maize Ecol. Appl. 2016 26 1211 1222 10.1890/15-1434 27509759 15. Raji S.N. Aparna G.N. Mohanan C.N. Subhash N. Proximal Remote Sensing of Herbicide and Drought Stress in Field Grown Colocasia and Sweet Potato Plants by Sunlight-Induced Chlorophyll Fluorescence Imaging J. Indian Soc. Remote Sens. 2016 45 463 475 10.1007/s12524-016-0612-3 16. Cendrero-Mateo M.P. Moran M.S. Papuga S. Thorp K. Alonso L. Moreno J. Ponce-Campos G. Rascher U. Wang G. Plant chlorophyll fluorescence: Active and passive measurements at canopy and leaf scales with different nitrogen treatments J. Exp. Bot. 2015 67 275 286 10.1093/jxb/erv456 26482242 17. Julitta T. Optical proximal sensing for vegetation monitoring, in Department of Earth and Environmental Sciences Ph.D. Thesis University of Milano-Bicocca Milan, Italy 2015 136 18. Guanter L. Zhang Y. Jung M. Joiner J. Voigt M. Berry J.A. Frankenberg C. Huete A.R. Zarco-Tejada P. Lee J.-E. Global and time-resolved monitoring of crop photosynthesis with chlorophyll fluorescence Proc. Natl. Acad. Sci. USA 2014 111 E1327 E1333 10.1073/pnas.1320008111 24706867 19. Zarco-Tejada P. Catalina A. González M. Martín P. Relationships between net photosynthesis and steady-state chlorophyll fluorescence retrieved from airborne hyperspectral imagery Remote Sens. Environ. 2013 136 247 258 10.1016/j.rse.2013.05.011 20. Zarco-Tejada P. González-Dugo V. Berni J. A. Fluorescence, temperature and narrow-band indices acquired from a UAV platform for water stress detection using a micro-hyperspectral imager and a thermal camera Remote Sens. Environ. 2012 117 322 337 10.1016/j.rse.2011.10.007 21. Lichtenthaler H.K. Miehé J.A. Fluorescence imaging as a diagnostic tool for plant stress Trends Plant Sci. 1997 2 316 320 10.1016/S1360-1385(97)89954-2 22. Moya G.G. Goulas Y. Remotely sensed blue and red fluorescence emission for monitoring vegetation ISPRS J. Photogramm. Remote Sens. 1992 47 27 10.1016/0924-2716(92)90033-6 23. Krause G.H. Weis E. Chlorophyll fluorescence as a tool in plant physiology Photosynth. Res. 1984 5 139 157 10.1007/BF00028527 24458602 24. da Silva Junior C.A. Teodoro P.E. Delgado R.C. Teodoro L.P.R. Lima M. de Andréa Pantaleão A. Baio F.H.R. De Azevedo G.B. de Oliveira Sousa Azevedo G.T. Capristo-Silva G.F. Persistent fire foci in all biomes undermine the Paris Agreement in Brazil Sci. Rep. 2020 10 16246 10.1038/s41598-020-72571-w 33004818 25. Chen X. Mo X. Hu S. Liu S. Relationship between fluorescence yield and photochemical yield under water stress and intermediate light conditions J. Exp. Bot. 2018 70 301 313 10.1093/jxb/ery341 26. Flexas J. Escalona J.M. Evain S. Gulías J. Moya I. Osmond C.B. Medrano H. Steady-state chlorophyll fluorescence (Fs) measurements as a tool to follow variations of net CO2 assimilation and stomatal conductance during water-stress in C3 plants Physiol. Plant. 2002 114 231 240 10.1034/j.1399-3054.2002.1140209.x 11903970 27. Fabrice Franck P.J. Popovic R. Resolution of the Photosystem I and Photosystem II contributions to chlorophyll fluorescence of intact leaves at room temperature Biochim. Biophys. Acta 2002 1556 239 246 10.1016/S0005-2728(02)00366-3 12460682 28. Meroni M. Rossini M. Guanter L. Alonso L. Rascher U. Colombo R. Moreno J. Remote sensing of solar-induced chlorophyll fluorescence: Review of methods and applications Remote Sens. Environ. 2009 113 2037 2051 10.1016/j.rse.2009.05.003 29. Malenovský Z. Mishra K.B. Zemek F. Rascher U. Nedbal L. Scientific and technical challenges in remote sensing of plant canopy reflectance and fluorescence J. Exp. Bot. 2009 60 2987 3004 10.1093/jxb/erp156 19465688 30. Maxwell K. Johnson G.N. Chlorophyll Fluorescence-a practical guide J. Exp. Bot. 2000 51 659 668 10.1093/jexbot/51.345.659 10938857 31. Krause G.H. Weis E. Chlorophyll fluorescence and photosynthesis:The Basics Ann. Rev. Plant Physiol. Plant Mol. Biol. 1991 42 313 349 10.1146/annurev.pp.42.060191.001525 32. Schreiber U. Groberman L. Vidaver W. Portable, solid-state fluorometer for the measurement of chlorophyll fluorescence induction in plants Rev. Sci. Instrum. 1975 46 538 542 10.1063/1.1134252 33. Dąbrowski P. Baczewska-Dąbrowska A.H. Bussotti F. Pollastrini M. Piekut K. Kowalik W. Wróbel J. Kalaji H.M. Photosynthetic efficiency of Microcystis ssp. under salt stress Environ. Exp. Bot. 2021 186 104459 10.1016/j.envexpbot.2021.104459 34. Pogrzeba M. Rusinowski S. Sitko K. Krzyżak J. Skalska A. Małkowski E. Ciszek D. Werle S. Mccalmont J. Mos M. Relationships between soil parameters and physiological status of Miscanthus x giganteus cultivated on soil contaminated with trace elements under NPK fertilisation vs. microbial inoculation Environ. Pollut. 2017 225 163 174 10.1016/j.envpol.2017.03.058 28365513 35. Kalaji H.M. Schansker G. Ladle R.J. Goltsev V. Bosa K. Allakhverdiev S.I. Brestic M. Bussotti F. Calatayud A. Dąbrowski P. Frequently asked questions about in vivo chlorophyll fluorescence: Practical issues Photosynth. Res. 2014 122 121 158 10.1007/s11120-014-0024-6 25119687 36. Damm A. Elbers J. Erler A. Gioli B. Hamdi K. Hutjes R. Kosvancova M. Meroni M. Miglietta F. Moersch A. Remote sensing of sun-induced fluorescence to improve modeling of diurnal courses of gross primary production (GPP) Glob. Chang. Biol. 2010 16 171 186 10.1111/j.1365-2486.2009.01908.x 37. Guanter L. Alonso L. Gómez-Chova L. Amorós-López J. Vila J. Moreno J. Estimation of solar-induced vegetation fluorescence from space measurements Geophys. Res. Lett. 2007 34 10.1029/2007GL029289 38. Govindjee Chlorophyll a fluorescence: A bit of basics and history Chlorophyll a Fluorescence: A Signature of Photosynthesis Papageorgiou G.C. Springer Dordrecht, The Netherlands 2004 1 42 39. Zarco-Tejada P.J. Miller J.R. Mohammed G.H. Noland T.L. Sampson P.H. Estimation of chlorophyll fluorescence under natural illumination from hyperspectral data Int. J. Appl. Earth Obs. Geoinf. 2001 3 321 327 10.1016/S0303-2434(01)85039-X 40. Stefan Maier K.P.G. Sun-Induced Fluorescence—A New Tool for Precision Farming John Wiley & Sons, Inc. Hoboken, NJ, USA 2001 41. Govindjee Sixty-three years since Kautsky: Chlorophyll a Fluorescence Aust. J. Plant Physiol. 1995 22 131 160 42. Alonso L. Gómez-Chova L. Vila-Frances J. Amoros-Lopez J. Guanter L. Calpe J. Moreno J. Improved Fraunhofer Line Discrimination Method for Vegetation Fluorescence Quantification IEEE Geosci. Remote Sens. Lett. 2008 5 620 624 10.1109/LGRS.2008.2001180 43. Plascyk J.A. The MK II Fraunhofer Line Discriminator (FLD-II) for Airborne and Orbital Remote Sensing of Solar-Stimulated Luminescense Opt. Eng. 1975 14 144339 10.1117/12.7971842 44. Plascyk J.A. Gabriel F.C. The Fraunhofer Line Discriminator MKII-An Airborne Instrument for Precise and Standardized Ecological Luminescence Measurement IEEE Trans. Instrum. Meas. 1975 24 306 313 10.1109/TIM.1975.4314448 45. Liu X. Liu L. Zhang S. Zhou X. New Spectral Fitting Method for Full-Spectrum Solar-Induced Chlorophyll Fluorescence Retrieval Based on Principal Components Analysis Remote Sens. 2015 7 10626 10645 10.3390/rs70810626 46. Meroni M. Busetto L. Colombo R. Guanter L. Moreno J. Verhoef W. Performance of Spectral Fitting Methods for vegetation fluorescence quantification Remote Sens. Environ. 2010 114 363 374 10.1016/j.rse.2009.09.010 47. van Leeuwen M. Kremens R.L. van Aardt J. Tracking diurnal variation in photosynthetic down-regulation using low cost spectroscopic instrumentation Sensors 2015 15 10616 10630 10.3390/s150510616 25951342 48. Zhang Y.-J. Zhao C.-J. Liu L.-Y. Wang J.-H. Wang R.-C. Chlorophyll Fluorescence Detected Passively by Difference Reflectance Spectra of Wheat (Triticum aestivum L.) Leaf J. Integr. Plant Biol. 2005 47 1228 1235 10.1111/j.1744-7909.2005.00154.x 49. Evans J. The Dependence of Quantum Yield on Wavelength and Growth Irradiance Funct. Plant Biol. 1987 14 69 79 10.1071/PP9870069 50. Van Wittenberghe S. Alonso L. Verrelst J. Hermans I. Valcke R. Veroustraete F. Moreno J. Samson R. A field study on solar-induced chlorophyll fluorescence and pigment parameters along a vertical canopy gradient of four tree species in an urban environment Sci. Total Environ. 2014 466–467 185 194 10.1016/j.scitotenv.2013.07.024 51. Rossini M. Nedbal L. Guanter L. Ač A. Alonso L. Burkart A. Cogliati S. Colombo R. Damm A. Drusch M. Red and far red Sun-induced chlorophyll fluorescence as a measure of plant photosynthesis Geophys. Res. Lett. 2015 42 1632 1639 10.1002/2014GL062943 52. Buschmann C. Variability and application of the chlorophyll fluorescence emission ratio red/far-red of leaves Photosynth. Res. 2007 92 261 271 10.1007/s11120-007-9187-8 17525834 53. Campbell P.E. Middleton E. Corp L. Kim M. Contribution of chlorophyll fluorescence to the apparent vegetation reflectance Sci. Total Environ. 2008 404 433 439 10.1016/j.scitotenv.2007.11.004 18164750 54. Tucker C.J. Red and photographic infrared linear combinations for monitoring vegetation Remote Sens. Environ. 1979 8 127 150 10.1016/0034-4257(79)90013-0 55. Gamon J.A. Peñuelas J. Field C.B. A narrow-waveband spectral index that tracks diurnal changes in photosynthetic efficiency Remote Sens. Environ. 1992 41 35 44 10.1016/0034-4257(92)90059-S 56. Gitelson A.A. Gamon J.A. Solovchenko A. Multiple drivers of seasonal change in PRI: Implications for photosynthesis 2. Stand level Remote Sens. Environ. 2017 190 198 206 10.1016/j.rse.2016.12.015 57. Gamon J.A. Huemmrich K.F. Wong C.Y.S. Ensminger I. Garrity S. Hollinger D.Y. Noormets A. Peñuelas J. A remotely sensed pigment index reveals photosynthetic phenology in evergreen conifers Proc. Natl. Acad. Sci. USA 2016 113 13087 13092 10.1073/pnas.1606162113 27803333 58. Zarco-Tejada P.J. Gonzalez-Dugo V. Williams L.E. Suárez L. Berni J.A.J. Goldhamer D. Fereres E. A PRI-based water stress index combining structural and chlorophyll effects: Assessment using diurnal narrow-band airborne imagery and the CWSI thermal index Remote Sens. Environ. 2013 138 38 50 10.1016/j.rse.2013.07.024 59. Zarco-Tejada P. Morales A. Testi L. Villalobos F. Spatio-temporal patterns of chlorophyll fluorescence and physiological and structural indices acquired from hyperspectral imagery as compared with carbon fluxes measured with eddy covariance Remote Sens. Environ. 2013 133 102 115 10.1016/j.rse.2013.02.003 60. Yves Goulas A.F. Daumard F. Champagne S. Ounis A. Marloie O. Moya I. Gross Primary Production of a Wheat Canopy Relates Stronger to Far Red Than to Red Solar-Induced Chlorophyll Fluorescence Remote Sens. 2017 9 97 10.3390/rs9010097 61. Wieneke S. Ahrends H. Damm A. Pinto F. Stadler A. Rossini M. Rascher U. Airborne based spectroscopy of red and far-red sun-induced chlorophyll fluorescence: Implications for improved estimates of gross primary productivity Remote Sens. Environ. 2016 184 654 667 10.1016/j.rse.2016.07.025 62. Rossini M. Meroni M. Celesti M. Cogliati S. Julitta T. Panigada C. Rascher U. Van Der Tol C. Colombo R. Analysis of Red and Far-Red Sun-Induced Chlorophyll Fluorescence and Their Ratio in Different Canopies Based on Observed and Modeled Data Remote Sens. 2016 8 412 10.3390/rs8050412 63. Zarco-Tejada P.J. Miller J.R. Mohammed G.H. Noland T.L. Sampson P.H. Vegetation stress detection through chlorophyll a + b estimation and fluorescence effects on hyperspectral imagery J. Environ. Qual. 2002 31 1433 1441 10.2134/jeq2002.1433 12371159 64. Zarco-Tejada P.J. Miller J.R. Mohammed G.H. Noland T.L. Sampson P.H. Chlorophyll Fluorescence Effects on Vegetation Apparent Reflectance: I. Leaf-Level Measurements and Model Simulation Remote Sens. Environ. 2000 74 596 608 10.1016/S0034-4257(00)00149-8 65. Meroni M. Colombo R. Leaf level detection of solar induced chlorophyll fluorescence by means of a subnanometer resolution spectroradiometer Remote Sens. Environ. 2006 103 438 448 10.1016/j.rse.2006.03.016 66. Rossini M. Panigada C. Cilia C. Meroni M. Busetto L. Cogliati S. Amaducci S. Colombo R. Discriminating Irrigated and Rainfed Maize with Diurnal Fluorescence and Canopy Temperature Airborne Maps ISPRS Int. J. Geo-Inf. 2015 4 626 646 10.3390/ijgi4020626 67. Palombi L. Cecchi G. Lognoli D. Raimondi V. Toci G. Agati G. A retrieval algorithm to evaluate the Photosystem I and Photosystem II spectral contributions to leaf chlorophyll fluorescence at physiological temperatures Photosynth. Res. 2011 108 225 239 10.1007/s11120-011-9678-5 21866392 68. Grossmann K. Frankenberg C. Magney T. Hurlock S.C. Seibt U. Stutz J. PhotoSpec: A new instrument to measure spatially distributed red and far-red Solar-Induced Chlorophyll Fluorescence Remote Sens. Environ. 2018 216 311 327 10.1016/j.rse.2018.07.002 69. McMurtrey J. Middleton E. Corp L. Campbell P. Butcher L. Daughtry C. Optical reflectance and fluorescence for detecting nitrogen needs in Zea mays L. Proceedings of the 2003 IEEE International Geoscience and Remote Sensing Symposium Toulouse, France 21–25 July 2003 70. Middleton E.M. Chappelle E.W. Cannon T.A. Adamse P. Britz S.J. Initial Assessment of Physiological Response to UV-B Irradiation Using Fluorescence Measurements J. Plant Physiol. 1996 148 69 77 10.1016/S0176-1617(96)80296-3 71. Atherton J. Liu W. Porcar-Castell A. Nocturnal Light Emitting Diode Induced Fluorescence (LEDIF): A new technique to measure the chlorophyll a fluorescence emission spectral distribution of plant canopies in situ Remote Sens. Environ. 2019 231 111137 10.1016/j.rse.2019.03.030 72. Romero J.M. Otero A. Lagorio M.G. Berger A.G. Cordon G.B. Canopy active fluorescence spectrum tracks ANPP changes upon irrigation treatments in soybean crop Remote Sens. Environ. 2021 263 112525 10.1016/j.rse.2021.112525

PMC009xxxxxx/PMC9099695.txt

==== Front Materials (Basel) Materials (Basel) materials Materials 1996-1944 MDPI 10.3390/ma15093155 materials-15-03155 Article Facile Preparation of a Novel Vanillin-Containing DOPO Derivate as a Flame Retardant for Epoxy Resins Chen Liping Luo Zhonglin https://orcid.org/0000-0003-4897-3920 Wang Biaobing * Laoutid Fouad Academic Editor Jiangsu Key Laboratory of Environmentally Friendly Polymeric Materials, Jiangsu Collaborative Innovation Center of Photovoltaic Science and Engineering, School of Materials Science and Engineering, Changzhou University, Changzhou 213164, China; chenliping526@163.com (L.C.); zhonglinluo@cczu.edu.cn (Z.L.) * Correspondence: biaobing@cczu.edu.cn; Tel./Fax: +86-0519-8633-0075 27 4 2022 5 2022 15 9 315512 3 2022 23 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). A novel bio-based flame retardant designated AVD has been synthesized in a one-pot process via the reaction of 9,10-dihydro-9-oxa-10-phospha-phenanthrene-10-oxide (DOPO), vanillin (VN), and 2- aminobenzothiazole (ABT). The structure of AVD was confirmed using Fourier transform infrared spectroscopy (FTIR), and 1H and 31P nuclear magnetic resonance spectroscopy (NMR). The curing process, thermal stability, flame retardancy, and mechanical properties of the epoxy resin (EP) modified with AVD have been investigated comprehensively. The extent of curing, the glass transition temperature and the crosslinking density of the blend decreased gradually with increasing AVD content. The thermogravimetric analysis (TGA) was used to demonstrate that the presence of AVD reduced the thermal decomposition rate for EP and enhanced the formation of carbon residue during resin decomposition. A blend of 7.5 wt% AVD (0.52% phosphorus) displays a UL-94V-0 rating and a LOI of 31.1%. Reduction of the peak heat release rate, total heat release rate and total smoke production was 41.26%, 35.70%, and 24.03%, respectively, as compared to the values for pure EP. The improved flame retardancy of the flame retardant epoxy (FREP) may be attributed to the formation of a compact and continuous protective char layer into the condensed phase as well as the release of non-combustible gases and phosphorus-containing radicals from the decomposition of AVD in the gas phase. AVD is a new and efficient biobased flame retardant for epoxy with great prospects for industrial applications. epoxy resin bio-based flame retardancy lower phosphorus content mechanism Postgraduate Research & Practice Innovation Program of Jiangsu ProvinceKYCX21-2808 This work was supported financially by Postgraduate Research & Practice Innovation Program of Jiangsu Province (KYCX21-2808). ==== Body pmc1. Introduction Epoxy resin (EP), as an important thermosetting resin which displays characteristics of chemical resistance, low curing shrinkage, outstanding adhesion, and great electrical insulation. It has been widely applied in aerospace, coatings, adhesives, and microelectronics [1,2,3]. However, intrinsic flammability restricts its applications in many high-tech areas. It is accordingly urgent to improve the flame retardancy of EP [4,5]. Traditional halogen-containing flame retardants (FRs) produce toxic gases such as dioxins during combustion, which are hazardous to human health and also have a negative effect on the environment [6]. Consequently, halogen-containing FRs have been forbidden in many applications, and the development of halogen-free FRs has become imperative. Recently, 9,10-dihydro-9-oxa-10-phospha-phenanthrene-10-oxide (DOPO) has become a promising phosphorus-based FR due to its non-toxic properties and high phosphorus content. Nevertheless, a satisfactory flame retardant effect is achieved only at high DOPO content in epoxy resin [7,8,9]. The presence of DOPO at these levels caused a deterioration of the mechanical properties of the resin. Fortunately, the highly reactive P-H bond of DOPO permits the introduction of other flame-retardant elements into its molecular structure, including nitrogen [10,11], sulfur [12,13], silicon [14,15], boron [16,17], etc. These DOPO-based derivatives display better flame retardant efficiency for EP due to cooperative effect of multiple flame retardant elements. To reduce dependence on petroleum resources and to reduce toxicity, new biobased flame retardants are being rapidly developed [18]. Some biobased materials have received much attention for the preparation of epoxy resin flame retardants, such as vanillin [19], isosorbide [20], tartaric acid [21], chitosan [22], furans [23], crop-based phenolics [24], glycerol/adipic acid hyperbranched poly (ester)s, [25] etc. All of the above biobased raw materials have great potential for application in the development of biobased phosphorus flame retardants. Daniel et al. [26] converted renewable isosorbide into the corresponding diacrylate and used it to synthesize four phosphorus-containing compounds that showed good flame retardant properties in epoxy resins. Two of them are stable at temperatures close to 400 °C and may be suitable flame retardants for polymers processed at high temperatures. Among them, vanillin is a promising starting material for the preparation of flame retardants due to the high reactivity of its aldehyde and the presence of phenolic hydroxyl group. A biobased reactive FR (VDG) has been synthesized in a one-pot reaction involving DOPO, vanillin, and 3,5-diamino-1,2,4-triazole flame retardant (VDG). A high LOI value of 37.0% and UL-94 V-0 rating were observed for the cured EP system containing 2.0 wt% VDG. This material also exhibits antibacterial effects toward E. coli. [27]. It has previously been demonstrated that the N/S-containing thiazole has great potential for the construction of efficient FRs [28,29,30]. A novel bio-based FR (marked as AVD) has synthesized from the reaction of DOPO, vanillin, and 2-aminobenzothiazole. The chemical structure of AVD was characterized using Fourier transform infrared spectroscopy (FTIR), and 1H and 31P nuclear magnetic resonance spectroscopy (NMR). EP/AVD thermosets with different AVD loading levels were produced and their curing behaviors, thermal stability, flammability and combustion behaviors were evaluated. Moreover, a flame retardant mode of action for AVD in EP has been proposed. 2. Experimental 2.1. Materials Epoxy resin with the epoxy value of 0.51 mol/100 g (commercial name: E-51) was purchased from CNOOC Changzhou Coating Chemical Research Institute Co., Ltd. (Changzhou, China). DOPO, vanillin (VN, 99%), 4,4′-Diaminodiphenyl methane (DDM, 98%) were acquired from Aladdin Reagents Co., Ltd. (Shanghai, China). 2-Aminobenzothiazole was purchased from Jiangsu Qiangsheng Functional Chemical Co., Ltd (Nanjing, China). Absolute ethanol was supplied by Sinopharm Chemical Reagent Co., Ltd (Shanghai, China). All raw materials were not purified and were used directly. 2.2. Synthesis of AVD The AVD was one-pot synthesized, and the synthetic route is illustrated in Scheme 1. To a 250 mL three-necked flask equipped with magnetic stirrer and reflux condenser, vanillin (VN) (0.06 mol, 9.129 g), 2-aminobenzothiazole (ABT) (0.06 mol, 6.0084 g) and anhydrous ethanol (100 mL) were added. After reacting at 80 °C for 5 h, DOPO (0.06 mol, 9.129 g) was added and stirred continuously for another 12 h. The crude product was collected by filtration, washed three times with anhydrous ethanol, and then the product was dried to constant weight in a vacuum oven at 70 °C. The pale-yellow powder was obtained (Yield: 76%, melting temperature: 205 °C). 2.3. Preparation of EP and Flame Retardant EPs (FREPs) The mole ratio of the amino group to epoxy group was 1:1 for all samples, and the formulations were shown in Table 1. Firstly, a transparent EP/AVD solution was obtained under magnetic stirring at 130 °C, and then cooled to 90 °C. Afterwards, DDM was introduced and kept stirring until it was completely dissolved. The mixture was then dumped into the preheated silicone rubber mold and cured at 100 °C for 2 h and 150 °C for 3 h. The Pure EP/DDM thermoset was prepared with the above-mentioned procedure. 2.4. Characterization FTIR spectra were obtained using a Perkin Elmer instrument (Waltham, MA, USA) over a spectral range of 4000–400 cm−1. All samples were milled with KBr and pressed into tablets. 1H NMR and 31P NMR spectra were collected on a Bruker Advance III-500 NMR spectrometer (Bruker, Waltham, MA, USA) using the DMSO-d6 as deuterated solvent. Differential scanning calorimetry (DSC) analysis was carried out on a Perkin-Elmer DSC 8000 (PE, Waltham, MA, USA) at different heating rate from 30 °C to 250 °C under N2 atmosphere. The weight of all samples was fixed at about 5 mg. Thermogravimetric analysis (TGA) was performed on a Perkin-Elmer TGA 4000 (Waltham, MA, USA) with a nitrogen flow rate of 20 mL·min−1. The specimen (about 8 mg) was placed in an alumina crucible and heated from 30 to 700 °C at a heating rate of 10 °C·min−1. The vertical burning (UL-94) test was measured by a CZF-3 instrument (Shine Ray Instrument Co. Ltd., Nanjing, China) according to ASTM D3801 standard. The limited oxygen index (LOI) value was measured using an HC-2 oxygen index meter (Jiang Ning Co. Ltd., Nanjing, China) according to ASTM D2863. The combustion behaviors were tested on a FTT cone calorimeter (Fire Testing Technology, East Grinstead, UK) according to the ISO 5660-1 standard at an external heat flux of 35 kW·m−2. Three replicates with dimensions of 100 × 100 × 3 mm3 and weight of about 36.5 g were tested, and their average values were collected as each point data. The microscopic morphologies of residual char were observed by a SUPRA55 scanning electron microscope (SEM) with an acceleration voltage of 5 kV. X-ray photoelectron spectroscopy (XPS) was determined by an ESCALAB 250Xi system (Thermo Fischer Scientific, Waltham, MA, USA), using Al Kα excitation radiation (hν = 1486.6 eV). Raman spectroscopy was collected through a DXR2xi laser Raman spectrometer (LRs) (Thermo Fischer Scientific, Waltham, MA, USA) in the range of 500–3000 cm−1 with an excitation wavelength of 532 nm. TG-IR spectroscopy was conducted using a combination system of a TGA 4000 thermogravimetric analyzer and a Spectrum II FTIR spectrophotometer. The sample (around 20 mg) was heated from 30 to 700 °C at 10 °C·min−1 with a nitrogen flow rate of 20 mL/min. A dynamic mechanical analysis (DMA) was carried out on a Perkin-Elmer DMA8000 (PE, Waltham, MA, USA). A three-point bending mold with an amplitude of 20 μm and a frequency of 1 Hz was selected. The experimental temperature interval was 30–260 °C at a rate of 10 °C·min−1. (Dimensions of all samples: 40 × 6 × 3 mm3). 3. Results and Discussion 3.1. Characterization of AVD The FTIR spectra of the target product (AVD) and raw materials (VN, ABT, and DOPO) are presented in Figure 1. With respect to the FTIR spectrum of AVD, the absorption peak at 3413 cm−1 is assigned to the stretching vibration of -OH in VN, and the disappearance of the characteristic absorption peak of -CHO observed at 1667 cm−1 indicates a complete reaction between VN and ABT [28]. Meanwhile, the typical P-H stretching vibration absorption peak of DOPO at 2437 cm−1 [29] disappears in the spectrum of AVD, and the double peaks (3396 cm−1 and 3272 cm−1) of -NH2 in ABT shifts to a single peak (3227 cm−1) of -NH in AVD [31]; the peak at 1377 cm−1 is ascribed to the stretching vibration of C-N [29]. All of these phenomena confirm that the addition reaction between the Schiff-base intermediate and DOPO proceeds successfully. Additionally, the absorption peaks at 1598 cm−1, 1449 cm−1, 1238 cm−1, and 1210 cm−1 are attributed to the benzene ring, C=N in the thiazole ring, P=O, and P-O-C stretching vibrations, respectively [28]. The above results confirm the initial chemical structure of AVD. Both 1H NMR and 31P NMR spectra were performed to further check the chemical structure of AVD. As shown in the 1H NMR spectrum of AVD (Figure 2a), the chemical shifts at 5.66 and 5.80 ppm are attributed to the hydrogen atom on the chiral carbon attached to the DOPO group [31]. The chemical shifts at 6.66–6.75 ppm, 6.85–8.18 ppm, and 9.05 ppm are assigned to N-H, p.roton hydrogen on the benzene ring (Ar-H), and the signal of -OH, respectively. The integral area ratio of the different proton chemical environments is in agreement with the theoretical values. Furthermore, the 31P NMR spectrum of AVD (Figure 2b) presents two signal peaks at 28.78 and 30.25 ppm. It indicates that the P element in AVD is in two different chemical environments, which might be due to the presence of the chiral carbon atom. Based on the above analysis, it is concluded that the target product AVD was synthesized successfully. 3.2. Curing Behaviors To investigate the effect of incorporation of AVD on the curing process of epoxy resin, the non-isothermal curing kinetics of the epoxy systems at different heating rates were performed by DSC, and the resultant DSC curves are shown in Figure 3a–d. As can be seen, the TP values of all samples shift toward higher temperatures as the heating rate increases. Moreover, the TP values gradually become greater with increasing AVD content at the same heating rate. This effect is mainly due to the steric hindrance of the rigid groups such as DOPO and benzothiazole in the AVD structure reduces the reactivity of ring-opening curing of epoxy resin. The apparent activation energy (Ea) of the epoxy systems are further calculated according to the Kissinger’s (Equation (1)) and Ozawa’s methods (Equation (2)) [32] and the fitted curves of ln(β/TP2) and lnβ versus 1/TP×103 are illustrated in Figure 3e,f, and the results are summarized in Table 2. (1) ln(β/TP2)=ln(AR/Ea)−Ea/RTP (2) lnβ=ln(AEa/R)−1.052Ea/RTP−5.331 wherein β is the heating rate, TP is the curing peak temperature, A is the pre-exponential factor and R is the ideal gas constant (8.314 J K−1 mol−1). All of the Ea values calculated from both Kissinger’s and Ozawa’s methods increase with the increasing of AVD content. It suggests that the addition of AVD enhances the energy barrier of the curing reaction, which further verifies the presence of the steric hindrance. 3.3. Thermal Stability The thermal stability of the pure EP and FREPs under nitrogen atmosphere was evaluated by TGA. The corresponding TG and DTG curves are depicted in Figure 4, and the related data are summarized in Table 3. Obviously, the AVD gives lower T5% (272.8 °C) and greater CR700 (30.1%) than the pure EP. With the addition of AVD, the T5% and Tmax of the cured FREPs decrease with the increase of AVD content, which suggests that the AVD promotes the decomposition of the epoxy matrix on advance. However, the Rmax decreases from 18.8%·min−1 for pure EP to 10.7%·min−1 for FREP-10, indicating that the presence of AVD delays the decomposition of the EP at a higher temperature. Meanwhile, the FREP-10 sample gives a CR700 of 24.5%, which is greater than that of the pure EP (19.7%). It implies that the incorporation of AVD improves the carbon formation ability of the cured FREPs. 3.4. Flame Retardancy of EP and FREPs LOI and UL-94 measurements were carried out to assess the flame retardancy, and the resultant LOI values and UL-94 rating are listed in Table 4. Apparently, the pure EP is a combustible polymer with the LOI value of 25% and fails to pass the UL-94 vertical burn testing. With the incorporation of 5 wt% AVD, the FREP-5 sample presents an LOI value of 30% and UL-94 V-1 testing. Furthermore, the FRSP containing 7.5 wt% AVD (the P content is 0.52 wt%) achieves the LOI value up to 31.3% and UL-94 V-0 rating. The results reveal that the AVD has high flame retardant efficiency for epoxy resin. 3.5. Analysis of Fire Behaviors The cone calorimetry test (CCT) is one of the most effective methods used in the laboratory to evaluate the combustion behavior of materials [33,34]. This method can provide a series of important parameters about material flammability, including time to ignition (TTI), peak heat release rate (PHRR), total heat release (THR), total smoke release (TSP), fire growth rate index (FIGRA), average effective combustion heat burn (av-EHC), average CO yield (av-COY), average CO2 yield (av-CO2Y) and char residue after combustion, all of which are summarized in Table 5. Figure 5 depicts some important curves, which include the heat release rate (HRR), total heat release (THR), total smoke release (TSP), smoke release rate (SPR), CO2 production rate, and residue mass over time. As compared with the pure EP, the FREPs have greater TTI values which tend to increase with the increasing AVD content. This can be ascribed to the fact that the presence of AVD advances the earlier decomposition of the EP matrix. This result is consistent with that of the TGA test. The heat release rate (HRR) is one of the key parameters to assess the burning intensity. As shown in Table 5, the PHRR and THR values of the pure EPare 1452.5 kW·m−2 and 67.3 MJ·m−2, respectively. The combustion intensity decreases significantly with increasing AVD content. For instance, the PHRR and THR values are 657.6 kW·m−2 and 57.4 MJ·m−2 for the FREP-10 sample, which are reduced by 54.7% and 14.7% with comparison to the pure EP, respectively. This demonstrates that the incorporation of AVD can effectively suppress the combustion intensity of epoxy composites. In addition, the fire growth rate index (FIGRA) was commonly used to assess the rate of fire growth during combustion and calculated based on HRR curves according to Equation (3) [35]. (3) FIGRA=PHRR/TPHRR The lower FIGRA value of the material indicates the higher fire safety performance. The results in Table 5 display that the FIGRA value decreases significantly after the incorporation of AVD, from 11.2 kW·m−2·s−1 for pure EP to 3.44 kW·m−2·s−1 for FREP-10, with a reduction of 69.2%. Therefore, it can be concluded that the FREPs containing AVD have excellent fire safety. It is well known that smoke is the cause of death for the majority of victims who lose their life due to respiratory injuries in fires. Therefore, the smoke suppression performance of flame retardant materials is a critical parameter. Obviously, the TSP value of FREP-10 (20.4 m2) is reduced by15.4% as compared that of the pristine EP (24.1 m2), illustrating that the AVD displays good smoke suppression on the epoxy resin. Furthermore, the average effective heat of combustion (av-EHC, HRR/MLR) is an essential parameter to measure the degree of combustion of volatile substances in the gas phase. Seen from Table 5, the av-EHC value [36] of the FREPs is decreased gradually with the increasing AVD loading level, indicating that AVD has a good gas-phase flame retardant effect. It is also shown in Table 5 that the FREPs display reduced av-CO2Y values and increased av-COY values. This is attributed to the occurrence of incomplete combustion, further verifying the gas-phase flame retardant effect of AVD. Moreover, much more carbon residuals were left for FREPs, confirming that the incorporation of AVD promotes the carbonization of the EP. This might be attributable to the fact that the decomposition of the compounds containing DOPO can generate polyphosphates which catalyze the dehydration and esterification of the EP matrix, thus facilitating the formation of a carbonaceous protective layer. 3.6. Morphology of Char Residues Figure 6 displays the digital photographs and SEM images of char residues after CCT. As shown in the images, the pure EP was burned almost completely and left a few broken and loose char residues (Figure 6a2). However, with the incorporation of AVD, much more and continuous char residues (Figure 6b2–d2) with greater expansion height (Figure 6b1–d1) were obtained after CCT. It is further evident from the SEM images that the pure EP exhibits a thin and friable char layer with many cracks on the surface and inside (Figure 6a3), which completely fails to protect the underlying substrate. Conversely, compact and continuous char layers (Figure 6b3–d3), which effectively isolate the underlying substrate from heat and oxygen, are observed for FREPs. This is mainly due to the fact that the phosphoric acid from the decomposition of AVD catalyzes the dehydration and carbonization of the EP substrate, and the benzothiazole group with better thermal stability also contributes to the production of char residues. 3.7. Chemical Component of Residual Char XPS was applied to analyze the compositional changes of the char residues of pure EP and FREPs. The XPS spectra with possible peak positions are presented in Figure 7, and the corresponding elemental contents are listed in Table 6. Compared with the neat EP, the FREP-7.5 sample displays a greater ratio of C/O and N/O. It means that the char layer of FREP-7.5 is rich in nitrogen heterocycles and aromatic compounds. Moreover, a low oxygen content is also found for FREP-7.5. The decrease of oxygen content is mainly due to the formation of PO∙ and PO2∙, which volatilize into the gas phase to play a flame retardant role. Furthermore, the presence of P and S elements in the char residues indicates that they can act as the flame retardant mode of action in the condensed phase. Figure 8 shows the C1s, N1s, O1s, and P2p spectra of the char residue of FREP-7.5. In Figure 8a, the C1s spectrum is decomposed to three bands at 284.3 eV (aliphatic and aromatic C-H and C-C), 286.0 eV (C-O-C and P-O-C) and 288.3 eV (carbonyl) [37]. In the N1s spectrum (Figure 8b), the bands at 397.7 eV and 399.4 eV are ascribed to C-N or P-N and N-H on the amine group, respectively [38]. For the O1s spectrum (Figure 8c), the peaks at 531.5 eV and 532.8 eV are attributed to C=O/P=O and -O- in the C-O-P group. For the P2p spectrum (Figure 8d), the P2p peak is split into two peaks at 132.4 eV and 133.4 eV, which are assigned to the P-O-C group in phosphate and P=O, respectively [39]. The above results suggest that the AVD can decompose to produce phosphate-containing compounds which promotine the dehydration and esterification of the EP substrate to form a high-quality protective layer. 3.8. Raman Characterization of the Char Residues Raman spectroscopy can be applied to characterize the degree of graphitization of carbon materials. Figure 9 plots the Raman spectra of the char residue of pure EP and FREPs. The spectra of all samples have two peaks belonging to the D-band (around 1360 cm−1) and the G-band (around 1600 cm−1), which represent disordered carbon and graphitized carbon, respectively [40]. The value of AD/AG (area ratio of D-band to G-band) can reflect the graphitization degree of the charcoal residue, and its lower value means the higher graphitization degree of the corresponding char layer. The AD/AG values of the char residues from pure EP, FREP-5, FREP-7.5, and FREP-10 were determined to be 2.57, 2.50, 2.30, and 2.11, respectively. It can be seen that the values of FREPs are all lower than that of the pure EP, and the lowest value is achieved for FREP-10, indicating that the addition of AVD enhanced the graphitization of the char layer, which facilitates the formation of a denser and continuous char layer that acts as a barrier to inhibit the further degradation of the substrate. 3.9. Analysis of Gaseous Products of Pyrolysis of EP Composites A TG-FTIR test was adopted to excavate the gas-phase volatiles generated during the pyrolysis of EP composites. Figure 10 demonstrated the characteristic spectra and 3D TG-FTIR spectra of the gas-phase pyrolysis products of the neat EP and FREP-7.5 at different temperatures. As can be seen, the pyrolysis product of FREP-7.5 appears earlier (349 °C) than that of pure EP (382 °C), suggesting that the earlier decomposition of the EP matrix is advanced by the introduction of AVD. Despite all this, the pyrolysis products of pure EP and flame retardant EP are almost the same at higher temperatures, including 3675 cm−1 (H2O), 2850 cm−1–3100 cm−1 (aliphatic C-H), 1337 cm−1 (C-N), 1252 cm−1 (C-O of bisphenol A), 1176 cm−1 (aliphatic C-O), 747 cm−1 (benzene C-H) [39]. With respective to FTIR spectra of FREP-7.5; nevertheless, some other bands occur at 1602 cm−1 (P-O-Ph), 1332 cm−1 (SO2), 1257 cm−1 (P=O), 1043 cm−1 (P-O-C), and 966 cm−1 (NH3). The variation of the peak intensities of these compounds reveals the flame retardant effect of AVD in the gas phase. In order to confirm the flame retardant effect of AVD in the gas phase, Figure 11 depicts the variation of the spectra absorbance of combustible volatiles (hydrocarbons, aromatic compounds, carbonyl compounds, and aliphatic ethers) with time. Obviously, the intensities of the corresponding peaks are reduced with the introduction of AVD. Since combustible volatiles provide a large amount of fuel for combustion [41,42], the significant reduction of their intensities is further evidence of the radical scavenging effect of AVD decomposition products. Simultaneously, the captured aromatic compounds can be used as a carbon source, thus improving the char yield of the FREPs. 3.10. Potential Flame-Retardant Mode of Action The results of the above tests indicated that AVD exerts a good flame retardant effect in both the gas phase and the condensed phase. Hence, the possible flame retardant mode of action of AVD is proposed as shown in Figure 12. AVD acts as a flame retardant in the gas phase by releasing non-combustible gases such as NH3, SO2, CO2, and PO∙, PO2∙. Non-combustible gases dilute the concentration of gases supporting combustion such as oxygen and carry away heat; phosphorus-containing radicals such as PO∙ and PO2∙ capture high-activity radicals (H∙ and OH∙) in the combustion area to break off the free radical chain reaction of combustion [43], and thus prevent further combustion of the matrix. Simultaneously, the decomposition of the AVD can produce polyphosphate, pyrophosphoric acid or metaphosphoric acid in the condensed phase, which can undergo an esterification reaction with the EP substrate. Dehydration and carbonization form dense char layers with a P-O-C structure. This continuous and dense char layers can block the transfer of heat and protect the EP matrix. 3.11. Mechanical Properties DMA was employed to study the dynamic thermomechanical behavior of epoxy resins. Figure 13 illustrates the curves of storage modulus (E′) and tan δ with temperature for pure EP and FREPs, and the obtained results from DMA are listed in Table 7. The storage modulus at 50 °C of the FREPs are higher than that of the pure EP and increase with the increasing AVD contents. This is mainly due to the presence of rigid DOPO and benzothiazole groups of the AVD. At temperatures above Tg, however, the AVD contents have converse influence on the storage modulus, which might be attributed to a lower crosslink density (υe) of the FREPs. As can be seen in Figure 13b, the occurrence of a single peak indicates good compatibility between the AVD and EP matrix. Moreover, the Tg values of the FREPs are decreased with the increasing AVD contents. This is attributed to the predominance of the crosslink density over the rigid groups. The crosslink density (υe) of the cured EP can be calculated from the equation derived from the theory of rubber elasticity [44]. (4) νe= E′/3RT E′: the storage modulus taken 40 °C above Tg, R: the ideal gas constant (8.314 J K−1 mol−1), T: the thermodynamic temperature at Tg + 40 °C. The calculated υe values for all samples are also summarized in Table 7. Since the presence of rigid DOPO and benzothiazole groups inhibits the motion of the molecular chains, the υe values are decreased with the increasing AVD contents. 4. Conclusions A novel bio-based flame retardant AVD was one-pot synthesized using DOPO, vanillin, and 2-aminobenzothiazole as raw materials. The introduction of AVD hindered the curing process and reduced the Tg and crosslinking density values. The AVD showed an opposite effect on the storage modulus at temperatures above or under Tg due to the competition between the rigidity and lowered crosslinking density. The TGA results demonstrated that the early decomposition of AVD decreased the T5% and Tmax values of the cured FREPs while retarding the further decomposition of the EP matrix at higher temperature. The doping of AVD, EP exhibits great flame retardant properties. At a AVD loading level of 7.5 wt% (P content only 0.52 wt%), the LOI value of 31.3% and UL-94 V-0 rating were achieved for FREP-7.5. Moreover, the PHRR, THR and TSP values of FREP-7.5 declined by 54.7%, 14.7% and 15.4%, respectively. The comprehensive analysis of the char residues after CCT demonstrated that the phosphate-containing compounds dehydrated and esterified the EP matrix to form compact and continuous protective layers with high quality, which acted as physical barriers to effectively isolate the underlying substrate from heat and oxygen in the condensed phase. Additionally, the analysis of the pyrolysis volatiles showed that the release of non-combustible gases and phosphorus-containing radicals prevented the further combustion of the matrix in the gas phase. In conclusion, AVD, as an efficient and environmentally friendly bio-based flame retardant, is consistent with the concept of sustainable development and has great potential application in many fields. The exploration of more bio-based raw materials for flame retardant modification is one of the very promising development directions in flame retardant research. Author Contributions Investigation, L.C.; Methodology, L.C.; Supervision, B.W.; Writing—original draft, L.C.; Writing—review & editing, B.W. and Z.L.; formal analysis, Z.L. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare that they have no potential conflict of interest. Figures, Scheme and Tables materials-15-03155-sch001_Scheme 1 Scheme 1 Synthesis of AVD. Figure 1 FTIR spectra of DOPO, VN, ABT and AVD. Figure 2 1H NMR of AVD (a) 31P NMR of AVD (b) and DOPO (c), (C* represents chiral carbon). Figure 3 DSC curves of EP (a), FREP-5 (b), FREP-7.5 (c), and FREP-10 (d) at different heating rates, fitting curves of Kissinger’s method (e) and Ozawa’s method (f). Figure 4 TG (a) and DTG (b) curves of AVD and epoxy thermosets in nitrogen. Figure 5 HRR (a), THR (b), SPR (c), TSP (d), CO2P (e), and Mass (f) curves of the sample during cone heat combustion. Figure 6 Digital photographs and SEM images of the residues char of EP (a1–a3), FREP-5 (b1–b3), FREP-7.5 (c1–c3) and FREP-10 (d1–d3). Figure 7 XPS total spectrum of residual char from EP and FREPs. Figure 8 C1s (a), N1s (b), O1s (c), and P2p (d) spectra of residual char of FREP-7.5. Figure 9 Raman spectra of char residues for EP (a), FREP-5 (b), FREP-7.5 (c) and FREP-10 (d). Figure 10 TG-FTIR spectra at different temperatures, EP (a) and FREP-7.5 (b) 3D TG-FTIR spectra of pyrolysis products for EP (c) and FREP-7.5 (d). Figure 11 Absorption intensity versus time for several typical volatiles of EP and FREP-7.5, hydrocarbons (a), carbonyl compounds (b), aromatic compounds (c), aliphatic ethers (d). Figure 12 The flame-retardant mode of action of AVD on FREPs. Figure 13 DMA curves of EP and FREPs, storage modulus (a), tan δ (b). materials-15-03155-t001_Table 1 Table 1 The formulations of EP and FREPs. Sample EP (g) DDM (g) AVD (g) AVD (wt%) P (wt%) EP 40 10.11 0 0 0 FREP-5 40 9.54 2.61 5 0.34 FREP-7.5 40 9.36 4.00 7.5 0.52 FREP-10 40 8.92 5.43 10 0.69 materials-15-03155-t002_Table 2 Table 2 Ea values calculated from non-isothermal DSC curves. Sample Kissinger’s Method Ea (kJ mol−1) Ozawa’s Method Ea (kJ mol−1) EP 49.27 53.71 FREP-5 53.36 57.63 FREP-7.5 55.14 59.31 FREP-10 55.14 59.33 materials-15-03155-t003_Table 3 Table 3 Thermal stability parameters of EP and FREPs. Sample T5% (°C) Tmax (°C) Rate (%·min−1) CR700 (wt%) EP 382.3 399.0 18.8 19.7 FREP-5 359.1 391.0 12.5 21.7 FREP-7.5 349.5 388.7 10.7 23.3 FREP-10 346.3 388.3 10.7 24.5 AVD 272.8 279.3, 410.6 3.5, 2.7 30.1 T5%: initial decomposition temperature; Tmax: maximum decomposition temperature Rmax: maximum decomposition rate; CR700 char residue at 700 °C. materials-15-03155-t004_Table 4 Table 4 UL-94 Flammability Ratings and LOI Values for EP Composites. Sample P (wt%) LOI (%) UL-94, 3 mm bar (b) t1 + t2 (s) Dripping Rating EP 0 25 ± 0.1 (a) BC Yes No FREP-5 0.34 30.0 ± 0.2 5.6 + 5.8 No V-1 FREP-7.5 0.52 31.3 ± 0.1 4.3 + 2.7 No V-0 FREP-10 0.69 32.5 ± 0.3 2 + 2 No V-0 (a): BC indicates burn to clamp; (b): t1, t2 are the average of the primary and secondary combustion times of the five sample strips, respectively. materials-15-03155-t005_Table 5 Table 5 Combustion data obtained from CCT. Sample EP FREP-5 FREP-7.5 FREP-10 TTI (s) 91 ± 2 73 ± 5 68 ± 3 56 ± 1 PHRR (kW/m2) 1453 ± 20 708 ± 14 658 ± 9 568 ± 16 THR (MJ/m2) 67.3 ± 2.1 59.7 ± 1.9 57.4 ± 2.4 55.8 ± 2.2 FIGRA (kW/m2·s) 11.2 ± 0.3 4.57 ± 0.6 4.11 ± 0.2 3.44 ± 0.4 TSP (m2) 24.1 ± 0.7 23.8 ± 0.3 23.1 ± 0.8 20.4 ± 0.5 av-EHC (MJ/kg) 22.1 ± 1.1 15.0 ± 0.7 14.3 ± 0.9 12.7 ± 1.3 av-COY (kg/kg) 0.28 ± 0.05 0.52 ± 0.07 0.58 ± 0.1 0.62 ± 0.04 av-CO2Y (kg/kg) 1.80 ± 0.07 1.36 ± 0.03 1.29 ± 0.08 1.27 ± 0.02 Char residue (%) 18.1 ± 0.1 20.1 ± 0.7 22.5 ± 0.3 23.1 ± 0.4 materials-15-03155-t006_Table 6 Table 6 Composition and content of char residues. Sample C (At%) O (At%) N (At%) P (At%) S (At%) EP 81.71 14.37 3.92 0 0 FREP-7.5 83.18 12.17 4.24 0.23 0.18 materials-15-03155-t007_Table 7 Table 7 Thermomechanical properties of EP and FREPs. Sample Tg (°C) E′ at 50 °C (MPa) E′ at Tg + 40 °C (MPa) υe (mol·m−3) EP 170.8 2786 63.56 5265 FREP-5 157.3 3157 58.91 5021 FREP-7.5 145.0 3216 52.40 4586 FREP-10 143.1 3790 51.42 4516 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Karaer Özmen F. Üreyen M.E. Koparal A.S. Cleaner production of flame-retardant-glass reinforced epoxy resin composite for aviation and reducing smoke toxicity J. Clean. Prod. 2020 276 124065 10.1016/j.jclepro.2020.124065 2. Yuan Y. Shi Y. Yu B. Zhan J. Zhang Y. Song L. Ma C. Hu Y. Facile synthesis of aluminum branched oligo(phenylphosphonate) submicro-particles with enhanced flame retardance and smoke toxicity suppression for epoxy resin composites J. Hazard. Mater. 2020 381 121233 10.1016/j.jhazmat.2019.121233 31557714 3. Wang W. Kan Y. Yu B. Pan Y. Liew K.M. Song L. Hu Y. Synthesis of MnO2 nanoparticles with different morphologies and application for improving the fire safety of epoxy Compos. Part A Appl. Sci. Manuf. 2017 95 173 182 10.1016/j.compositesa.2017.01.009 4. Zhang Y. Synthesis and characterization of PEDSCD and its application as a flame retardant in epoxy resins RSC Adv. 2021 11 34849 34859 10.1039/D1RA06116J 35494757 5. Jin S. Qian L. Qiu Y. Chen Y. Xin F. High-efficiency flame retardant behavior of bi-DOPO compound with hydroxyl group on epoxy resin Polym. Degrad. Stab. 2019 166 344 352 10.1016/j.polymdegradstab.2019.06.024 6. Xu Y. Nie S. Dai G. Yang J. Dong X. Feng X. Effect of phosphorus-modified nickel phyllosilicates on the thermal stability, flame retardancy and mechanical property of epoxy composites J. Polym. Res. 2022 29 1 15 10.1007/s10965-021-02843-0 7. Huo S. Song P. Yu B. Ran S. Chevali V.S. Liu L. Fang Z. Wang H. Phosphorus-containing flame retardant epoxy thermosets: Recent advances and future perspectives Prog. Polym. Sci. 2021 114 101366 10.1016/j.progpolymsci.2021.101366 8. Zhang Q. Wang J. Yang S. Cheng J. Ding G. Hu Y. Huo S. Synthesis of a P/N/S-based flame retardant and its flame retardant effect on epoxy resin Fire Saf. J. 2020 113 102994 10.1016/j.firesaf.2020.102994 9. Zhao J. Dong X. Huang S. Tian X. Song L. Yu Q. Wang Z. Performance comparison of flame retardant epoxy resins modified by DPO-PHE and DOPO-PHE Polym. Degrad. Stab. 2018 156 89 99 10.1016/j.polymdegradstab.2018.08.007 10. Shen D. Xu Y.J. Long J.W. Shi X.H. Chen L. Wang Y.Z. Epoxy resin flame-retarded via a novel melamine-organophosphinic acid salt: Thermal stability, flame retardance and pyrolysis behavior J. Anal. Appl. Pyrolysis 2017 128 54 63 10.1016/j.jaap.2017.10.025 11. Fang Y. Miao J. Yang X. Zhu Y. Wang G. Fabrication of polyphosphazene covalent triazine polymer with excellent flame retardancy and smoke suppression for epoxy resin Chem. Eng. J. 2020 385 123830 10.1016/j.cej.2019.123830 12. Wagner J. Deglmann P. Fuchs S. Ciesielski M. Fleckenstein C.A. Döring M. A flame retardant synergism of organic disulfides and phosphorous compounds Polym. Degrad. Stab. 2016 129 63 76 10.1016/j.polymdegradstab.2016.03.023 13. Jian R. Wang P. Xia L. Zheng X. Effect of a novel P/N/S-containing reactive flame retardant on curing behavior, thermal and flame-retardant properties of epoxy resin J. Anal. Appl. Pyrolysis 2017 127 360 368 10.1016/j.jaap.2017.07.014 14. Qiu Y. Qian L. Feng H. Jin S. Hao J. Toughening Effect and Flame-Retardant Behaviors of Phosphaphenanthrene/Phenylsiloxane Bigroup Macromolecules in Epoxy Thermoset Macromolecules 2018 51 9992 10002 10.1021/acs.macromol.8b02090 15. Feng Y. He C. Wen Y. Ye Y. Zhou X. Xie X. Mai Y.W. Improving thermal and flame retardant properties of epoxy resin by functionalized graphene containing phosphorous, nitrogen and silicon elements Compos. Part A Appl. Sci. Manuf. 2017 103 74 83 10.1016/j.compositesa.2017.09.014 16. Tang S. Qian L. Qiu Y. Dong Y. High-performance flame retardant epoxy resin based on a bi-group molecule containing phosphaphenanthrene and borate groups Polym. Degrad. Stab. 2018 153 210 219 10.1016/j.polymdegradstab.2018.04.037 17. Yang S. Zhang Q. Hu Y. Synthesis of a novel flame retardant containing phosphorus, nitrogen and boron and its application in flame-retardant epoxy resin Polym. Degrad. Stab. 2016 133 358 366 10.1016/j.polymdegradstab.2016.09.023 18. Wu Q. Andersson R.L. Holgate T. Johansson E. Gedde U.W. Olsson R.T. Hedenqvist M.S. Highly porous flame-retardant and sustainable biofoams based on wheat gluten and in situ polymerized silica J. Mater. Chem. A. 2014 2 20996 21009 10.1039/C4TA04787G 19. Liu J. Dai J. Wang S. Peng Y. Cao L. Liu X. Facile synthesis of bio-based reactive flame retardant from vanillin and guaiacol for epoxy resin Compos. Part B Eng. 2020 190 107926 10.1016/j.compositesb.2020.107926 20. Howell B.A. Lazar S T. Isosorbide as a platform for the generation of new biobased organophosphorus flame retardants Ind. Eng. Chem. Res. 2020 1 1 3 10.33552/ICBC.2020.01.000509 21. Howell B.A. Sun W. Biobased flame retardants from tartaric acid and derivatives Polym. Degrad. Stab. 2018 157 199 211 10.1016/j.polymdegradstab.2018.10.006 22. Chen R. Luo Z.J. Yu X.J. Tang H. Zhou Y. Zhou H. Synthesis of chitosan-based flame retardant and its fire resistance in epoxy resin Carbohydr. Polym. 2020 245 116530 10.1016/j.carbpol.2020.116530 32718634 23. Howell B.A. Han X.R. Effective biobased phosphorus flame retardants from starch-derived bis-2,5-(Hydroxymethyl)furan Molecules 2020 25 592 10.3390/molecules25030592 32013211 24. Howell B.A. Ostrander E.A. Oberdorfer K.L. Phosphorus flame retardants from crop plant phenolic acids Sustainability & Green Polymer Chemistry Volume 1: Green Products and Processes American Chemical Society Washington, DC, USA 2020 Volume 1372 199 208 10.1021/bk-2020-1372.ch011 25. Howell B.A. Lazar S.T. Biobased plasticizers from glycerol/adipic acid hyperbranched poly(ester)s Ind. Eng. Chem. Res. 2019 58 17227 17234 10.1021/acs.iecr.9b03869 26. Daniel Y.G. Howell B.A. Phosphorus flame retardants from isosorbide bis-acrylate Polym. Degrad. Stab. 2018 156 14 21 10.1016/j.polymdegradstab.2018.07.027 27. Nabipour H. Wang X. Batool S. Song L. Hu Y. A phosphaphenanthrene-containing vanillin derivative as co-curing agent for flame-retardant and antibacterial epoxy thermoset Polymers 2021 217 123460 10.1016/j.polymer.2021.123460 28. Cheng J. Wang J. Yang S. Zhang Q. Hu Y. Ding G. Huo S. Aminobenzothiazole-substituted cyclotriphosphazene derivative as reactive flame retardant for epoxy resin React. Funct. Polym. 2020 146 104412 10.1016/j.reactfunctpolym.2019.104412 29. Wang P. Xia L. Jian R. Ai Y. Zheng X. Chen G. Wang J. Flame-retarding epoxy resin with an efficient P/N/S-containing flame retardant: Preparation, thermal stability, and flame retardance Polym. Degrad. Stab. 2018 149 69 77 10.1016/j.polymdegradstab.2018.01.026 30. Zou J. Duan H. Zhang J. Cao J. Zhang C. Wan C. Ma H. Multi-element heterocyclic compound derived from DOPO and thiadiazole toward flame-retardant epoxy resin with satisfactory mechanical properties J. Appl. Polym. Sci. 2021 139 52036 10.1002/app.52036 31. Zou J. Duan H. Chen Y. Ji S. Cao J. Ma H. A P/N/S-containing high-efficiency flame retardant endowing epoxy resin with excellent flame retardance, mechanical properties and heat resistance Compos. Part B Eng. 2020 199 108228 10.1016/j.compositesb.2020.108228 32. Wang P. Chen L. Xiao H. Flame retardant effect and mechanism of a novel DOPO based tetrazole derivative on epoxy resin J. Anal. Appl. Pyrolysis 2019 139 104 113 10.1016/j.jaap.2019.01.015 33. Zhang Y. Yan H. Feng G. Liu R. Yang K. Feng W. Zhang S. He C. Non-aromatic Si, P, N-containing hyperbranched flame retardant on reducing fire hazards of epoxy resin with desirable mechanical properties and lower curing temperature Compos. Part B Eng. 2021 222 109043 10.1016/j.compositesb.2021.109043 34. Huang C. Zhao Z.Y. Deng C. Lu P. Zhao P.P. He S. Chen S.W. Lin W. Facile synthesis of phytic acid and aluminum hydroxide chelate-mediated hybrid complex toward fire safety of ethylene-vinyl acetate copolymer Polym. Degrad. Stab. 2021 190 109659 10.1016/j.polymdegradstab.2021.109659 35. Guo Y. Chen X. Cui J. Guo J. Zhang H. Yang B. Effect of ionic liquid octyltriphenylphosphonium-chelated orthoborates on flame retardance of epoxy Polym. Adv. Technol. 2021 32 1579 1596 10.1002/pat.5195 36. Wei Z. Gu X. Wu J. Wei M. Yu Q. Xiujuan T. Wang Z. Performance comparison of epoxy resins modified with diphenylphosphine oxide and DOPO Fire Mater. 2019 43 892 902 10.1002/fam.2749 37. Chu F. Ma C. Zhang T. Xu Z. Mu X. Cai W. Zhou X. Ma S. Zhou Y. Hu W. Renewable vanillin-based flame retardant toughening agent with ultra-low phosphorus loading for the fabrication of high-performance epoxy thermoset Compos. Part B Eng. 2020 190 107925 10.1016/j.compositesb.2020.107925 38. Luo H. Rao W. Zhao P. Wang L. Liu Y. Yu C. An efficient organic/inorganic phosphorus–nitrogen–silicon flame retardant towards low-flammability epoxy resin Polym. Degrad. Stab. 2020 178 109195 10.1016/j.polymdegradstab.2020.109195 39. Wang D. Liu Q. Peng X. Liu C. Li Z. Li Z. Wang R. Zheng P. Zhang H. High-efficiency phosphorus/nitrogen-containing flame retardant on epoxy resin Polym. Degrad. Stab. 2021 187 109544 10.1016/j.polymdegradstab.2021.109544 40. Chen L. Jiang Q. Luo Z. Wang B. Facile synthesis of a reactive P/N/S-containing compound toward highly effective flame retardancy of epoxy resin with high transparency and improved mechanical strength Fire Saf. J. 2021 126 103472 10.1016/j.firesaf.2021.103472 41. Kong F.B. He Q.L. Peng W. Nie S.B. Dong X. Yang J. Eco-friendly flame retardant poly(lactic acid) composites based on banana peel powders and phytic acid: Flame retardancy and thermal property J. Polym. Res. 2020 27 10.1007/s10965-020-02176-4 42. Rao W. Zhao P. Yu C. Zhao H.B. Wang Y.Z. High strength, low flammability, and smoke suppression for epoxy thermoset enabled by a low-loading phosphorus-nitrogen-silicon compound Compos. Part B Eng. 2021 211 108640 10.1016/j.compositesb.2021.108640 43. Varganici C.D. Rosu L. Lehner S. Hamciuc C. Jovic M. Rosu D. Mustata F. Gaan S. Semi–interpenetrating networks based on epoxy resin and oligophosphonate: Comparative effect of three hardeners on the thermal and fire properties Mater. Design. 2021 212 110237 10.1016/j.matdes.2021.110237 44. Chen Y. Duan H. Ji S. Ma H. Novel phosphorus/nitrogen/boron-containing carboxylic acid as co-curing agent for fire safety of epoxy resin with enhanced mechanical properties J. Hazard. Mater. 2021 402 123769 10.1016/j.jhazmat.2020.123769 33254780

PMC009xxxxxx/PMC9099696.txt

==== Front Materials (Basel) Materials (Basel) materials Materials 1996-1944 MDPI 10.3390/ma15092981 materials-15-02981 Review Recent Progress in Single and Combined Porosity-Evaluation Techniques for Porous Materials https://orcid.org/0000-0001-5522-8860 Wang Yuqing https://orcid.org/0000-0003-4109-6750 Zhou Bo * Babarao Ravichandar Academic Editor Camblor Miguel A. Academic Editor Jiangsu Key Laboratory of Micro and Nano Heat Fluid Flow Technology and Energy Application, School of Environmental Science and Engineering, Suzhou University of Science and Technology, Suzhou 215009, China; 1911024003@post.usts.edu.cn * Correspondence: zhoub@mail.usts.edu.cn 20 4 2022 5 2022 15 9 298125 2 2022 18 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). The accurate determination of the porosity and specific surface area of porous materials such as shale and cement plays a key role in gas-energy-storage estimation and exploitation, building-heat and humidity-transfer investigation, and permeability-characteristics evaluation. Therefore, it is crucial to select appropriate measurement methods to accurately study the porosity, as well as other properties, of porous materials. In this review, various porosity-measurement methods are discussed. The most recent research findings and progress in combined methodologies are introduced and summarized. The measurement medium and chemical composition of the sample affect the porosity-measurement results. Therefore, depending on the measurement properties of different methods and the characteristics of the sample, an appropriate method can be selected. Furthermore, various methods can be combined to obtain more accurate measurement results than individual methods. porosity measurement SANS CT SEM NMR MIP WIP adsorption gas expansion ==== Body pmc1. Introduction Over the past decades, methane, hydrogen, and other gaseous energy carriers have been extensively developed and widely utilized. Consequently, porous materials have attracted significant research interest for gas storage and transport. Natural media and artificial porous materials are essential for gas storage. Natural media such as coal and shale are primary energy sources [1]. Shale gas predominantly exists in adsorbed phases in organic matter and inorganic minerals and in free phases in fractures and intergranular pores [2,3]. Traditional measurement methods, such as mercury intrusion porosimetry (MIP), are typically used to characterize porosity [4,5]. Understanding the pore structure of shale is important for evaluating the storage of free gas. The porosity of artificial porous materials affects hydrate formation and their gas-storage capacity [6]. Moreover, artificial porous materials are suitable for high-pressure gas storage because of their large specific surface area (SSA). Metal-organic frameworks (MOFs), covalent organic frameworks (COFs), and porous aromatic frameworks (PAFs) have been extensively examined as CO2, CH4, and H2 gas-storage materials owing to their large pore volumes and surface areas [7,8,9]. The structures of these nanoporous materials must be fully described to elucidate the relationship between good structural qualities and gas-storage properties [10]. These artificial materials are typically characterized via adsorption analysis. In both natural and artificial porous materials, gases are stored as adsorbed and free gases. Quantitative porosity measurements of these materials help evaluate their free-gas content and gas-storage capacity. In addition to fluid-storage capacity, the porosity of a material also significantly affects its transport behavior, including fluid flow, component diffusion, and heat transfer. Porous catalysts contain numerous ion/mass-absorption sites, abundant surface-reaction active sites, and sufficient mass-charge transfer channels, significantly increasing the ion-transport rate [11]. Accurate characterization of the SSA and porosity of materials is essential to evaluate their catalytic effects and improve chemical-reaction rates. Structured porous metals (e.g., Ni, Cu, Zn, and their oxides) have been used as electrodes in lithium-ion batteries or fuel cells owing to their advantages such as high conductivity, high porosity/SSA, controllable structure, and light weight, which significantly impact the porous-medium flow performance by affecting their fluid distribution and transport, hydrothermal properties, and electrical resistance. Furthermore, cementitious materials are typical porous structures employed in the field of architecture. The heat and moisture transfer of the building envelope are very complex processes that affect the buildings’ energy consumption and durability, as well as the thermal comfort of residents [12,13]. The pore structure of cementitious materials determines their macrophysical properties (e.g., permeability) and affects their durability [14,15,16,17]; hence, it is crucial to study the porosity of cementitious materials. The selection of appropriate experimental methods is an important factor for low-porosity and low-permeability materials. Porosity-measurement methods are classified into radiation-detection and fluid-intrusion methods. Most fluid-intrusion methods, such as MIP, water-immersion porosimetry (WIP), adsorption analysis, and gas-expansion/immersion methods, have been widely used in previous porosity-measurement studies [5,18,19,20]. Recently, several radiation-detection methods have been introduced in the field of pore-structure characterization, including low-field nuclear magnetic resonance (LF-NMR), small-angle scattering technology, electron microscopy, and computed tomography (CT) scanning [21,22,23,24]. Although the application of various measurement methods has become well-established, and their measurement accuracy has been recognized to a certain extent, the following challenges still exist: (i) for fluid-intrusion methods, the molecular size of the measuring medium determines the diameter range of the open pores (connected to the outer surface) that can be penetrated; therefore, some tiny pores will be excluded; (ii) adsorption affects the gas-filling amount, resulting in inaccurate skeleton volume measurements; (iii) for CT and electron microscopy, high resolution limits the field of view and leads to a time-consuming image-processing step. Several scholars have combined methods to compare and verify the measurement results of different methods, increasing their credibility [24,25,26,27]. In this review, currently used porosity-measurement methods are classified into radiation-detection and fluid-intrusion-measurement methods. Each of these is briefly introduced, and their characteristics and measurement accuracies are compared. The implementation and conclusions of a combination of multiple methods are introduced. The advantages of the LF-NMR combined with other methods are discussed and the applicability of various measurement methods is clarified. In general, the present review provides resources for new researchers in related fields to gain insight into existing experimental methods and help them select suitable experimental methods. 2. Radiation-Detection Methods Radiation-detection methods for porous media are based on the refraction, transmission, or scattering of rays by a solid skeleton. Among the various radiation-detection methods, electron microscopy uses electron beams, CT scanning utilizes X-rays, small-angle scattering uses X-rays or neutron rays, and NMR is based on the excitation of hydrogen atoms via electromagnetic radiation. 2.1. Electron Microscopy Transmission electron microscopy (TEM) was designed according to the principle of optical projection microscopy. The samples must be thin enough (thickness of 100–500 nm) to be transparent to the electron beam. At present, the resolution of a 400–kV TEM instrument can reach 0.2 nm [21]. Furthermore, TEM should be used in conjunction with other methods when characterizing organic and carbon samples [28,29]. Scanning electron microscopy (SEM) is used to produce amplified images by replacing light waves with electrons; it can scan the sample surface with a resolution below 1 nm and an amplification of more than 4 × 106 times [21]. Compared to traditional optical microscopy, SEM has a higher resolution, dynamic amplification range, and depth of field. It can also better analyze samples in detail when connected to an X-ray analyzer [30]. Figure 1 shows the SEM morphology slice-and-view process of cement-paste samples. The cement-paste samples were prepared with Type I Portland cement and deionized water with a water-to-binder ratio of 0.45. Then, the cement paste was cut using a low-speed diamond saw into 10 × 10 × 2 mm3 samples. 1000 slices of 2D SEM images were aligned and stacked into a 3D bounding box with dimensions 10 × 10 × 20 μm. More details can be found in Lim et al [31]. TEM and SEM can directly observe the pore morphology and identify pore types [32]. Figure 2 shows the schematic of TEM and SEM optical designs. They have been proven to be effective methods for characterizing the pore morphology and structure as follows: (i) Although TEM can probe nanoscale features, it requires thin samples that are transparent to the electron beam. Preparing a thin sample requires destructing the bulk sample and polishing the surface, which may cause artificial cracks and pores, in turn resulting in inaccurate imaging results [33,34]. Currently, many laboratories use argon ion-milling powder to complete the polishing process and maintain the true mineral texture and pore structure [35]. (ii) Owing to the high magnification involved in SEM, it can only provide a local pore morphology with a narrow area [27,36]. To overcome this drawback, multiple images can be spliced or reconstructed in 3D. However, the reconstruction process is complicated and time-consuming [37]. (iii) Sufficient image segmentation is required when using focused ion-beam (FIB)-SEM tomography to quantitatively analyze 3D structures. The segmentation method of choice is related to the accuracy of the quantitative porosity analysis. Čalkovský et al. compared the signal-processing algorithm, Otsu’s method, Darwinian particle swarm optimization (DPSO), harmony search optimization (HSO), and fuzzy c-means threshold algorithms, and found that all these algorithms slightly underestimated the true pore size. The derived criteria for selecting the intensity threshold at the pore–polymer interface were proposed for accuracy improvement [38]. 2.2. Small-Angle Scattering Technology Small-angle scattering technology is based on quantitatively interpreting the microstructure and porosity of samples using the relationship between the scattering radiation intensity and scattering angle obtained via neutron or X-ray irradiation. Two basic approaches are currently being adopted in the research of small-angle scattering technology: small-angle X-ray scattering (SAXS) and small-angle/ultra-small angle neutron scattering (SANS/USANS) [22]. Figure 3 shows the steps involved in SANS data analysis. SANS/USANS can characterize nanopore structures and their confined fluid behavior. Owing to their high permeability of neutrons, SANS/USANS can detect the interior of samples and provide information on closed pores and pore-size structures of 1 nm–10 μm compared with X-rays [33,39]. Neutron scattering is more sensitive to the position of hydrogen and its isotopes, making it suitable for studying the contents of hydrogen and other hydrogen elements (solid and liquid). Figure 4 shows the scattering profiles for shale samples named QD1-L3, QD1-L4, WX2-8, WX2-33, WX2-49, and WX2-54, from which the pore-size distribution (PSD) can be obtained. In Figure 4a, Q and I(Q) can be respectively defined as:(1) Q=4πλsinθ (2) I(Q)=N(Δρ*)2∫V2(r)f(r)P(Q,r)dr where λ is the neutron wavelength; θ is the Bragg angle, which is the half of the scattering angle; N is the pore number density; (Δρ*)2 is the scattering contrast, which is equal to (ρs* − ρp*)2, that is, the square of the difference between the scattering length density (SLD) of the matrix and that of the pores (generally taken to be zero); V(r) is the spherical volume; f(r) is the PSD; r is the spherical pore radius; and P(Q, r) is the spherical form factor. Here, the pore size can be estimated by Bragg’s law with Q as r = π/Q in radius or d = 2π/Q in diameter. The approximately linear relationship between log(I/(Q)) and log(Q) in Figure 4a indicates that the six samples tested have a fractal pore structure. Figure 4b shows that the PSD has a peak at approximately 2 nm for each shale sample, which indicates either the existence of inaccessible pores or heterogeneity at a pore size of approximately 2 nm for the samples tested. Several challenges still exist that limit the application of small-angle scattering technology, as follows: (i) Scattering techniques, including SAXS and SANS, allow the characterization of open pores (connected to the outer surface) and closed pores (contrary to open pores) but provide limited information on pore morphology [27]; (ii) SANS cannot provide full-scale porous information on the sample; (iii) Therefore, the pore-skeleton two-phase hypothesis model was chosen to complete the porosity measurement in SANS [40]. However, for samples that are rich in minerals and organic matter, the mineral/mineral-phase scattering is affected, making the selected two-phase model inaccurate for measurements. 2.3. Computed Tomography (CT) Scanning CT scanning utilizes the interaction between X-rays and materials for porous-material characterization or imaging, with the most common scanning mode being based on attenuation scanning for X-rays. Figure 5 shows the CT scan procedure. According to the spatial resolution, these methods can be classified as macro-, micro-, and nano-CT. Macro-CT is used to scan samples with sizes of 10 cm, while the scanning object range of micro-CT is 1 cm–10 μm. Even high-resolution micro-CT cannot capture the entire pore range of 10 nm–10 µm when the pore microstructure is studied [41]. Nano-CT can further characterize the pore microstructure at the submicron scale and compensate for the micro-CT data to a certain extent [23]. However, even the most advanced nano-CT cannot capture the entire sample pore range except for the lower boundary of the capillary pore [42,43]. CT has proven to be an effective method for characterizing the pore morphology and structure of samples [44,45]. The CT method can be applied to study crack evolution after compression, create 3D image tomography, and analyze the sample microstructure [46,47,48,49,50]. Its shortcomings include temporal and spatial-resolution limitations and the problem of distinguishing between material components with similar attenuation coefficients. Moreover, high-resolution applications require a longer display time and a smaller representative sample size, which increases the calculation time, similar to SEM [51]. Figure 6 shows the CT images of Portland cement under different stresses. The aggregate in the CT image is marked in white, while the cracks and pores are marked in black or nearly black. When the stress was up to 36.08 MPa, a crack gradually enlarged. When the axial stress was 12.7 MPa, the specimen entered the damage stage, in which mesocracks propagated and rapidly converged. 2.4. Low-Field Nuclear Magnetic Resonance (LF-NMR) The spin-precession movement of a nucleus exhibits a specific resonance frequency under an external electromagnetic field. When the magnetization vector of the nuclei is disturbed in a direction different from the magnetic field direction, it gradually relaxes towards the latter. The relaxation time of the nuclei depends on the pore structure characteristics of a porous sample because of the interactions between the nuclei and pore surfaces. The magnetization relaxation processes can be polarized and detected by external radio frequency (RF) pulses; thus, microscopic pores are characterized. Compared to expensive solid-state NMR techniques, LF-NMR systems are more suitable for detecting pore-filling fluids (containing protons in the fluid molecules) in many porous materials, and they involve static magnetic fields of the order of a few Tesla and operate at frequencies between 10 and 50 MHz. A limitation of NMR is that the tested sample and fluid must not contain a large concentration of matter, such as ferromagnetic metals and minerals. This significantly affects the external magnetic field. Figure 7 shows a schematic of the LF-NMR experimental setup for detecting pore-filling methane. The LF-NMR techniques are classified into the following groups: (i) relaxometry; (ii) imaging (NMR imaging, T1 or T2 spin-echo imaging, or spin-density mapping); and (iii) NMR relaxation tomography, where T1 and T2 are relaxation times for the longitudinal and transverse directions to the magnetic field, respectively [53]. LF-NMR instruments utilize a shorter spin echo to detect finer pore structures such as micro/nanoscale pore spaces. In a laboratory setting, echoes of 20–100 µs spacing were achieved [54]. Two-dimensional NMR techniques, including T1-T2 and D-T2 (diffusion-T2) mapping, have been developed. T1-T2 mapping is more sensitive to molecular motion in the frequency range between the Larmor frequency (approximately 2 MHz) and extremely low frequencies. Therefore, the T1/T2 ratio can be used as a parameter to reflect the free and restricted states of the molecules in the fluid [56]. T1-T2 mapping can also be utilized as a unique probe to distinguish oil-filled pores from organic and inorganic mineral pores, while D-T2 mapping can be used to distinguish between oil and water in core samples [54]. However, crystal water in the sample cannot be accurately distinguished because LF-NMR measurements require the sample to be completely saturated with water [57]. Examples of typical LF-NMR results are shown in Figure 8. Figure 8a demonstrates an example of the application of D-T2 technique, where the horizontal line represents the diffusion coefficient of water, and the diagonal line is the “oil line,” where the oil signals can be found. The signal is clearly along the oil line, indicating the presence of light oil in the sample. The top and right panels are projections along their respective dimensions [54]. In Figure 8b, the T2 spectrum is divided into four parts, P1, P2, P3, and P4, which can be defined as (i) the adsorbed methane in micropores, (ii) the porous-medium-confined methane, (iii) the interparticle free methane in the interparticle space of powdered shale, and (iv) bulk methane in the space between shale particles and the inner wall of the sample cell, respectively [2]. The diffusion process of CO2 in the n-tetradecane is shown in Figure 8c. In this process, the initial pressure was 5000 kPa and the temperature was 30 °C. It is clear from the image that CO2 diffuses gradually towards the bottom in the porous medium, and the concentration eventually tends to be consistent. In addition, the interface gradually increased due to the increase in the volume of the liquid phase when CO2 dissolved into the liquid phase. 2.5. Summary Most radiation-detection methods are used to observe the microscopic morphology of a sample, except for LF-NMR, which is, in principle, an indirect measurement method based on fluid saturation. However, unlike other radiation-detection methods, LF-NMR is applicable for large samples, where the sample scale is restricted by the sizes of the RF coils and the auxiliary fluid-saturation system. A core challenge in using LF-NMR for porous-medium characterization is that even though the signals of fluid in nanoscale pores can be detected, identifying the type, phase, and state of the fluid in the pores for the relaxation-time spectra is difficult; this is because the spectra are determined by both the properties of the fluid and the PSD of the material. Different radiation-detection methods are compared in Table 1. 3. Fluid-Intrusion Methods Fluid-intrusion methods require the sample to be fully immersed in gas or liquid, and the measurement is generally conducted in a series of equilibrium states. The core concept is to transfer the measurement of the pore geometry to the quantification of the pore-filling fluid. Fluid-intrusion methods include MIP, gas expansion/intrusion methods, WIP, and adsorption analysis. These methods have a wide application range in porosity measurement, and their measurement results are considered reliable [1,59,60]. 3.1. Mercury Intrusion Porosimetry (MIP) MIP is based on the Wasllbum equation and capillary phenomenon, providing two types of measurements: high-pressure and constant-speed MIP [40,61]. High-pressure MIP can enter the pore-throat space (>2 nm) and has a wide range of microscopic pore-structure characterization [1,4]. However, the amount of mercury in the throat and pores cannot be identified separately, and the experimental results are generally greater than the real values in high-pressure mercury porosimetry experiments owing to the pore-space enlargement caused by the high mercury-injection pressure and wetting lag phenomenon [18]. In constant-speed MIP, mercury enters pores at a constant speed, and its experimental state is closer to the real mercury-injection state. The average porosity and permeability measured by constant-speed MIP were lower than those measured by gas, indicating that the pore volume and PSD of the inaccessible part still have a significant influence. MIP is used for the high-precision measurement of open pores. It is a promising technique for various materials and has several applications [18,26,27,52]. However, it is not suitable for detecting micropores (<2 nm) because mercury does not fill every pore [1,62]. Furthermore, the MIP application is impeded by the following drawbacks: (i) The measurement results of crushed powder are larger than those of plug samples because of the destruction of the pore structure during the crushing process [5,63]; (ii) Small pores are measured under high pressures, which may damage the samples [64]. 3.2. Gas-Expansion/Intrusion Method The gas-expansion method is derived from Boyle’s law. The pore volume of a sample is calculated by measuring the pressure change during gas expansion. Helium is used as the medium because it is the smallest nonadsorptive gas molecule and can successfully penetrate the sample’s entire structure [19,65]. The helium-expansion method is used to measure connected pores. The gas-expansion method is a reliable, well-established, and commercialized measurement method. Various measuring instruments have been developed, such as true volume and density-measurement instruments, porosity-measuring instruments, and pore-size analyzers, with an accuracy of ±0.03%. These instruments use helium, nitrogen, and other inert gases as filling gas. Furthermore, they have a high degree of automation, are easy to operate, and are essential in the field of porosity measurement. Fu et al. measured the total porosity of shale with a volumetric analyzer and a solid densitometer and discussed the factors affecting the measurement results [66]. Sun et al. used a porometer to measure the helium porosity and density of shale samples. The results were compared with those obtained from gas (CO2 and N2) adsorption and SANS [25]. Zhou et al. used an AP608 automated permeameter-porosimeter to measure the helium porosity and air permeability of coal for comparison with CT and MIP results [45]. Compared with WIP and MIP, the gas-expansion method has a shorter analysis time, simpler operation, and weaker influence on the samples; therefore, it is convenient for repeated measurements [66]. Compared with crushed samples, plug samples underestimate shale porosity [62,67]. In general, this volume-calculation process is based only on the initial and final pressures and has high requirements for temperature control during measurement. Repeated vacuumization and inflation are required for multiple measurements. The opening process of the balance valve inevitably increases the volume of the tube system, which causes systematic measurement errors. 3.3. Water-Immersion Porosimetry (WIP) WIP uses the weight of the sample under dry conditions and the weight difference between air and water after the sample is fully immersed in water to indirectly calculate porosity. This method is suitable for measuring samples with low porosity (<5%) [5]. It is called keroseimmersion porosimetry (KIP) when kerosene is used, and dual-liquid porosimetry (DLP) when both kerosene and water are used [68]. The solutions usually comply with the following conditions [5]: (i) low surface tension and viscosity, and high moisture; (ii) high vapor pressure and low evaporation rate; (iii) does not easily react with samples; (iv) stable composition and density; (v) are harmless and can be disposed of safely. Tomasz et al. [68] used DLP on shale samples from the Podhale and Baltic Basins to quickly measure clay-bound water (CBW) at 40–80% relative humidity (RH) without crushing. CBWmin (40% RH) provides the bound-water value under high hydrocarbon saturation, and CBWmax (80% RH) represents the maximum water content of the bound water. The results indicated that WIP is suitable for shale with low density, a high diagenesis degree, and strong cementation, and can be used to calculate hydrocarbon reserves. DLP complements WIP and KIP and obtains the CBW range of rock debris, providing more useful information for formation evaluation. WIP can keep the samples intact and thereby does not change the sample’s composition [69]. It also has the advantages of low measurement cost, repeatability, and high reliability. However, this method requires the sample to be dried at 200 °C until it reaches a constant weight; this may remove the crystal water in the sample and change the pore structure, causing increased porosity [70]. Simultaneously, the sample must be fully saturated, which takes a long time. Similar to NMR, the clay minerals and organic matter in shale are prone to irreversible chemical reactions with intrusive water phases, damaging the samples and resulting in large measurement results [64]. 3.4. Adsorption Analysis Adsorption analysis is based on the capillary aggregation phenomenon and the principle of volume-equivalent substitution. For conventional test fluids such as carbon dioxide and nitrogen, isothermal adsorption tests are usually conducted under low-temperature conditions with the test pressures under the corresponding saturation pressure. Under the assumption that the pore shape is cylindrical and tubular, a capillary-aggregation model was established to estimate the PSD characteristics and pore volume (PV) of the sample [71]. The volumetric method for adsorption was applied by measuring the pressure change caused by capillary aggregation. Figure 9 shows a typical system diagram for nitrogen-adsorption analysis. This volumetric method can effectively characterize the PSD of micropores and mesopores (2–50 nm) in the sample compared to the MIP method [72]. Overall, adsorption-analysis techniques are well-established and widely applied for material characterization. When adsorption tests are conducted at high pressures and high temperatures (HTHP, for the critical states of the test fluids), the pore-filling fluid usually exists in the gaseous phase; hence, the isotherms cannot provide PSD or PV analysis without capillary aggregation. However, high-pressure and high-temperature adsorption tests show the real state of the fluid in the porous material in application scenarios. In addition to the volumetric method, the gravimetric method, which uses a high-precision balance to measure the gravity change of a sample due to adsorption under various pressures, is also widely applied in HTHP adsorption tests. The gravimetric method is widely used to determine the gas adsorption and sorption capacities of coal and shale [73,74,75,76]. Similar to other fluid-immersion methods, the adsorption analysis also presents the following challenges. (i) Shrinkage and swelling are induced when the fluid is in contact with the porous structure. Most of these effects are irreversible and increase the pore volume, although there are some exceptions—for example, the shrinkage and swelling caused by the contact of CO2 and the coal molecular structure are reversible [19]. (ii) Adsorption analysis is a method used to characterize the pores in principle, whose final measurement results are based on a hypothetical pore model [14]. (iii) Accurate nitrogen-adsorption tests require the sample to be dried and vacuumed before the experiment, which may alter the pore structures of some samples [40]. Different fluid-intrusion methods are compared in Table 2. 4. Combination of Various Measurement Methods In recent years, many scholars have combined different porosity-measurement methods to conduct further comparative analyses of various aspects of the samples and clarified the characteristics of different methods and what to consider when selecting methods. Table 3 summarizes part of the measurement work conducted in recent years with respect to sample type and size, measurement method and conditions, and measurement results and conclusions. 4.1. Combination of LF-NMR and Other Methods As introduced previously, the LF-NMR technique, as a noncontact method, can obtain the distribution of hydrogen-containing components in the sample, but the amount of substance identified by NMR is related to the chemical composition of the material and the PSD, which requires special investigation for different materials. In recent years, many studies have focused on various materials, adopting a combination of NMR and fluid-intrusion methods to further understand the NMR measurement results. Coal and shale are complex natural-gas-storage materials that have been well-studied by combining LF-NMR and other methods. The combination of NMR and adsorption analyses can distinguish the type of methane contained in coal and shale, providing a quantitative amount of methane and some information on methane migration. An NMR method to characterize the adsorption capacity of coal for methane was established by Yao et al., and the measurement results were compared with the HTHP volumetric adsorption-analysis results [55]. They found that the adsorption amount measured via NMR was less than that measured via methane-adsorption (MA) analysis. The explanation was that the methane in coal exists on the pore surface and in solid solutions, which LF-NMR cannot detect [79,80]. In 2018, Yao’s team used NMR and methane/nitrogen adsorption to evaluate and compare the gas content of shale and reached similar conclusions. They discovered that NMR could identify the type of methane gas (free or adsorbed) and quantify free methane in shale. However, MA measurement only quantified the amount of adsorbed methane instead of free gas [2]. In recent years, many scholars have used NMR to monitor the adsorption and gas flow of methane in coal, demonstrating that NMR can provide more detailed methane-flow information [2,81,82]. When NMR is combined with WIP, it can detect free water inside a dry sample, and WIP can determine the mass change before and after water immersion. A recent study by Zhao et al. supports that the measurement results of the two methods are consistent [24]. Before the water-immersion process began, the NMR signals of all samples were not equal to zero, proving that free water still existed in the samples after drying. NMR provides more pore information than MIP. In the measurement of cement by Zhao et al. ([18]), the MIP data show that there is no pore size smaller than 10 nm, which were the predicted result that refers to the true gel pores or pores generated by pressure damage. By contrast, the NMR method is suitable for the distinction of capillary (<100 nm) and gel pores (200–1000 nm) based on the T2 curve, where the first two T2 peaks (with an increasing relaxation time) represent the capillary and gel pores, respectively. The PSD of three groups of coal-cutting samples with different particle sizes was measured by Chang et al. using MIP and NMR [26]. The MIP results for cuttings of different particle sizes differed significantly, owing to the influence of interparticle voids and intrusion pressure. NMR is independent of the sample size and shape and thus provides more accurate porosity information than MIP. When the particle size of the cuttings was large (≥1 mm), most of the pores inside the cuttings were intact, providing reasonable porosity-analysis results. Combining the LF-NMR method with fluid-intrusion methods is an effective approach to understanding the component morphology inside the sample and provides evidence to explain the LF-NMR T2 peaks related to the physical properties of the fluid in pores, which improves the reliability and accuracy of LF-NMR measurements. In addition to combining with fluid-intrusion methods, NMR can also be combined with radiation methods, such as SANS, to obtain more information about the sample. Further work in this area is required. 4.2. Combination of Different Fluid-Intrusion Methods Fluid-intrusion methods are well-established measurement methods; however, when any one is used alone, it provides limited pore-size information. The measurement results obtained using different fluid-intrusion methods are not identical for the same sample. Therefore, it is critical to use different fluid-intrusion methods to conduct a comprehensive material analysis. A more complete PSD range can be obtained by combining different fluid-intrusion methods than only using an individual method. At present, there is a consensus that the molecular size of the medium affects the measurement results. Mercury cannot enter tiny pores owing to the influence of surface tension, and the measurement result is usually the smallest. Sun et al. ([25]) concluded that the helium-expansion method measures open pores with a diameter smaller than 0.2 nm, the CO2-adsorption method (CA) measures open pores with a diameter of 0.3–1.4 nm, and the N2-adsorption method (NA) quantifies openings with a diameter of 1.4–300 nm. The porosity result of the helium-expansion method was slightly larger than the NA/CA result, owing to the small size of the helium molecules. Similarly, the case reported by Wang et al. ([4]) indicates that the total porosity results obtained via helium expansion are more accurate than those of CA and mercury intrusion. The CA, NA, and MIP data can be used to define the microporous, mesoporous, and macroporous PSDs, respectively. A combination of the three methods can be used to obtain the entire PSD range. Wang et al. ([27]) held the same opinion, considering that CA is suitable for characterizing the porosity and SSA of micropores (<2 nm), NA and SANS are suitable for mesopores (2–50 nm), and MIP is suitable for macropores (>50 nm). Mastalerz et al. ([78]) used NA to calculate the mesoporosity of shale and CA to analyze micropores; their results showed that the SSA obtained via CA was larger than that of NA. In addition to the measurement medium, which affects the porosity measurement results, the chemical composition of the sample also affects the accuracy of various measurement methods. When CO2 and methane molecules come in contact with coal, shrinkage and swelling occur. Rodrigues et al. found that the volume of adsorbed CO2 tends to be much higher than the free-gas volume when in contact with the coal structure [19]. The shrinkage and swelling effects of carbon dioxide on the coal structure were completely reversible. Methane also induces shrinkage and swelling when it comes into contact with the molecular structure of the coal. Although these effects are smaller than those of carbon dioxide, they are irreversible and increase the coal volume. In the study by Mastalerz et al., for coal, the SSA measurement result was NA < SANS < CA, and for shale, NA < CA < SANS, which may be due to the expansion of organic matter in the coal sample during the CO2-adsorption process [78]. For cementitious materials, intrusion water can react with hydration products when using WIP, increasing the total mass of the intrusion water and improving the measurement results. Recent cases reported by Qian et al. ([14]) indicate that the porosity-measurement results of cementitious materials follow MIP < GIP < WIP. Inaccessible porosity should be considered when choosing an appropriate porosity measurement method. In the experiment by Wang et al. ([27]), the PSDs obtained from SANS, NA, and MIP seem to be reasonably consistent for most of the tested shale samples. However, for QD1-L3, WX2-8, and WX2-33, the results show that the PSD estimated by SANS is larger than that estimated by MIP in the pore size range of 100–300 nm, indicating the existence of inaccessible porosity. In summary, when using fluid-intrusion methods, an appropriate measurement medium and method should be selected according to the sample composition. If necessary, various methods can be combined to obtain a complete sample PSD. 5. Conclusions This paper reviews various experimental methods for measuring porosity and porous media. These experimental methods can be classified as radiation-detection and fluid-intrusion methods. Each of these methods can obtain the pore-size information of a sample, but their combined use can provide more accurate information than individually, which is of great significance for understanding the gas-storage and transport mechanisms in porous materials. Based on the review presented herein, the following conclusions are drawn:MIP, the gas-expansion/intrusion method, and adsorption analysis are the most developed measurement technologies. Several developed commercial instruments based on these methods are available. Adsorption analysis mainly focuses on qualitative characterization, and the accuracy of quantitative results for pore volume is questionable. Fluid-intrusion methods can change the sample characteristics to different degrees, among which MIP is the most prominent, and gas-adsorption/intrusion analysis causes marginal damage. SEM and CT are widely used for material analysis and characterization. The quantitative measurement of pore volume requires numerous slices, multi-angle measurement, and 3D reconstruction, and has low measurement efficiency and accuracy. SANS can provide information regarding pore sizes of 1 nm–10 μm but provides limited information on pore morphology. The SEM, CT, and SANS measurement results are limited by the measurement scale. LF-NMR can quantitatively characterize the material itself and the hydrogen-containing fluid in the pores of the material; however, the explanation of the T2 spectrum or other two-dimensional spectra needs to be interpreted in combination with other methods. Typically, LF-NMR is used in combination with MIP to clarify the meaning of the T2 spectrum. The pore volume results of a material obtained by different methods are not consistent. Some methods can yield similar results, while others show clear differing trends depending on both the physical mechanism of the method and the properties of the material. Based on comparative studies on the porosity measurement methods listed in Table 1, the following conclusions can be drawn: For shale and cement, (WIP/NMR/MA/MIP/NA) < CA < SANS < He; For coal, (MIP/WIP/NMR-NA/SANS) < MA/CA < He. For the measurement properties of different methods, suitable methods can be selected according to the characteristics of the samples, or multiple methods can be compared or combined to obtain more accurate measurement results. At present, various joint experiments are combined with NMR experiments or adsorption analysis and other methods. However, there is still scope for development. Acknowledgments The authors gratefully acknowledge the support provided by the National Natural Science Foundation of China No. 51806147. Author Contributions Conceptualization, B.Z. and Y.W.; methodology, Y.W.; writing—original draft preparation, Y.W.; writing—review and editing, B.Z.; supervision, B.Z.; funding acquisition, B.Z. All authors have read and agreed to the published version of the manuscript. Funding This research was funded by National Natural Science Foundation of China, Grant Number 51806147. Institutional Review Board Statement Not applicable. Informed Consent Statement Not applicable. Data Availability Statement Not applicable. Conflicts of Interest The authors declare no conflict of interest. Figure 1 SEM morphology slice-and-view process of cement-paste samples [31]. (a) Milling and imaging, (b) 2D image-stack registration, (c) 3D reconstruction. Figure 2 The schematic the optical designs of (a) TEM and (b) SEM. Figure 3 Schematic diagram of SANS (modified from [22]). Figure 4 (a) Scattering profiles in log−log plots: I(Q) versus Q; (b) PSD of shale samples ([27].). Figure 5 X-ray CT scan procedure (modified from [41]). Figure 6 2D CT image of Portland cement under different stresses ([52]). Figure 7 Schematic of the LF-NMR experimental setup (modified from [55]). Figure 8 Examples of NMR results: (a) D-T2 mapping measured at 2 MHz for a shale sample from the Smackover formation (from [54]) (b) T2 spectra of methane in shale under different pressures (from [2]); the porosity can be calculated according to the calculation of the shaded area. (c) 2D NMR images of the CO2-diffusion process in porous media saturated with n-tetradecane (from [58]). Figure 9 Schematic of the volumetric adsorption method. materials-15-02981-t001_Table 1 Table 1 Comparisons of different radiation detection methods. Method Using Ray Disadvantages Advantages Electron microscopy Electron beams Limitation of thin slices; complex image processing Direct observation of the sample. SANS X-rays or neutron rays Little information on pore morphology. Characterization of open pores and closed pores. CT X-rays Limitation of temporal resolution and spatial resolution 3D image tomography LF-NMR Excitation of hydrogen atoms by electromagnetic radiation Constraints of identifying the type, phase, and state of the fluid in the pores Detection of large samples materials-15-02981-t002_Table 2 Table 2 Comparisons of different fluid-intrusion methods. Method Common Measurement Medium Disadvantages Advantages MIP Mercury Destruction of the pore structure Characterization of macropore and mesopore Gas expansion/intrusion method He High requirements for temperature and leakage control; volume increment caused by valve-opening process Easy operation; repeatability WIP Water Limitation for samples containing clay minerals and organic matter Low measurement cost; repeatability Adsorption analysis N2, CH4, CO2 Shrinkage and swelling caused by CH4 and CO2 Characterization of micropores and mesopores materials-15-02981-t003_Table 3 Table 3 Joint-method literature review. Authors Sample Sample Size Method Test Condition Result and Conclusion Chang et al., 2020 [26] Coal from the Qinshui and the Junggar basins, China. 1:Plugs:D: 2.5 cm,               L: 6 cm 2:Particle:   1.00–1.70 mm;   2.36–3.35 mm;   4.75–6.70 mm NMR; MIP; Expansion(He) NMR: 0.53 T magnetic          Strength, 23 MHz. MIP:200 MPa ϕHe > ϕNMR The measurement results of NMR and MIP are consistent Yao et al. 2014 [55] Coal from southeastern Qinshui Basin, China Powder with a 60–80 mesh size NMR; MA NMR: CPMG sequences of 18,000 echoes, echo spacing 0.3 ms, trains 64. PA:25 °C, 6 MPa Adsorption capacity: NMR < MA Yao et al., 2019 [2] 1. Shale from Hunan province in south China. 2. Shale from Sichuan province of southwestern China. n/a NMR; MA NMR: 23.15 MHz,          0.54 T magnetic          strength Adsorption: n/a Adsorption capacity: NMR < MA Wang et al., 2017 [4] Shale from Shihui Trough, eastern Qaidam Basin, China. Ultrafine particle size NA; CA; MIP; SEM; Expansion(He) NA/CA: 77 K MIP: 3000 psi Expansion: 28 °C, 1.2 MPa Porosity < 10 um, mainly mesoporous ϕHe > ϕNA/CA/MIP; the expansion result is more accurate than other methods Wang et al., 2020 [27] 1. Shale in northeast Chongqing near the edge of the Sichuan Basin, China. 2. Shale in northeast Yunnan near the southwestern edge of the UYP, China. Plugs: D: 1 cm, L: 1 cm MIP; SANS; CA; NA          MIP: Pressures          from 0.14 to 413          Mpa CA/NA:77 and 273 K PSD result: The measurement results of four methods are consistent. Zhao et al., 2021 [24] 1. White Portland cement (WPC) 2. Standard sand 1.WPC: SSA of 380 m2/kg; Density of 3100 kg/m3 2. Standard sand: Particle sizes ranging from 0.08 mm to 2.0 mm NMR; WIP NMR:0.5 T magnetic strength, 21.3 MHz frequency Water-absorption capacity: The results of LF-NMR imaging measurements are consistent with the WIP method Zhao et al., 2018 [18] PII 52.5 Portland cement n/a MIP; NMR NMR:0.42 T magnetic strength, 18 MHz frequency Pore-size range: The measurement range and the order of magnitude are consistent. Zuena et al., 2019 [57] Limestone MIP:1 × 1 × 4 cm3 NMR:5 × 10 × 2 cm3 MIP; NMR n/a There is a good correlation between the quantitative results obtained by MIP and the qualitative ones observed with NMR. Sun et al., 2017 [25] Shale from the northwest of Guizhou province, Southwest China SANS:10 × 10 × 2 mm3                   Expansion:                   Plugs: D: 2.5                   cm, L: 3 cm Adsorption:40–80 mesh SANS; Expansion(He); NA; CA n/a ϕNA/CA < ϕHe Mergia et al., 2010 [77] A self-sintering carbon mesophase powder based on petroleum residues Particle size:1μm Expansion (He); NA; SANS n/a ϕSANS and ϕNA are consistent, He expansion is smaller than SANS. Mastalerz et al., 2012 [78] 1. Shale from New Albany in Indiana. 2. Coal from the Petersburg Formation in America. n/a NA; CA; (U)SANS NA:77.35 K, 101.3 kPa CA:273.1 K             SANS: λ = 4.8 Å,             0.002 < Q < 0.7             Å−1 USANS:λ = 2.4 Å, 5 × 10−5 < Q < 0.003 Å−1 SSA result: For coal, NA < SANS < CA For shale, NA < CA < SANS Shi et al., 2020 [46] Coal from the Qinshui Basin, China Plugs: H: 5 cm, L: 2 cm CT; MIP n/a n/a Qian et al., 2021 [14] Cement paste n/a Expansion(N2); MIP; WIP n/a ϕMIP < ϕN2 < ϕWIP Liu et al., 2019 [40] Shale n/a SANS; NA; MIP             SANS: λ = 0.53             nm (Δλ/λ = 18%)           NA: 77 K, P/P0           range in 0.01–0.99             MIP: injection             pressure 0–60,000             psia SSA result: NA < SANS < MIP CA: CO2-adsorption method. NA: N2-adsorption method. MA: CH4-adsorption method. SSA: Specific surface area. PSD: Pore-size distribution. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Sang Q. Zhang S. Li Y. Dong M. Bryant S. Determination of organic and inorganic hydrocarbon saturations and effective porosities in shale using vacuum-imbibition method Int. J. Coal Geol. 2018 200 123 134 10.1016/j.coal.2018.10.010 2. Yao Y. Liu J. Liu D. Chen J. Pan Z. A new application of NMR in characterization of multiphase methane and adsorption capacity of shale Int. J. Coal Geol. 2019 201 76 85 10.1016/j.coal.2018.11.018 3. Curtis J.B. Fractured shale-gas systems AAPG Bull. 2002 86 1921 1938 4. Wang M. Yu Q. Pore structure characterization of carboniferous shales from the eastern qaidam basin, China: Combining helium expansion with low-pressure adsorption and mercury intrusion J. Pet. Sci. Eng. 2017 152 91 103 10.1016/j.petrol.2017.02.007 5. Kuila U. Mccarty D.K. Derkowski A. Fischer T.B. Prasad M. Total porosity measurement in gas shales by the water immersion porosimetry (WIP) method Fuel 2014 117 1115 1129 10.1016/j.fuel.2013.09.073 6. Ladjevardi S.M. Asnaghi A. Izadkhast P.S. Kashani A.H. Applicability of graphite nanofluids in direct solar energy absorption Sol. Energy 2013 94 327 334 10.1016/j.solener.2013.05.012 7. An L. Liu X. Deng B. Jiang H. Cheng G.J. Liquid metal nanolayer-linked mof nanocomposites by laser shock evaporation Matter 2021 4 3977 3990 10.1016/j.matt.2021.10.012 8. Chen L. Wang Y. Wang H. Yang Y. Li J. Theoretical study of thiophene-inserted COF as high performance anode material for Li-ion battery Chem. Phys. Lett. 2021 783 139042 10.1016/j.cplett.2021.139042 9. Tozzini V. Pellegrini V. Prospects for hydrogen storage in graphene PCCP 2013 15 80 89 10.1039/C2CP42538F 23165421 10. Mukhtar A. Mellon N.B. Bustam M.A. Saqib S. Lee S.-P. Kareem F.A. Ullah S. Impact of amine functionality on the selective CO2/CH4 adsorption behavior of porous covalent triazine adsorbent J. Nat. Gas. Sci. Eng. 2020 83 103582 10.1016/j.jngse.2020.103582 11. Mao H. Zhang F. Du M. Dai L. Qian Y. Pang H. Review on synthesis of porous TiO2-based catalysts for energy conversion systems Ceram. Int. 2021 47 25177 25200 10.1016/j.ceramint.2021.06.039 12. Rogers J.A. Kaviany M. Funicular and evaporative-front regimes in convective drying of granular beds Int. J. Heat Mass Transf. 1992 35 469 480 10.1016/0017-9310(92)90283-X 13. Yang L. Zhang Y. Liu Z. Zhao P. Liu C. In-situ tracking of water transport in cement paste using X-ray computed tomography combined with cscl enhancing Mater. Lett. 2015 160 381 383 10.1016/j.matlet.2015.08.011 14. Qian R. Shi J. Liu C. Liu G. Liu Z. She W. Zhang Y. Zhang Y. Liang Y. Investigations on pore-structure in cementitious materials using gas intrusion porosimetry Measurement 2021 171 108816 10.1016/j.measurement.2020.108816 15. Lv Y. Zhang W. Wu F. Li H. Zhang Y. Xu G. Influence of initial damage degree on the degradation of concrete under sulfate attack and wetting–drying cycles Int. J. Concr. Struct. Mater. 2020 14 47 10.1186/s40069-020-00422-z 16. Liu C. Liu Z. Zhang Y. A multi-scale framework for modelling effective gas diffusivity in dry cement paste: Combined effects of surface, knudsen and molecular diffusion Cem. Concr. Res. 2020 131 106035 10.1016/j.cemconres.2020.106035 17. Qian R. Zhang Y. Liu C. Yang L. Liu G. She W. Quantitative characterization of three-dimensional pore structure in hardened cement paste using X-ray microtomography combined with centrifuge driven metal alloy intrusion Mater. Charact. 2018 145 277 283 10.1016/j.matchar.2018.08.047 18. Zhao H. Qin X. Liu J. Zhou L. Tian Q. Wang P. Pore structure characterization of early-age cement pastes blended with high-volume fly ash Constr. Build. Mater. 2018 189 934 946 10.1016/j.conbuildmat.2018.09.023 19. Rodrigues C.F. Sousa M. The measurement of coal porosity with different gases Int. J. Coal Geol. 2002 48 245 251 10.1016/S0166-5162(01)00061-1 20. Ajaxon I. Maazouz Y. Ginebra M.-P. Öhman C. Persson C. Evaluation of a porosity measurement method for wet calcium phosphate cements J. Biomater. Appl. 2015 30 526 536 10.1177/0885328215594293 26163278 21. Kwiecińska B. Pusz S. Valentine B.J. Application of electron microscopy tem and SEM for analysis of coals, organic-rich shales and carbonaceous matter Int. J. Coal Geol. 2019 211 103203 10.1016/j.coal.2019.05.010 22. Sun M. Zhao J. Pan Z. Hu Q. Yu B. Tan Y. Sun L. Bai L. Wu C. Blach T.P. Pore characterization of shales: A review of small angle scattering technique J. Nat. Gas. Sci. Eng. 2020 78 103294 10.1016/j.jngse.2020.103294 23. Bossa N. Chaurand P. Vicente J. Borschneck D. Levard C. Aguerre-Chariol O. Rose J. Micro- and nano-X-ray computed-tomography: A step forward in the characterization of the pore network of a leached cement paste Cem. Concr. Res. 2015 67 138 147 10.1016/j.cemconres.2014.08.007 24. Zhao H. Wu X. Huang Y. Zhang P. Tian Q. Liu J. Investigation of moisture transport in cement-based materials using low-field nuclear magnetic resonance imaging Mag. Concr. Res. 2021 73 252 270 10.1680/jmacr.19.00211 25. Sun M. Yu B. Hu Q. Zhang Y. Li B. Yang R. Melnichenko Y.B. Cheng G. Pore characteristics of longmaxi shale gas reservoir in the northwest of Guizhou, China: Investigations using small-angle neutron scattering (SANS), helium pycnometry, and gas sorption isotherm Int. J. Coal Geol. 2017 171 61 68 10.1016/j.coal.2016.12.004 26. Chang Y. Yao Y. Liu Y. Zheng S. Can cuttings replace cores for porosity and pore size distribution analyses of coal? Int. J. Coal Geol. 2020 227 103534 10.1016/j.coal.2020.103534 27. Wang Y. Zhu Y. Zhang R. Anovitz L.M. Chen S. SANS coupled with fluid invasion approaches for characterization of overall nanopore structure and mesopore connectivity of organic-rich marine shales in china Int. J. Coal Geol. 2020 217 103343 10.1016/j.coal.2019.103343 28. Jurkiewicz K. Pawlyta M. Burian A. Structure of carbon materials explored by local transmission electron microscopy and global powder diffraction probes C-J. Carbon Res. 2018 4 68 10.3390/c4040068 29. Hower J.C. Berti D. Hochella M.F. Jr. Rimmer S.M. Taulbee D.N. Submicron-scale mineralogy of lithotypes and the implications for trace element associations: Blue gem coal, knox county, kentucky Int. J. Coal Geol. 2018 192 73 82 10.1016/j.coal.2018.04.006 30. Zhang G.Q. Hirasaki G.J. House W.V. Effect of internal field gradients on NMR measurements Petrophysics 2001 42 37 47 31. Lim S. Lee H.-S. Kawashima S. Pore structure refinement of cement paste incorporating nanosilica: Study with dual beam scanning electron microscopy/focused ion beam (SEM/FIB) Mater. Charact. 2018 145 323 328 10.1016/j.matchar.2018.08.045 32. Zhao Y. Liu X. Chen B. Yang F. Zhang Y. Wang P. Robinson I. Three-dimensional characterization of hardened paste of hydrated tricalcium silicate by serial block-face scanning electron microscopy Materials 2019 12 1882 10.3390/ma12121882 31212729 33. Xu H. Probing nanopore structure and confined fluid behavior in shale matrix: A review on small-angle neutron scattering studies Int. J. Coal Geol. 2020 217 103325 10.1016/j.coal.2019.103325 34. Anovitz L.M. Cole D.R. Characterization and analysis of porosity and pore structures Rev. Mineral. Geochem. 2015 80 61 164 10.2138/rmg.2015.80.04 35. Erdman N. Drenzek N. Integrated preparation and imaging techniques for the microstructural and geochemical characterization of shale by scanning electron microscopy AAPG Mem. 2013 102 7 14 36. Kelly S. El-Sobky H. Torres-Verdín C. Balhoff M.T. Assessing the utility of FIB-SEM images for shale digital rock physics Adv. Water Resour. 2016 95 302 316 10.1016/j.advwatres.2015.06.010 37. Heath J.E. Dewers T.A. McPherson B.J.O.L. Petrusak R. Chidsey T.C. Jr. Rinehart A.J. Mozley P.S. Pore networks in continental and marine mudstones: Characteristics and controls on sealing behavior Geosphere 2011 7 429 454 10.1130/GES00619.1 38. Čalkovský M. Müller E. Meffert M. Firman N. Mayer F. Wegener M. Gerthsen D. Comparison of segmentation algorithms for FIB-SEM tomography of porous polymers: Importance of image contrast for machine learning segmentation Mater. Charact. 2021 171 110806 10.1016/j.matchar.2020.110806 39. Liu X. Zhang D. A review of phase behavior simulation of hydrocarbons in confined space: Implications for shale oil and shale gas J. Nat. Gas. Sci. Eng. 2019 68 102901 10.1016/j.jngse.2019.102901 40. Liu K. Ostadhassan M. Sun L. Zou J. Yuan Y. Gentzis T. Zhang Y. Carvajal-Ortiz H. Rezaee R. A comprehensive pore structure study of the Bakken shale with SANS, N2 adsorption and mercury intrusion Fuel 2019 245 274 285 10.1016/j.fuel.2019.01.174 41. Chung S.-Y. Kim J.-S. Stephan D. Han T.-S. Overview of the use of micro-computed tomography (micro-CT) to investigate the relation between the material characteristics and properties of cement-based materials Constr. Build. Mater. 2019 229 116843 10.1016/j.conbuildmat.2019.116843 42. Han T.-S. Zhang X. Kim J.-S. Chung S.-Y. Lim J.-H. Linder C. Area of lineal-path function for describing the pore microstructures of cement paste and their relations to the mechanical properties simulated from μ-CT microstructures Cem. Concr. Compos. 2018 89 1 17 10.1016/j.cemconcomp.2018.02.008 43. Gallucci E. Scrivener K. Groso A. Stampanoni M. Margaritondo G. 3D experimental investigation of the microstructure of cement pastes using synchrotron X-ray microtomography (μCT) Cem. Concr. Res. 2007 37 360 368 10.1016/j.cemconres.2006.10.012 44. Li A. Ding W. He J. Dai P. Yin S. Xie F. Investigation of pore structure and fractal characteristics of organic-rich shale reservoirs: A case study of lower cambrian qiongzhusi formation in malong block of eastern yunnan province, south china Mar. Pet. Geol. 2016 70 46 57 10.1016/j.marpetgeo.2015.11.004 45. Zhou H. Zhong J. Ren W. Wang X. Yi H. Characterization of pore-fracture networks and their evolution at various measurement scales in coal samples using X-ray μCT and a fractal method Int. J. Coal Geol. 2018 189 35 49 10.1016/j.coal.2018.02.007 46. Shi R. Liu J. Wang X. Elsworth D. Liu Z. Wei M. Liu X. Wang Z. Experimental observations of heterogeneous strains inside a dual porosity sample under the influence of gas-sorption: A case study of fractured coal Int. J. Coal Geol. 2020 223 103450 10.1016/j.coal.2020.103450 47. Farokhian D. Azin R. Ranjbar A. Application of medical and dental CT-scan technologies for determining porosity distribution of the persian gulf coastal zone and zagros basin core samples J. Afr. Earth. Sci. 2019 150 96 106 10.1016/j.jafrearsci.2018.10.009 48. Zhang M. He Y. Ye G. Lange D.A. van Breugel K. Computational investigation on mass diffusivity in portland cement paste based on X-ray computed microtomography (μCT) image Constr. Build. Mater. 2012 27 472 481 10.1016/j.conbuildmat.2011.07.017 49. Liu F. Zhang G. Luo T. Xin J. Jiang W. Study on pore development and water migration regularity in the process of strength formation of hydraulic concrete Measurement 2021 183 109811 10.1016/j.measurement.2021.109811 50. Salcedo I.R. Cuesta A. Shirani S. León-Reina L. Aranda M.A.G. Accuracy in cement hydration investigations: Combined X-ray microtomography and powder diffraction analyses Materials 2021 14 6953 10.3390/ma14226953 34832365 51. Żaba K. Puchlerska S. Książek M. Sitek R. Wiśniewski P. Mizera J. Investigations of model multilayer ceramic casting molds in a raw state by nondestructive methods Materials 2021 14 7761 10.3390/ma14247761 34947353 52. Tian W. Han N. Analysis on meso-damage processes in concrete by X-ray computed tomographic scanning techniques based on divisional zones Measurement 2019 140 382 387 10.1016/j.measurement.2019.04.026 53. Golsanami N. Sun J. Zhang Z. A review on the applications of the nuclear magnetic resonance (NMR) technology for investigating fractures J. Appl. Geophys. 2016 133 30 38 10.1016/j.jappgeo.2016.07.026 54. Song Y.Q. Kausik R. NMR application in unconventional shale reservoirs—A new porous media research frontier Prog. Nucl. Magn. Reson. Spectrosc. 2019 112 17 33 10.1016/j.pnmrs.2019.03.002 31481157 55. Yao Y. Liu D. Xie S. Quantitative characterization of methane adsorption on coal using a low-field NMR relaxation method Int. J. Coal Geol. 2014 131 32 40 10.1016/j.coal.2014.06.001 56. Chen J.J. Hürlimann M. Paulsen J. Freed D. Mandal S. Song Y.Q. Dispersion of T1 and T2 nuclear magnetic resonance relaxation in crude oils ChemPhysChem 2014 15 2676 2681 10.1002/cphc.201402077 24919743 57. Zuena M. Tomasin P. Alberghina M.F. Longo A. Marrale M. Gallo S. Zendri E. Comparison between mercury intrusion porosimetry and nuclear magnetic resonance relaxometry to study the pore size distribution of limestones treated with a new consolidation product Measurement 2019 143 234 245 10.1016/j.measurement.2019.05.021 58. Zhao Y. Chen J. Yang M. Liu Y. Song Y. A rapid method for the measurement and estimation of CO2 diffusivity in liquid hydrocarbon-saturated porous media using mri Magn. Reson. Imaging 2016 34 437 441 10.1016/j.mri.2015.12.024 26707850 59. Chen M. Hosking L.J. Sandford R.J. Thomas H.R. Dual porosity modelling of the coupled mechanical response of coal to gas flow and adsorption Int. J. Coal Geol. 2019 205 115 125 10.1016/j.coal.2019.01.009 60. Li D. Liu Q. Weniger P. Gensterblum Y. Busch A. Krooss B.M. High-pressure sorption isotherms and sorption kinetics of CH4 and CO2 on coals Fuel 2010 89 569 580 10.1016/j.fuel.2009.06.008 61. Zhao H. Ning Z. Wang Q. Zhang R. Zhao T. Niu T. Zeng Y. Petrophysical characterization of tight oil reservoirs using pressure-controlled porosimetry combined with rate-controlled porosimetry Fuel 2015 154 233 242 10.1016/j.fuel.2015.03.085 62. Sun J. Dong X. Wang J. Schmitt D.R. Xu C. Mohammed T. Chen D. Measurement of total porosity for gas shales by gas injection porosimetry (GIP) method Fuel 2016 186 694 707 10.1016/j.fuel.2016.09.010 63. Zou Y. Malzbender J. Development and optimization of porosity measurement techniques Ceram. Int. 2016 42 2861 2870 10.1016/j.ceramint.2015.11.015 64. Maria S. Methods for porosity measurement in lime-based mortars Constr. Build. Mater. 2010 24 2572 2578 10.1016/j.conbuildmat.2010.05.019 65. Tracz T. Zdeb T. Effect of hydration and carbonation progress on the porosity and permeability of cement pastes Materials 2019 12 192 10.3390/ma12010192 66. Fu Y. Jiang Y. Xia G. Optimization of gri porosity determination method for marine shale Nat. Gas. Ind. 2020 40 20 28 67. Luffel D.L. Guidry F.K. New core analysis methods for measuring reservoir rock properties of devonian shale J. Pet. Technol. 1992 44 1184 1190 10.2118/20571-PA 68. Topór T. Derkowski A. Kuila U. Fischer T.B. McCarty D.K. Dual liquid porosimetry: A porosity measurement technique for oil- and gas-bearing shales Fuel 2016 183 537 549 10.1016/j.fuel.2016.06.102 69. Kolias S. Investigation of the possibility of estimating concrete strength by porosity measurements Mater. Struct. 1994 27 265 272 10.1007/BF02473043 70. Testamanti M.N. Rezaee R. Determination of NMR T2 cut-off for clay bound water in shales: A case study of carynginia formation, perth basin, western australia J. Pet. Sci. Eng. 2017 149 497 503 10.1016/j.petrol.2016.10.066 71. Busch A. Gensterblum Y. CBM and CO2-ECBM related sorption processes in coal: A review Int. J. Coal Geol. 2011 87 49 71 10.1016/j.coal.2011.04.011 72. Chalmers G.R. Bustin R.M. Lower cretaceous gas shales in northeastern British Columbia, Part I: Geological controls on methane sorption capacity Bull. Can. Petrol. Geol. 2008 56 1 21 10.2113/gscpgbull.56.1.1 73. Busch A. Gensterblum Y. Krooss B.M. High-pressure sorption of nitrogen, carbon dioxide, and their mixtures on argonne premium coals Energy Fuels 2007 21 1640 1645 10.1021/ef060475x 74. Busch A. Gensterblum Y. Krooss B.M. Siemons N. Investigation of high-pressure selective adsorption/desorption behaviour of CO2 and CH4 on coals: An experimental study Int. J. Coal Geol. 2006 66 53 68 10.1016/j.coal.2005.07.003 75. Harpalani S. Prusty B.K. Dutta P. Methane/CO2 sorption modeling for coalbed methane production and CO2 sequestration Energy Fuels 2006 20 1591 1599 10.1021/ef050434l 76. Siemons N. Busch A. Measurement and interpretation of supercritical CO2 sorption on various coals Int. J. Coal Geol. 2007 69 229 242 10.1016/j.coal.2006.06.004 77. Mergia K. Stefanopoulos K.L. Ordás N. García-Rosales C. A comparative study of the porosity of doped graphites by small angle neutron scattering, nitrogen adsorption and helium pycnometry Microporous Mesoporous Mater. 2010 134 141 149 10.1016/j.micromeso.2010.05.019 78. Mastalerz M. He L. Melnichenko Y.B. Rupp J.A. Porosity of coal and shale: Insights from gas adsorption and SANSUSANS techniques Energy Fuels 2012 26 5109 5120 10.1021/ef300735t 79. Alexeev A. Ulyanova E. Starikov G. Kovriga N. Latent methane in fossil coals Fuel 2004 83 1407 1411 10.1016/j.fuel.2003.07.001 80. Alexeev A. Vasylenko T. Ul’Yanova E. Phase states of methane in fossil coals Solid State Commun. 2004 130 669 673 10.1016/j.ssc.2004.03.034 81. Kang J. Fu X. Li X. Liang S. Nitrogen injection to enhance methane and water production: An experimental study using the LF-NMR relaxation method Int. J. Coal Geol. 2019 211 103228 10.1016/j.coal.2019.103228 82. Liu J. Yao Y. Liu D. Elsworth D. Experimental evaluation of CO2 enhanced recovery of adsorbed-gas from shale Int. J. Coal Geol. 2017 179 211 218 10.1016/j.coal.2017.06.006

PMC009xxxxxx/PMC9099697.txt

==== Front Cells Cells cells Cells 2073-4409 MDPI 10.3390/cells11091505 cells-11-01505 Article miR-126-3p and miR-21-5p as Hallmarks of Bio-Positive Ageing; Correlation Analysis and Machine Learning Prediction in Young to Ultra-Centenarian Sicilian Population https://orcid.org/0000-0003-3565-9529 Accardi Giulia 1† https://orcid.org/0000-0003-3119-4445 Bono Filippa 2† Cammarata Giuseppe 34† https://orcid.org/0000-0003-2593-3221 Aiello Anna 1* https://orcid.org/0000-0003-4953-1446 Herrero Maria Trinidad 5 https://orcid.org/0000-0002-9935-1040 Alessandro Riccardo 36 Augello Giuseppa 3 https://orcid.org/0000-0002-6985-4907 Carru Ciriaco 7 https://orcid.org/0000-0002-2470-7030 Colomba Paolo 3 https://orcid.org/0000-0003-3377-6123 Costa Maria Assunta 8 De Vivo Immaculata 9 Ligotti Mattia Emanuela 13 Lo Curto Alessia 3 Passantino Rosa 8 https://orcid.org/0000-0002-2192-8938 Taverna Simona 34 Zizzo Carmela 3 Duro Giovanni 3‡ https://orcid.org/0000-0001-8004-2363 Caruso Calogero 1‡ Candore Giuseppina 1‡ Penalva Luiz Otavio Academic Editor 1 Laboratory of Immunopathology and Immunosenescence, Department of Biomedicine, Neurosciences and Advanced Technologies, University of Palermo, 90134 Palermo, Italy; giulia.accardi@unipa.it (G.A.); mattiaemanuela.ligotti@unipa.it (M.E.L.); calogero.caruso@unipa.it (C.C.); giuseppina.candore@unipa.it (G.C.) 2 Department of Economics, Business and Statistics, University of Palermo, Viale delle Scienze, Building N. 13, 90128 Palermo, Italy; filippa.bono@unipa.it 3 Institute for Biomedical Research and Innovation, National Research Council, 90146 Palermo, Italy; giuseppe.cammarata@cnr.it (G.C.); riccardo.alessandro@unipa.it (R.A.); giuseppa.augello@irib.cnr.it (G.A.); paolo.colomba@irib.cnr.it (P.C.); alessia.l_86@hotmail.it (A.L.C.); simona.taverna@cnr.it (S.T.); carmela.zizzo@irib.cnr.it (C.Z.); giovanni.duro@irib.cnr.it (G.D.) 4 Institute of Translational Pharmacology, National Research Council, 90146 Palermo, Italy 5 Clinical and Experimental Neuroscience, Institute for Aging Research, Biomedical Institute for Bio-Health Research of Murcia, School of Medicine, University of Murcia, Campus Mare Nostrum, 30100 Murcia, Spain; mtherrer@um.es 6 Section of Biology and Genetics, Department of Biomedicine, Neuroscience and Advanced Diagnostics, University of Palermo, 90127 Palermo, Italy 7 Department of Biomedical Sciences, University of Sassari, 07100 Sassari, Italy; carru@uniss.it 8 Institute of Byophysics, National Research Council, 90146 Palermo, Italy; mariaassunta.costa@cnr.it (M.A.C.); rosa.passantino@ibf.cnr.it (R.P.) 9 Department of Epidemiology, Harvard T.H. Chan School of Public Health, Boston, MA 021382, USA; nhidv@channing.harvard.edu * Correspondence: anna.aiello@unipa.it † These authors contributed equally to this work. ‡ These authors contributed equally to this work. 30 4 2022 5 2022 11 9 150521 2 2022 22 4 2022 © 2022 by the authors. 2022 https://creativecommons.org/licenses/by/4.0/ Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). Human ageing can be characterized by a profile of circulating microRNAs (miRNAs), which are potentially predictors of biological age. They can be used as a biomarker of risk for age-related inflammatory outcomes, and senescent endothelial cells (ECs) have emerged as a possible source of circulating miRNAs. In this paper, a panel of four circulating miRNAs including miR-146a-5p, miR-126-3p, miR-21-5p, and miR-181a-5p, involved in several pathways related to inflammation, and ECs senescence that seem to be characteristic of the healthy ageing phenotype. The circulating levels of these miRNAs were determined in 78 healthy subjects aged between 22 to 111 years. Contextually, extracellular miR-146a-5p, miR-126-3p, miR-21-5p, and miR-181a-5p levels were measured in human ECs in vitro model, undergoing senescence. We found that the levels of the four miRNAs, using ex vivo and in vitro models, progressively increase with age, apart from ultra-centenarians that showed levels comparable to those measured in young individuals. Our results contribute to the development of knowledge regarding the identification of miRNAs as biomarkers of successful and unsuccessful ageing. Indeed, they might have diagnostic/prognostic relevance for age-related diseases. ageing inflamm-ageing endothelial senescence longevity miRNAs This research received no external funding. ==== Body pmc1. Introduction microRNAs (miRNAs) are single-stranded and non-coding RNA molecules of 21–23 nucleotides that negatively regulate gene expression, which are involved in a wide range of physiological and pathological conditions. They have been observed to freely circulate in human body fluids such as blood, saliva, cerebrospinal and amniotic fluids, breast milk, and urine or in extracellular vesicles (EVs), and are selectively packaged and secreted by cells [1,2,3]. Recent findings demonstrate that senescent endothelial cells (ECs) are a source of circulating miRNAs, so they have been suggested as potential targets to counteract the ageing process [1,4]. The modulation of miRNAs transcription or their circulating levels have been also associated with age-related diseases [4,5]. The studies relating miRNA profiles to human longevity are relatively few. Since miRNAs pathways are evolutionary conserved, it is possible to formulate deep insights into the topic thanks to experimental models. Studies on C. elegans and mammals have demonstrated the role of miRNAs in modulating ageing, longevity, and cellular senescence. In C. elegans the modulation of miRNA maturation and functions can modify the lifespan and altered longevity phenotypes [6,7]. Changes in circulating miRNA levels could be considered a measure of biological ageing; accordingly, some miRNAs are proposed as new potential biomarkers of ageing and age-related diseases [4,8,9,10,11,12,13,14,15]. Recent findings indicate that several miRNAs are involved in inflamm-ageing, an ageing-related state characterized by systemic chronic low-grade inflammation that in turn supports a biological-background inducing predisposition to age-related diseases. Specific miRNAs such as miR-126-3p, miR-21-5p, miR-146a-5p, and miR-181a-5p are widely recognised as senescence-associated miRNAs (SA-miRNAs) and regulate the expression of genes involved in ageing and longevity [16,17]. miR-126-3p is one of the most described SA-miRNAs and is considered as an endothelial cell-specific miRNA that regulates vascular integrity and angiogenesis [18]. Circulating miR-126-3p, released by ECs, exerts protective mechanisms against vascular endothelial dysfunctions involved in age-related vascular diseases [19,20]. miR-21 is an inflamma-miRNA that orchestrates the tuning of the inflammatory response. It can be considered as an miRNA that is able to modulate the “switch on/off” of inflammation at appropriate times [21]. Since ageing is associated with the alteration of redox signalling and increased circulating inflammatory cytokines, miR-21 circulating levels could be useful biomarkers of age-related diseases and their complications [22]. miR-146a is also considered an inflamm-miRNA and imparts several beneficial effects for healthy ageing. It was demonstrated that circulating miR-146a levels decline in healthy older people. Both miR-146a and miR-21 are able to track physiological and pathological ageing trajectories, so the monitoring of these miRNAs can contribute to the promotion of healthy ageing and prevent or postpone age-related diseases onset [23]. miR-181a-5p is known as ageing-related mito-miRNA since it can modulate mitochondrial activity. Its modulation could mediate the loss of mitochondrial integrity and function in ageing cells, contributing to the inflammatory response [24]. In the present study, the levels of miR-126-3p, miR-21-5p, miR-146a-5p, and miR-181a-5p were analyzed in the plasma samples of different age-cohorts, from young to ultra-centenarians, of a population from Western Sicily. The aim of the paper was to identify specific miRNAs as potential biomarkers of healthy ageing and their correlation with age and sex as well as with smoking habits, and other biological parameters. miRNA levels have also been evaluated in the conditioned media of human umbilical vein endothelial cells (HUVECs) undergoing replicative senescence. Furthermore, since some specific biological features seem to be age-associated, it was possible to find the relation between these features and people’s age using machine learning (ML) techniques. ML is a branch of Artificial Intelligence focused on imitating the way human beings are capable of learning by taking experimental data and examples as inputs to infer a specific result or output [25]. ML often builds mathematical models based on sample data to make predictions without being explicitly programmed for such purposes. In general, ML models are considered as black boxes where inputs are transformed into outputs by combining such inputs with a set of adjusted data obtained as a result of a training process [26]. Although it provides many benefits in different areas of application, the lack of interpretability could be a problem in some circumstances. In the context of this work, the age prediction problem is also addressed by considering several biological features, since it is interesting to study which factors affect the population from young to ultra-centenarians. Finding such relations in an interpretable manner is important; therefore, it is necessary to use ML techniques with a high degree of interpretability. Consequently, a decision-tree classifier has been selected for this purpose, since such an ML method presents a high rate of accuracy, great robustness and the results are very interpretable as long as they are short. 2. Materials and Methods 2.1. Study Design, Participants, and Anamnestic Data The participants were recruited from June 2017 to March 2020 within the project “Discovery of molecular and genetic/epigenetic signatures underlying resistance to age-related diseases and comorbidities (DESIGN, 20157ATSLF)”, funded by the Italian Ministry of Education, University and Research. The Ethics Committee of Palermo University Hospital (Palermo, Sicily, Italy) approved the study protocol (Nutrition and Longevity, No. 032017). The study was conducted in accordance with the Declaration of Helsinki and its amendments. Study participants (or their caregivers) gave their written informed consent prior to enrolment. All recruited subjects were Sicilians aged between 22 and 111 years. We excluded people with chronic invalidating diseases, such as neoplastic and autoimmune diseases, as well as with acute diseases, such as infectious diseases, and individuals with severe dementia. To respect privacy, all donors were identified with an alphanumeric code and the data were managed using a database accessible exclusively by researchers involved in the project. A team composed of demographers, statistics, biologists, and physicians, from University of Palermo, disseminated a detailed questionnaire to collect demographic, clinical, and anamnestic data of interest from participants as well as functional and cognitive information. The enrolment was conducted at University of Palermo for young adults, adults, and older adults, using social networks and word of mouth for recruitment, whereas it was conducted at home for the ultra-centenarians. For more information about the recruited population, please see [27]. A total of 78 healthy donors (females: 42; males: 36) were randomly selected from this large database. The population was divided in four age groups, i.e., young adults (22–50 years, n = 19), adults (51–70 years, n = 28), older adults (71–99 years, n = 20), and ultra-centenarians (100–111 years, n = 11). The age, gender, smoking habits and some anthropometric, hematological, haematochemical, oxidative, and molecular parameters of the recruited subjects are reported in Supplemental Table S1. 2.2. Plasma Sample Acquisition and RNA Isolation Overnight fasting blood samples were obtained in the morning and processed as previously reported [27]. Total RNA was extracted and purified from plasma, using an miRNeasy®Mini kit (Qiagen, CA, cat.No. 217004), according to standard protocol. The RNA concentration was assessed, using the RNA Nano 6000 Assay Kit of the Agilent Bio-analyzer 2100 System (Agilent Technologies, Palo Alto, CA, USA). RNA quality was assessed with the Eppendorf biophotometer D30 (Eppendorf, Hamburg, Germany). For this study, we used only RNA with a ratio of A260/280 from 1.9 to 2. 2.3. TaqMan RT-qPCR miRNA Assays The isolated miRNAs were retro-transcripted using an miScript Single Cell qPCR kit (Qiagen, Hilden, Germany), according to the manufacturer’s protocol. The expression levels of miRNAs were evaluated with a SYBR green-based Real-Time quantitative PCR (RT-qPCR), using Step one plus (Applied Biosystem, Waltham, MA, USA). For the amplification, we used an miScript SYBR green PCR kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. The 20 μL PCR mixture included 2 μL of reverse transcription product, 10 μL of QuantiTect SYBR Green PCR Master Mix, 2 μL of miScript Universal Primer, 4 μL of RNase-free water and 2 μL of Primer Assay specific for each microRNAs (miScript Primer Assays miR-21-5p Lot. N°201803230019, miR-126-3p Lot. N°20150817013s, miR-146a-5p Lot. N°201709120071, miR-181a-5p Lot. N°20161222018 Qiagen, Hilden, Germany). The reaction mixtures were incubated at 95 °C for 15 min, followed by 40 amplification cycles of 94 °C for 15 s, 55 °C for 30 s, and 70 °C for 30 s. Triplicate samples and inter-assay controls were used. Therefore, for the normalization of RT-qPCR data, using the 2-DCT method, we used miR-30a (miScript Primer Assay Lot. N°201709270012 Qiagen, Hilden, Germany). Linear fold changes were calculated and plotted on scatter plots using Prism (GraphPad Prism Software, San Diego, CA, USA). 2.4. Cell Cultures HUVECs (ATCC®CRL-1730™, Manassas, VA, USA) were grown in Medium 200 (GIBCO) supplemented with hydrocortisone (1 µg/mL), human epidermal growth factor (10 ng/mL), basic fibroblast growth factor (3 ng/mL), heparin (10 µg/mL), gentamicin (10 µg/mL), amphotericin (0.25 µg/mL), and fetal bovine serum (2% v/v), at 37 °C, in a 5% CO2 atmosphere, at 95% humidity. For cell-replicative senescence, first passage cryopreserved HUVECs were grown and serially passed until they reached senescence [28]. The number of population doublings (PDs) was calculated by using the following formula: PD = [ln (number of cells harvested)–ln (number of cells seeded)]/ln2. For the experiments, cells were used at different cumulative population doublings (CPDs) which is the sum of all PD. Cells studied in early passage (CPD < 20) were regarded as young cells, whereas those that have been passed more times were regarded as intermediate-age (CPD < 32) or old, i.e., senescent (CPD > 56) endothelial cells (ECs). 2.5. Collection of Cell-Conditioned Media and miRNA Isolation The conditioned media, containing all factors secreted by the cell line, were harvested from the cultures, centrifuged for 4 min at 300× g at 4 °C × 10 min, and stored at −20 °C until required for extraction. Total RNA was extracted by phenol/guanidine-based lysis of samples and a silica-membrane-column-based purification using an miRNeasy® Mini kit (Qiagen, CA, cat.No. 217004), according to standard protocol. The RNA concentration was assessed using RNANano 6000 Assay Kit of the Agilent Bio-analyzer 2100 System (Agilent Technologies, Palo Alto, CA, USA). 2.6. Senescence-Associated β-Galactosidase Staining HUVECs were grown in a 4-well chamber slide (Nunc™ Lab-Tek™ II, Thermo Fisher Scientific, Fremont, CA, USA), at a density of 6 × 104 cells/well, in 500 µL of culture medium, for 24 h. At the end of the different times of doubling, HUVECs were fixed and stained for galactosidase activity using a Senescence Cell Staining kit according to the manufacturer instructions (Sigma-Aldrich, St. Louis, MO, USA). The percentage of senescence-associate gal positive cells was determined by counting the number of blue cells within a sample, using a Zeiss Axioskop microscope (Carl Zeiss, Göttingen, Germany) with an X20 lens. Ten random fields were photographed for each passage, and the percentage of SA-β-gal-positive cells were calculated. 2.7. MTS Assay Cell viability was evaluated by using CellTiter 96 Aqueous One Solution Cell Proliferation Assay (Promega, Madison, WI, USA), a colorimetric method for determining the number of viable cells in proliferation, according to manufacturer instructions. 2.8. miRNA Targeted Gene Prediction and KEGG Pathway Analyses by miRWalk We utilized miRWalk, an miRNA target gene prediction database, to select predicted and validated targets, and to analyze enriched KEGG pathways [29]. The miRWalk prediction database integrated other programs including miRanda, miRDB, RNA22, and Targetscan. Moreover, KEGG is an exhaustive database for the functional interpretation and practical application of genomic information and integrates macromolecular datasets from genome sequencing and other high-throughput experimental techniques [30,31]. 2.9. Statistical Analyses of miRNA Levels All calculations were made by using Stata 16.1 Software (StataCorp, College Station, TX, USA). Descriptive statistics were calculated for all the data considered in the study. Data from in vitro experiments were expressed as means and standard error of the means from at least three independent experiments. Comparison among multiple groups, in data with normal distribution and with a homogeneity of variance, was analysed by one-way analysis of variance (ANOVA). Kruskal–Wallis and Dunn tests with Bonferroni correction were considered to compare groups when data were not normally distributed and in the presence of heterogeneity. A robust multiple regression analysis was considered to look at the relationship between miRNAs and some individual and lifestyle characteristics of sample units. The robust estimation method corrects the standard errors of estimated coefficients in the presence of heteroskedasticity. The Wald test based on the robustly estimated variance matrix was considered to evaluate the significance of coefficients and R2 as a goodness-of-fit statistic. In this study, statistical significance was assumed when p < 0.05. 2.10. ML Techniques The dataset was randomly divided into training and test sets. Age prediction was carried out to see which factors affect the population from young to ultra-centenarians [32]. A training and evaluation ratio of 7:3 was used. The ML algorithm used was the decision tree classifier. The decision tree algorithm belongs to the family of supervised ML algorithms [26]. The decision tree construction process consists of the selection of characteristics. Each attribute or characteristic is calculated following certain standards that allow for the selection of the most important characteristic that are characteristic of the partition samples each time. One of the main advantages of this algorithm is its high classification precision and great robustness [33,34]. To obtain reliable data, all data provided in the database were revised. Once each optimal model was found, they were fitted to the entire training set and tested on the test set. The ML algorithm was developed using the Python 3.9 (Wilmington, DE, USA) programming language with Scikit-learn, a free software ML library [35]. 3. Results 3.1. Plasma Values of miRNAs miRNAs were collected from the plasma of 78 healthy donors, randomly selected, and grouped into four age classes: young adults (22–50 years, n = 19), adults (51–70 years, n = 28), older adults (71–99 years, n = 20), and ultra-centenarians (100–111 years, n = 11). Figure 1 shows the distribution of miRNAs levels in relation to age. For miR-21-5p and miR-126-3p, it is apparent that the maximum values were found in the age ranges 51–70 and 71–99 years, while the ultra-centenarians showed lower levels and the smallest variability. For the miR-146a-5p and miR-181-5p, in all classes of age, it was observed that the median value is around 5 but ultra-centenarians show a higher variability than the other classes of age. Individual points are considered as outliers if, when defining the upper adjacent value U = q3 + (3/2) (q3 − q1) and the lower adjacent value L = q1 − (3/2) (q3 − q1), the point is higher than U or lower than L. The upper (U) and lower (L) adjacent values are defined by Tukey (1977). Regarding significance, the Kruskal–Wallis test clearly demonstrates that all the four miRNAs show significant age-related variability (Table 1). All data (mean, median and standard deviation of the total samples and of the samples divided by gender) and the p-values of the comparisons between the different age groups are presented in Table 1. Given the relatively small number of the various groups, there was no significance found between groups of age by sex. The Table 1 shows mean, median, and standard deviation (SD) of the four miRNAs in different population age groups and for gender. The letters a, b, c, and d indicate, respectively, young adults (22–49), adults (50–69), older adults (70–98), and ultra-centenarians (100–111). The Kruskal–Wallis test and Dunn test with Bonferroni adjustment are used to test significant differences among age groups for each miRNA. Data show that for the miR-21-5p values there was no significant difference between the ultra-centenarian and young people (data not shown) while the ultra-centenarian values were lower and significantly different from those of adults and older adults. For miR-126-3p, the values for ultra-centenarians were significantly lower than those of other groups that were not significantly different from each other. For the miR-146a-5p and miR-181-5p values, there was no significant difference between ultra-centenarian values and the other group due to the high heterogeneity of the centenarian measures. The plasma levels of miRNA of the young adults were lower than those of the other adult groups but the significance with both groups was obtained only for miR-181-5p. 3.2. Correlation of Plasma Values of miRNAs with Some Parameters Robust multiple regression models were considered to estimate the relationship between each miRNA with some individual and lifestyle-variables, i.e., gender, age, BMI, and smoking habits. The estimation method was used to correct the standard errors of estimated coefficients in the presence of heteroskedasticity. Results of the estimated coefficients are reported in Table 2 and Table 3. Figure 2 reports the scatterplots of the estimated models and the observed data for each miRNA by age. As the relationship between miR-21-5p and age seems to follow a quadratic function, a quadratic model for age was estimated:(1) miR−21−5p=b0+b1Age+b2Age2+b3BMI+b4Smoke+b5Gender The miR-21-5p increases with age, with a decreasing average rate of change. A higher level of miR-21-5p was observed around the age 60–65 years (Figure 1. at the top left) and after these ages it tended to shrink to the same levels as in younger participants. No significant effects, at a 5% significance level, were found for gender, BMI, and smoking habits. For the other miRNAs, linear relationships with age were estimated as follows:miR−126−3p=b0+b1AgeClass+b2BMI+b3Smoke+b4Gender miR−146−5p=b0+b1AgeClass+b2BMI+b3Smoke+b4Gender (2) miR−181a−5p=b0+b1AgeClass+b2BMI+b3Smoke+b4Gender The estimated model for the miR-126-3p shows that in the centenarians compared to young cells, the level of miR-126-3p reduced by a mean of −3.63. The category of ho has never smoked vs. who smoked show an increasing mean level of 2.91. No significant effect exists for gender and BMI. For the miR-146a-5p, no significant effect was noted for all the analysed variables. The levels of the miR-181a-5p increased in ex-smokers and in never smokers compared with smokers and no difference exists for gender in all miRNAs. 3.3. miRNAs Levels in HUVECs Undergoing Replicative Senescence Since the ECs senescence is known to be involved in the ageing process, we hypothesised that the observed age-related alterations of the miRNA levels in study could be linked to EC senescence. Because it is difficult to study EC senescence ex vivo, to investigate age-related changes of miRNA, HUVECs were used as an in vitro model. In cultured HUVECs undergoing replicative senescence, miRNA levels were measured in CM. Cell senescence was defined based on cumulative population doubling (CPD). We considered CPD > 56 for old (senescent) cells, CPD < 32 for intermediate age cells, and CPD < 20 for young cells. Senescence was estimated by SA-β-gal activity and measured as the percentage of positive cells (40 ± 10 in old, 10 ± 4 in intermediate age and 2 ± 1 in young cells, data not shown). As shown in Figure 3, the CM of old ECs showed higher amounts of the four miRNAs than for young cells. The same results were obtained with the CM of intermediate age ECs for two out of four of the studied miRNAs (miR-21-5p and miR-126-3p) that were present in higher amounts than in the CM of young cells. In addition, the CM of old ECs showed higher amounts of miR-146a-5p and miR-181a-5p than CM of intermediate cells. These results indicate that in vitro EC miRNA levels correlate with the replicative senescent state and suggest that variations observed ex vivo in adult and older adults might be linked to the EC senescence process (Figure 3). 3.4. Enriched KEGG Pathway Clustered by Validated Targets of miR-21-5p, miR-126-3p, miR-146a-5p, and miR-181a-5p and Corresponding Target Genes In order to investigate the possible regulation mechanisms involved in the ageing process of the four miRNAs under study we utilized an online bioinformatics database, called miRWalk, to select plausible targets and validated targets of these miRNAs. Considering that miR-21-5p and miR-126-3p display similar patterns while miR-146a-5p and miR-181a-5p show a linear trend, we analyzed the two pairs of miRNAs separately. The potential target genes of miR-21-5p and miR-126-3p were 448 and 271, respectively. A total of 27 genes were recognized as common targets of both miR-21-5p and miR-126-3p by overlapping the analyses. Results of the enrichment KEGG pathway indicated that the targeted genes regulated by miR-21-5p and miR-126-3p were involved in 105 pathways of which the top 15 are shown in Supplementary Table S2. Interestingly, among the most significant pathways, two are related to longevity regulating pathways. In total, 1492 and 1475 target genes of miR-146a-5p and miR-181a-5p were, respectively, predicted and 52 genes were co-regulated by miR-146a-5p and miR-181a-5p. The target genes of miR-146a-5p and miR-181a-5p were significantly enriched in 207 KEGG pathways. Notably, among the first 15 pathways, three are related to the Sumoylation process, an essential post-translational modification that has evolved to regulate intricate networks within emerging complexities of eukaryotic cells [36]. See Supplementary materials Figures S1 and S2 and Tables S2–S5 for the complete analyses. 3.5. ML Analyses The results of ML are represented in Figure 4. For calculation, miRNAs, C reactive protein, telomeres, paraoxonase (PON), trolox equivalent antioxidant capacity, and malondialdehyde values were considered according to age-ranges. Overall, 100% of subjects under 50 years old have a telomere length of above 0.903, an miR-181a-5p of below 4.606, a miR-21-5p of below 17.04, and an miR-126 -3p of less than 13.586. Individuals with a telomere length greater than 0.903 and a miR-181a-5p greater than 4.606 were between 50 and 70 years. Individuals with a telomere length below 0.903, a miR-21-5p greater than 5.164, a miR-181a-5p less than 11.441 and a PON less than 140.813 were between 71 and 99 years. For ultra-centenarians, the telomere length is always less than 0.903. Additionally, they have an miR-21-5p below 5.164. We also found ultra-centenarians with an miR-21-5p higher than 5.164 but all of them would have had an miR-181a-5p above 11.441. 4. Discussion Ageing is a series of physiological events that are usually ineluctable. It becomes a risk and accelerator factor for most age-related diseases, including cardiovascular ones [37]. In turn, senescence is the biological ageing of cells and an irreversible form of long-term cell-cycle arrest, caused by excessive intracellular or extracellular stress or damage. ECs senescence impairs vascular functions, thereby enhancing the ageing of tissues and organs. Several stimuli, including reactive oxygen species, inflammatory cytokines, and telomere dysfunction, can increase their senescence [17]. Recently, senescent ECs have emerged as a possible source of circulating miRNAs that are short non-coding RNAs that generally either induce the degradation of mRNA or repress the translation of target transcripts. It has already been established that they are important regulators of the ageing process and modulators of longevity. Different miRNAs directly influence the duration of life through the modulation of various ageing pathways which represent adaptive mechanisms aimed at maintaining the homeostasis of the organism including inflammatory responses [4]. In the present study, we analysed the blood levels of four circulating miRNAs, namely miR-146a-5p, miR-126-3p, miR-21-5p, and miR-181a-5p, which are involved in several pathways related to inflammation and EC senescence, in 78 healthy subjects aged between 22 to 111 years. Most studies on centenarians describe them as the best models of successful ageing, representing selected people in which the appearance of major age-related diseases, such as cancer and cardiovascular diseases, has been consistently delayed or escaped. However, extreme longevity is often characterized by a non-unique and unambiguous phenotype, as demonstrated by the centenarian population in the world. Not all centenarians are similar, although all can represent a model of “positive biology” [38,39]. The data reported that for miR-126-3p and miR-21-5p, the maximum values were found between 51 and 99 years while the ultra-centenarians showed lower levels. For miR-146a-5p and miR-181a-5p, the highest mean values were observed in the group of individuals over 100 years old, but the differences with the values of younger people were not significant due to their heterogeneity as previously stated. The Kruskal–Wallis test showed that all four miRNAs had significant age-related variability. Then, robust multiple regressions were performed to estimate the relationship between each miRNA with some parameters in addition to age, i.e., gender, BMI, and smoking habits. No effect of gender and BMI was observed in relation to all circulating miRNAs levels, unlike the results of Ameling et al., who found a significant association of some miRNAs with BMI, likely due to a different composition of the study population with a wider range of BMI values [40]. Never-smoking is instead related to an incremental effect of miR-126-3p and miR-181-a-5p. This datum is not surprising because several miRNAs have been shown to be under regulated in smokers [41]. Finally, a significant effect with age was observed for miR-126a-3p and miR-21-5p. For the levels of miR-181-a-5p no significant age effect was observed. In previous reports in other samples of Sicilian populations, results have demonstrated that in addition to miR-181a-5p the blood levels of miR-223-5p and let-7a-5p also did not show significant differences between centenarians and adult females (18–64 years old) [42]. Unfortunately, in that paper, older women (65–99 years old) were not included. In another report, mean levels of miR-126-3p, carried in small extracellular vesicles (EV) isolated from plasma, increased significantly with age, with the highest levels observed in the nineties [2]. Thus, we assumed that the age-related miRNAs variations in adult and older adults could be linked to ECs senescence. Based on the evidence that in vitro replicative senescence of ECs mimics the progressive age-related changes of endothelial functions described ex vivo, we used HUVECs undergoing replicative senescence as an in vitro model [43]. We observed that HUVECs released miRNA at progressively higher levels with increasing senescence processes except for miR-146a-5p and miR-181a-5p, whose intermediate values were not significantly different from those observed in younger cells. In a previous paper, HUVECs also released EVs-miR-126-3p in progressively higher levels with increasing senescence process. Overall, these data suggest that variations observed ex vivo from young to older people might be related to the EC senescence process [38]. Finally, results from ML highlighted the significant role, in people over 70 year, played by miR-21-5p and miR-181a-5p, in addition to PON and telomere length [44,45]. Circulatory miRNAs change with age and with the development of age-related diseases. This suggests their possible use as non-invasive “biomarkers of ageing”, which can predict successful ageing and longevity [8]. Some other studies have investigated the differential expression of miRNA between young, old and centenarian populations. Balzano et al. performed a study comparing circulating miRNA levels in three cohorts, i.e., centenarians, patients with rheumatoid arthritis in treatment with corticosteroids, and young and middle-aged healthy subjects as controls [13]. Their results showed that miR-425-5p, miR-21, and miR-212 were significantly decreased in centenarians and in patients with rheumatoid arthritis treated with corticosteroids compared to controls. The authors suggest that this miRNAs pattern could protect against inflammation for centenarians and also found that this occurs in patients treated with glucocorticoids. Serna et al. studied the miRNA expression profiles in mononuclear cells of centenarians, octogenarians, and younger Spaniards [14] The results showed a significant overlap of the profiles of centenarians with those of the young but not with the octogenarians. A longitudinal study was performed by Smith-Vikos et al. using serum samples collected from short-lived (58–75 years) and long-lived (76–92 years) participants of the Baltimore Longitudinal Study of Ageing [15]. A total of 24 miRNAs were found to be significantly upregulated and 73 were found to be significantly downregulated in long-lived individuals. Six of these miRNAs, not analysed in the present study, were both differentially expressed and correlated with individual lifespan. Recent data on inflammation-related miRNAs in successful and unsuccessful ageing outline a complex scenario characterised by an altered expression of specific miRNAs which mainly target the nuclear factor κB (NF-κB) pathway, the master regulator of inflammation. In physiological conditions, their transcription is at baseline levels. However, the initiation of pro-inflammatory signalling results in the strong co-induction of their expression through a mechanism that is largely NF-κB-dependent [46]. Furthermore, miRNAs have been reported to show age-dependent changes in blood levels, and to be markers of inflamm-ageing, i.e., the low-grade inflammatory status of ageing, in successful and unsuccessful ageing [11,12]. miR-146a-5p was the first to be identified as an NF-κB-dependent miRNA, being up regulated in response to various immune stimuli. In turn, miR-146a downregulates relevant proteins which are also in the canonical NF-κB pathway. miR-146a exerts several beneficial effects on successful ageing, via a fine tuning of canonical and non-canonical NF-κB pathways [23,47]. A decline in miR-146a-5p levels would be expected to trigger the release of inflammatory cytokines [23]. Accordingly, the downregulation of miR-146a has been observed in patients with diabetes, obesity, and hypertension [48]. An increased expression of miR-146a in HUVECs and in aortic and coronary ECs during replicative senescence occurs, thus demonstrating that NF-κB activation and cell senescence can be modulated by the miRNAs described. Thus, cellular senescence and inflammation signalling activation may be closely interconnected and share common regulators [11]. Furthermore, miR-21-5p, induced by several pro-inflammatory molecules, is able to activate NF-κB and Nod Like Receptor family pyrin domain containing 3 (NLRP3), hence orchestrating the fine tuning of the inflammatory response through direct and indirect activities on NF-κB and NLRP3 pathways in a context-dependent manner. In vitro and in vivo studies of animal models confirmed the essential role played by miR-21 in regulating this inflammatory switch, both promoting or inhibiting NF-κB/NLRP3 pathways. For all these reasons, miR-21 can be considered as an miRNA that is able to modulate the “switch on/off” of inflammation at appropriate times [11]. Accordingly, miR-21-5p is down-regulated in healthy older people and up-regulated in patients affected by many age-associated disorders, representing an excellent candidate with which to track the ageing trajectories. A significant miR-21-5p increase observed in Alzheimer’s patients with severe cognitive impairment supports this hypothesis [49]. miR-21 also lies at the intersection of senescence, inflammation, and age-related diseases [12]. It has been demonstrated that the age-related increase in plasma miR-126-3p was paralleled by a 5/6-fold increase in intra/extracellular miR-126-3p in cultured HUVECs undergoing senescence. Indeed, it has a role in the maintenance of endothelial integrity, enhancing endothelial functions and promoting blood vessel formation. Senescence-associated miR-126-3p up-regulation is likely a senescence-associated compensatory mechanism. It is also a modulator of inflammation and the innate immune response, targeting some components of the NF-kB pathway [50]. Finally, a target prediction and pathway analysis of miR-181a-5p showed overlapping involvement of the inflammatory pathways, consistent with the relationship between chronic inflammation and ageing [51]. The upregulation of miR-181a may be associated with the homeostatic response to inflammatory stimuli [52]. In a previous paper, we presented a case report of a centenarian woman in relatively good health, despite laboratory signs of atrophic gastritis and a chronic status of inflammation [42]. Her levels of miR-181a-5p, miR-223-5p, and let-7a-5p, which play a role in the control of innate immunity and inflammation, were higher than those of female centenarians. These data suggested a possible epigenetic modulation, likely with anti-inflammatory effects that confer protection against tissue damage. A decreased expression of miR-181a has been observed in patients with coronary artery disease and it has been suggested to have an antiatherogenic effect through blocking NF-κB activation and vascular inflammation [53]. Furthermore, the target pathway analysis of miR-126-3p and miR-21-5p also demonstrated their involvement in the nutrient sensing pathway that is linked to insulin and insulin growth factor-1, which has an important role as “gatekeeper” by balancing the cell response to oxidative stress and nutrient availability. Downstream of this pathway there is the Forkhead box O3 (FOXO3) A transcription factor. Notably, the FOXO family transactivate genes are involved in resistance to oxidative stress, energy metabolism, DNA damage repair, glucose metabolism, autophagy and the protection of proteins by chaperones, favouring survival and longevity [54]. Instead, a target pathway analysis of miR-146a-5p and miR-181a-5p also illustrates their involvement in the sumoylation process. Sumoylation plays critical roles in cellular senescence. Enhancing the global sumoylation or inhibiting the desumoylation process seems to promote senescence. In sumoylation-mediated cellular senescence, the p53 and RB proteins are SUMO substrates and have been identified as important molecules in this senescence process [55,56]. Overall, the data suggest that the analyzed miRNAs can be active components of the senescent ECs secretome and may modulate the rate of inflammation at the cellular and systemic level [12]. Tissue and circulating miRNAs could contribute to restraining the activity of the senescent cell secretome and modulate the destruction induced by the activation of the inflammatory response [12,42,50]. With age, cellular senescence, visceral obesity, microbiota, and the stimulation of the immune system contribute to inducing and perpetuating inflamm-ageing over time, inducing an upregulation of these miRNAs, also called inflamma-miRs, that lead to the excessive activation of the inflammatory pathways [10,11,12]. Thus, the up-regulation of inflamma-miRs in blood circulation occurs in healthy older individuals. However, as our data show, this increase is not significant (although with a great heterogeneity of response, see above) in ultra-centenarians since they are able to control inflammation. On the other hand, their values are greater in patients with age-related diseases. So, circulating levels could be useful biomarkers of these diseases and of their complications. The main sources of inflamma-miR circulating in successful and unsuccessful ageing should be both immune cells and ECs [9,11,23,57]. Accordingly, our data show that variation observed ex vivo from younger to older people might be related to the ECs senescence process. 5. Conclusions Our data provides further evidence of these four miRNAs acting as biomarkers of successful and unsuccessful ageing, although a note of caution should be added considering the small sample size and the imbalance of the number of samples for each age group. We confirmed that miR-126-3p and miR-21-5p are correlated with ageing and it is therefore possible to consider them hallmarks of bio-positive ageing. Moreover, ML data for miR-21-5p and miR-181a-5p suggest that these miRNAs can be ideal indicators of longevity along with telomere length. Indeed, correlation analyses show that miR-181a-5p is lower in never-smoked subjects. Thus, circulating inflamma-miRNAs might have diagnostic/prognostic relevance for age-related diseases. Acknowledgments We thank Nieves Pavón (Polytechnic University of Cartagena) for providing support during the review process and University of Murcia that specifically funded MTH. Supplementary Materials The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/cells11091505/s1. Table S1. Anthropometric, hematological, hematochemical, molecular, oxidative stress parameters and smoking habits. Table S2. KEGG pathway of miR-21-5p and miR-126-3p. Table S3. REACTOME pathway of miR-21-5p and miR-126-3p. Table S4. KEGG pathway of miR-146a-5p and miR-181a-5p. Table S5. REACTOME pathway of miR-146a-5p and miR-181a-5p. Figure S1. The top 15 enriched KEGG and REACTOME pathways from predicted target genes of miR-21-5p and miR-126-3p searched by miRWalk. Figure S2. The top 15 enriched KEGG and REACTOME pathways from predicted target genes of miR-146a-5p and miR-181a-5p searched by miRWalk. Click here for additional data file. Author Contributions Conceptualization, C.C. (Calogero Caruso), G.D., G.C. (Giuseppina Candore) and G.C. (Giuseppe Cammarata); Formal analysis, F.B., and G.C. (Giuseppe Cammarata); Funding acquisition, C.C. (Calogero Caruso) and G.D.; Investigation, G.A. (Giulia Accardi), A.A., G.A. (Giuseppa Augello), P.C., M.A.C., M.E.L., A.L.C., R.P., S.T. and C.Z.; Methodology, G.A. (Giulia Accardi), A.A., M.T.H., R.A., C.C. (Ciriaco Carru) and I.D.V.; Project administration, G.D.; Supervision, A.A., C.C. (Calogero Caruso), G.C. (Giuseppina Candore), G.C. (Giuseppe Cammarata); Writing original draft, G.A. (Giulia Accardi), A.A., C.C. (Calogero Caruso), G.C. (Giuseppe Cammarata) and F.B. Writing, review and editing, G.A. (Giulia Accardi), A.A., C.C. (Calogero Caruso), G.C. (Giuseppe Cammarata) and F.B. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement The study was conducted according to the guidelines of the Declaration of Helsinki. The Institutional Ethics Committee (“Paolo Giaccone”, University Hospital) approved the study protocol (Nutrition and Longevity, No. 032017). Informed Consent Statement Informed consent was obtained from all subjects involved in the study. Data Availability Statement The data presented in this study are available in [insert article or supplementary material here]. Conflicts of Interest The authors declare that they have no competing interest. Figure 1 Box-and-Whisker plots of miRNAs levels by class of age. Box-and-Whisker plots report minimum, quartiles (q1, q2 and q3) and maximum levels of each miRNA by age class. The individual points (or dots) plotted are outliers. The spacings between different parts of the box indicate dispersion and skewness in the data and shows a graphical measure of interquartile range (q3–q1) and range (max–min). Figure 2 Scatter Plot of observed and fitted models of correlation of plasma values for miRNAs with age. Figure 3 miRNAs levels variation in HUVECs undergoing senescence. Bar charts show extracellular miR-146a-5p, miR-21-5p, miR-181a-5p, miR-126-3p levels in young, intermediate, and old (senescent) HUVECs. Data were calculated by qRT-PCR and represent mean ± SD of three different experiments analysed. CTs (cycle thresholds) resulting from qRT-PCR analysis were normalised with miR-30a; levels were calculated with 2-DCT method and expressed as folds, with respect to lowest value registered. Comparisons among multiple groups were analysed by one-way analysis of variance, followed by Bonferroni’s post hoc test. * p < 0.05. Figure 4 Figure shows the machine learning results. For the interpretation of the decision tree, it is necessary to know that each node of a decision tree contains a condition. This condition can be true or false. If the condition is true, we descend to the next left node. If the condition is false, we descend to the next right node. The different colors represent the age groups (light blue < 50, orange 50–70, green 71–99, pink/lilac > 99 years). cells-11-01505-t001_Table 1 Table 1 Plasma miRNA levels in the 4 age-groups of Sicilian population. Please, note that a, b, c, and d are the letters used to indicate the different groups: a = Young Adults; b = Adults; c = Older Adults; d = Ultracentenarians. miRNA (a) Young Adults (22–50 y.o.) N = 19, M = 7, W = 12 (b) Adults (51–70 y.o.) N = 28, M = 14, W = 14 (c) Older Adults (71–99 y.o.) N = 20, M = 13, W = 7 (d) Ultracentenarians (100–111 y.o.) N = 11, M = 2, W = 9 p-Value of Kruskal–Wallis Test p-Value Bonferroni Test mean median SD mean median SD mean median SD mean median SD p = 0.0003 (a vs. b), p = 0.0028; (a vs. c), p = 0.0284; (b vs. d), p = 0.0024; (c vs. d), p = 0.0171. miR-21-5p N 8.22 4.96 6.45 18.87 17.34 11.06 16.73 17.47 10.74 5.73 4.01 4.58 M 7.71 4.96 6.35 19.56 18.07 13.24 15.40 17.32 8.75 3.05 3.05 1.36 W 8.51 7.50 6.78 18.18 17.34 8.83 19.19 27.77 14.19 6.33 4.43 4.88 miR-126-3p N 5.70 5.10 2.16 7.58 6.33 4.54 10.77 9.36 7.33 2.67 1.80 1.62 p = 0.0002 (a vs. d), p = 0.0238; (b vs. d), p = 0.0005; (c vs. d), p = 0.0001. M 6.27 6.06 2.19 6.98 5.84 4.80 11.85 12.36 8.21 3.72 3.72 1.46 W 5.37 4.49 2.17 8.18 6.71 4.35 8.76 6.35 5.30 2.44 1.68 1.64 miR-146a-5p N 3.34 3.57 0.94 3.34 3.57 0.94 5.67 3.79 5.63 6.23 3.85 5.63 p = 0.0491 (a vs. b), p = 0.0177. M 3.11 3.57 0.78 5.41 3.71 4.92 5.05 3.43 6.23 4.60 4.60 2.43 W 3.48 3.32 1.03 6.13 5.66 2.47 6.83 5.41 4.52 6.60 3.85 6.17 miR-181a-5p N 3.11 3.28 1.03 7.22 5.41 5.52 5.27 5.07 2.85 9.30 4.93 10.36 p = 0.0024 (a vs. b), p = 0.0006; (a vs. c), p = 0.0358. M 2.87 2.82 0.78 6.30 4.83 4.73 4.99 4.03 3.12 4.65 4.65 0.40 W 3.25 3.43 1.16 8.15 6.62 6.24 5.80 5.19 2.39 10.34 10.47 11.30 cells-11-01505-t002_Table 2 Table 2 Coefficients and significance estimates of miR-21-5p. R2 = 0.346 Coefficient p Age 1.46 <0.0005 Age2 −0.01 <0.0005 BMI −0.20 =0.54 Smoke Smoker (reference) Ex-smokers 1.94 =0.57 Never smoked 5.24 =0.06 Gender M (reference) F −2.02 =0.40 b0 (constant) −23.32 =0.01 cells-11-01505-t003_Table 3 Table 3 Coefficients and significance estimates of other miRNAs. miR-126-3p miR-146a-5p miR-181a-5p Coefficient p Coefficient p Coefficient p Age Class 22–50 (Reference) 51–70 2.08 =0.14 2.84 <0.0005 4.23 0.00 71–99 4.37 =0.05 2.39 =0.16 1.61 0.30 100–111 −3.63 <0.0005 1.15 =0.34 5.28 0.13 BMI 0.07 =0.61 −0.05 =0.74 −0.01 0.95 Smoke Smoker (Reference) Ex-smoker 2.05 =0.25 2.14 =0.10 4.13 0.00 Never smoked 2.91 =0.04 1.36 =0.06 1.93 0.01 Gender M (Reference) F −0.85 =0.49 0.72 =0.46 1.63 0.16 b 0 2.50 =0.49 2.95 =0.44 0.68 0.87 R2 = 0.2779 R2 = 0.116 R2 = 0.2169 Our elaboration was conducted with Stata Software 16.1. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Jung H.J. Suh Y. Circulating miRNAs in ageing and ageing-related diseases J. Genet. Genom. 2014 41 465 472 10.1016/j.jgg.2014.07.003 25269672 2. Lo Curto A. Taverna S. Costa M.A. Passantino R. Augello G. Adamo G. Aiello A. Colomba P. Zizzo C. Zora M. Can Be miR-126-3p a Biomarker of Premature Aging? An Ex Vivo and In Vitro Study in Fabry Disease Cells 2021 10 356 10.3390/cells10020356 33572275 3. Pucci M. Reclusa Asiáin P. Duréndez Sáez E. Jantus-Lewintre E. Malarani M. Khan S. Fontana S. Naing A. Passiglia F. Raez L.E. Extracellular Vesicles As miRNA Nano-Shuttles: Dual Role in Tumor Progression Target Oncol. 2018 13 175 187 10.1007/s11523-018-0551-8 29374391 4. Kinser H.E. Pincus Z. MicroRNAs as modulators of longevity and the aging process Hum. Genet. 2020 139 291 308 10.1007/s00439-019-02046-0 31297598 5. miRbase: The microRNA database Home Page Available online: https://www.miRbase.org/index.shtml (accessed on 20 February 2022) 6. Kato M. Chen X. Inukai S. Zhao H. Slack F.J. Age-associated changes in expression of small, noncoding RNAs, including microRNAs, in C. elegans RNA 2011 17 1804 1820 10.1261/rna.2714411 21810936 7. Smith-Vikos T. Slack F.J. MicroRNAs and their roles in aging J. Cell Sci. 2012 125 7 17 10.1242/jcs.099200 22294612 8. Kumar S. Vijayan M. Bhatti J.S. Reddy P.H. MicroRNAs as Peripheral Biomarkers in Aging and Age-Related Diseases Prog. Mol. Biol. Transl. Sci. 2017 146 47 94 10.1016/bs.pmbts.2016.12.013 28253991 9. Ogonowski N. Salcidua S. Leon T. Chamorro-Veloso N. Valls C. Avalos C. Bisquertt A. Rentería M.E. Orellana P. Duran-Aniotz C. Systematic Review: microRNAs as Potential Biomarkers in Mild Cognitive Impairment Diagnosis Front. Aging Neurosci. 2022 13 807764 10.3389/fnagi.2021.807764 35095478 10. Olivieri F. Capri M. Bonafè M. Morsiani C. Jung H.J. Spazzafumo L. Viña J. Suh Y. Circulating miRNAs and miRNA shuttles as biomarkers: Perspective trajectories of healthy and unhealthy aging Mech. Ageing Dev. 2017 165 162 170 10.1016/j.mad.2016.12.004 27986629 11. Olivieri F. Prattichizzo F. Giuliani A. Matacchione G. Rippo M.R. Sabbatinelli J. Bonafè M. miR-21 and miR-146a: The microRNAs of inflammaging and age-related diseases Ageing Res. Rev. 2021 70 101374 10.1016/j.arr.2021.101374 34082077 12. Olivieri F. Rippo M.R. Procopio A.D. Fazioli F. Circulating inflamma-miRs in aging and age-related diseases Front. Genet. 2013 4 121 10.3389/fgene.2013.00121 23805154 13. Balzano F. Deiana M. Dei Giudici S. Oggiano A. Pasella S. Pinna S. Mannu A. Deiana N. Porcu B. Masala A.G.E. MicroRNA Expression Analysis of Centenarians and Rheumatoid Arthritis Patients Reveals a Common Expression Pattern Int. J. Med. Sci. 2017 14 622 628 10.7150/ijms.18972 28824293 14. Serna E. Gambini J. Borras C. Abdelaziz K.M. Belenguer A. Sanchis P. Avellana J.A. Rodriguez-Mañas L. Viña J. Centenarians, but not octogenarians, up-regulate the expression of microRNAs Sci. Rep. 2012 2 961 Erratum in: Sci. Rep. 2013, 3, 1575. Mohammed, Kheira [corrected to Abdelaziz, Kheira M] 10.1038/srep01575 23233880 15. Smith-Vikos T. Liu Z. Parsons C. Gorospe M. Ferrucci L. Gill T.M. Slack F.J. A serum miRNA profile of human longevity: Findings from the Baltimore Longitudinal Study of Aging (BLSA) Aging 2016 8 2971 2987 10.18632/aging.101106 27824314 16. Cammarata G. Duro G. Chiara T.D. Curto A.L. Taverna S. Candore G. Circulating miRNAs in Successful and Unsuccessful Aging. A Mini-review Curr. Pharm. Des. 2019 25 4150 4153 10.2174/1381612825666191119091644 31742494 17. Yamakuchi M. Hashiguchi T. Endothelial Cell Aging: How miRNAs Contribute? J. Clin. Med. 2018 7 170 10.3390/jcm7070170 18. Venkat P. Cui C. Chopp M. Zacharek A. Wang F. Landschoot-Ward J. Shen Y. Chen J. MiR-126 Mediates Brain Endothelial Cell Exosome Treatment-Induced Neurorestorative Effects After Stroke in Type 2 Diabetes Mellitus Mice Stroke 2019 50 2865 2874 10.1161/STROKEAHA.119.025371 31394992 19. Wang S. Aurora A.B. Johnson B.A. Qi X. McAnally J. Hill J.A. Richardson J.A. Bassel-Duby R. Olson E.N. The endothelial-specific microRNA miR-126 governs vascular integrity and angiogenesis Dev. Cell 2008 15 261 271 10.1016/j.devcel.2008.07.002 18694565 20. Jebari-Benslaiman S. Galicia-García U. Larrea-Sebal A. Olaetxea J.R. Alloza I. Vandenbroeck K. Benito-Vicente A. Martín C. Pathophysiology of Atherosclerosis Int. J. Mol. Sci. 2022 23 3346 10.3390/ijms23063346 35328769 21. Borja-Gonzalez M. Casas-Martinez J.C. McDonagh B. Goljanek-Whysall K. Inflamma-miR-21 Negatively Regulates Myogenesis during Ageing Antioxidants 2020 9 345 10.3390/antiox9040345 22. Ren S. Lin P. Wang J. Yu H. Lv T. Sun L. Du G. Circular RNAs: Promising Molecular Biomarkers of Human Aging-Related Diseases via Functioning as an miRNA Sponge Mol. Ther. Methods Clin. Dev. 2020 18 215 229 10.1016/j.omtm.2020.05.027 32637451 23. Olivieri F. Lazzarini R. Babini L. Prattichizzo F. Rippo M.R. Tiano L. Di Nuzzo S. Graciotti L. Festa R. Brugè F. Anti-inflammatory effect of ubiquinol-10 on young and senescent endothelial cells via miR-146a modulation Free Radic. Biol. Med. 2013 63 410 420 10.1016/j.freeradbiomed.2013.05.033 23727324 24. Rippo M.R. Olivieri F. Monsurrò V. Prattichizzo F. Albertini M.C. Procopio A.D. MitomiRs in human inflamm-aging: A hypothesis involving miR-181a, miR-34a and miR-146a Exp. Gerontol. 2014 56 154 163 10.1016/j.exger.2014.03.002 24607549 25. Bi Q. Goodman K.E. Kaminsky J. Lessler J. What is Machine Learning? A Primer for the Epidemiologist Am. J. Epidemiol. 2019 188 2222 2239 10.1093/aje/kwz189 31509183 26. Khatri S. Arora A. Prakash A. Supervised Machine Learning Algorithms for Credit Card Fraud Detection: A Comparison Proceedings of the 2020 10th International Conference on Cloud Computing, Data Science & Engineering (Confluence) Noida, India 29–31 January 2020 680 683 10.1109/Confluence47617.2020.9057851 27. Aiello A. Accardi G. Aprile S. Caldarella R. Carru C. Ciaccio M. De Vivo I. Gambino C.M. Ligotti M.E. Vasto S. Age and Gender-related Variations of Molecular and Phenotypic Parameters in A Cohort of Sicilian Population: From Young to Centenarians Aging Dis. 2021 12 1773 1793 10.14336/AD.2021.0226 34631220 28. Boisen L. Drasbek K.R. Pedersen A.S. Kristensen P. Evaluation of endothelial cell culture as a model system of vascular ageing Exp. Gerontol. 2010 45 779 787 10.1016/j.exger.2010.06.003 20600781 29. miRWalk Home—miRWalk Available online: uni-heidelberg.de (accessed on 20 February 2022) 30. KEGG Pathway KEGG Overview Available online: genome.jp (accessed on 20 February 2022) 31. Kanehisa M. Sato Y. Furumichi M. Morishima K. Tanabe M. New approach for understanding genome variations in KEGG Nucleic Acids Res. 2019 47 D590 D595 10.1093/nar/gky962 30321428 32. Fuentes P. Amador S. Lucas-Ochoa A.M. Cuenca-Bermejo L. Fernández-Villalba E. Raparelli V. Norris C. Kautzky-Willer A. Kublickiene K. Pilote L. Sex, rurality and socioeconomical status in Spanish centennial population (2017) Aging 2021 13 22059 22077 10.18632/aging.203563 34565717 33. Quinlan J.R. Induction of Decision Trees Mach. Learn. 1986 1 81 106 10.1007/BF00116251 34. Zhou H. Zhang J. Zhou Y. Guo X. Ma Y. A feature selection algorithm of decision tree based on feature weight Expert Syst. Appl. 2021 164 113842 10.1016/j.eswa.2020.113842 35. Pedregosa F. Varoquaux G. Gramfort A. Michel V. Thirion B. Grisel O. Blondel M. Prettenhofer P. Weiss R. Dubourg V. Scikit-learn: Machine Learning in Python J. Mach. Learn. Res. 2011 12 2825 2830 36. Sahin U. de Thé H. Lallemand-Breitenbach V. Sumoylation in Physiology, Pathology and Therapy Cells 2022 11 814 10.3390/cells11050814 35269436 37. Caruso C. Accardi G. Ligotti M.E. Vasto S. Candore G. Pathobiology of aging: An introduction to age-related diseases Human Aging Caruso C. Candore G. Academic Press Cambridge, MA, USA 2021 35 74 9780128225691 38. Accardi G. Ligotti M.E. Candore G. Phenotypic Aspects of Longevity Centenarians Caruso C. Springer Cham, Switzerland 2019 978-3-030-20762-5 39. Caruso C. Passarino G. Puca A. Scapagnini G. “Positive biology”: The centenarian lesson Immun. Ageing 2012 9 5 10.1186/1742-4933-9-5 22524297 40. Ameling S. Kacprowski T. Chilukoti R.K. Malsch C. Liebscher V. Suhre K. Pietzner M. Friedrich N. Homuth G. Hammer E. Associations of circulating plasma microRNAs with age, body mass index and sex in a population-based study BMC Med. Genom. 2015 8 61 10.1186/s12920-015-0136-7 41. Banerjee A. Luettich K. MicroRNAs as potential biomarkers of smoking-related diseases Biomark. Med. 2012 6 671 684 10.2217/bmm.12.50 23075247 42. Aiello A. Accardi G. Aprile S. Caldarella R. Cammarata G. Carru C. Caruso C. Ciaccio M. De Vivo I. Gambino C.M. Pro-inflammatory status is not a limit for longevity: Case report of a Sicilian centenarian Aging Clin. Exp. Res. 2021 33 1403 1407 10.1007/s40520-020-01628-7 32577916 43. Ong J. Woldhuis R.R. Boudewijn I.M. van den Berg A. Kluiver J. Kok K. Terpstra M.M. Guryev V. de Vries M. Vermeulen C.J. Age-related gene and miRNA expression changes in airways of healthy individuals Sci. Rep. 2019 9 3765 10.1038/s41598-019-39873-0 30842487 44. Lescai F. Marchegiani F. Franceschi C. PON1 is a longevity gene: Results of a meta-analysis Ageing Res. Rev. 2009 8 277 284 10.1016/j.arr.2009.04.001 19376276 45. Arbeev K.G. Verhulst S. Steenstrup T. Kark J.D. Bagley O. Kooperberg C. Reiner A.P. Hwang S.J. Levy D. Fitzpatrick A.L. Association of Leukocyte Telomere Length With Mortality Among Adult Participants in 3 Longitudinal Studies JAMA Netw. Open 2020 3 200023 10.1001/jamanetworkopen.2020.0023 46. Boldin M.P. Baltimore D. MicroRNAs, new effectors and regulators of NF-κB Immunol. Rev. 2012 246 205 220 10.1111/j.1600-065X.2011.01089.x 22435557 47. Olivieri F. Lazzarini R. Recchioni R. Marcheselli F. Rippo M.R. Di Nuzzo S. Albertini M.C. Graciotti L. Babini L. Mariotti S. MiR-146a as marker of senescence-associated pro-inflammatory status in cells involved in vascular remodelling Age 2013 35 1157 1172 10.1007/s11357-012-9440-8 22692818 48. Roganović J. Downregulation of microRNA-146a in diabetes, obesity and hypertension may contribute to severe COVID-19 Med. Hypotheses 2021 146 110448 10.1016/j.mehy.2020.110448 33338955 49. Giuliani A. Gaetani S. Sorgentoni G. Agarbati S. Laggetta M. Matacchione G. Gobbi M. Rossi T. Galeazzi R. Piccinini G. Circulating Inflamma-miRs as Potential Biomarkers of Cognitive Impairment in Patients Affected by Alzheimer’s Disease Front. Aging Neurosci. 2021 13 647015 10.3389/fnagi.2021.647015 33776746 50. Bonafè M. Olivieri F. Circulating microRNAs in aging Oncotarget 2015 6 6 1340 10.18632/oncotarget.3175 51. Noren Hooten N. Fitzpatrick M. Wood W.H. III De S. Ejiogu N. Zhang Y. Mattison J.A. Becker K.G. Zonderman A.B. Evans M.K. Age-related changes in microRNA levels in serum Aging 2013 5 725 740 10.18632/aging.100603 24088671 52. Xie W. Li Z. Li M. Xu N. Zhang Y. miR-181a and inflammation: miRNA homeostasis response to inflammatory stimuli in vivo Biochem. Biophys. Res. Commun. 2013 430 647 652 10.1016/j.bbrc.2012.11.097 23220232 53. Su Y. Yuan J. Zhang F. Lei Q. Zhang T. Li K. Guo J. Hong Y. Bu G. Lv X. MicroRNA-181a-5p and microRNA-181a-3p cooperatively restrict vascular inflammation and atherosclerosis Cell Death Dis. 2019 10 365 10.1038/s41419-019-1599-9 31064980 54. Flachsbart F. Dose J. Gentschew L. Geismann C. Caliebe A. Knecht C. Nygaard M. Badarinarayan N. ElSharawy A. May S. Identification and characterization of two functional variants in the human longevity gene FOXO3 Nat. Commun. 2017 8 2063 10.1038/s41467-017-02183-y 29234056 55. Princz A. Tavernarakis N. The role of SUMOylation in ageing and senescent decline Mech. Ageing Dev. 2017 162 85 90 10.1016/j.mad.2017.01.002 28088449 56. Gong L. Sun Q. Li D.W. Sumoylation in Cellular Senescence and Aging Curr. Mol. Med. 2017 16 871 876 10.2174/1566524016666161223104915 28017139 57. Olivieri F. Bonafè M. Spazzafumo L. Gobbi M. Prattichizzo F. Recchioni R. Marcheselli F. La Sala L. Galeazzi R. Rippo M.R. Age- and glycemia-related miR-126-3p levels in plasma and endothelial cells Aging 2014 6 771 787 10.18632/aging.100693 25324472